PHOTO-SKETCH SYNTHESIS USING AUTOENCODER BASED

GENERATIVE MODELING

A PROJECT REPORT

Submitted by

JACK ANDRE J (team member names in alphabetical order) RA2011026020147

ABISHEK RAJ M (Initials at the last) RA2011026020131
CHARUDEVE KS RA2011026020139

Under the guidance of

Mrs. Angeline R
Assistant Professor (Selection Grade), //designation
Department of Computer Science and Engineering //Supervisor’s dept.

in partial fulfillment for the award of the degree of

BACHELOR OF TECHNOLOGY
in

COMPUTER SCIENCE AND ENGINEERING

with specialization in

ARTIFICIAL INTELLIGENCE AND MACHINE INTELLIGENCE

of

FACULTY OF ENGINEERING AND TECHNOLOGY

SCHOOL OF COMPUTERSCIENCE ENGINEERING

SRM INSTITUTE OF SCIENCE AND TECHNOLOGY

RAMAPURAM, CHENNAI -600089

May 2025
 SRM INSTITUTE OF SCIENCE AND TECHNOLOGY
(Deemed to be University U/S 3 of UGC Act, 1956)

BONAFIDE CERTIFICATE

Certified that this project report titled “PHOTO-SKETCH SYNTHESIS USING

AUTOENCODER BASED GENERATIVE MODELING” is the bonafide work of
JACK ANDRE J [REG NO: RA2011026020147], ABISHEK RAJ M [REG NO:
RA2011026020131], CHARUDEVE KS [REG NO: RA2011026020139] who
carried out the project work under my supervision. Certified further, that to the best of
my knowledge, the work reported herein does not form any other project report or
dissertation on the basis of which a degree or award was conferred on an occasion on
this or any other candidate.

SIGNATURE SIGNATURE

Name of the Supervisor Name of the HoD

Designation Professor/Associate Professor and Head
Name of the Department, Name of the Department,
S R M Institute of Science and Technology, S R M Institute of Science and Technology,
Chennai-89. Chennai-89.

Submitted for the project viva-voce held on___________at S R M

Institute of Science and Technology, Chennai -600089.

INTERNAL EXAMINER EXTERNAL EXAMINER

ii
 SRM INSTITUTE OF SCIENCE AND TECHNOLOGY
RAMAPURAM, CHENNAI - 89

DECLARATION

We hereby declare that the entire work contained in this project report
titled “PHOTO-SKETCH SYNTHESIS USING AUTOENCODER
BASED GENERATIVE MODELING” has been carried out by JACK
ANDRE [REG NO: RA2011026020147], ABISHEK RAJ M [REG
NO: RA2011026020131], CHARUDEVE KS [REG NO:
RA2011026020139] at SRM Institute of Science and Technology,
Ramapuram, Chennai- 600089, under the guidance of Ms. Angeline R,
[Link]., (Ph.D)., Assistant Professor (Selection Grade), Department of
Computer Science and Engineering.

Place: Chennai
Date: JACK ANDRE J

ABISHEK RAJ M

CHARUDEVE KS

iii
 Own Work Declaration

Department of Computer Science and Engineering

SRM Institute of Science and Technology

Own Work Declaration form

Degree/ Course : B. Tech Computer Science and Engineering with Specialization in

Artificial Intelligence and Machine Learning
Student Name : JACK ANDRE J, ABISHEK RAJ M, CHARUDEVE KS
Registration Number : RA2011026020147, RA2011026020131, RA2011026020139
Title of Work : PHOTO-SKETCH SYNTHESIS USING AUTOENCODER BASED
GENERATIVE MODELING
We hereby certify that this assessment complies with the University’s Rules and Regulations relating to
Academic misconduct and plagiarism, as listed in the University Website, Regulations, and the Education
Committee guidelines.
We confirm that all the work contained in this assessment is our own except where indicated, and that We have
met the following conditions:
● Clearly references / listed all sources as appropriate
● Referenced and put in inverted commas all quoted text (from books, web, etc)
● Given the sources of all pictures, data etc. that are not our own
● Not made any use of the report(s) or essay(s) of any other student(s) either past or present
● Acknowledged in appropriate places any help that we have received from others (e.g., fellow students,
technicians, statisticians, external sources)
● Compiled with any other plagiarism criteria specified in the Course handbook / University website
We understand that any false claim for this work will be penalised in accordance with the University policies and
regulations.

DECLARATION:
We are aware of and understand the University’s policy on Academic misconduct and
plagiarism and we certify that this assessment is our own work, except where indicated by
referring, and that we have followed the good academic practices noted above.

RA2011026020147 RA2011026020131 RA2011026020139

iv

ACKNOWLEDGEMENT
 We place on record our deep sense of gratitude to our honourable Chairman
Dr. R. SHIVAKUMAR, MBBS., MD., for providing us with the requisite
infrastructure throughout the course.
We take the opportunity to extend our hearty and sincere thanks to our
Dean, Dr. M. SAKTHI GANESH., Ph.D., for maneuvering us into
accomplishing the project.
We take the privilege to extend our hearty and sincere gratitude to the
Professor and Chairperson, Dr. K. RAJA, Ph.D., for his suggestions, support and
encouragement towards the completion of the project with perfection.
We thank our honorable Head of the department [Link]. Designation,
Department for his/her constant motivation and unwavering support.
We express our hearty and sincere thanks to our guide
[Link]. Designation, Department for his/her encouragement, motivation and
constant guidance throughout this project work.
Our thanks to the teaching and non-teaching staff of the Department of
Computer Science and Engineering of SRM Institute of Science and Technology,
Chennai, for providing necessary resources for our project

v
 ABSTRACT

The transformation of images into sketches is a critical area in computer vision with wide-

ranging applications from artistic rendering to enhancing visual content. This project introduces

an innovative method for synthesizing sketches from photos using an autoencoder-based

generative model. Unlike conventional methods that modify input images to create sketch-like

appearances, our system employs autoencoders to generate sketches from scratch. The system

is structured around an encoder-decoder architecture built upon autoencoders. When presented

with a photo, the encoder extracts crucial features and representations, capturing the image's

underlying structure and content. These extracted features are then used by the decoder to

produce a sketch, employing generative modeling techniques to simulate hand-drawn strokes,

lines, and textures. This photo-sketch synthesis system presents significant benefits for law

enforcement agencies by aiding them in suspect identification, complement facial recognition

technology, assist in cold case investigations, verify witness testimonies and support

undercover operations.

