CODE FEST’ 25 HACKATHON 2.

Problem Statements
TEAM CODE THEME PROBLEM STATEMENT & OUTCOME

Problem Statement:

The proliferation of bots and fake accounts on social media platforms is a

significant cybersecurity issue. These automated accounts are used to
manipulate online discourse, spread spam, amplify disinformation, and
manipulate trends or public opinion. AI-based systems must be designed
CF2502 Cyber Security to detect and neutralize these threats while maintaining the integrity of
the platform’s ecosystem.

Requirements Based on the Problem Statement:

1. Bot Sophistication: Bots are becoming increasingly

sophisticated, with the ability to simulate human-like
interactions, engage in meaningful conversations, and even learn
and adapt over time. Detecting these bots without introducing
significant delays in the platform’s operation is a major
challenge.
2. Fake Account Creation and Behaviour Mimicry: Fake accounts
often mimic legitimate user behaviour, including posting,
commenting, and interacting with content in ways that seem
natural. AI systems must discern subtle differences between
human users and fake accounts based on patterns in their
interactions and engagement.
Scale and Speed: Social media platforms handle billions of interactions
daily. Identifying and stopping large-scale bot operations in real time
requires AI systems that can scale to handle massive amounts of data
without compromising performance
Problem Statement:

Many corporations struggle with bias in hiring processes, leading to

unfair selections, lack of diversity, and missed opportunities for
talented candidates. Traditional hiring methods often Favor certain
demographics due to unconscious bias, and companies lack AI-driven
Artificial Intelligence tools to ensure fair, transparent, and skill-based recruitment.
CF2504 for Corporate and
Social Good Requirements Based on the Problem Statement:

Build an AI-powered hiring assistant that ensures fair and unbiased

recruitment by evaluating candidates purely on skills and
qualifications, while leveraging external validation from professional
platforms. The system should:

• Resume Screening AI – Develop an NLP-based model that

analyses resumes without revealing personal details (e.g.,
name, gender, photo) and ranks candidates based on skills.
• Skill-Based AI Interview Bot – Create an AI-driven chatbot that
conducts initial technical screenings, evaluates answers, and
generates candidate scores.
• Certificate & Social Profile Validation – Integrate AI to verify
candidates' certifications, LinkedIn profiles, GitHub
 contributions, or other professional credentials to ensure
authenticity and expertise.
• Real-Time Candidate Analysis Dashboard – Display AI-
generated rankings, strengths, and improvement areas for
each candidate.
• Data Privacy & Compliance – Ensure GDPR-compliant data
handling, securing candidate information with encryption.

AI-Powered Smart Tutoring Assistant

Problem Statement:

Traditional tutoring is not scalable, and students often struggle to

get real-time doubt-solving due to a lack of personalized
assistance. Online resources are static and don’t provide adaptive
learning experiences.
CF2505 Artificial Intelligence
for Corporate and Requirements Based on the Problem Statement:
Social Good.
Develop an AI-driven tutoring assistant that provides interactive,
personalized learning with the following key features:

1. NLP-Based Doubt Solving:

2. Computer Vision for Handwritten Query Recognition:
3. Reinforcement Learning for Teaching Styles:
4. Interactive Explanations & Gamification:
5. Multilingual Support:
6. Integration with Existing Learning Platforms

Expected Impact:
• Increased accessibility to high-quality tutoring.
• Personalized learning paths for students.
• Automated doubt-solving in real-time.
 Problem Statement:

Many students drop out of school or college due to mental health

struggles, academic stress, or emotional burnout. However, institutions
lack real-time systems to detect at-risk students before it’s too late. A
Machine learning real-time AI-driven mental health and stress prediction system is needed
CF2506 and Ai in Education to detect burnout signs early and provide personalized interventions.

Requirements Based on the Problem Statement:

1. AI-Powered Emotion & Sentiment Analysis from Student

Interactions: Analyze student discussion forums, emails, and
assignments using NLP Sentiment Models (BERT, RoBERTa,
TextBlob) to detect stress-related language.
2. Live Dashboard for Educators & Counselors: Use Apache
Superset + Streamlit for real-time visualization of student mental
health trends.
3. Voice & Facial Stress Recognition: Use Deep Learning (Whisper
+ OpenFace) to analyze speech stress levels and
microexpressions.
4. AI-Powered Intervention & Chatbot Therapy: Deploy a mental
health conversational AI assistant using LLMs (GPT-4 + RAG-
based Therapy Models) to provide immediate emotional support.
5. Suicide & Self-Harm Risk Prediction: Implement Multi-Modal
AI Models combining social media analysis, survey responses,
and physiological data.

