AURA PULSE: AN AI-POWERED SYSTEM FOR REAL-

TIME EMOTIONAL SUPPORT AND PERSONALIZED

RECOMMENDATIONS
(A final year project report)
Submitted in partial fulfilment of the requirements
for the degree
of

BACHELOR OF TECHNOLOGY
in
COMPUTER SCIENCE & DESIGN
by

Nensi Varshney (2101431650023)

Prince Kumar (2101431650027)
Shivang Rastogi (2101431650036)
Vivek Chaudhary (2101431650046)

under the Supervision of

DR. MONIKA NAGAR
(Assistant Professor)

Department of Computer Science and Design

IMS Engineering College
NH-09, Adhyatmik Nagar, Near Dasna,
Distt: Ghaziabad, UP
May, 2025
 VISION OF THE INSTITUTE
To make IMSEC an Institution of Excellence for empowering students through
technical education, incorporating human values, and developing engineering acumen
for innovations and leadership skills to upgrade society.

MISSION OF THE INSTITUTE

 To promote academic excellence by continuous learning in core and emerging
Engineering domains using innovative teaching and learning methodologies.
 To inculcate values and ethics among the learners.
 To promote industry interactions and cultivate young minds for
entrepreneurship.
 To create a conducive learning ecosystem and research environment on a
perpetual basis to develop students as technology leaders and entrepreneurs
who can address tomorrow’s societal needs.

VISION OF THE DEPARTMENT

To produce technocrats for designing technology-based solutions with sound ethics,
innovative mindset and leadership skills required for the sustainable development of
the Nation.

MISSION OF THE DEPARTMENT

M1: To develop understanding of design principles and techniques in the domain
of computer science and their applicability for developing innovative
solution to real world problems.
M2: To provide an environment of collaborative learning for developing technical
and leadership skills among graduates.
M3: To produce graduates who are competent in their profession, exhibit ethics
with a responsible attitude towards the nation.

ii
 PROGRAMME EDUCATIONAL OBJECTIVES (PEOs)
1. To produce graduates who are employable in industries/public sector/research
organizations or work as an entrepreneur.

2. To produce graduates who can provide solutions to challenging problems in their

profession by applying computer engineering theory and practices.

3. To develop ability to demonstrate team work with the ability of leadership,

analytical reasoning for solving time critical problems and strong human values
for responsible professional.

PROGRAMME SPECIFIC OUTCOME (PSOs)

1. Empowering the students for continuous learning and develop efficient solutions
for emerging challenges in the computation domain.
2. Preparing graduates who are able to apply standard software engineering practices
in software development and management process using suitable programming
languages and platforms.

PROGRAM OUTCOMES
Engineering Graduates will be able to:
1. Engineering knowledge: Apply the knowledge of mathematics, science,
engineering fundamentals, and an engineering specialization to the solution of
complex engineering problems.

2. Problem analysis: Identify, formulate, review research literature, and analyze

complex engineering problem searching substantiated conclusions using first
principles of mathematics, natural sciences, and engineering sciences.

3. Design/development of solutions: Design solutions for complex engineering

problems and design system components or processes that meet the specified
needs with appropriate consideration for the public health and safety, and the
cultural, societal, and environmental considerations.

iii
4. Conduct investigations of complex problems: Use research-based knowledge and
research methods including design of experiments, analysis and interpretation of
data, and synthesis of the information to provide valid conclusions.

5. Modern tool usage: Create, select, and apply appropriate techniques, resources,
and modern engineering and IT tools including prediction and modeling to
complex engineering activities with an understanding of the limitations.

6. The engineer and society: Apply reasoning informed by the contextual knowledge
to assess societal, health, safety, legal and cultural issues and the consequent
responsibilities relevant to the professional engineering practice.

7. Environment and sustainability: Understand the impact of the professional

engineering solutions in societal and environmental contexts, and demonstrate the
knowledge of, and need for sustainable development.

8. Ethics: Apply ethical principles and commit to professional ethics and

responsibilities and norms of the engineering practice.

9. Individual and team work: Function effectively as an individual, and as a member

or leader in diverse teams, and in multidisciplinary settings.

10. Communication: Communicate effectively on complex engineering activities

with the engineering community and with society at large, such as, being able to
comprehend and write effective reports and design documentation, make effective
presentations, and give and receive clear instructions.

11. Project management and finance: Demonstrate knowledge and understanding of

the engineering and management principles and apply these to one’s own work, as
a member and leader in a team, to manage projects and in multidisciplinary
environments.

12. Life-long learning: Recognize the need for, and have the preparation and ability
to engage in independent and life-long learning in the broadest context of
technological change.

iv
 CO-PO-PSO MAPPING FOR ACADEMIC SESSION 2024-25

Course Name: Project AKTU Course Code: KCS851

Semester / Year: VIII/ 4th NBA Code:C411
Subject Coordinator: Dr. Shalendra Kumar, Mr. Awdhesh
Kumar

Course Outcomes:
COGNITIVE
CO. No. LEVEL
DESCRIPTION
(BLOOMS
TAXONOMY)
C411.1 Analyze and understand the real-life problem and apply their
K4 , K5
knowledge to get programming solution.
C411.2Engage in the creative design process through the integration and
application of diverse technical knowledge and expertise to K4 , K5
meet customer needs and address social issues.
C411.3 Use the various tools and techniques, coding practices for
K5 , K6
developing real life solution to the problem.
C411.4 Find out the errors in software solutions and establishing the
K4 , K5
process to design maintainable software applications
C411.5 Write the report about what they are doing in project and
K5, K6
learning
the team working skills

CO-PO-PSO Mapping:

PO PO PO
PO1 PO2 PO3 PO4 PO5 PO6 PO7 PO8 PO9 PSO1 PSO2
10 11 12
C411.1 3 3 3 3 3 1 1 2 2 3 3 3 3 3

C411.2 3 3 3 3 3 1 1 2 2 3 3 3 3 3

C411.3 3 3 3 3 3 1 1 2 2 3 3 3 3 3

C411.4 3 3 3 3 3 1 1 2 2 3 3 3 3 3

C411.5 3 3 3 3 3 1 1 3 3 3 3 3 3 3

C411
3 3 3 3 3 1 1 2.2 2.2 3 3 3 3 3

v
 DECLARATION
I hereby declare that the work, which is being presented in the Project, entitled
“Aura Pulse : An Ai-Powered System For Real-Time Emotional Support And
Personalized Recommendations” in partial fulfillment for the award of Degree of
“Bachelor of Technology” in Computer Science and Design, and submitted to the
Department of Computer Science and Design, IMS Engineering College,
Ghaziabad, affiliated to Dr. A.P.J Abdul Kalam Technical University, Uttar Pradesh,
Lucknow is a record of my own investigations carried under the Guidance of DR.
MONIKA NAGAR, IMS Engineering College, Ghaziabad.

I have not submitted the matter presented in this Project anywhere for the award of
any other Degree.

Signature:
Name: Nensi Varshney
Roll No: 2101431650023
Date:

Signature:
Name: Prince Kumar
Roll No: 2101431650027
Date:

Signature:
Name: Shivang Rastogi
Roll No: 2101431650036
Date:

Signature:
Name: Vivek Chaudhary
Roll No: 2101431650046
Date:

vi
 CERTIFICATE

I hereby certify that the work which is being presented in the project report entitled
“Aura Pulse : An Ai-Powered System For Real-Time Emotional Support And
Personalized Recommendations” by “ NENSI VARSHNEY (2101431650023),
PRINCE KUMAR (2101431650027), SHIVANG RASTOGI (2101431650036),
VIVEK CHAUDHARY (2101431650046) ” in partial fulfillment of requirements for
the award of degree of [Link]. (CSD) submitted in the Department of Computer
Science and Design at “IMS Engineering College” under A.P.J. ABDUL KALAM
TECHNICAL UNIVERSITY, LUCKNOW is an authentic record of work carried out
under the supervision of DR. MONIKA NAGAR

SUPERVISOR
Name: DR. MONIKA NAGAR
Signature:
Date:

vii
 ACKNOWLEDGEMENT

I would like to place on record my deep sense of gratitude to Dr. Monika Nagar Dept.
of Computer Science and Design, IMSEC, Ghaziabad, India for his generous
guidance, help and useful suggestions.

I express my sincere gratitude to Prof. (Dr.) Sonia Juneja, HoD in Department of

Computer Science and Design, IMSEC, Ghaziabad, for her stimulating guidance,
continuous encouragement and supervision throughout the course of present work.

I am extremely thankful to Prof. (Dr.) Manas Kumar Mishra, Director, IMSEC,

Ghaziabad, for providing me infrastructural facilities to work in, without which this
work would not have been possible.

Signature:
Name: Nensi Varshney
Roll No: 2101431650023
Date:

Signature:
Name: Prince Kumar
Roll No: 2101431650027
Date:

Signature:
Name: Shivang Rastogi
Roll No: 2101431650036
Date:

Signature:
Name: Vivek Chaudhary
Roll No: 2101431650046
Date:

viii
 TABLE OF CONTENT

Vision and Mission........................................................................................................ii