vi
 TABLE OF CONTENTS

Page. No

ABSTRACT vi

LIST OF FIGURES x

LIST OF TABLES xi

LIST OF ACRONYMS AND ABBREVIATIONS xii

1 INTRODUCTION 1
1.1 Problem Statement 12
1.2 Aim of the Project 13
1.3 Project Domain 13
1.4 Scope of the Project 13
1.5 Methodology 14
1.6 Organization of the Report 14

2 LITERATURE REVIEW 16

3 PROJECT DESCRIPTION 19
3.1 Existing System 19
3.2 Proposed System 19
3.2.1 Advantages 19
3.3 Feasibility Study 20
3.3.1 Economic Feasibility 20

vii
 3.3.2 Technical Feasibility 20
3.3.3 Social Feasibility 20
3.4 System Specification 21
3.4.1 Hardware Specification 21
3.4.2 Software Specification 21
3.4.3 Standards and Policies 21

4 PROPOSED WORK 22
4.1 General Architecture 22
4.2 Design Phase 22
4.2.1 Data Flow Diagram 23
4.2.2 UML Diagram 24
4.2.3 Use Case Diagram 25
4.2.4 Sequence Diagram 25
4.3 Module Description 26
4.3.1 Module 1: Image Processing 26
4.3.2 Module 2: Feature Extraction 26
4.3.3 Module 3: Image to Sketch Synthesis 27
4.3.4 Step 2: Processing of Data 27
4.3.5 Step 3: Split the Data 28
4.3.6 Dataset Sample 28
4.3.7 Step 4: Building the Model 29
4.3.8 Step 5: Compiling and Training the Model 30

5 IMPLEMENTATION AND TESTING 31

5.1 Input and Output 31
5.1.1 Image of the Subject 31
5.1.2 Predicted Sketch Synthesis of Subject 31

viii
 5.2 Testing 32
5.2.1 Types of Testing 32
5.2.2 Unit testing 32
5.2.3 Integration testing 34
5.2.4 Functional testing 35
5.2.5 Test Result 35

6 RESULTS AND DISCUSSIONS

6.1 Efficiency of the Proposed System
6.2 Comparison of Existing and Proposed System

7 CONCLUSION AND FUTURE ENHANCEMENTS

7.1 Conclusion
7.2 Future Enhancements
7.3 Results

8 SOURCE CODE & POSTER PRESENTATION

8.1 Sample Code 39

References

A. Sample screenshots

B. Proof of Publication/Patent filed/ Conference Certificate

ix
 LIST OF FIGURES

4.1 Architecture Diagram 23

4.2 Data Flow Diagram 24
4.3 Uml Diagram 25
4.4 Use Case Diagram 26
4.5 Sequence Diagram 26
4.6 Pre-Processing Of Data 28
4.7 Dataset Of Photos 29
4.8 Dataset Of Sketches 29
4.9 Model Summary 30
4.10 Code To Compile Model And Fit Model To Dataset 31

5.1 Original Image Of The Subject 32

5.2 Predicted Sketch Of The Subject 33
5.3 Image Augmentation Using Opencv-Python 34
5.4 Code For Printing Images And Their Corresponding Sketches 35
5.5 Sketches And Their Image Printed As Output 35
5.6 Training And Testing Split 36
5.7 Model Is Being Compiled And Trained 36

7.1 Prediction And Sketch Matched 38

7.2 Model Prediction And Sketch Matched 38

8 Sample Code 40

x
 LIST OF TABLES

Table No. Table Name Page No.

1 Literature survey 16

xi
LIST OF ACRONYMS AND ABBREVIATIONS

API APPLICATION PROGRAM INTERFACE

CBIR CONTENT BASED IMAGE RETRIEVAL
CNN CONVOLUTIONAL NEURAL NETWORK
CUHK CHINESE UNIVERSITY OF HONG KONG
SBIR SKETCH BASED IMAGE RETRIEVAL

xii
 CHAPTER 1

INTRODUCTION

There are many studies in the learning field that investigated the ways of applying machine learning
techniques for various educational purposes. One of the focuses of these studies is to identify high-risk
students, as well as to identify features which affect the performance of students. Students are the major
strength for numerous universities. Universities and students play a significant part in producing
graduates of superior calibers with its academic performance accomplishment. However, academic
performance achievement changes as various sort of students may have diverse degree of performance
achievement. Machine learning is the ability of a system to consequently gain from past experience and
improve performance. Nowadays machine learning for education gains more attention. Machine
learning is used for analyzing information based on past experience and predicting future performance.

1.1 PROBLEM STATEMENT

Student performance and academic outcomes are influenced by a variety of complex, interrelated
factors. Traditional educational systems often lack the tools to accurately predict student success or
identify those at risk of underperforming. As a result, timely interventions are rarely implemented,
leading to higher dropout rates and reduced academic achievement. Educators face challenges in
analyzing vast amounts of student data manually, which limits data-driven decision-making. There is a
pressing need for intelligent systems that can uncover hidden patterns and provide actionable insights.
Existing solutions are often generic and fail to address individual learning needs. Our project, Edu
Predict, aims to bridge this gap by using AI to forecast academic performance and support personalized
learning paths. By leveraging machine learning algorithms, we enhance early intervention strategies
and academic planning. This approach promotes student success, retention, and overall institutional
performance.

1.2 AIM OF THE PROJECT

● To design an AI-powered system that accurately predicts student academic performance.

● To identify students at risk of underperformance and enable timely interventions.
12
 ● To support personalized learning strategies through data-driven insights.
● To assist educators in making informed academic and support decisions.
● To enhance overall student success, retention, and institutional outcomes.

1.3 PROJECT DOMAIN

The domain of this project is Artificial Intelligence in Education (AIED), a transformative subfield of
AI that focuses on enhancing teaching and learning processes through intelligent technologies. This
project specifically falls within the intersection of predictive analytics, educational data mining, and
student success systems.

1.4 SCOPE OF THE PROJECT

This project focuses on developing an AI-driven system to predict student academic performance using
historical and behavioral data. It enables early identification of at-risk students and supports timely
interventions by educators. The system offers predictive insights, personalized feedback, and
performance trends. It is scalable for use in schools, colleges, and universities. The solution aims to
enhance student success, retention, and institutional outcomes through data-informed strategies.