Problem Statement:
Students and workers in rural areas face limited access to quality
education and skill development opportunities. Traditional online
courses are not tailored to rural contexts, making it diLicult for
individuals to acquire relevant skills. AI can personalize learning
experiences and make education more accessible.
Requirements:
CF2507 Machine learning 1. AI-Based Personalized Learning Paths – Develop an adaptive
and Ai in Education AI system that customizes educational content based on the
learner’s pace and interests.
2. AI-Enabled Rural Job Training Programs – Use AI to analyse
job market trends and recommend relevant training courses.
3. Multi-Language AI Tutor – Implement an AI-powered tutor
that provides explanations in regional languages.
4. AI-Powered Career Guidance System – Develop an AI system
that suggests career paths based on a learner’s skills and
interests.
5. AI-Based Skill Certification Verification – Use AI to verify and
validate rural workers’ skills through an automated
assessment system.
 Problem Statement:

Developing a web-based application that captures an image of a word

puzzle containing a grid
of jumbled alphabets presents a challenge in accurately identifying user-
circled letters. The
system must detect and extract these marked letters, recognize the
CF2509 Open Innovation corresponding characters,
Challenge and generate the maximum number of meaningful words from the
jumbled selection. This
requires effective implementation of image processing, character
recognition, and word
formation algorithms to ensure precise detection and accurate word
extraction.

Consider a word puzzle where a user is given a 5×5 grid of jumbled

letters and is required to
form words by circling specific letters. The web application should
capture an image of this grid,
detect the circled letters, and extract potential meaningful words.
Example Grid:
TAPLE
BORAD
SIDPN
ELGRA
CATHM
User-circled letters: T, A, B, L, E
The application should:
1. Detect these circled letters.
2. Recognize them as T, A, B, L, E.
3. Generate meaningful words from the jumbled set:
TABLE
ABLE
TALE
BAT
TAB
The system should use image processing techniques to identify
circled letters, OCR (Optical
Character Recognition) to extract characters, and word formation
algorithms to generate the
maximum number of valid words.
 Problem Statement:

Develop an AI-driven Emotional Intelligence Tutor (EIT) that detects

student emotions, stress levels, and mental fatigue in real time, providing
adaptive interventions such as personalized motivation, mental well-
CF2511 Machine learning being strategies, and AI-assisted mentoring. The system should integrate
and Ai in Education psychological models, sentiment analysis, and cognitive-behavioural AI
to foster a mentally healthy learning environment.

Requirements Based on the Problem Statement:

1. AI-Powered Sentiment & Emotion Recognition: Monitors

student frustration, anxiety, and motivation.
2. Psychological AI Interventions: Provides mental well-being
exercises, stress reduction techniques, and AI-driven mentoring.
3. Adaptive Learning Pace Control: Adjusts workload based on
student emotional and cognitive readiness.
4. Conversational AI for Mental Health Support: Acts as a virtual
counsellor with explainable AI-based guidance.
5. Bias-Free Emotional AI: Prevents stereotyping and ensures fair
emotional assessment across demographics.

Problem Statement:

Develop an advanced AI -based Smart Traffic & Driver Monitoring

System that integrates camera functions, AI-powered real-time object
detection, predictive analytics, driver state monitoring.
CF2512 Artificial Intelligence
for Corporate and Core AI Functional Requirements:
Social Good
1. Detect, classify, and track traffic signals, road signs, trees,
pedestrians, animals, and dynamic road conditions.
2. Analyse real-time traffic scenarios and predict potential
hazards.
3. Overlay AI visualizations to enhance situational awareness.
4. Monitor driver attentiveness, fatigue, and dangerous driving
behaviour for improved road safety.
 .Problem Statement:

With the rapid advancement of AI&ML-driven content generation,

deepfake technology has become a significant threat to digital security,
misinformation control, and identity verification. Deepfake videos,
images, and audio can be used for fraudulent activities, political
CF2513 Open Innovation propaganda, and social manipulation, making it increasingly difficult to
Challenge distinguish between real and artificially generated content. Traditional
detection methods often fail to keep up with evolving deepfake
techniques. Therefore, there is a need for an AI-powered deepfake
detection system that can accurately analyse and classify media content
to identify deepfake manipulations in real-time.