CO-PO-PSO Mapping....................................................................................................v
Candidate’s Declaration................................................................................................vi
Certificate.....................................................................................................................vii
Acknowledgement.......................................................................................................viii
Table of contents...........................................................................................................ix
List of Figures..............................................................................................................xii
ABSTRACT..................................................................................................................1
Chapter 1 INTRODUCTION......................................................................................2
1.1 Problem Identification............................................................................2
1.2 Detailed Problem Definition...................................................................2
1.3 Current Market Survey...........................................................................3
1.4 Advantages offered by Current Project..................................................4
1.5 Current Challenges.................................................................................4
Chapter 2 LITERATURE SURVEY..........................................................................5
2.1 Runfang Guo et al. (2024).....................................................................5
2.2 Kranthi Kiran et al. (2024)....................................................................5
2.3 Padmavati K et al. (2024)......................................................................5
2.4 Moeez Rajjan et al. (2023)....................................................................5
2.5 M. Athavle et al. (2021)........................................................................5
2.6 Mrinmoy Bhaumik et al. (2021)...........................................................6
2.7 Metilda Florence et al. (2020)...............................................................6
Chapter 3 METHODOLOGY.....................................................................................7
3.1 Data Sources...........................................................................................7
3.2 Preprocessing Steps...................................................................................7
3.3 Machine Learning Models Facial Emotion Recognition........................7
3.4 Recommendation Engine Algorithms.....................................................8
3.5 Deployment Infrastructure......................................................................8
3.6 Design Flow Chart................................................................................10
3.7 AuraPulse Workflow Explanation.........................................................11

ix
Chapter 4 TECHNICAL REQUIRMENTS.............................................................15
4.1 Programming Languages........................................................................15
4.2 Libraries And Frameworks......................................................................15
4.3 APIs......................................................................................................16
4.4 Database Management...........................................................................16
4.5 Data Processing And Storage..................................................................16
4.6 Security And Compliance........................................................................17
4.7 User Experience And Accessibility..........................................................17
Chapter 5 PROJECT OVERVIEW AND DESCRIPTION...................................18
5.0 Home Page.............................................................................................18
5.1 Sign in....................................................................................................19
5.2 Log in....................................................................................................20
5.2.1 Login option................................................................................21
5.3 System Overview...................................................................................22
Chapter 6 FACE DETECTION................................................................................23
6.0 Recommendation Based on Different Moods............................................23
6.0.1 Neutral Mood..............................................................................23
6.0.2 Happy Mood...............................................................................24
6.0.3 Suprise Mood..............................................................................24
6.0.4 Sad Mood...................................................................................25
6.1 Playlist recommendation (Happy).............................................................26
6.2 Lifestyle recommendation (Happy)..........................................................26
6.3 Playlist recommendation (Neutral)...........................................................27
6.4 Lifestyle recommendation (Neutral).........................................................28
6.5 Playlist recommendation (Sad).................................................................29
6.6 Lifestyle recommendation (Sad)...............................................................30
6.7 Playlist recommendation (Fear)................................................................31
6.8 Lifestyle recommendation (Fear)..............................................................32
6.9 Playlist recommendation (Angry).............................................................33
6.10 Lifestyle recommendation (Angry).........................................................34
6.11 Playlist recommendation (Suprise).........................................................35
6.12 Lifestyle recommendation (Suprise)........................................................36

x
Chapter 7 CODE OVERVIEW.................................................................................37
7.0 Code Overview......................................................................................37
7.1 For Register...........................................................................................38
7.2 For Login...............................................................................................39
7.3 For Storage............................................................................................40
7.4 [Link]..........................................................................................41
7.5 [Link]...............................................................................................42
7.6 [Link]...................................................................................43
7.7 [Link].............................................................................................44
7.8 [Link]..................................................................................................45
Chapter 8 RESULTS & DISCUSSION....................................................................46
8.1 Results Real-Time Mood Detection.....................................................46
Chapter 9 CONCLUSION.........................................................................................50
Chapter 10 FUTURE SCOPE...................................................................................53

References...................................................................................................................58

Research Paper

xi
 LIST OF FIGURES

Sr. No. Title of the figure Page No

Flow Chart
3.6.1 10
Home Page
5.0 18
Sign in Page
5.1 19
Login Page
5.2 20
Login Option Page
5.2.1 21
Camera Permission
5.3 22
Neutral Mood
6.0.1 23
Happy Mood
6.0.2 24
Surprise Mood
6.0.3 24
Sad Mood
6.0.4 25
Playlist for Happy Mood
6.1 26
Lifestyle recommendation for Happy Mood
6.2 26
Playlist for Neutral Mood
6.3 27
Lifestyle recommendation for Neutral Mood
6.4 28
Playlist recommendation for Sad Mood
6.5 29
Lifestyle recommendation for Sad Mood
6.6 30
Playlist recommendation for Fear Mood
6.7 31
Lifestyle recommendation for Fear Mood
6.8 32
Playlist recommendation for Angry Mood
6.9 33
Lifestyle recommendation for Angry Mood
6.10 34
Playlist recommendation for Suprise Mood
6.11 35
Lifestyle recommendation for Suprise Mood
6.12 36
[Link]
7.0 37
[Link]
7.1 38
[Link]
7.2 39
[Link]
7.3 40
[Link]
7.4 41
[Link]
7.5 42
[Link]
7.6 43
[Link]
7.7 44

xii
 [Link]
7.8 45

xiii
 ABSTRACT

Aura Pulse is a cutting-edge AI-powered platform developed to provide real-time

emotional support, tackling the growing challenges of stress, anxiety, and burnout in
today’s fast-moving digital era. Utilizing advanced facial expression analysis, Aura
Pulse interprets visual cues to create a comprehensive emotional profile of the user.
This instant emotional evaluation enables the platform to deliver personalized
suggestions aligned with the user’s mood and mental state, promoting overall well-
being and emotional awareness.

One of the core features of Aura Pulse is music-based therapy, which suggests
mood-specific music to uplift spirits, reduce stress, and create a calming
atmosphere. Additionally, it offers comprehensive health tips, including
mindfulness exercises, meditation techniques, and self-care strategies that
promote mental resilience. The real-time analysis feature ensures continuous
monitoring of emotional states, dynamically adapting its recommendations based
on the user’s shifting moods. This makes it an intelligent and responsive
companion for mental health management.
To maximize accessibility, Aura Pulse is designed for cross-platform
compatibility, ensuring seamless use across web, mobile, and wearable devices.
Whether a user is at home, at work, or on the go, they can access their emotional
insights and receive tailored recommendations with ease. The platform also
provides data-driven insights, tracking emotional trends and patterns over days,
weeks, and months. This empowers users with valuable information about their
emotional well-being, enabling them to identify triggers and take proactive steps
toward mental health management.
Integration with third-party apps such as Spotify, Apple Health, and mental health
resources ensures a holistic and enriching experience. Through these integrations,
users can seamlessly connect their emotional well-being with their daily routines,
making mental health management more effortless and intuitive.
Beyond these features, Aura Pulse also incorporates stress management tools,
such as interactive breathing exercises and guided meditation sessions, to help
users regain a sense of calm and balance. The emotion-tracking dashboard
visually represents fluctuations in emotional states, helping users understand their
emotional patterns over time. This feature fosters self-awareness and encourages
users to take informed steps toward improving their mental health.

To cultivate a supportive and interactive space, Aura Pulse features a community

engagement platform where users can share their experiences, seek guidance, and
connect with others facing similar challenges. This sense of belonging promotes
empathy, mutual understanding, and collective well-being, ensuring that emotional
support remains both accessible and tailored to individual needs.

1
 CHAPTER-1
INTRODUCTION
1.1 Problem Identification

Mental health challenges, including stress, anxiety, and depression, are increasing
globally. Existing emotional support methods often lack personalization and
immediate accessibility, limiting their effectiveness in addressing real-time emotional
needs. According to the World Health Organization, over 300 million people globally
suffer from depression, with the numbers steadily rising. Technology-driven solutions
like Aura Pulse offer a promising avenue to address these challenges effectively.

1.2 Detailed Problem Definition

In today’s fast-paced digital era, individuals are increasingly vulnerable to mental

health issues such as stress, anxiety, and emotional burnout. While there is a growing
awareness of mental well-being, the tools and systems currently available for
emotional support remain inadequate in delivering timely, personalized, and context-
aware interventions.

Traditional mood assessment methods—such as self-reporting questionnaires or text-

based sentiment analyses—are inherently limited in scope. These approaches depend
heavily on user input and often fail to capture the nuanced, subconscious expressions
of emotional states. As a result, the feedback and recommendations provided tend to
be generic, delayed, or misaligned with the user’s actual emotional needs.

Moreover, most existing systems lack real-time adaptability and multimodal

integration. They do not leverage advanced technologies such as facial expression
recognition or voice tone analysis, which are critical in detecting subtle emotional
cues. Without these capabilities, current platforms are unable to offer dynamic
emotional insights or proactive mental health support.

Aura Pulse addresses these gaps by integrating AI-driven facial and voice emotion
recognition to deliver real-time emotional assessments. This allows for the generation
of a detailed emotional profile, which is then used to offer personalized therapeutic
recommendations, such as mood-based music playlists, mindfulness exercises, and

2
mental health resources. Unlike traditional platforms, Aura Pulse not only detects a
user’s emotional state with greater accuracy but also adapts its responses to changing
moods and emotional patterns.

The goal is to bridge the gap between emotional awareness and mental health action
by creating a seamless, intelligent, and accessible platform that enhances user well-
being. Aura Pulse aims to empower individuals by fostering emotional intelligence
and providing timely interventions, thereby promoting a healthier and more balanced
lifestyle through technology.

1.3 Current Market Survey

The market for emotion recognition technologies is growing rapidly, with applications
in entertainment, healthcare, and consumer engagement. Companies like Affectiva
and IBM Watson are developing systems for emotion detection, but their solutions
often focus on specific use cases such as marketing analytics. Aura Pulse
differentiates itself by combining emotion detection with actionable
recommendations, offering a holistic approach to emotional well-being.

Unlike existing platforms that primarily analyse emotional cues for commercial
benefit or isolated feedback, Aura Pulse is purpose-built for mental health
management. It goes beyond emotion recognition by integrating personalized
therapeutic interventions such as music-based mood enhancement, guided
mindfulness exercises, and community-driven emotional support. The platform’s real-
time adaptability and multi- modal input—through facial expressions and voice
analysis—make it uniquely positioned to address individual emotional needs
dynamically. This positions Aura Pulse not just as a diagnostic tool, but as an
intelligent emotional companion, aligning with the growing demand for proactive and
user-centric mental health solutions.

In addition, Aura Pulse leverages advanced AI to continuously learn from user

interactions, enabling more refined emotional insights and tailored support over time.
By using secure data practices and ethical AI frameworks, the platform ensures user
privacy while fostering trust—an essential element in mental health applications just a
technology, but a trusted ally in the journey toward emotional resilience.

3
1.4 Advantages Offered By Current Project

 Real-time emotion detection using facial and voice recognition.

Uses facial and voice recognition to instantly detect emotional states, allowing
for timely and relevant support.
 Personalized Therapeutic Recommendations:
Suggests music, mindfulness activities, and self-care practices tailored to the
user's current mood.
 Emotion Tracking and Insights:
Provides visual dashboards to help users monitor emotional patterns over time,
promoting self-awareness.
 Cross-Platform Compatibility:
Accessible on web, mobile, and wearable devices, ensuring support is
available anytime, anywhere.
 Ethical AI and data privacy at its core.
Aura Pulse adheres to strict data protection standards, ensuring user
information remains confidential. Transparent algorithms and consent-driven
data collection build trust and encourage consistent usage.
 Secure and Privacy-First Design:
Prioritizes user data protection with ethical AI implementation and secure
emotion data handling.