13
 1.5 METHODOLOGY

The methodology for this project involves a systematic approach to designing and implementing an AI-
powered student performance prediction system. Initially, historical data is collected from educational
institutions, including academic scores, attendance records, behavioral logs, and demographic details.
This data is then preprocessed to handle missing values, normalize formats, and ensure overall
consistency. Key features that significantly impact student performance are selected through statistical
and domain-driven analysis. Appropriate machine learning algorithms such as Decision Trees, Random
Forest, or Neural Networks are chosen based on the nature of the data. The model is trained on a labeled
dataset to learn patterns and correlations related to student success. After training, the model is
evaluated using metrics like accuracy, precision, recall, and F1-score to ensure its reliability. Once
validated, the model is used to predict student outcomes and flag at-risk students. A user-friendly
dashboard is developed to present insights to educators, along with personalized recommendations for
academic support. The system is tested extensively before deployment, and continuous feedback from
users is incorporated to refine and improve the model over time

1.6 ORGANIZATION OF THE REPORT

Chapter 1: Introduction – Introduces the project, its objectives, scope, and significance in the field of
education.
Chapter 2: Literature Survey – Reviews existing research, related work, and technologies relevant to AI
in education and student performance prediction.
Chapter 3: System Analysis – Defines the problem statement, identifies system requirements, and
describes the feasibility study.
Chapter 4: System Design and Methodology – Explains the architecture, data flow, algorithm selection,
and step-by-step methodology used for model development.
Chapter 5: Implementation – Details the development process, tools and technologies used, and
integration of machine learning models.
Chapter 6: Testing and Evaluation – Describes the testing strategies, evaluation metrics, and
performance analysis of the system.
Chapter 7: Results and Discussion – Presents the outcomes of the project, discusses model accuracy,
and interprets the predictions.
Chapter 8: Conclusion and Future Work – Summarizes the project, its contributions.

14
]]]

CHAPTER 2

LITERATURE REVIEW

The integration of Artificial Intelligence (AI) and Machine Learning (ML) into the educational domain
has significantly advanced the ability to predict student performance, identify at-risk students, and
enhance academic outcomes. Numerous studies have explored various ML techniques for modeling and
improving student success.

Smith and Patel [1] conducted one of the foundational studies in this domain, exploring the application
of diverse ML algorithms for predicting student performance. Their findings highlighted the
effectiveness of decision trees and support vector machines in accurately classifying student outcomes.
Similarly, Zhang and Wei [2] focused on the early prediction of at-risk students in higher education,
emphasizing the potential of early interventions informed by educational data mining techniques.

Johnson and Zhao [3] built upon this foundation by comparing multiple ML models, including logistic
regression, random forests, and neural networks, for predicting academic success among college
students. Their research identified neural networks as a strong candidate due to its ability to model
complex, non-linear relationships in educational data.

Kumar and Singh [4] presented a comprehensive survey of methods used to predict student dropout in
higher education. They reviewed a range of ML models and highlighted ensemble approaches as
particularly effective due to their improved accuracy and robustness over single-model methods.

Tan and Lim [5] implemented ensemble machine learning models to predict academic performance,
confirming that combining multiple algorithms, such as gradient boosting and random forests, yields
higher prediction accuracy. Their study demonstrated the value of ensemble learning in educational
contexts.

15
Gupta and Arora [6] explored the use of deep neural networks to model student performance in online
courses. Their work addressed the unique challenges posed by e-learning environments, including
limited interaction and variable engagement, and demonstrated that neural networks could effectively
learn from complex behavioral data.

Lee and Park [7] applied ML techniques for the early identification of at-risk high school students. Their
results showed that early intervention, based on predictive analytics, could significantly reduce dropout
rates and improve academic engagement.

Singh and Chauhan [8] conducted a comparative analysis of different ML algorithms used for student
performance prediction. They evaluated metrics such as accuracy, precision, recall, and F1-score, and
concluded that no single algorithm is universally superior, emphasizing the importance of context-
specific model selection.

Gupta and Sharma [9] expanded the scope by proposing AI-based predictive models aimed at enhancing
educational outcomes. Their research focused on integrating these models into real-world educational
platforms, offering a practical roadmap for AI deployment in academic institutions.

Finally, Roy and Kumar [10] introduced a deep learning approach for real-time prediction of student
success. Their model leverages recurrent neural networks (RNNs) to process time-series educational
data, enabling dynamic prediction and timely feedback mechanisms for students and educators alike.

In summary, the existing literature demonstrates the growing sophistication and applicability of AI
and ML techniques in the educational sector. These studies collectively establish a strong foundation for
the development of intelligent systems that support student success through proactive and personalized
interventions.

16
 CHAPTER 3

PROJECT DESCRIPTION

3.1 EXISTING SYSTEM

The current systems used in educational institutions largely rely on traditional methods of student
assessment and monitoring. Performance evaluations are typically based on periodic exams,
assignments, and attendance, offering limited insight into overall student progress. These systems lack
predictive capabilities and fail to identify at-risk students in advance. Academic counseling is often
reactive, initiated only after performance has declined. Some institutions use basic Learning
Management Systems (LMS) like Moodle or Blackboard, but these platforms offer limited data
analytics. While some AI tools exist, they are either costly, not customizable, or focus narrowly on
specific subjects. Most existing solutions do not integrate multiple data points such as behavior,
engagement, and historical trends for holistic analysis. There is a lack of automation in generating early
alerts or personalized interventions. Furthermore, educators receive little support in making data-driven
decisions. These gaps create the need for a more intelligent, scalable, and proactive system. Edu Predict
addresses these limitations using AI-powered predictive modeling and actionable insights.

3.2 PROPOSED SYSTEM

The proposed system, Edu Predict, is an AI-based solution designed to predict student academic
performance and support timely interventions. It collects and analyzes diverse data such as grades,
attendance, participation, and behavioral patterns. The system uses machine learning algorithms to
identify at-risk students before academic issues arise. It provides predictive insights and personalized
recommendations to educators and mentors. A user-friendly dashboard displays student progress and
risk levels for easy monitoring. Edu Predict supports real-time data updates and adaptive learning
strategies. The system aims to improve retention rates and academic outcomes through proactive
decision-making. It is scalable, secure, and customizable for institutions of various sizes. By
integrating multiple data points, Edu Predict delivers a holistic view of student performance. This
intelligent platform empowers educators to take informed, timely, and impactful actions.