Requirements Based on the Problem Statement:

1. AI-Based Detection Model – Develop a machine learning or

deep learning model (e.g., CNN, Transformer, or GAN-based
analysis) to identify deepfake content with high accuracy.
2. Multi-Modal Analysis – Detect deepfakes in videos, images, and
audio by analysing inconsistencies in facial expressions, lip-sync,
lighting, and speech patterns.
3. Real-Time Processing – Enable fast and efficient detection to
analyse media content in real-time.
4. Explainability & Confidence Scores – Provide reasons for
detection results, including probability scores and highlighted
tampered regions.
5. Scalability & API Integration – Develop an API for easy
integration into social media platforms, video streaming services,
and authentication systems.
6. User-Friendly Interface – Build a web-based or mobile
application for users to upload and analyse media files for
deepfake detection.
7. Robust Against Adversarial Attacks – Improve detection models
to counter adversarial deepfake techniques designed to bypass
detection systems.
8. Continuous Model Updating – Implement an adaptive system
that updates itself with new deepfake techniques through active
learning.
9. Forensic Analysis & Reporting – Generate detailed reports for
forensic investigations and legal purposes.
10. Ethical & Privacy Compliance – Ensure responsible use of AI in
detection while maintaining user privacy and ethical
considerations.
 Problem Statement:

Developing a web-based Augmented Reality (AR) application that

allows users to interactively select and visualize 3D models
superimposed on a real-world image target presents a significant
challenge. The system should dynamically generate and overlay 3D
models based on user inputs while maintaining accurate positioning
CF2514 Virtual Reality and and tracking on the same reference image.
Augmented Reality
Key Features & Requirements:
1. Image Target Recognition – The application should recognize a
predefined image (e.g., a hotel menu card or an application UI). 2.
User Input Selection – Users should be able to select items from the
menu (e.g., a food item in a restaurant menu). 3. Dynamic 3D Model
Overlay – Based on the selected item, the corresponding 3D model
(e.g., a burger, pizza, or dish) should be rendered and overlaid in AR.
4. Real-time Interaction – Users should be able to view the 3D model
from diLerent angles, zoom in/out, an interact with it. 5. Web-Based
Implementation – The solution should be accessible through a
browser without requiring additional app installations. 6. Accurate
Model Placement – The system should ensure precise alignment of
the 3D model over the image target without drift or misalignment
Problem Statement:

Many students drop out of school or college due to mental health

struggles, academic stress, or emotional burnout. However, institutions
lack real-time systems to detect at-risk students before it’s too late. A
real-time AI-driven mental health and stress prediction system is needed
to detect burnout signs early and provide personalized interventions.

CF2515 Machine learning Requirements Based on the Problem Statement:

and Ai in Education
6. AI-Powered Emotion & Sentiment Analysis from Student
Interactions: Analyze student discussion forums, emails, and
assignments using NLP Sentiment Models (BERT, RoBERTa,
TextBlob) to detect stress-related language.
7. Real-Time Behavioral & Engagement Monitoring: Track
attendance patterns, assignment delays, and course progress
using Time-Series Anomaly Detection (LSTMs + ARIMA).
8. Voice & Facial Stress Recognition: Use Deep Learning (Whisper
+ OpenFace) to analyze speech stress levels and
microexpressions.
9. AI-Powered Intervention & Chatbot Therapy: Deploy a mental
health conversational AI assistant using LLMs (GPT-4 + RAG-
based Therapy Models) to provide immediate emotional support.
10. Suicide & Self-Harm Risk Prediction: Implement Multi-Modal
AI Models combining social media analysis, survey responses,
and physiological data.
11. Live Dashboard for Educators & Counselors: Use Apache
Superset + Streamlit for real-time visualization of student mental
health trends.
 Problem Statement:

Develop an AI-driven Emotional Intelligence Tutor (EIT) that detects

student emotions, stress levels, and mental fatigue in real time, providing
adaptive interventions such as personalized motivation, mental well-
being strategies, and AI-assisted mentoring. The system should integrate
psychological models, sentiment analysis, and cognitive-behavioural AI
Machine learning to foster a mentally healthy learning environment.
CF2516 and Ai in Education
Requirements Based on the Problem Statement:

1. AI-Powered Sentiment & Emotion Recognition: Monitors

student frustration, anxiety, and motivation.
2. Psychological AI Interventions: Provides mental well-being
exercises, stress reduction techniques, and AI-driven mentoring.
3. Adaptive Learning Pace Control: Adjusts workload based on
student emotional and cognitive readiness.
4. Conversational AI for Mental Health Support: Acts as a virtual
counsellor with explainable AI-based guidance.
5. Bias-Free Emotional AI: Prevents stereotyping and ensures fair
emotional assessment across demographics.

Problem Statement:

Traditional vehicle insurance claim processing is slow, requiring manual

inspection to assess damage. This leads to delays in claim approvals and
inaccurate cost estimates.

CF2520 Artificial Intelligence Requirements Based on the Problem Statement:

for Corporate and
Social Good. Develop an AI-based vehicle damage assessment system using OpenCV
and Deep Learning to analyse vehicle images and provide automated
damage detection and cost estimation.

1. Damage Detection & Classification

2. Automated Cost Estimation
3. OCR for License Plate & Insurance Details
4. Multi-View Image Analysis
5. Mobile App Integration
 Problem Statement:

The proliferation of bots and fake accounts on social media platforms is a

significant cybersecurity issue. These automated accounts are used to
manipulate online discourse, spread spam, amplify disinformation, and
manipulate trends or public opinion. AI-based systems must be designed
to detect and neutralize these threats while maintaining the integrity of
CF2522 the platform’s ecosystem.
Cyber Security
Requirements Based on the Problem Statement:

3. Bot Sophistication: Bots are becoming increasingly

sophisticated, with the ability to simulate human-like
interactions, engage in meaningful conversations, and even learn
and adapt over time. Detecting these bots without introducing
significant delays in the platform’s operation is a major
challenge.
4. Fake Account Creation and Behaviour Mimicry: Fake accounts
often mimic legitimate user behaviour, including posting,
commenting, and interacting with content in ways that seem
natural. AI systems must discern subtle differences between
human users and fake accounts based on patterns in their
interactions and engagement.
Scale and Speed: Social media platforms handle billions of interactions
daily. Identifying and stopping large-scale bot operations in real time
requires AI systems that can scale to handle massive amounts of data
without compromising performance.
Problem Statement:

Farmers in rural areas struggle with low productivity due to a lack of

real-time guidance on crop selection, pest control, and weather
conditions. Existing agricultural advisory services are often generic
and not tailored to individual farm conditions. There is a need for an
AI-powered software solution that provides personalized
recommendations based on data analysis.
Requirements:
CF2523 Agricultural and 1. AI-Powered Crop Recommendation System – Develop an
Rural Development NLP-based chatbot that analyzes past yield data, soil reports,
and climate conditions to suggest optimal crops.
2. AI-Driven Pest & Disease Detection – Implement an AI model
that processes text-based symptoms reported by farmers to
diagnose crop diseases and recommend treatments.
3. Automated Weather & Climate Advisory – Integrate an AI
model that analyzes weather forecasts and provides farmers
with real-time farming strategies.
4. AI-Based Market Trend Analysis – Use AI to predict demand
for crops, helping farmers choose high-profit crops for
cultivation.
 5. Multi-Language AI Chatbot for Rural Farmers – Provide AI-
driven farming assistance in regional languages to ensure
accessibility.