1.5 Current Challenges

 Ensuring high accuracy in real-time facial and voice emotion detection.

 Protecting user privacy and securing sensitive emotional data.

 Limited availability of diverse and standardized emotion datasets.

 Technical complexity in integrating multiple modules and APIs.

 Sustaining user engagement through intuitive and non-intrusive design.

4
 CHAPTER-2

LITERATURE SURVEY

2.1 Runfang Guo et al. (2024)

Explores the integration of emotion recognition with intelligent devices, focusing on

the challenges posed by dynamic environments. Their research highlights the
necessity of robust algorithms for real-time applications, ensuring accurate and
reliable emotion detection in varying conditions.

2.2 Kranthi Kiran et al. (2024)

Demonstrates the effectiveness of AI-driven emotion-based music

recommendations, achieving 80% accuracy using facial recognition. The study
underscores the potential of artificial intelligence in enhancing user experiences
by providing personalized music therapy based on detected emotions.

2.3 Padmavati K et al. (2024)

Utilizes convolutional neural networks (CNNs) for facial expression recognition,

leading to effective music therapy outcomes. Their research emphasizes the
critical role of machine learning in emotion detection and its applications in
mental health support through personalized music recommendations.

2.4 Moeez Rajjan et al. (2023)

Combines micro-expression detection with music recommendations, achieving a

62.1% accuracy rate. This study provides valuable insights into the complexities
of integrating multiple data streams for enhanced emotion analysis and highlights
the technical challenges in real-time emotion recognition systems.

2.5 M. Athavle et al. (2021)

Implements Bollywood music recommendations based on real-time facial

expressions using CNNs. Their research illustrates the cultural adaptability of

5
emotion-based systems, demonstrating how AI-driven music therapy can cater to
diverse user preferences across different regions.

2.6 Mrinmoy Bhaumik et al. (2021)

Explores the use of generative adversarial networks (GANs) for emotion-based

music recommendations. Their approach improves the diversity and relevance
of suggested tracks, enhancing the personalization and effectiveness of AI-driven
music therapy solutions. Additionally, the study investigates the use of GANs in
refining emotion classification accuracy by generating synthetic facial expression
data to enhance training models.

2.7. Metilda Florence et al. (2020)

Proposes a system capable of detecting user emotions with over 80% accuracy
through real-time facial expression recognition. This study provides a strong
foundation for real-world applications, demonstrating the feasibility of using AI-
powered emotion detection for mental health support and user engagement.
Furthermore, the research examines the scalability of the system across different
devices and explores potential integrations with wearable technology for
continuous emotion monitoring.

6
 CHAPTER-3

METHODOLOGY AND DESIGN

3.1 Data Sources

 Facial Emotion Datasets: The Face Expression Recognition dataset by Jonathan

Oheix (Kaggle) consists of approximately 56.5 MB of image data organized into two
main directories—train and validation—each containing seven subdirectories
corresponding to different facial expressions. This dataset supports classification tasks
by providing structured emotional image data for model training and evaluation.

3.2 Preprocessing Steps

 Noise Removal: Images often contain noise. Technique such as Gaussian filters for
images is used to clean the data.

 Normalization: This step involves adjusting the data to ensure consistency across
the dataset. For images, pixel values are normalized.

 Balancing Datasets: Ensuring that datasets are balanced in terms of the number of
samples per class is crucial for model accuracy. Techniques like oversampling,
undersampling, and synthetic data generation (using SMOTE) are employed.

 Standardization of Facial Images: Images are resized to a standard resolution of

48x48 pixels. This ensures uniformity in training data and reduces computational
complexity.

3.3 Machine Learning Models Facial Emotion Recognition

 Convolutional Neural Networks (CNNs): CNNs are well-suited for image

recognition tasks. A multi-layered CNN architecture is implemented to detect facial
emotions. Key layers include convolutional layers, pooling layers, and fully ix
connected layers.

7
 Data Augmentation: Techniques such as rotation, flipping, and zooming are used
to augment the dataset and prevent overfitting.

 Transfer Learning: Pre-trained models like VGG16 and ResNet are fine tuned on
the facial emotion dataset to leverage their learned features and improve performance.

3.4 Recommendation Engine Algorithms

 Collaborative Filtering: Both user-based and item-based collaborative filtering

techniques are used. User-based filtering suggests items based on users with similar
preferences, while item-based filtering recommends items similar to what the user has
previously liked.

 Matrix Factorization: Techniques such as Singular Value Decomposition (SVD)

are employed to decompose the user-item interaction matrix and uncover latent
features.

 Content-Based Filtering: This approach analyzes the content of items (e.g., song
features, articles) to recommend similar items. Features such as genre, tempo, and
mood descriptors for music, and topic keywords for mental health resources, are
considered.

 Hybrid Approaches: Combining collaborative and content-based filtering can

enhance recommendation accuracy. For instance, weighted hybrid methods or model
based hybrids like Factorization Machines are used.

 Continuous Learning: The recommendation engine is designed to learn from user

interactions continuously. Reinforcement learning techniques are applied to optimize
the recommendation policy based on user feedback and engagement metrics.

3.5 Deployment Infrastructure

 Cloud Platforms: The system is deployed on scalable cloud platforms such as

AWS or Google Cloud. Services like Amazon EC2, Lambda, and Google Kubernetes

8
Engine

9
(GKE) ensure high availability and scalability.

 Load Balancing: Load balancers distribute incoming traffic across multiple

instances to ensure system reliability and performance.

 Database Management: Managed database services like Amazon RDS and Cloud
SQL are used for efficient data storage and retrieval. User Interfaces

 Web Interface: Developed using [Link], the web interface offers a responsive
and interactive experience. Key features include real-time emotion detection, music
recommendations, and mental health resources.

 Component-Based Architecture: React components facilitate modular

development and ease of maintenance.

 State Management: Redux or Context API is used for managing application state
and ensuring a seamless user experience.

 Mobile Interface: Built using frameworks like React Native or Flutter, the mobile
interface provides a consistent experience across iOS and Android devices.

 Offline Capabilities: Enabling offline access ensures users can benefit from the
system even without an internet connection.

 Push Notifications: Implementing push notifications keeps users engaged by

providing timely recommendations and updates.

Security and Privacy

 Data Encryption: Ensuring data privacy and security is paramount. Techniques

such as SSL/TLS for data transmission and encryption at rest protect user
information.

 User Authentication: Implementing secure authentication mechanisms like

OAuth2.0 and multi-factor authentication (MFA) ensures only authorized access to
the system.

10
3.6 DESIGN: FLOW CHART

Fig. 3.6.1: Flow Chart

11
3.7 : AuraPulse Workflow Explanation of fig. 3.6.1

i. Overview

AuraPulse is an intelligent web-based platform designed to analyze user emotions and

offer personalized suggestions, such as music tracks and health tips, based on detected
emotional states. The application involves a structured and interactive process
beginning from user authentication (registration/login) to emotion detection and finally,
the delivery of emotion-based suggestions.

ii. User Authentication Flow

a. Registration Process
The registration module ensures new users can securely sign up to access AuraPulse’s
features.

Steps:

 Start Registration: Triggered when a user clicks on “Register” from the main
interface.

 Enter Details: Users input their credentials (e.g., email, password, name).

 Validate Details:
o The application checks if inputs follow proper format and if the email
is unique.

o If validation fails, the user is prompted to correct the errors.

 Registration Successful:
o On successful validation, the user data is stored in
Firebase Authentication.

o The user is redirected to the emotion detection module.

12
 b. Login Process

Returning users authenticate through the login flow.

Steps:
 Start Login: Initiated when a user selects the “Login” option.

 Enter Credentials: User provides registered email and password.

 Validate Credentials:
o The system verifies the entered information against Firebase Auth
records.
o If invalid, an error is shown, prompting re-entry.

 Login Successful:
o On successful verification, the user is allowed access to emotion
detection.

iii. Emotion Detection Module

Once authenticated, the next phase is emotion recognition using image processing and AI
models.

a. Detect Emotion

The system captures the user's image through a webcam or uploaded photo. This image is
then processed using the following:

 Face Detection:

o OpenCV’s Haar Cascade classifier is used to locate faces in the image.

 Preprocessing:

o The face area is cropped and resized to 48x48 pixels.

13
 o Converted to grayscale for compatibility with the trained CNN model.

 Prediction:

o A Keras-based pre-trained model (emotiondetector.h5) is loaded.

o It predicts one of the predefined emotions: angry, disgust, fear, happy,
neutral, sad, surprise.

b. Emotion Detected

Upon successful prediction, the system stores the detected emotion in the session context
and forwards the result to the suggestion engine.

iv. Suggestion System

The final module offers personalized suggestions based on the detected emotional state.

a. Suggest Music

Using a mapped folder structure or playlist database:

 Each emotion corresponds to a curated folder of mood-matching songs (e.g.,

“happy” maps to upbeat tracks).

 The system selects random or predefined songs from the relevant folder.

 These are returned as part of the suggestion response to the frontend.

b. Suggest Health Tips

In addition to music:
 Simple well-being suggestions are fetched based on emotion.

14
 o For example:
 Sad → “Take a walk outside”, “Talk to a friend”.
 Angry → “Practice deep breathing”, “Try meditation apps”.

 These tips are fetched from a static JSON or lightweight database.

c. Display Suggestions

Finally, both song links and health suggestions are presented in a UI component on the
screen. This allows users to either play music directly or read actionable advice.

15
 CHAPTER-4

TECHNICAL REQUIREMENTS

4.1 Programming Languages

 Python (Backend): Python is chosen for backend development due to its

simplicity, readability, and extensive support for machine learning and data
processing libraries. It allows for quick prototyping and development of robust,
scalable backend services.

 JavaScript (Frontend): JavaScript, along with frameworks like [Link], is used

for frontend development. Its versatility and widespread adoption make it ideal for
creating interactive and responsive user interfaces.

4.2 Libraries and Frameworks

 TensorFlow: TensorFlow is an open-source machine learning framework

developed by Google. It provides comprehensive tools for building and training
machine learning models, particularly neural networks. TensorFlow supports both
CPU and GPU computing, making it efficient for large-scale data processing and
model training.