17
3.2.1 ADVANTAGES

Early Identification of At-Risk Students: The system proactively detects students likely to
underperform, enabling timely academic support and intervention.
Data-Driven Decision Making: Educators and administrators can make informed decisions
based on accurate predictive insights rather than assumptions.
Personalized Learning Support: Recommendations are tailored to each student’s learning
needs, helping improve engagement and performance.
Improved Academic Outcomes: Timely interventions lead to better student success rates,
higher retention, and overall academic improvement.
Real-Time Monitoring: Continuous data tracking allows for dynamic updates and immediate
response to changes in student behaviour or performance.
Scalability and Adaptability: The system can be easily scaled across institutions and customized
to fit different curricula, student groups, and data sources.

3.3 FEASIBILITY STUDY

The Edu Predict system is technically feasible using readily available tools like Python, scikit-learn, and
educational datasets. It is operationally practical, offering a user-friendly interface for educators with
minimal training required. Economically, it is cost-effective due to the use of open-source technologies and
existing institutional infrastructure. The system is legally and ethically compliant, ensuring data privacy
and fairness in predictions. The project timeline is realistic, with each phase planned for smooth and timely
execution.

3.3.1 ECONOMIC FEASIBILITY

The Edu Predict system is economically feasible as it utilizes open-source tools and existing
institutional resources, minimizing development and deployment costs. No high-end hardware or
expensive software licenses are required. The long-term benefits, such as improved student

18
performance and reduced dropout rates, outweigh the initial investment. This makes the project a
cost-effective and sustainable solution for educational institutions.

3.3.2 TECHNICAL FEASIBILITY

The Edu Predict system is technically feasible as it leverages widely available technologies such as
Python, machine learning libraries (e.g., scikit-learn, pandas), and data visualization tools. The required
hardware and software are easily accessible and compatible with most educational institution
infrastructures. Additionally, the technical skills needed for development, deployment, and maintenance
are common among developers, ensuring smooth implementation and scalability.

3.3.3 SOCIAL FEASIBILITY

The Edu Predict system is socially feasible as it aims to enhance student success by providing
personalized support and early interventions, promoting equity in education. By helping identify at-risk
students, the system empowers educators to take proactive measures, ultimately improving student
retention and outcomes. The project is aligned with the increasing adoption of technology in education
and will be positively received by educators, students, and stakeholders focused on academic
improvement.

3.4 SYSTEM SPECIFICATION

The system requires a minimum of 4GB RAM, Intel i5 processor, and 100GB storage. Software
dependencies include Python 3.x, scikit-learn, pandas, and a web framework like Flask or Django. It
supports Windows, Linux, and macOS environments.

3.4.1 HARDWARE SPECIFICATION

Processor: Intel Core i5 or equivalent (minimum)

RAM: 4 GB (8 GB recommended for faster processing)
Hard Disk: 100 GB of available storage space
Display: 14-inch monitor or larger with 1366x768 resolution or higher

19
 Input Devices: Standard keyboard and mouse
Network: Internet connectivity for dataset access and updates.

3.4.2 SOFTWARE SPECIFICATION

Libraries & Frameworks:

scikit-learn (for machine learning algorithms)
pandas & NumPy (for data manipulation)
matplotlib & seaborn (for data visualization)
Flask or Django (for web-based interface, if applicable)
Database: SQLite or MySQL (for storing student data)
IDE/Code Editor: VS Code
Browser: Google Chrome

20
 CHAPTER 4

PROPOSED WORK

4.1 SYSTEM ARCHITECTURE

The system architecture of Edu Predict outlines the workflow of the AI-based prediction model designed
to assess and enhance student performance. The process begins with the user interacting through a user
interface, where student-related data is entered or accessed. This data is directed to the input module,
which communicates with the student database containing historical academic records and related
information

Fig:4.1: System Architecture

In Fig 4.1, The system then performs two main analysis steps using machine learning algorithms:

1. Naive Bayes (NB) is used for student performance analysis, helping identify patterns and classify
students based on existing data.

2. K-Nearest Neighbors (KNN) is applied for predicting student performance, based on similarities with
previously recorded data points.

Following these steps, the system carries out a probability calculation to determine the likelihood of

21
 academic success or risk for each student. These results are then used to generate insights and support.

4.2 BLOCK DIAGRAM

Figure 4.2: Block Diagram

This Figure 4.2 shows the best and perfect content explaining the Machine Learning Workflow for Edu Predict
ideal for your project report:

The diagram illustrates the complete machine learning pipeline used in the Edu Predict system to
forecast student success. It begins with a Student Dataset, which undergoes Data Preprocessing to
ensure quality and consistency. This preprocessing includes:

Data Cleaning: Removing irrelevant, duplicate, or missing entries.

Data Selection: Choosing relevant features and records that influence student performance.

Data Transformation: Normalizing or encoding data into formats suitable for machine learning.

Following preprocessing, the refined data moves into Feature Engineering, where new meaningful
features are created or selected to enhance model accuracy. This data is then sent to Data Storage, ready
for model development.

In the Machine Learning Model Development phase, the data is split into:

22
Training Set (80%): Used to train the model using suitable algorithms.

Testing Set (20%): Used to evaluate model performance and validate accuracy.

The chosen Machine Learning Algorithm (such as Naive Bayes, KNN, etc.) is applied during the
training phase. After Model Implementation, the system generates the Result, which predicts student
performance and identifies those who may need academic support.

This structured approach ensures accurate, data-driven predictions and supports better decision-making
in education.

4.2.1 USECASE DIAGRAM

Figure 4.2.1: Usecase Diagram

The Figure 4.2.1 is a Based on the use case diagram you've shared, here's a well-written and
detailed description of the use case model for your major project report on Edu Predict: Leveraging AI
for Student Success:

The use case diagram illustrates the core functionality of the Edu Predict system, designed to
predict student performance using machine learning techniques. The system is structured around the
interaction between external actors (users) and the internal processes that contribute to predictive
analytics in education.
23
 1. Actor1 (Data Provider / Admin / Faculty):

Responsible for collecting and uploading student-related data into the system.