Problem Statement:
Developing a web-based Augmented Reality (AR) application that
allows users to interactively select and visualize 3D models
superimposed on a real-world image target presents a significant
challenge. The system should dynamically generate and overlay 3D
models based on user inputs while maintaining accurate positioning
and tracking on the same reference image.
Virtual Reality and
CF2524 Augmented Reality Key Features & Requirements:
1. Image Target Recognition – The application should recognize a
predefined image (e.g., a hotel menu card or an application UI). 2.
User Input Selection – Users should be able to select items from the
menu (e.g., a food item in a restaurant menu). 3. Dynamic 3D Model
Overlay – Based on the selected item, the corresponding 3D model
(e.g., a burger, pizza, or dish) should be rendered and overlaid in AR.
4. Real-time Interaction – Users should be able to view the 3D model
from diLerent angles, zoom in/out, an interact with it. 5. Web-Based
Implementation – The solution should be accessible through a
browser without requiring additional app installations. 6. Accurate
Model Placement – The system should ensure precise alignment of
the 3D model over the image target without drift or misalignment.
 Problem Statement:

Farmers in rural areas struggle with low productivity due to a lack of

real-time guidance on crop selection, pest control, and weather
conditions. Existing agricultural advisory services are often generic
and not tailored to individual farm conditions. There is a need for an
AI-powered software solution that provides personalized
recommendations based on data analysis.
CF2526 Agricultural and Requirements:
Rural Development 1. AI-Powered Crop Recommendation System – Develop an
NLP-based chatbot that analyzes past yield data, soil reports,
and climate conditions to suggest optimal crops.
2. AI-Driven Pest & Disease Detection – Implement an AI model
that processes text-based symptoms reported by farmers to
diagnose crop diseases and recommend treatments.
3. Automated Weather & Climate Advisory – Integrate an AI
model that analyzes weather forecasts and provides farmers
with real-time farming strategies.
4. AI-Based Market Trend Analysis – Use AI to predict demand
for crops, helping farmers choose high-profit crops for
cultivation.
5. Multi-Language AI Chatbot for Rural Farmers – Provide AI-
driven farming assistance in regional languages to ensure
accessibility.

Problem Statement:

Virtual Reality Problem Statement Web-Based Virtual Reality (VR)

Therapy for Mental Health Mental health therapy often requires
innovative and engaging approaches to help individuals cope with
stress, anxiety, PTSD, and other psychological challenges. Traditional
CF2527 therapy methods can sometimes be limited in accessibility,
Open Innovation engagement, and immersive eLectiveness. A Web-Based Virtual
Challenge Reality (VR) Therapy System can oLer a safe, interactive, and
immersive environment to support mental health therapy by creating
calm, controlled, and guided experiences that help individuals relax,
manage emotions, and improve mental well-being.

Requirements:
1. Immersive Virtual Therapy Environments – Provide calming VR
spaces (e.g., nature landscapes, underwater worlds, starry night
skies) to reduce stress and anxiety. 2. Guided Meditation &
Mindfulness – OLer breathing exercises, guided meditations, and
relaxation techniques within VR to help users focus and manage
emotions. 3. Cognitive Behavioral Therapy (CBT) Simulations – Use
interactive VR scenarios to help users confront fears (e.g., social
anxiety, phobias) in a safe, controlled space. 4. Web-Based
Accessibility – The application should run on a web browser without
 requiring high-end VR hardware, supporting WebVR, WebXR, or A-
Frame for a flexible experience. 5. Real-Time Feedback & Emotion
Tracking – The system should provide visual and audio feedback
based on user reactions, guiding them towards a calmer state. 6.
Adaptive Therapy Sessions – The experience should adapt based on
user responses, helping personalize mental health support over
time. 7. Remote Therapy & Support – Enable therapist-patient virtual
interaction to allow for guided therapy sessions and discussions. 8.
Cross-Device Compatibility – Support VR headsets (Oculus Quest,
HTC Vive), desktop, and mobile devices using WebVR/WebXR for
greater accessibility. Example Use Cases: 1. Stress & Anxiety Relief
Users enter a virtual nature environment (e.g., beach, forest, or
mountain) where they can listen to calming sounds and practice
deep breathing exercises. VR mindfulness training helps users focus
on the present moment and regain emotional balance.