 Keras: Keras is a high-level neural networks API written in Python, capable of

running on top of TensorFlow. It allows for easy and fast prototyping, and supports
convolutional and recurrent networks, making it suitable for both image and audio
data processing.

 OpenCV: OpenCV (Open Source Computer Vision Library) is used for real-time
computer vision tasks. It provides tools for image processing, including facial
recognition, object detection, and feature extraction. OpenCV is essential for
preprocessing facial images and detecting emotions in real-time.

16
4.3 APIs

 Google Vision API: The Google Vision API provides powerful image analysis
capabilities. It can detect and extract various features from images, such as facial
expressions, which can be used to enhance the accuracy of mood detection.

 Spotify API: The Spotify API allows for integration with the Spotify music
streaming service. It enables access to a vast library of music tracks and provides
functionalities for personalized music recommendations based on the detected
emotional state of the user.

4.4 Database Management

 MongoDB: A NoSQL database like MongoDB is chosen for its flexibility in

handling unstructured and semi-structured data. It is used to store user profiles,
emotional state data, music preferences, and interaction history.

 PostgreSQL: A relational database like PostgreSQL is used for structured data

management, ensuring data integrity and supporting complex queries. It is used for
storing metadata, user authentication data, and system logs.

4.5 Data Processing and Storage

 Apache Kafka: Apache Kafka is a distributed streaming platform used for building
real-time data pipelines. It ensures efficient data ingestion and processing, handling
large volumes of data from various sources like user interactions, emotion detection
models, and recommendation engines.

 Hadoop HDFS: Hadoop Distributed File System (HDFS) is used for storing large
datasets. It provides fault tolerance and high throughput access to data, making it
suitable for storing raw and processed data for machine learning models.

17
4.6 Security and Compliance

 Encryption: Data encryption both at rest and in transit is implemented using

protocols like SSL/TLS for data transmission and AES for storage.

 Authentication and Authorization: OAuth 2.0 is used for secure user

authentication, along with Multi-Factor Authentication (MFA) for enhanced security.
Role-based access control (RBAC) ensures that users have appropriate permissions.

 Compliance: Adherence to data protection regulations such as GDPR and CCPA is

crucial. Regular audits and compliance checks are conducted to ensure that the system
meets legal and ethical standards.

4.7 User Experience and Accessibility

 Responsive Design: Ensuring that the user interface is responsive and adapts to
various screen sizes is essential for a seamless experience across devices.

 Accessibility Features: Implementing accessibility features such as screen reader

support, keyboard navigation, and high-contrast modes ensures that the system is
usable by individuals with disabilities.

 User Feedback Mechanisms: Incorporating user feedback mechanisms such as

surveys, ratings, and feedback forms helps in continuously improving the system
based on user input.

18
 CHAPTER-5

PROJECT OVERVIEW AND DESCRIPTION

5.0 Home Page

The image is a modern and visually appealing login/sign-up interface for a platform
called AuraPulse, which offers mood-based music exploration. The left side features
a vibrant blue background with an illustration of a person sitting beside a rocket,
symbolizing an emotional or musical journey. A call-to-action encourages new users
to sign up with the text: “Discover music that matches your mood.” On the right side,
existing users can log in using their email and password or opt for a Google sign-in.
The design is clean, minimalist, and user-friendly, clearly targeted at a tech-savvy,
emotionally attuned audience.

Figure 5.1: Home page

19
5.1 Sign in

This image displays the Google sign-in process for the AuraPulse platform, a mood-
based music service. The original login/signup interface remains visible in the
background, showcasing the clean UI split between a sign-up prompt with an
illustration on the left and login inputs on the right. Overlaid is a Google OAuth
popup window prompting the user to choose a Google account to continue to
aurapulse- [Link].

Two Google account options are listed, allowing for seamless one-click
authentication. This illustrates how AuraPulse integrates third-party login to enhance
user convenience and streamline the onboarding process.

a) Registration / Account Creation Page: The Aura Pulse registration page

provides a seamless and user-friendly interface for account creation. Users can sign
up using their email and password or opt for Google authentication for a quicker
process.

Figure 5.1: Sign in Page

20
5.2 Login

This image shows a continuation of the Google sign-in and login process for
AuraPulse, a mood-based music discovery platform. The background retains the sleek
dual-pane interface, with the left side featuring an illustration of a person beside a
rocket and a "Sign Up" option, while the right side provides login fields and a button
for Google sign-in. The foreground displays the Google account selection window,
offering the user two account
options—princeraja01082002@[Link] and princekumar01082002@[Link]—
to authenticate via OAuth through [Link].

This highlights a user-friendly integration of Google login for seamless access.

Figure 5.2: Login Page

21
5.2.1 Login option

This image is shows us the option like we can login or signin via google email id in
fig 5.2.1,

Figure 5.2.1: Login option page

22
5.3 System Overview

The user interface is designed to prioritise simplicity, clarity, and accessibility. Built
using ReactJS, it allows users to easily navigate through symptom selection, view
prediction results, and get matched with doctors. The figures 4., showcase. the main
pages and the functionalities of the system frontend, which highlight the ease of
navigation and interaction.

a) Camera permission

Figure 5.3: Camera Permission

23
 CHAPTER-6

FACE DETECTION

6.0 : Recommendation Based on Different Moods

6.0.1 : Neutral Mood

The image displays in fig.5.8.4 below the user interface of Aura Pulse, a web-based
mood detection system using facial recognition. The application identifies the user's
mood in real time, as shown with a detected "neutral" emotion. It features a "Detect
Mood" button and provides personalized recommendations, including "Vibe with a
Playlist" for mood-based music suggestions and "Lifestyle Recommendation" for
well- being tips. The interface is designed for a seamless user experience with secure
login/logout functionality.

Figure 6.0.1: Neutral Mood

24
6.0.2 : Happy Mode
The detected mood in this instance is "happy", influencing personalized
recommendations such as music playlists and lifestyle recommendation shown in
fig.6.0.2

Figure 6.0.2 Happy Mood

6.0.3 : Surprise Mood

The detected mood is "surprise", which influences the system’s personalized
recommendations, such as music playlists and lifestyle recommendation shown in
fig.6.0.3

Figure 6.0.3: Surprise Mood

25
6.0.4 : Sad Mood

In this next case, the detected mood is "sad", allowing the system to provide tailored
music playlists and lifestyle recommendations to enhance emotional well-being
shown in fig.6.0.4

Figure 6.0.4: Sad Mood

26
6.1 Playlist recommendation (HAPPY)

The Aura Pulse web application detects the user's facial expression and identifies the
mood as "happy." Based on this mood, it recommends an upbeat and energetic
playlist featuring Bollywood party and celebration songs. The interface allows the
user to play songs directly and enhance their joyful state. This personalized
recommendation system improves user experience by aligning music with emotions
shown in fig.6.1.

Figure 6.1: Playlist for Happy Mood

6.2 Lifestyle recommendation (HAPPY)

The Aura Pulse web application detects the user's happy mood using facial
recognition and provides lifestyle recommendations to maintain positivity. The
suggestions encourage spreading joy, practicing gratitude, and celebrating small
achievements. This feature helps users reinforce their positive emotions and lead a
more fulfilling life shown in fig.6.2

Figure 6.2: Lifestyle recommendation for Happy Mood

27
6.3 Playlist recommendation (NEUTRAL)

This image shows the interface of the Mood Detection Hub feature within the
AuraPulse web application.
On the left side, the application uses facial recognition to analyze the user's
expression and identifies the mood as "neutral."

A "Detect Mood" button allows users to initiate the analysis process. On the right
side, a recommended playlist is displayed, tailored to the detected neutral mood,
featuring calm and balanced tracks such as "Hasi Ban Gaye," "Ilahi," and
"Khairiyat." Playback controls are provided next to each song for immediate
listening. A "Go Back" and "Logout" button offer user navigation and session control,
demonstrating a user-centric design for personalized music experiences.

Figure 6.3: Playlist for Neutral Mood

28
6.4 Lifestyle recommendation (NEUTRAL)

This image displays the Mood Detection Hub interface of the AuraPulse web
application, focusing on lifestyle recommendations based on the detected emotional
state. The left panel features a facial recognition section where the system has
identified the user's mood as "neutral." Users can trigger mood analysis using the
"Detect Mood" button. On the right panel, the application offers practical Lifestyle
Recommendations aimed at maintaining emotional well-being, including tips such as
balancing work and rest, staying hydrated, maintaining a regular sleep schedule, and
engaging in uplifting activities. A "Go Back" button provides navigation, while a
"Logout" button is located at the bottom center for secure session management.

Figure 6.4: Lifestyle recommendation for Neutral Mood

29
6.5 Playlist recommendation (SAD)

This image presents the Mood Detection Hub of the AuraPulse web application,
where the system utilizes facial recognition to determine the user's emotional state. In
this instance, the detected mood is "sad," as shown on the left panel after the user
interacts with the "Detect Mood" button. Based on this mood, the right panel displays
a curated Recommended Playlist containing emotionally resonant tracks such as
"Agar Tum Saath Ho," "Channa Mereya," and "Tera Yaar Hoon Main." These
selections aim to complement or soothe the user’s emotional state. Navigation options
like "Go Back" and "Logout" ensure a user-friendly and secure experience.

Figure 6.5 : Playlist recommendation for Sad Mood

30
6.6 Lifestyle recommendation (SAD)

This image displays the Mood Detection Hub interface of the AuraPulse web
application, focusing on lifestyle recommendations based on the detected emotional
state. The left panel features a facial recognition section where the system has
identified the user's mood as "Sad." Users can trigger mood analysis using the "Detect
Mood" button. On the right panel, the application offers practical Lifestyle
Recommendations aimed at maintaining emotional well-being, including tips such as
balancing work and rest, staying hydrated, maintaining a regular sleep schedule, and
engaging in uplifting activities. A "Go Back" button provides navigation, while a
"Logout" button is located at the bottom center for secure session management.

Figure 6.6: Lifestyle recommendation for Sad Mood

31
6.7 Playlist recommendation (Fear)

This image showcases the Mood Detection Hub of the AuraPulse application, where
the facial recognition system has identified the user's emotional state as "fear." Based
on this analysis, the system generates a Recommended Playlist intended to soothe or
empower the user, featuring tracks like "Khoon Chala" from Rang De Basanti, "Phir
Le Aya Dil (Reprise)," and "Tu Kuja" from Highway. These emotionally evocative
songs are selected to resonate with the user's mood, helping provide comfort or
motivation. The interface maintains clarity with an accessible "Detect Mood" button
and intuitive navigation features like "Go Back" and "Logout" for a smooth user
experience.