Interacts with the system during the initial data collection, preprocessing, and feature extraction
stage.

2. Actor2 (Student / Faculty / End User):

Receives the predicted student performance output.

May also interact with the system to view reports or analytics based on predictions.

Use Cases:

1. Collect Data:

Actor1 initiates this use case.

Involves gathering data such as attendance records, grades, participation metrics, and other relevant

Academic information

2. Pre-processing:

This step ensures data is cleaned and transformed into a usable format.

Tasks may include handling missing values, normalization, and encoding categorical variables.

3. Feature Extraction:

Important characteristics or indicators of student performance are identified and extracted.

These features serve as inputs for the machine learning models.

4. Apply Machine Learning Algorithms:

The system uses algorithms such as decision trees, random forest, neural networks, or support
vector

machines to learn from the data.

24
 Actor1 can configure or initiate this process based on defined criteria.

5. Predict Student Performance:

Provides insights such as risk of failure, performance trends, or suggestions

4.2.2 ACTIVITY DIAGRAM

Figure 4.2.2 : Activity Diagram

In Fig 4.2.2, The activity diagram outlines the sequential flow of activities involved in the
EduPredict system, which aims to predict student performance using AI-driven machine learning
models. It provides a holistic view of the input sources, preprocessing stages, and final prediction
process.

1. Student Interaction and Data Capture

Student: The process begins with the student, who is a primary source of various data points.

Mark Attendance: Daily attendance data is collected, which helps assess consistency, engagement.

Extra-Curricular Activities: Participation in clubs, sports, or other non-academic activities is noted.

Ragging & Stress: Any behavioral, psychological, or social issues affecting student mental health.

Staff: Teachers and mentors assess the student based on observed behavior and performance.

25
 Staff Opinion on Students: Staff feedback is documented, offering qualitative insights.

Student Database: All collected data is stored in a centralized student database.

Preprocessing: The raw data is cleaned and standardized to handle missing values, inconsistencies.

Feature Extraction: Important attributes influencing academic performance

Apply Machine Learning Algorithms: Advanced ML models are applied to the refined dataset

4.3 MODULE DESCRIPTION

The Edu Predict system leverages machine learning models to analyze and predict student academic
performance with the goal of improving educational outcomes. The model is built using a well-
structured dataset containing historical student performance data, including attendance, internal
assessment scores, background factors, and other relevant attributes.

4.3.1 MODULE1 : Data Preprocessing and Feature Engineering Module

This module is responsible for preparing raw student data for model training. It includes tasks like
data cleaning (removing missing or inconsistent values), data transformation (normalization or
encoding), and feature selection. Feature engineering is also performed to extract meaningful
attributes that directly influence student performance. This ensures that the input data is accurate,
relevant, and optimized for machine learning.

4.3.2 MODULE 2 : Machine Learning Model Development Module

This module handles the development and training of machine learning algorithms such as Naive
Bayes and K-Nearest Neighbors (KNN). The dataset is split into training and testing sets, where the
model learns from historical data patterns. The goal is to accurately predict student outcomes such as
academic risk levels or performance categories. Model evaluation metrics are used to fine-tune and
validate the accuracy of predictions.

4.3.3 MODULE 3: Prediction and Visualization Module

26
 Once the model is trained, this module is responsible for deploying the model and providing real-
time predictions based on new student data. It presents results in a user-friendly interface with
visualizations like charts or dashboards. Educators can use these insights to identify at-risk students
early and take proactive actions. This module bridges technical analysis with practical application in
academic environments.

4.3.4 PROCESSING OF DATA

Data processing is a crucial step in the Edu Predict system that ensures the quality, reliability, and
usability of student-related data before it is fed into machine learning models. The process begins
with data collection from academic records, attendance logs, test scores, and other relevant sources.
Once gathered, the data undergoes a preprocessing phase which includes:
Data Cleaning: Removal of missing values, duplicates, and irrelevant entries to ensure consistency.
Data Transformation: Converting raw data into structured formats suitable for analysis, such as
normalizing numeric values and encoding categorical data.
Data Selection: Choosing the most relevant features that significantly influence student
performance, such as attendance, grades, and participation.
Feature Engineering: Creating new informative features or modifying existing ones to improve
model accuracy and predictive power.
After preprocessing, the refined dataset is split into training and testing subsets, ensuring balanced
learning. This clean and structured data is then used to train machine learning algorithms, enabling
the system to predict future academic outcomes effectively.

27
 CHAPTER 5

IMPLEMENTATION AND TESTING

5.1 INPUT AND OUTPUT

5.1.1 FEATURE IMPORTANCE ANALYSIS EDU PREDICT

● The correlation heatmap illustrates the relationship between various features in the student
dataset, highlighting how different variables interact and contribute to student success. The
values range from -1 to 1, where values closer to 1 represent a strong positive correlation and
values closer to 0 indicate weak or no correlation.

28
 Figure 5.1: Feature Importance Analysis Edu Predict

Activities Engaged and Days Studied show a very strong positive correlation (0.91).

Total Clicks has a high correlation with Activities Engaged (0.92) and Days Studied (0.80),
indicating that more platform interaction reflects increased study effort.
Assessments Completed is moderately correlated with both Final Result (0.36) and Average
Assessment Score (0.55), implying that frequent assessment participation contributes to better
outcomes.
Average Assessment Score correlates well with Final Result (0.46), validating its importance as a
performance indicator.
Code Module has weaker correlations with most features, indicating it's a more categorical or
independent factor.

5.1.2 FEATURE IMPORTANCE ANALYSIS

29
The feature importance graph illustrates the contribution of individual input variables in
determining the success of students, based on the model's learning during training. These scores,
measured using the F-score, reflect how frequently and effectively a feature was used to make
accurate splits in the decision trees of the model. A higher F-score indicates a greater impact on the
predictive accuracy of the model.

Figure 5.2: Feature Analysis

Assessments Completed – F-score: 10.10

This feature dominates the graph, making it the most influential predictor. It indicates that
the number of assessments a student has completed has a strong positive correlation with
their overall academic success. Students who consistently complete their assessments
demonstrate discipline, engagement, and knowledge retention, which naturally leads to
better results. The model heavily relies on this feature when making predictions.