Problem Statement:

Develop an AI-powered irrigation system that optimizes water usage

based on Google Weather API data, crop type, and historical irrigation
patterns. The system should predict the optimal amount of water needed
and automatically adjust irrigation schedules to conserve water while
ensuring healthy crop growth

CF2528 Agricultural and Requirements Based on the Problem Statement:

Rural Development
1. Integrate Google Weather API to fetch real-time and forecasted
weather conditions.
2. Develop a machine learning model to recommend irrigation
schedules based on temperature, rainfall, and crop type.
3. Ensure the system works with minimal historical data and adapts
to uncertain weather conditions.
4. Robust Against Adversarial Attacks – Improve detection models
to counter adversarial deepfake techniques designed to bypass
detection systems.
5. Continuous Model Updating – Implement an adaptive system
that updates itself with new deepfake techniques through active
learning.
6. Forensic Analysis & Reporting – Generate detailed reports for
forensic investigations and legal purposes.
7. Ethical & Privacy Compliance – Ensure responsible use of AI in
detection while maintaining user privacy and ethical
considerations.
 Problem Statement:

Develop an advanced AI -based Smart Traffic & Driver Monitoring

System that integrates camera functions, AI-powered real-time object
detection, predictive analytics, driver state monitoring.

CF2529 Core AI Functional Requirements:

Open Innovation
Challenge 5. Detect, classify, and track traffic signals, road signs, trees,
pedestrians, animals, and dynamic road conditions.
6. Analyse real-time traffic scenarios and predict potential
hazards.
7. Overlay AI visualizations to enhance situational awareness.
8. Monitor driver attentiveness, fatigue, and dangerous driving
behaviour for improved road safety.

Problem Statement:

The proliferation of bots and fake accounts on social media platforms is a

significant cybersecurity issue. These automated accounts are used to
manipulate online discourse, spread spam, amplify disinformation, and
manipulate trends or public opinion. AI-based systems must be designed
CF2530 to detect and neutralize these threats while maintaining the integrity of
Open Innovation the platform’s ecosystem.
Challenge
Requirements Based on the Problem Statement:

1. Bot Sophistication: Bots are becoming increasingly

sophisticated, with the ability to simulate human-like
interactions, engage in meaningful conversations, and even learn
and adapt over time. Detecting these bots without introducing
significant delays in the platform’s operation is a major
challenge.
2. Fake Account Creation and Behaviour Mimicry: Fake accounts
often mimic legitimate user behaviour, including posting,
commenting, and interacting with content in ways that seem
natural. AI systems must discern subtle differences between
human users and fake accounts based on patterns in their
interactions and engagement.
Scale and Speed: Social media platforms handle billions of interactions
daily. Identifying and stopping large-scale bot operations in real time
requires AI systems that can scale to handle massive amounts of data
without compromising performance.
 Problem Statement:

AI-Powered Stock Prediction & Recommendation App Using Big Data

Analytics

Stock market investing is highly volatile, influenced by millions of

CF2533 Data Visualization transactions, global economic trends, and investor sentiment.
Traditional investment strategies rely on technical analysis,
fundamental data, and human intuition, which can be slow,
inaccurate, and biased.

With the rise of big data analytics and AI, there is an opportunity to
build an intelligent stock prediction and recommendation system
that can:
• Process large-scale real-time financial data to
identify hidden patterns.
• Predict short-term and long-term stock price
movements with high accuracy.
• Analyse market sentiment using news, social media,
and investor behaviour.
• Provide personalized stock recommendations based
on user risk profiles and investment goals.
• Visualize complex financial data for better decision-
making.

This app will leverage big data frameworks, AI-driven forecasting

models, and real-time analytics to help both beginner and
professional investors make data-driven investment decisions with
greater confidence.
Problem Statement :
Developing a web-based Augmented Reality (AR) application that
allows users to interactively select and visualize 3D models
superimposed on a real-world image target presents a significant
challenge. The system should dynamically generate and overlay 3D
models based on user inputs while maintaining accurate positioning
and tracking on the same reference image.
Pec Open Innovation
Challenge Key Features & Requirements:

1. Image Target Recognition – The application should recognize a

predefined image (e.g., a hotel menu card or an application UI). 2.
User Input Selection – Users should be able to select items from the
menu (e.g., a food item in a restaurant menu). 3. Dynamic 3D Model
Overlay – Based on the selected item, the corresponding 3D model
(e.g., a burger, pizza, or dish) should be rendered and overlaid in AR.
4. Real-time Interaction – Users should be able to view the 3D model
from diLerent angles, zoom in/out, an interact with it. 5. Web-Based
Implementation – The solution should be accessible through a
browser without requiring additional app installations. 6. Accurate
Model Placement – The system should ensure precise alignment of
the 3D model over the image target without drift or misalignment.