Figure 6.7: Playlist recommendation for Fear Mood

32
6.8 Lifestyle recommendation (FEAR)

This screenshot illustrates the Mood Detection Hub of the AuraPulse application,
where facial recognition has identified the user's mood as "fear." In response, the
system provides tailored Lifestyle Recommendations on the right side of the
interface. These include advice such as gradually facing fears, using grounding
techniques like the 5-4-3-2-1 method, and engaging in calming hobbies like reading,
art, or meditation. The goal is to help users manage and reduce anxiety through
constructive actions. The interface is clean, offering intuitive navigation with "Detect
Mood," "Go Back," and "Logout" buttons for ease of use.

Figure 6.8 : Lifestyle recommendation for Fear Mood

33
6.9 Playlist recommendation (ANGRY)

This screenshot captures the Mood Detection Hub interface of the AuraPulse
application, where the user's facial expression has been recognized as "angry." In
response, the application recommends a high-energy and motivational playlist
featuring tracks like Apna Time Aayega and Azadi from Gully Boy, and Challa from
URI. These empowering songs are curated to help the user channel their anger into
motivation and action. The interface remains consistent with a clear layout, allowing
users to interact easily via the "Detect Mood," "Go Back," and "Logout" buttons for a
smooth experience.

Figure 6.9: Playlist recommendation for Angry Mood

34
6.10 Lifestyle recommendation (ANGRY)

This screenshot shows the Mood Detection Hub detecting the user's facial expression
as "angry." Instead of music, the interface provides Lifestyle Recommendations
tailored to help manage anger. The suggestions include mindfulness exercises, deep
breathing techniques, and engaging in physical activities like walking or exercise.
These tips aim to help the user calm down, regain control, and process emotions
constructively. The consistent interface layout ensures usability with options to detect
mood again or return via the "Go Back" button.

Figure 6.10 : Lifestyle recommendation for Angry Mood

35
6.11 : Playlist recommendation (SURPRISE)

The Mood Detection Hub web application recognizes the user's facial expression as
"surprised" and tailors its response accordingly. For this mood, the system may
recommend playlists or lifestyle suggestions that promote grounding and help the user
process unexpected emotions. This includes engaging in light activities, journaling
thoughts, or taking a moment to pause and reflect. The interface is designed to support
users in moments of heightened alertness or unexpected reactions, offering tools to
bring calm and clarity.

Figure 6.11: Playlist recommendation for Surprise Mood

36
6.12: Lifestyle recommendation (SURPRISE)
 Based on the detected mood (surprise), the application provides three lifestyle
suggestions:

o Adapt to surprises positively and keep an open mind.

o Turn unexpected events into learning opportunities.
o Maintain flexibility in your daily life to handle surprises smoothly.

Figure 6.12 : Lifestyle recommendation for Suprise Mood

37
 CHAPTER-7

CODE OVERVIEW

7.0 : Code Overview

The image shows a React component ([Link]) implementing route protection using
Firebase authentication. It imports essential modules like react-router-dom for
routing, react hooks (useEffect, useState), and Firebase's auth object for managing
user state. The ProtectedRoute component checks if a user is authenticated by
listening to auth state changes using [Link], and shows a loading
message while checking. If a user is authenticated, it renders the child components;
otherwise, it redirects to the home route using <Navigate to="/" />.

Figure 7.0: [Link]

38
7.1 : FOR REGISTER

This image displays a React component ([Link]) that handles user registration
using Firebase Authentication. It includes several useState hooks to manage form
inputs, validation errors, and UI states like loading and password visibility. Firebase
functions createUserWithEmailAndPassword and signInWithPopup are imported to
support email/password registration and Google sign-in. Additionally, the component
uses react-toastify for toast notifications and react-router-dom's useNavigate for
navigation after successful registration. The form validation logic begins with the
validateForm function to ensure proper input before submission.

Figure 7.1: [Link]

39
7.2 : FOR LOGIN

This React component implements a login screen for an application using Firebase
Authentication. It supports both email/password login and Google OAuth login via
Firebase. The component manages form inputs using React useState hooks and
provides validation for email format and password length. Upon successful login,
users are redirected to the /emotion-detection route, while errors trigger toasts and are
displayed to the user. It also includes UI features like toggling password visibility and
displaying loading indicators. The react-toastify library is used for showing success or
error messages, and react-icons provides icons for UI elements.

Figure 7.2 : [Link]

40
7.3 : FOR STORAGE

This code initializes a Firebase app using a specific configuration and sets up Firebase
Authentication and Firestore database services. It imports necessary modules from
Firebase, creates an app instance with initializeApp, and configures authentication
with multiple third-party providers such as Google, Facebook, GitHub, and Twitter.
Firestore is also initialized to enable cloud-based data storage and retrieval. Finally,
the app instance, authentication service, Firestore database, and all auth providers are
exported for use throughout the application, enabling features like user login and data
handling.

Figure 7.3 : [Link]

41
7.4 : [Link]

The [Link] file shown in the image is the configuration file for a frontend
project built with Vite and React. It defines the project metadata (like name, version,
and module type), scripts to run the app (dev, build, lint, and preview), and the
project's dependencies. These include core libraries like react, react-dom, and
firebase, as well as additional tools such as:

axios: for making HTTP requests,

[Link]: for face recognition/detection features,
react-router-dom: for routing between
components/pages, react-toastify: for toast notifications,
react-icons and react-webcam: for icons and webcam integration.

This file ensures that all necessary packages are installed and provides scripts for
development and production builds using Vite.

Figure 7.4 : [Link]

42
7.5 : [Link]

This Flask backend application powers an emotion detection system that receives
image data (typically from a webcam via a React frontend), detects human faces using
OpenCV’s Haar Cascade classifier, and classifies the detected emotions using a pre-
trained Keras model loaded from JSON and H5 files. Once an emotion is identified
(like "happy" or "sad"), it fetches mood-specific songs from corresponding
directories, enabling dynamic music suggestions based on user emotions. The server
also exposes multiple API endpoints (/api/data, /api/recognize, and
/api/songs/<mood>) and uses Flask-CORS to allow secure communication with the
React frontend.

Figure 7.5 : [Link]

43
7.6 : [Link]

This image shows the [Link] file of a React project named "frontend",
version 0.1.0. It contains metadata about the project's dependencies and ensures
consistent installs across environments. The dependencies section lists packages like
react, react-dom, firebase, axios, react-router-dom, react-toastify, and react-webcam,
which support features like routing, authentication, notifications, and webcam
integration. It also includes development tools such as @testing-library for testing and
fontawesome for icons. This file helps lock exact versions of dependencies to avoid
compatibility issues during builds or deployments.

Figure 7.6 : [Link]

44
7.7 : [Link]

This image shows the [Link] file located in the public directory of a React
project. It serves as the main HTML template for the React app created using Create
React App. The file includes metadata like charset, viewport settings, theme color,
and a description. It references assets such as the favicon, Apple touch icon, and a
manifest file using the %PUBLIC_URL% placeholder, which gets replaced with the
correct path during the build process. The page title is set to "React App", and there's
a <noscript> tag prompting users to enable JavaScript, as it's required for the app to
function.

 Template for React App: It serves as the base HTML file for the React app
created with Create React App, into which the React components are injected.
 Asset References: It uses %PUBLIC_URL% to reference public assets like
the favicon, logo, and manifest, which get resolved correctly during the build.
 Metadata Setup: Includes important meta tags for viewport settings, theme
color, and app description to enhance mobile and SEO performance.

Figure 7.7: [Link]

45
7.8 : [Link]

The [Link] file defines the foundational styling for the React application by
importing the "Poppins" font from Google Fonts and applying a global reset to
margin, padding, and box-sizing.

It sets a consistent font for the body and input elements, styles the main container to
occupy the full viewport height with a white background and hidden overflow, and
uses absolute positioning in the .forms-container class to ensure form elements are
properly layered and fully responsive within the layout.

Figure 7.8 : [Link]

46
 CHAPTER-8

RESULT ANALYSIS

8.1 Results Real-Time Mood Detection

 Accuracy: The emotion detection models achieved an accuracy of 82% of facial

emotion recognition. This high accuracy indicates that the models effectively identify
users' emotional states.

 Latency: The average latency for real-time mood detection was found to be under
200 milliseconds, ensuring prompt responses and seamless user experience.

 Confusion Matrix: Analysis of the confusion matrix revealed that the models
performed exceptionally well in distinguishing between common emotions such as
happiness, sadness, and anger. However, subtle emotions like surprise and fear had
slightly lower recognition rates.

a. Personalized Music Recommendations

 User Engagement: Users reported an increase in engagement and satisfaction with

personalized music recommendations. Surveys indicated that 85% of users felt that
the recommended music matched their emotional state.

 Continuous Learning: The system's ability to learn from user interactions led to
continuous improvements in recommendation accuracy over time. Monthly
evaluations showed a gradual increase in user satisfaction scores.

b. Mental Health Support

 Resource Utilization: Users frequently accessed mental health resources, with a

significant number of sessions devoted to mindfulness exercises, stress reduction
techniques, and self-reflection prompts.

47
 Impact on Well-being: Preliminary surveys indicated that users experienced a
positive impact on their well-being, reporting reduced stress levels and improved
emotional regulation after using the system's mental health support features.

c. User-Friendly Interface

 Usability Scores: The user interface received high usability . This indicates that
users found the interface intuitive, engaging, and easy to navigate.

 Accessibility: Implementing accessibility features, such as screen reader support and

high-contrast modes, ensured that users with disabilities could effectively use the
system. Feedback from users with disabilities was overwhelmingly positive.

d. Discussion Real-Time Mood Detection

 Strengths: The real-time mood detection module demonstrated high accuracy and
low latency, making it reliable for real-time applications. The use of advanced
machine learning models, such as CNNs and RNNs, significantly contributed to the
system's performance.

 Challenges: Despite the high accuracy, there were challenges in recognizing subtle
emotions, which may require further refinement of the models. Additionally, ensuring
consistent performance across diverse demographics remains an ongoing goal.