Average Assessment Score – F-score: 4.16

Average scores across assessments provide direct insight into a student’s understanding of
the subject matter. This feature ranks second in importance, emphasizing that not just
completion but consistent performance in assessments is vital. High average scores signal
strong academic potential and mastery, making this an essential indicator of success.

Days Studied – F-score: 2.91

30
 The number of days a student actively studies or accesses learning material reflects their
study routine. While not as dominant as assessments, this feature plays a meaningful role.
Regular study habits contribute to better preparation, concept clarity, and overall
performance, which is why it holds a moderate level of importance.

Activities Engaged – F-score: 2.18

This feature measures student interaction with learning activities such as quizzes,
assignments, or discussions. Higher engagement levels indicate better involvement in the
learning process. Although its importance score is relatively lower, it still contributes to
understanding student dedication and active participation, which are essential for long-term
academic success.

Number of Previous Attempts – F-score: 2.15

This represents how many times a student has attempted the same course or subject. While it
may indicate perseverance, it can also be a sign of struggling students. The model uses this
feature to understand historical performance and learning curves, but its influence is
moderate compared to others.

Total Clicks – F-score: 2.12

Total clicks measure the interaction of students with the online learning system. Although it
provides some insights into how frequently students use the platform, it doesn’t always
equate to effective learning. Thus, it has the lowest F-score among all features in this
analysis, suggesting it's a supplementary indicator rather than a primary one.

5.2 MODEL PERFORMANCE

In Fig 5.3, The bar chart above depicts the ratio of classes correctly labeled by the model, providing a

clear overview of how well the AI-based system, Edu Predict, performs in categorizing students into
their respective academic outcomes Pass, Fail, Distinction, and Withdrawn. This evaluation is crucial in
understanding the model’s reliability and its real-world applicability in predicting student performance.

31
 Fig 5.3 : Model Classification Accuracy Analysis

Distinction – Accuracy ≈ 0.88

The model demonstrates the highest accuracy in predicting students who achieve Distinction. This
suggests that students who perform exceptionally well have distinct and easily identifiable patterns in
their academic behavior — such as high assessment scores, frequent engagement, consistent study
habits, and full assessment completion. These strong signals enable the model to differentiate
distinction-level students with confidence.
Fail – Accuracy ≈ 0.76
The second most accurate classification is the Fail category. Students at risk of failing often exhibit clear
risk indicators such as low scores, minimal engagement, missed assessments, and fewer study days. The
model is well-trained to recognize these patterns and flag such students early, making it a valuable tool
for academic intervention and support planning.
Withdrawn – Accuracy ≈ 0.69
Students who withdraw from courses are somewhat accurately classified by the model. While some
withdrawal patterns are predictable — such as sudden drops in activity or engagement — others may be
influenced by external, non-academic factors (e.g., personal, financial, or health reasons), which the
model may not fully capture. Hence, the slightly lower prediction accuracy in this class.
Pass – Accuracy ≈ 0.68

32
The model shows relatively lower accuracy for students who simply pass without distinction. This may
be due to overlapping characteristics with students who are borderline between pass and fail. Since their
behavior is not as distinctly marked as those who excel or struggle, the model finds it more challenging
to classify this group perfectly. Improving the model with finer feature engineering or additional
behavioral indicators could enhance performance in this category.

CHAPTER 6
RESULTS AND DISCUSSIONS

6.1 KPI METRICS OVERVIEW

The Edu Predict application dashboard serves as a smart, interactive, and user-centric interface
designed to assist students in tracking academic performance, identifying strengths, and receiving
personalized educational and career recommendations using AI.

33
 Fig 6.1 : Mobile Application Dashboard

1. The KPI Dashboard offers a snapshot of the student’s key academic indicators:
Current GPA (9.7): Reflects the overall academic achievement. This metric is dynamically
updated using predictive analytics based on ongoing performance.

Skill Assessment (85): AI evaluates the student's proficiency across critical domains and suggests
improvement areas.
Program Success (82): Predicts the likelihood of successfully completing the current academic
program using machine learning models.
Career Match (78): Shows how closely the student’s skills and academic progress align with
suitable career paths, generated using AI-based recommendation systems.
These KPIs are derived from real-time data and predictive algorithms to help students make data-
informed decisions.

The proposed system can achieve good performance with relatively smaller datasets compared to more
34
data-hungry models, reducing data requirements and training time. It is also well-suited for tasks such
as image denoising, image compression, anomaly detection, and image-to-image translation, providing
versatility in application areas. The proposed system can also generalize well to unseen data and is
capable of adapting to different domains with minor adjustments to training strategies.

6.2 GPA TREND ANALYTICS

The GPA Trend Chart under the “Analytics” tab provides a visual representation of the student’s
academic progression over time:
Tracks semester-wise or term-wise GPA changes.
Helps identify periods of academic growth or decline.
AI analyses the trend and can trigger alerts or recommendations if a significant drop is predicted.
This feature supports early intervention and motivates students to stay consistent in their academic
journey.

35
 Fig 6.2 : GPA Trend Analysis

Helps identify periods of academic growth or decline.

AI analyses the trend and can trigger alerts or recommendations if a significant drop is predicted.
This feature supports early intervention and motivates students to stay consistent in their academic
journey.

6.3 QUICK ACTION CENTER

The “Quick Actions” section provides easy access to key student resources:

Higher Education: Guides students with insights and suggestions on further studies based on their
academic profile.

Career Paths: Offers career roadmaps and AI-generated career suggestions based on student
interests and skill assessments.

36
 Fig 6.3 : Quick Action Center

Connect Educator: Facilitates direct interaction with academic advisors or mentors for personalized
guidance and support.
This module makes the application not just predictive, but also proactive and supportive in student
development.

6.4 RECENT ACTIVITY TRACKER

The Recent Activities Panel logs all major interactions by the student:
Skill Assessment Completed
Program Recommendation Updated
Career Path Simulation Viewed

37
 Fig 6.4: Recent Activities Overview

This feature allows students and administrators to track engagement levels and ensure that the system
recommendations are being utilized effectively. The Recent Activities section in the Edu Predict
application serves as a detailed activity log that tracks and displays the latest interactions performed by
the student within the platform. This component plays a critical role in both user engagement and
behavioral analysis for the AI model.

The logged data is directly fed into the model’s behavioral learning engine, which improves prediction
accuracy and recommendation relevance over time. The AI interprets activity frequency and type to
assign engagement scores and update prediction weights.