 Future Improvements: Future work could focus on enhancing the recognition of

subtle emotions and improving the system's robustness across different demographic
groups. Incorporating additional data sources, such as physiological signals (e.g.,
heart rate), could further enhance accuracy.

e. Personalized Music Recommendations

 Effectiveness: The personalized music recommendation system effectively

increased user engagement and satisfaction. The hybrid approach, combining
collaborative and

48
content-based filtering, proved to be highly effective in curating relevant music tracks.

 Challenges: One challenge was managing the diversity of user preferences, as music
tastes can vary widely. Additionally, the cold start problem for new users required
innovative solutions to provide relevant recommendations from the outset.

 Future Improvements: Enhancing the system's ability to handle the cold start
problem and incorporating more sophisticated user profiling techniques could further
improve recommendation accuracy. Exploring the integration of social network data
and contextual information could also provide richer recommendations.

f. Mental Health Support

 Impact: The mental health support features had a positive impact on users' well
being, demonstrating the value of integrating mental health resources into the system.
Users appreciated the availability of well-researched advice and exercises.

 Challenges: Ensuring the accuracy and relevance of mental health content was a
challenge, as it required continuous updates and validation from mental health
professionals.

 Future Improvements: Future work could focus on expanding the range of mental
health resources and incorporating real-time feedback mechanisms to tailor support to
individual users' needs. Collaborating with mental health experts to ensure the quality
and accuracy of content is also essential.

g. User-Friendly Interface

 Strengths: The user interface was well-received, with high usability scores
indicating that it was intuitive and engaging. The inclusion of customization and
accessibility features ensured a positive experience for a diverse user base.

49
 Challenges: Continuous user feedback is crucial to maintaining and improving the
interface. Addressing diverse user preferences and accessibility needs requires
ongoing effort and iteration.

 Future Improvements: Regular user testing and feedback collection will help in
refining the interface. Implementing advanced personalization features, such as
adaptive interfaces that adjust based on user behavior, could further enhance the user
experience.

50
 CHAPTER-9
CONCLUSION

9.1 Conclusion

Aura Pulse has demonstrated significant potential in enhancing users' emotional well-
being through advanced real-time mood detection and personalized recommendations.
By leveraging state-of-the-art machine learning techniques, such as Convolutional
Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), the system
effectively detects emotions from both facial expressions and voice tones.

The integration of these technologies allows Aura Pulse to provide a tailored user
experience that adapts to individual emotional states. The personalized music
recommendation engine has shown to be highly effective, with users reporting
increased engagement and satisfaction. The inclusion of mental health support
features further underscores the system's holistic approach to improving user well-
being, offering valuable resources and exercises to manage stress and enhance
emotional regulation.

The user-friendly interface, designed with accessibility and customization in mind, has
received positive feedback, ensuring that Aura Pulse is inclusive and easy to navigate
for a diverse user base. The robust security measures and compliance with data
protection regulations emphasize the system's commitment to user privacy and data
security.

9.2 Future Prospects

Looking ahead, there are several exciting prospects for the development and enhancement
of Aura Pulse:

1. Advanced Emotion Detection

 Multimodal Emotion Recognition: Future iterations of Aura Pulse can explore

multimodal emotion recognition, combining facial expressions, voice tones,

51
physiological signals (e.g., heart rate), and text analysis (e.g., sentiment analysis from user
messages) to achieve a more comprehensive understanding of user emotions.

 Improved Subtle Emotion Recognition: Enhancing the models to better recognize

subtle and complex emotions, such as nostalgia or confusion, will further refine the
system's accuracy and effectiveness.

2. Enhanced Personalization

 Context-Aware Recommendations: Incorporating contextual information, such as

time of day, location, and user activities, can lead to even more relevant and timely
xxii recommendations.

 Social Network Integration: Integrating social network data can provide deeper
insights into user preferences and social interactions, enabling more personalized and
socially aware recommendations.

3. Expanded Mental Health Support

 Real-Time Feedback Mechanisms: Implementing real-time feedback mechanisms,

such as mood tracking and journaling features, can help users gain better self
awareness and actively manage their emotional states.

 Collaborations with Mental Health Professionals: Partnering with mental health

experts can ensure that the resources and advice provided by Aura Pulse are accurate,
up-to-date, and comprehensive.

4. Technological Advancements

 AI and Machine Learning Innovations: Staying abreast of the latest advancements in

AI and machine learning will enable continuous improvement of the system's
capabilities. This includes exploring new architectures, algorithms, and techniques.

52
 Scalability and Performance: Enhancing the scalability and performance of Aura
Pulse to handle a growing user base and increasing data volumes will be essential.
This involves optimizing infrastructure, load balancing, and distributed computing
strategies.

5. User Experience Enhancements

 Adaptive User Interfaces: Developing adaptive user interfaces that adjust based on
user behavior and preferences can further enhance the user experience.

 Gamification and Engagement: Incorporating gamification elements, such as

achievements and rewards, can increase user engagement and motivation to use the
system regularly.

6. Ethical and Social Considerations

 Bias Mitigation: Continuously monitoring and addressing biases in the models and
data to ensure fairness and inclusivity for all users.

 Ethical AI Practices: Adhering to ethical AI practices, including transparency,

accountability, and user empowerment, will be crucial in maintaining user trust and
ensuring the responsible use of technology.

53
 CHAPTER-10
FUTURE SCOPE

10.1 Future Scope

The future scope of Aura Pulse extends beyond its current capabilities, aiming to
evolve into a highly adaptive and intelligent emotional well-being assistant. The
next iterations of the application will focus on integrating advanced technologies
and user-centric features that enhance emotional insight, responsiveness, and
accessibility.

One of the most promising directions is the implementation of voice analysis for
emotion detection, which leverages tone, pitch, and speech patterns to gauge
emotional states. This will complement facial recognition, enabling a more
holistic and context-sensitive understanding of a user's mood. Alongside this,
multimodal emotion recognition will be a central enhancement—combining
inputs such as facial expressions, vocal attributes, and physiological data (like
heart rate or skin conductance) collected via wearable devices to achieve higher
accuracy in emotional interpretation.

To make emotional well-being tracking more meaningful, Aura Pulse will integrate
real-time feedback systems that provide instant responses or suggestions based
on the user’s detected emotional state. This could include grounding exercises,
calming audio, or personalized motivational content. Furthermore, mood
journaling and visualization tools will help users reflect on emotional patterns
over time, enabling greater self-awareness and proactive mental health
management.

Context-aware recommendations will become increasingly sophisticated, tailoring

content and interventions not only based on current mood but also on behavioral
history, time of day, and social interactions. Integration with social networks or
community support platforms can provide emotional validation and peer
support, fostering a sense of connection and belonging. However, privacy and

54
ethical considerations will be strictly upheld to protect sensitive user data.

The collaboration with mental health professionals and therapists will also be
prioritized, ensuring that all advice and tools offered align with psychological best
practices. Through such partnerships, the app can incorporate cognitive-
behavioral techniques, mindfulness practices, and crisis support functionalities.

Another major enhancement will be the deployment of AI-driven adaptive

learning algorithms. These systems will continuously learn from individual user
behavior and feedback to refine emotional predictions and personalize
interventions with increased precision. Over time, Aura Pulse will become more
attuned to subtle behavioral cues and user preferences.

The integration with wearable technology (such as smartwatches or fitness

bands) is also on the roadmap. This allows continuous monitoring of biometric
data, enabling Aura Pulse to provide proactive emotional insights and wellness
suggestions before issues escalate. These interventions might include breathing
exercises when stress is detected or nudges to take breaks when signs of burnout
emerge.

To ensure global reach, Aura Pulse will aim to support multi-language

interfaces and cultural sensitivity models, making the platform inclusive and
effective across diverse populations. Emotional expressions can vary across
cultures, so adapting recognition models accordingly is crucial to ensuring
accuracy and respectfulness.

Lastly, the platform could evolve into a comprehensive wellness ecosystem,

integrating modules for nutrition tracking, sleep analysis, meditation, and
productivity to create a 360-degree support system for user well-being. With
advancements in neurotechnology, future versions might even explore non-
invasive neural feedback mechanisms for emotion detection and enhancement.
In summary, Aura Pulse is envisioned to become a state-of-the-art emotional
companion—intelligent, adaptive, and empathetic. By incorporating multimodal
AI, real-time monitoring, cultural adaptability, and professional collaboration, it

55
aims to redefine how technology can support mental health and emotional
resilience in an increasingly complex world.

10.2 : Key Points:

 Integration of voice analysis and multimodal emotion recognition.

 Real-time emotional feedback, journaling, and mood tracking features.
 AI-driven adaptive learning for smarter emotional recommendations.
 Wearable device integration for biometric emotion detection.
 Cultural and linguistic adaptability for global accessibility.
 Collaboration with mental health professionals for credibility.
 Evolution into a comprehensive well-being ecosystem.

Detailed Future Enhancements:

1. Voice Analysis & Emotion Detection

 Analyze speech tone, pitch, and pacing to detect emotional states.
 Enhance emotion detection when visual input is limited or unavailable.

2. Multimodal Emotion Recognition

 Combine:
o Facial expressions
o Voice modulation
o Physiological signals (e.g., heart rate, skin temperature)
 Provide a more accurate and holistic view of user emotion.

3. Real-time Feedback & Mood Tracking

 Offer immediate, context-aware suggestions:
o Breathing exercises
o Music or affirmations
o Motivational prompts
 Enable users to track mood patterns over time using visual dashboards.

56
4. Context-Aware Recommendations
 Use behavioral history, time of day, and user habits to tailor:
o Content
o Daily check-ins
o Mental health tips

5. Social Network Integration

 Provide options for:
o Peer support groups
o Sharing mood status (optional and secure)
o Receiving encouragement or tips from friends

6. Professional Mental Health Collaboration

 Collaborate with licensed psychologists to:
o Validate emotional analysis methods
o Provide structured therapeutic content
o Add guided mental health exercises (CBT, mindfulness)

7. AI-Driven Adaptive Learning

 Implement personalized AI that:
o Learns from user behavior over time
o Optimizes recommendations based on feedback
o Becomes increasingly accurate and helpful

8. Wearable Device Integration

 Link with smartwatches or bands to track:
o Heart rate variability
o Sleep quality
o Stress levels
 Enable proactive intervention (before user realizes emotional shift).

9. Language and Cultural Expansion

 Support multiple languages and regional behaviors.
 Adapt models to interpret emotions as expressed differently across cultures.