CHAPTER 7
38
 CONCLUSION AND FUTURE ENHANCEMENTS

7.1 CONCLUSION

The Edu Predict project represents a significant advancement in the application of Artificial
Intelligence for academic success and student performance prediction. By integrating powerful
machine learning models with educational datasets, the system provides actionable insights,
personalized recommendations, and predictive analytics that empower students to make informed
academic decisions.
Throughout the project, multiple modules were developed to collect, process, and analyze student
behaviour, such as assessment performance, learning activity engagement, and skill progression. The
system's dashboard offers a user-friendly interface that visualizes key performance indicators (KPIs)
like GPA trends, skill assessment scores, and program success rates—making data-driven learning
more accessible and engaging.
With high accuracy in classifying students into categories such as Pass, Fail, Distinction, and
Withdrawn, the model proves to be a reliable academic companion. It not only assists students in
identifying areas of improvement but also helps institutions in early intervention and personalized
guidance.
In conclusion, Edu Predict bridges the gap between data and success by creating a smart learning
environment that is adaptive, supportive, and future-ready.

7.2 FUTURE ENHANCEMENTS

To further elevate the impact and usability of Edu Predict, several future enhancements can be
implemented. One major improvement would be the integration of real-time analytics, allowing
students and educators to receive instant feedback and progress updates based on live data from
learning platforms. Additionally, incorporating an AI-powered career recommendation engine could
guide students toward suitable professional paths by analyzing their academic strengths, interests, and
skill assessment results. The inclusion of a smart chatbot for 24/7 academic support would also enhance
user experience, providing instant assistance for queries related to coursework, deadlines, and subject

39
matter. Expanding the system with dedicated portals for parents and educators can encourage
collaborative efforts in monitoring student growth and planning personalized interventions. Another
potential enhancement involves the development of a predictive dropout prevention model using
behavioral and engagement data to identify at-risk students early. Gamifying learning tasks through
badges, rewards, and milestones can boost student motivation and engagement. Lastly, enabling
multilingual support and improving accessibility features will ensure that Edu Predict remains
inclusive, user-friendly, and widely adaptable across diverse learning environments.

40
 CHAPTER 8

SOURCE CODE & POSTER PRESENTATION

8.1 SAMPLE CODE

<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0"/>
<title>EDU-PREDICT - AI Student Success Predictor</title>
<script src="[Link] crossorigin="anonymous"></script>
<style>
:root {
--primary: #007bff;
--secondary: #28a745;
--light-bg: #f0fdf4;
--text-dark: #333;
--text-light: #fff;
--gradient: linear-gradient(to right, #007bff, #28a745);
--font-heading: 'Segoe UI', sans-serif;
--font-body: 'Arial', sans-serif;
}

body {
font-family: var(--font-body);
margin: 0;
background: #e6f7f9;
}

.container {
max-width: 700px;
41
 margin: 40px auto;
background: #fff;
padding: 30px;
border-radius: 15px;
box-shadow: 0 8px 20px rgba(0, 0, 0, 0.1);
animation: fadeIn 0.8s ease-out;
}

header {
text-align: center;
margin-bottom: 30px;
}

header h1 {
font-family: var(--font-heading);
font-size: 2.5rem;
color: var(--primary);
}

header p {
font-size: 1.1rem;
color: #444;
margin-top: 10px;
font-style: italic;
}

label {
display: block;
margin: 15px 0 5px;
font-weight: bold;
color: var(--text-dark);
}

input[type="number"] {
width: 100%;
padding: 10px;
border: 2px solid var(--secondary);
border-radius: 6px;
42
 font-size: 1rem;
}

button {
margin-top: 25px;
width: 100%;
padding: 12px;
background: var(--gradient);
color: white;
font-size: 1.1rem;
border: none;
border-radius: 6px;
cursor: pointer;
transition: background 0.3s;
}

button:hover {
background: linear-gradient(to right, #0056b3, #1e7e34);
}

.page {
display: none;
}

.[Link] {
display: block;
}

.result-box {
background: var(--light-bg);
padding: 25px;
border-radius: 12px;
text-align: center;
color: var(--text-dark);
animation: popIn 0.6s ease-out;
}

.result-box h2 {
43
 color: var(--primary);
font-size: 2rem;
}

.result-box p {
font-size: 1.2rem;
margin-top: 15px;
}

.icon {
font-size: 3rem;
color: var(--secondary);
margin-bottom: 10px;
}

@keyframes fadeIn {
from { opacity: 0; transform: translateY(20px); }
to { opacity: 1; transform: translateY(0); }
}

@keyframes popIn {
from { opacity: 0; transform: scale(0.9); }
to { opacity: 1; transform: scale(1); }
}

.back-button {
margin-top: 20px;
display: inline-block;
color: var(--primary);
text-decoration: underline;
cursor: pointer;
}

</style>
</head>
<body>

<!-- PAGE 1: Input -->

44
<div class="container page active" id="formPage">
<header>
<h1> EDU-PREDICT</h1>
<p>AI-based Student Success Predictor</p>
</header>

<form id="studentForm">
<label>Physics (out of 100):</label>
<input type="number" id="physics" required />

<label>Chemistry (out of 100):</label>

<input type="number" id="chemistry" required />

<label>Maths (out of 100):</label>

<input type="number" id="maths" required />

<label>Attendance (out of 10):</label>

<input type="number" id="attendance" required />

<label>Discipline (out of 10):</label>

<input type="number" id="discipline" required />

<label>Student Rating (out of 10):</label>

<input type="number" id="rating" required />

<label>Extra Curricular (out of 10):</label>

<input type="number" id="extra" required />

<button type="submit"><i class="fas fa-rocket"></i> Predict Now</button>

</form>
</div>

<!-- PAGE 2: Results -->

<div class="container page" id="resultPage">
<header>
<h1> Prediction Results</h1>
<p>Your personalized success insights</p>
</header>
45
 <div class="result-box">
<div class="icon" id="resultIcon"><i class="fas fa-star"></i></div>
<h2 id="behaviorResult">Behavior: </h2>
<p id="suggestionResult">Suggestion: </p>
</div>

<div class="back-button" onclick="goBack()"><i class="fas fa-arrow-left"></i> Back to Form</div>