57
10. Comprehensive Wellness Ecosystem

1. Incorporate:
o Nutrition tracking
o Sleep analysis
o Guided meditation
o Productivity tools

2. Offer all-in-one support for mental, emotional, and physical health.

58
 REFERENCES

1. Ashwini Jadhav, Nikita Bhaise, Mihir Narwade, Ruchita Phalke, Yash Talele,
“Emotion based Music Recommendation System using CNN”, International
Journal of Research Publication and Reviews, Vol. 4, pp. 3288-3292, Issue
May 2023.

2. Athavle, M., Mudale, D., Shrivastav, U., & Gupta, M.(2021). "Music
recommendation based on face emotion recognition. "Journal of Informatics
Electrical and Electronics Engineering, Volume02, Issues:02.

3. Bhaumik M., Attah, P. U., & Javed, F. (2021). "Emotion Integrated Music
Recommendation System Using Generative Adversarial Networks." SMU
Data Science Review, Vol. 5, No. 3, Article 4. Available at: SMU Data
Science Review.

4. Florence,S.M., & Uma, M. (2020). “Emotional Detection and Music

Recommendation System based on User Facial Expression.” IOP Conf. Series:
Materials Science and Engineering, 912 (2020)

5. G. Tirumala, M. Niharika, S. Shailu, M. Manaswini, G. Venga Vinodini,

“Music Recommendation System based on Emotions using CNN”, Journal of
Engineering Sciences, Vol. 14, Issue 2023.

6. Guo, R., Guo, H., Wang, L., Chen, M., Yang, D., & Li, B. (2024).
“Development and application of emotion recognition technology - a systematic
literature review.” BMC Psychology. Available at: BMC Psychology .

7. Kranthi Kiran, B., Shanthan, P., Ram, K. S., & Usha, K. (2024). “Emotion
Based Music Recommendation System using VGG16-CNN Architecture.”

59
International Journal for Research in Applied Science & Engineering
Technology (IJRASET), Volume 12 Issue VI.

8. Maduri Athavle, Deepali Mudale, Upasana Shrivastav, Megha Gupta, “Music

Recommendation based on Face Emotion Recognition”, ICAI, Vol. 02, Sno.
018, pp. 1-11, Issue 2021.

9. Padmavati, K., Amruth Kumar, D. K., Ganavi, S. M., Keerthi, K. C., &
Sourabha, Y. N. (2024). “SoulSound: Enhancing Musical Therapy through
Facial Expression Recognition with Machine Learning.” International
Research Journal of Engineering and Technology (IRJET).

10. Phaneendra, Madhusmitha Muduli, Siri Lakshmi Reddy, R. Veenasree,

“EMUSE – An Emotion based Music Recommendation System”, IRJMETS,
Volume 04, Issue 05 May 2022.

11. Rajjan, M., Deore, P., Mohite, Y., & Desai, Y. (2023). “Harmonic Fusion: AI-
Driven Music Personalization via Emotion-Enhanced Facial Expression
Recognition.” International Journal of Innovative Science and Research
Technology, Volume 8, Issue 12.

60
AURA PULSE : AN AI-POWERED SYSTEM FOR REAL- TIME
EMOTIONAL SUPPORT AND PERSONALIZED
RECOMMENDATIONS

Nesnsi Varshney1, Prince Kumar2, Shivang Rastogi3, Vivek Chaudhary4, Dr. Monika Nagar5
1234
Student, Deparment of Computer Science and Design, IMS Engineering College
Ghaziabad,India
5
Assistant Professor, Deparment of Computer Science and Design, IMS Engineering College
Ghaziabad,India

ABSTRACT

Aura Pulse is a cutting-edge AI-powered platform developed to provide real-time emotional support,
tackling the growing challenges of stress, anxiety, and burnout in today’s fast-moving digital era.
Utilizing advanced facial expression analysis, Aura Pulse interprets visual cues to create a
comprehensive emotional profile of the user. This instant emotional evaluation enables the platform to
deliver personalized suggestions aligned with the user’s mood and mental state, promoting overall
well-being and emotional awareness.

One of the core features of Aura Pulse is music-based therapy, which suggests mood-specific
music to uplift spirits, reduce stress, and create a calming atmosphere. Additionally, it offers
comprehensive health tips, including mindfulness exercises, meditation techniques, and self-care
strategies that promote mental resilience. The real-time analysis feature ensures continuous
monitoring of emotional states, dynamically adapting its recommendations based on the user’s
shifting moods. This makes it an intelligent and responsive companion for mental health
management.
To maximize accessibility, Aura Pulse is designed for cross-platform compatibility, ensuring
seamless use across web, mobile, and wearable devices. Whether a user is at home, at work, or on
the go, they can access their emotional insights and receive tailored recommendations with ease.
The platform also provides data-driven insights, tracking emotional trends and patterns over days,
weeks, and months. This empowers users with valuable information about their emotional well-
being, enabling them to identify triggers and take proactive steps toward mental health management.
Integration with third-party apps such as Spotify, Apple Health, and mental health resources ensures
a holistic and enriching experience. Through these integrations, users can seamlessly connect their
emotional well-being with their daily routines, making mental health management more effortless
and intuitive.
Beyond these features, Aura Pulse also incorporates stress management tools, such as interactive
breathing exercises and guided meditation sessions, to help users regain a sense of calm and balance.
The emotion-tracking dashboard visually represents fluctuations in emotional states, helping
users understand their emotional patterns over time. This feature fosters self-awareness and
encourages users to take informed steps toward improving their mental health.

To cultivate a supportive and interactive space, Aura Pulse features a community engagement platform
where users can share their experiences, seek guidance, and connect with others facing similar
challenges. This sense of belonging promotes empathy, mutual understanding, and collective well-
being, ensuring that emotional support remains both accessible and tailored to individual needs.

Keywords: AI-driven Platform, Music-based therapy, Stress management, Emotion-tracking

61
 I. INTRODUCTION

Mental health challenges, including stress, anxiety, and depression, are increasing globally,
affecting millions of people. According to the World Health Organization, over 300 million
individuals suffer from depression, with numbers steadily rising. Traditional emotional support
methods often lack personalization and immediate accessibility, making them less effective in
addressing real-time emotional needs. Conventional mood assessment methods, such as self-
reporting or text-based analyses, fail to capture the full spectrum of human emotions, often leading
to inaccurate or delayed interventions. Many existing systems do not provide timely support or adapt
to a user’s unique emotional state. Aura Pulse addresses this gap by integrating advanced facial
recognition technology to detect emotions in real time. It delivers personalized music
recommendations and mental health resources, creating a seamless user experience that enhances
emotional well-being and self-awareness. The market for emotion recognition technology is
expanding rapidly, with applications in entertainment, healthcare, and consumer engagement.
Companies like Affectiva and IBM Watson are developing emotion detection systems, primarily
focusing on specific use cases such as marketing analytics. However, Aura Pulse differentiates itself
by combining emotion detection with actionable recommendations, providing a holistic and user-
centric approach to emotional well-being rather than merely analyzing emotions for commercial
purposes. Offering real-time emotion detection through facial recognition, personalized music
recommendations, and access to mental health resources, Aura Pulse enhances user engagement and
emotional intelligence while providing a non-invasive and accessible solution for mental health
support.

Helping user to adopt proactive mental health strategies. By fostering self-awareness and emotional
intelligence, Aura Pulse empowers users to take control of their mental well-being, promoting a
healthier and more balanced lifestyle in an increasingly fast-paced digital world.
This research explores the integration of AI-driven emotion recognition with intelligent systems,
focusing on the challenges of real-time emotion detection in dynamic environments. It emphasizes
the importance of robust machine learning algorithms to improve accuracy and reliability in
emotion-based applications, enabling seamless user experiences in personalized mental health
support. Aura Pulse presents an AI-powered music recommendation system based on facial emotion
recognition. By analyzing facial expressions, the system identifies users' emotional states and
recommends personalized music to enhance mood and well-being. The research highlights the
effectiveness of AI in creating emotionally adaptive entertainment experiences. Aura Pulse
investigates the role of CNNs in facial expression recognition for music therapy applications. It
demonstrates how deep learning techniques can accurately classify emotions and generate tailored
music suggestions to support emotional well-being. The study underscores the importance of dataset
diversity and model optimization in achieving reliable emotion recognition. Aura Pulse focuses on
region-specific music recommendations using CNN-based facial expression recognition. It
examines how cultural preferences influence emotional responses to music, showcasing the
adaptability of AI-driven systems in delivering personalized recommendations based on diverse user
demographics.

II. LITERATURE SURVEY

Runfang Guo et al. (2024); Their research examines the incorporation of emotion recognition into
smart devices, emphasizing the difficulties presented by ever-changing environments. It underscores
the need for strong algorithms capable of real-time processing to guarantee precise and dependable
emotion detection across diverse conditions.(6)

B. Kranthi Kiran et al. (2024); The study showcases the efficiency of AI-powered music
recommendations based on emotions, attaining an 80% accuracy rate through facial recognition. It
highlights the role of artificial intelligence in improving user experiences by delivering personalized
music therapy tailored to detected emotional states.. (7)

Padmavati K et al. (2024); Their research employs convolutional neural networks (CNNs) for
recognizing facial expressions, resulting in successful music therapy outcomes. It highlights the essential

62
role of machine learning in identifying emotions and its significance in mental health support through
customized music recommendations.. (9)

Moeez Rajjan et al. (2023); The study integrates micro-expression detection with music
recommendations, reaching an accuracy rate of 62.1%. It offers valuable insights into the challenges of
merging multiple data sources for improved emotion analysis and emphasizes the technical difficulties
in real-time emotion recognition systems.. (11)

M. Athavle et al. (2021The study utilizes CNNs to provide Bollywood music recommendations based
on real-time facial expressions. It showcases the cultural flexibility of emotion-based systems,
highlighting how AI-powered music therapy can accommodate diverse user preferences across various
regions.. (2)

Mrinmoy Bhaumik et al. (2021); The study examines the application of generative adversarial networks
(GANs) in emotion-driven music recommendations. This method enhances the variety and relevance of
song suggestions, improving the personalization and efficiency of AI-powered music therapy.
Additionally, it explores how GANs enhance emotion classification accuracy by generating synthetic
facial expression data to optimize training models.(3)