</div>

<script>
const form = [Link]("studentForm");

[Link]("submit", function(e) {
[Link]();

const physics = parseFloat([Link]("physics").value);

const chemistry = parseFloat([Link]("chemistry").value);
const maths = parseFloat([Link]("maths").value);
const attendance = parseFloat([Link]("attendance").value);
const discipline = parseFloat([Link]("discipline").value);
const rating = parseFloat([Link]("rating").value);
const extra = parseFloat([Link]("extra").value);

const totalScore = physics + chemistry + maths +

(attendance * 10) + (discipline * 10) +
(rating * 10) + (extra * 10);

const percentage = (totalScore / 700) * 100;

let behavior = "";

let icon = "";
if (percentage > 80) {
behavior = "Excellent";
icon = "fas fa-thumbs-up";
} else if (percentage > 60) {
behavior = " Good";
icon = "fas fa-smile";
46
 } else {
behavior = "⚠ Needs Improvement";
icon = "fas fa-exclamation-circle";
}

let suggestion = "";

if ([Link]("Excellent") && physics > 80 && chemistry > 80 && maths > 80) {
suggestion = " Suggested Course: Engineering (Great command over PCM)";
} else if ([Link]("Good") && physics > 60 && chemistry > 60 && maths > 60) {
suggestion = " Suggested Course: [Link]. or Technical Degree";
} else if (extra > 7) {
suggestion = " Great in Extra Curricular! Explore creative fields like Media, Sports, or Design.";
} else if (discipline > 8 && attendance > 8) {
suggestion = " Strong discipline & attendance! Consider Civil Services or Management.";
} else if (rating > 7) {
suggestion = " Great personality! Consider Communication, PR, or Teaching.";
} else {
suggestion = " Keep improving and explore Diploma or Skill-Based programs. You're on the right path!";
}

[Link]("behaviorResult").textContent = behavior;
[Link]("suggestionResult").textContent = suggestion;
[Link]("resultIcon").innerHTML = <i class="${icon}"></i>;

[Link]("formPage").[Link]("active");
[Link]("resultPage").[Link]("active");
});

function goBack() {
[Link]("resultPage").[Link]("active");
[Link]("formPage").[Link]("active");
}
</script>
</body>
</html>

47
 REFERENCES

1.J. Smith and R. Patel, "Predicting student performance using machine learning techniques,"
Int. J. Educ., vol. 45, no. 3, pp. 123-134, 2020.
2.L. Zhang and J. Wei, "Early prediction of at-risk students in higher education," Educ.
Data Mining Conf., pp. 85-94, 2021.
3. K. Johnson and Q. Zhao, "Machine learning models for predicting academic success of
college students," J. Educ. Res., vol. 61, no. 2, pp. 200-210, 2021.
4. A. Kumar and P. Singh, "A survey on predicting student dropout in higher education
using ML," Educ. Technol. Res., vol. 32, no. 4, pp. 301-310, 2022.
5. Y. Tan and J. Lim, "Predicting academic performance of students using ensemble
machine learning models," J. Learn. Anal., vol. 9, no. 1, pp. 45-56, 2022.
6. R. Gupta and S. Arora, "Predictive modeling of student performance in online courses
using neural networks," J. Online Learn., vol. 5, no. 2, pp. 112-124, 2022.
7. H. Lee and S. Park, "Application of machine learning for early identification of at-risk
students in high school," Int. Conf. Artif. Intell., pp. 128-137, 2023.
8. S. Singh and A. Chauhan, "A comparative study of machine learning algorithms for
student performance prediction," Educ. Technol. Rev., vol. 37, no. 2, pp. 211-223, 2023.
9. S. Gupta and N. Sharma, "Enhancing educational outcomes with AI-based predictive
models," Springer [Link]., vol. 16, pp. 202-213, 2023.
10. P. Roy and S. Kumar, "Towards real-time prediction of student success:A deep learning
approach," IEEE Trans. Educ., vol. 67, no. 1, pp. 58-69, 2024.

48
49
 SRM INSTITUTE OF SCIENCE AND TECHNOLOGY
(Deemed to be University u / s 3 of UGC Act, 1956)

Office of Controller of Examinations

REPORT FOR PLAGIARISM CHECK ON THE DISSERTATION / PROJECT REPORT FOR UG / PG
PROGRAMMES
(To be attached in the dissertation / project report)

1 Name of the JACK ANDRE J

Candidate (IN ABISHEK RAJ M
BLOCK CHARUDEVE KS
LETTERS)
Bharathi Salai, Chennai -600089
2 Address of Candidate Mobile Number: +91 98401 64161,
+91 91508 90969,
+91 97911 50465
RA2011026020147, RA2011026020131, RA2011026020139
3 Registration Number

4 Date of Birth 17/06/2002, 25/03/2002, 06/08/2002

5 Department Computer Science and Engineering with

specialization in Artificial Intelligence and Machine
Learning

6 Faculty Engineering and Technology

7 Title of the Dissertation / TITLE

Project

Individual or group: Group

(Strike whichever is not applicable)

8 Whether the above a) If the project / dissertation is done in group, then how many
project / dissertation is students together completed the project : 03
done by b) Mention the Name and Register number of
other candidates:
JACK ANDRE J [RA2011026020147], ABISHEK RAJ M
[RA2011026020131], CHARUDEVE KS [RA2011026020139]

ADDRESS OF GUIDE
9 Name and address
of the Supervisor / Mail ID :
Guide
Mobile Number :

50
 NA
10 Name and address of
the Co- Mail ID: NA Mobile Number: NA
Supervisor /Guide

11 Software Used Turnitin

12 Date of Verification DD/MM/YYYY

13 Plagiarism Details: (to attach the final report from the software)

Percentage Percentag % of plagiarism after

C of similarity e of excluding Quotes,
h Title of the Report index similarity Bibliography, etc.,
a (including index
pt self citation) (Excludin
er g self
citation)
TITLE
1 NA NA 10%

Appendices NA NA NA

We declare that the above information has been verified and found true to the best of our knowledge.

Name and Signature of the Staff

Signature of the Candidate ( Who uses the plagiarism check software )

Name and Signature of the Supervisor / Name and Signature of the Co-Supervisor / Co-Guide
Guide

Dr. XXX
Name and Signature of the HOD

51
.

52
53
54
55
56
57