S. Metilda Florence et al. (2020); proposes a system capable of detecting user emotions with over
80% accuracy through real-time facial expression recognition. This study provides a strong
foundation for real-world applications, demonstrating the feasibility of using AI-powered emotion
detection for mental health support Furthermore, the research examines the scalability of the system
across different devices and explores potential integrations with wearable technology for continuous
emotion monitoring. (4)

Phaneendra, et al. (2022); EMUSE is a music recommendation system that utilizes facial emotion
recognition alongside the Spotify API to deliver tailored song suggestions. It features an intuitive
interface where users can view their detected emotions along with recommended tracks, enriching their
listening experience. However, a key limitation is its exclusive dependence on facial expressions,
which may miss deeper emotional nuances that influence recommendation accuracy. Furthermore, the
constant monitoring of facial features raises concerns regarding user privacy and the feasibility of
implementing such a system. (10)

Maduri Athavle et al. (2021); This system employs Convolutional Neural Networks (CNNs) to detect
facial emotions and generate music recommendations accordingly. By leveraging advanced machine
learning techniques, it enhances emotion analysis, resulting in personalized song suggestions.
However, the system struggles to accurately interpret subtle emotions, which may impact the precision
of its recommendations. Additionally, ongoing facial monitoring raises privacy concerns, and the study
does not provide comprehensive details on the evaluation methods used to measure the system’s
effectiveness. (8)

G. Tirumala, et al. (2023)This research investigates a music recommendation system that utilizes
CNNs trained on audio features extracted from songs to enhance emotion representation. By directly
analyzing audio signals, the system strives to better capture the emotional essence of music. However,
it depends on Support Vector Machines (SVMs) for song recommendations, which may be less
effective compared to methods like collaborative filtering. Furthermore, the absence of detailed
explanations about the CNN architecture and the specific audio features used restricts a thorough
understanding of the system’s capabilities (5)

Ashwini Jadhav, et al. (2023)This system leverages CNNs to process music using Mel-Frequency
Cepstral Coefficients (MFCCs) for emotion-driven recommendations. By focusing on audio features, it
provides a distinct method for interpreting the emotional aspects of music. However, the study does not
offer detailed information about the CNN architecture used and fails to elaborate on the
recommendation approach, restricting a deeper understanding of its functionality and possible
enhancements.(1)

63
 III. PROPOSED METHODOLOGY

The methodology of this research focuses on data collection, preprocessing, model implementation,
deployment, and security considerations. For data collection, widely recognized facial emotion
datasets like FER2013 are utilized, containing thousands of labeled facial images representing
diverse expressions and emotions. In the preprocessing phase, noise removal techniques such as
Gaussian filters for images are applied to enhance data quality. Normalization ensures consistency
across the dataset by adjusting pixel values for uniformity Additionally, facial images are
standardized by resizing them to a uniform resolution of 48x48 pixels, reducing computational
complexity while preserving essential features.

Fig.1. Flowchart of the model

Convolutional neural networks (CNNs) are used for facial emotion recognition because of their
efficiency in image analysis. A deep CNN architecture incorporating convolutional, pooling, and fully
connected layers is designed to ensure precise emotion detection. To improve model performance and
minimize overfitting, data augmentation methods such as rotation, flipping, and zooming are applied.
Furthermore, transfer learning is employed by fine-tuning pre-trained models like VGG16 and ResNet
on facial emotion datasets, utilizing their learned features to enhance accuracy and efficiency.

The system is designed as a web application, developed using [Link] to provide an interactive and
responsive experience. Key features include real-time emotion detection, personalized music
recommendations, and access to mental health resources. The web application is optimized for
cross-device compatibility, ensuring accessibility from different screen sizes without compromising
functionality.

Beta testing is conducted for this project to gather real-world user feedback and improve accuracy,
usability, and overall user experience. Testers interact with the web application in various scenarios,
helping identify potential
issues and areas for enhancement. Insights from beta testing are utilized to refine the emotion

64
detection model, optimize recommendations, and ensure a seamless user experience before the final
deployment.

IV. WORKING MODULES

a.) Login Module :

The Aura Pulse login page features a simple and user-friendly design, allowing users to sign in using
email and password or Google authentication. It ensures secure access while maintaining a clean and
intuitive interface.

Fig.2. Login Page

b.) Registration / Account Creation Page:

The Aura Pulse registration page provides a seamless and user-friendly interface for account creation.
Users can sign up using their email and password or opt for Google authentication for a quicker
process.

Fig. 3. Account Creation Page

b) Recommendation Based on Different Moods

The image displays in fig.4 below the user interface of Aura Pulse, a web-based mood detection system
using facial recognition. The application identifies the user's mood in real time, as shown with a
detected "neutral" emotion. It features a "Detect Mood" button and provides personalized
recommendations, including "Vibe with a Playlist" for mood-based music suggestions and "Lifestyle
Recommendation" for well-being tips. The interface is designed for a seamless user experience with
secure login/logout functionality.

65
 Fig. 4. Neutral Mood

The detected mood in this instance is "happy", influencing personalized recommendations such as
music playlists and lifestyle recommendation shown in fig.5

Fig. 5. Happy Mood

The detected mood is "surprise", which influences the system’s personalized recommendations, such as
music playlists and lifestyle recommendation shown in fig.6

Fig. 6. Surprise Mood

In this next case, the detected mood is "sad", allowing the system to provide tailored music playlists
and lifestyle recommendations to enhance emotional well-being shown in fig.7

66
 Fig. 7. Sad Mood
The Aura Pulse web application detects the user's facial expression and identifies the mood as "happy."
Based on this mood, it recommends an upbeat and energetic playlist featuring Bollywood party and
celebration songs. The interface allows the user to play songs directly and enhance their joyful state.
This personalized recommendation system improves user experience by aligning music with emotions
shown in fig.8.

Fig.8. Songs Playlist for Happy Mood

The Aura Pulse web application detects the user's happy mood using facial recognition and provides
lifestyle recommendations to maintain positivity. The suggestions encourage spreading joy, practicing
gratitude, and celebrating small achievements. This feature helps users reinforce their positive
emotions and lead a more fulfilling life shown in fig.9.

Fig.8. Lifestyle Recommendations for Happy Mood

67
 V. CONCLUSION

In summary, Aura Pulse has shown great promise in improving users' emotional well-being through
advanced real-time mood detection and personalized recommendations. Utilizing cutting-edge machine
learning techniques like Convolutional Neural Networks (CNNs), the system accurately identifies
emotions from facial expressions, enabling a customized experience that adapts to users' emotional
states. The personalized music recommendation engine has been highly effective, with users reporting
greater engagement and satisfaction. Moreover, the integration of mental health support features
underscores the system's comprehensive approach to well-being by providing essential resources and
exercises for stress management and emotional regulation. Designed with accessibility and
customization in mind, the intuitive interface has received positive feedback, ensuring that Aura Pulse
remains inclusive and easy to use for a diverse audience.

VI. FUTURE SCOPES

Future enhancements of Aura Pulse will include voice analysis for emotion detection, improving
accuracy and personalization. Multimodal emotion recognition, integrating facial expressions, voice
tones, and physiological signals, will provide a more comprehensive emotional assessment. Context-
aware recommendations and social network integration will refine user experiences, while features like
real-time feedback, mood tracking, and journaling will enhance mental health support. Collaborations
with mental health professionals will ensure accurate and reliable resources, positioning Aura Pulse as
an advanced and intelligent emotional well-being companion. Additionally, AI-driven adaptive
learning will enable the system to refine recommendations based on user behavior over time.
Integration with wearable devices can further enhance real-time emotion tracking, offering proactive
interventions. Expanding language and cultural adaptability will also make Aura Pulse more accessible
to a diverse global audience.

REFERENCES

1. Ashwini Jadhav, Nikita Bhaise, Mihir Narwade, Ruchita Phalke, Yash Talele, “Emotion based
Music Recommendation System using CNN”, International Journal of Research Publication
and Reviews, Vol. 4, pp. 3288-3292, Issue May 2023.
2. Athavle, M., Mudale, D., Shrivastav, U., & Gupta, M.(2021). "Music recommendation based
on face emotion recognition. "Journal of Informatics Electrical and Electronics Engineering,
Volume02, Issues:02.
3. Bhaumik M., Attah, P. U., & Javed, F. (2021). "Emotion Integrated Music Recommendation
System Using Generative Adversarial Networks." SMU Data Science Review, Vol. 5, No. 3,
Article 4. Available at: SMU Data Science Review.
4. Florence,S.M., & Uma, M. (2020). “Emotional Detection and Music Recommendation
System based on User Facial Expression.” IOP Conf. Series: Materials Science and
Engineering, 912 (2020)
5. G. Tirumala, M. Niharika, S. Shailu, M. Manaswini, G. Venga Vinodini, “Music
Recommendation System based on Emotions using CNN”, Journal of Engineering Sciences,
Vol. 14, Issue 2023.
6. Guo, R., Guo, H., Wang, L., Chen, M., Yang, D., & Li, B. (2024). “Development and
application of emotion recognition technology - a systematic literature review.” BMC
Psychology. Available at: BMC Psychology .
7. Kranthi Kiran, B., Shanthan, P., Ram, K. S., & Usha, K. (2024). “Emotion Based Music
Recommendation System using VGG16-CNN Architecture.” International Journal for
Research in Applied Science & Engineering Technology (IJRASET), Volume 12 Issue VI.
8. Maduri Athavle, Deepali Mudale, Upasana Shrivastav, Megha Gupta, “Music
Recommendation based on Face Emotion Recognition”, ICAI, Vol. 02, Sno. 018, pp. 1-11,
Issue 2021.
9. Padmavati, K., Amruth Kumar, D. K., Ganavi, S. M., Keerthi, K. C., & Sourabha, Y. N.
(2024). “SoulSound: Enhancing Musical Therapy through Facial Expression Recognition with
Machine Learning.” International Research Journal of Engineering and Technology (IRJET).
10. Phaneendra, Madhusmitha Muduli, Siri Lakshmi Reddy, R. Veenasree, “EMUSE – An
Emotion based Music Recommendation System”, IRJMETS, Volume 04, Issue 05 May 2022.

68
11. Rajjan, M., Deore, P., Mohite, Y., & Desai, Y. (2023). “Harmonic Fusion: AI-Driven Music
Personalization via Emotion-Enhanced Facial Expression Recognition.” International Journal
of Innovative Science and Research Technology, Volume 8, Issue 12.

69
CERTIFICATES

70
71