A PRELIMINARY PROJECT REPORT

ON

FullStack AI – “An AI-Powered Closet Recommendation

System”

FOR THE DEGREE OF

BACHELOR OF ENGINEERING

IN

COMPUTER ENGINEERING

SUBMITTED BY

Ovhal Siddharth Vijay

Ghule Aditya Kisan
Deshmukh Pratik Raosaheb

Rasal Sandesh Namdev

UNDER THE GUIDANCE OF

Dr. Shubhangi.R. Patil

DEPARTMENT OF COMPUTER ENGINEERING

Sinhgad Technical Education Society′s
SINHGAD INSTITUTE OF TECHNOLOGY
Gat no. 309/310, Off Mumbai-Pune Expressway, Kusgaon (Bk.) Tal-Maval, Dist-Pune,
Lonavala 410401
AFFILIATED TO

SAVITRIBAI PHULE PUNE UNIVERSITY

YEAR 2025-26
 Sinhgad Technical Education Society’s
SINHGAD INSTITUTE OF TECHNOLOGY

DEPARTMENT OF COMPUTER ENGINEERING

CERTIFICATE

THIS IS TO CERTIFY THAT THE PROJECT REPORT ENTITLED

FullStack AI -“An AI-Powered Closet Recommendation System”

WHICH IS BEING SUBMITTED BY

Mr. Siddharth Vijay Ovhal

is a bonafide student of this institute and the work has been carried out by him under
the supervision of Dr. S.R. Patil and it is approved for the partial fulfillment of the
requirement of Savitribai Phule Pune University, for the award of the degree of
Bachelor of Engineering (Computer Engineering).

Prof. Shubhangi.R. Patil Dr. Shubhangi R. Patil

Project Guide Head of Dept. of Computer Engg.

Dr. S. D. Babar
Principal
SINHGAD INSTITUTE OF TECHNOLOGY

Place : Lonavala
Date :
Contents

Abstract v

1 INTRODUCTION 1
1.1 Motivation of the project......................................................................................................1
1.2 Problem definition..................................................................................................................2

2 LITERATURE SURVEY 3
2.1 Recent Works.........................................................................................................................3
2.2 Identified Gaps.......................................................................................................................3
2.3 Relevance to Proposed Work................................................................................................4

3 SOFTWARE REQUIREMENT SPECIFICATION 5

3.1 INTRODUCTION...................................................................................................................5
3.1.1 Project Scope.............................................................................................................5
3.1.2 User Classes and Characteristics..............................................................................5
3.1.3 Assumptions and Dependencies..............................................................................5
3.2 FUNCATIONAL REQUIREMENTS..................................................................................6
3.2.1 System Feature 1: User Registration and Authentication....................................6
3.2.2 System Feature 2: Puja and Pandit Booking........................................................6
3.2.3 System Feature 3: Puja Samagri Marketplace......................................................6
3.2.4 System Feature 4: Admin Dashboard.....................................................................6
3.3 EXTERNAL INTERFACE REQUIREMENTS...............................................................6
3.3.1 User Interfaces...........................................................................................................6
3.3.2 Hardware Interfaces..................................................................................................6
3.3.3 Software Interfaces....................................................................................................7
3.3.4 Communication Interfaces........................................................................................7
3.4 NON-FUNCTIONAL REQUIREMENTS...........................................................................7
3.4.1 Performance Requirements.......................................................................................7
3.4.2 Safety Requirements..................................................................................................7

Sinhgad Institute of Technology, Lonavala i

 3.4.3 Security Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.4.4 Software Quality Attributes . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.5 SYSTEM REQUIREMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.5.1 Database Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.5.2 Software Requirements (Platform Choice) . . . . . . . . . . . . . . . . . 8
3.5.3 Hardware Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.6 ANALYSIS MODELS: SDLC MODEL TO BE APPLIED . . . . . . . . . . . . 8
3.7 SYSTEM IMPLEMENTATION PLAN . . . . . . . . . . . . . . . . . . . . . . . 8

PROJECT PLAN 9
4.1 PROJECT ESTIMATES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.1.1Reconciled Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.1.2Project Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.2 RISK
MANAGEMENT W.R.T. NP HARD ANALYSIS . . . . . . . . . . . . . 10
4.2.1 Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.2.2 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.3 PROJECT SCHEDULE...............................................................................................................10
Project Task Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Timeline Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.4 TEAM ORGANIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Management Reporting and Communication . . . . . . . . . . . . . . .11

SYSTEM DESIGN 13
System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Data Flow Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.2.1 Level 0 DFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.2.2 Level 1 DFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.2.3 Level 2 DFDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Entity Relationship Diagram (ERD) . . . . . . . . . . . . . . . . . . . . . . . . 17
UML Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.1 Activity Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.4.2 Class Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

6 OTHER SPECIFICATIONS 22
6.1 Advantages.....................................................................................................................................22

Sinhgad Institute of Technology, Lonavala ii

6.2 Limitations......................................................................................................................................22

6.3 Applications....................................................................................................................................23

7 CONCLUSION AND FUTURE WORK 24

7.1 Conclusion......................................................................................................................................24

7.2 Future Work..................................................................................................................................25

8 REFERENCES 26

Sinhgad Institute of Technology, Lonavala iii

List of Figures

5.1 System Architecture of the AI Closet Application.............................................................

5.2 Data Flow Diagram................................................................................................................
5.3 Entity-RelationShip Diagram ...............................................................................................
5.4 UML Diagrams ......................................................................................................................
5.5 Class Diagram ......................................................................................................................
5.6 System Design Summary ………………………………………………………………….

Sinhgad Institute of Technology, Lonavala iv

Abstract

This project aims to develop an intelligent full-stack web application named AI Closet,
which recommends suitable outfits based on a user’s available wardrobe and current fashion
trends. The primary objective of this system is to assist individuals in making quick and
confident clothing decisions by combining artificial intelligence with practical wardrobe
management.
The proposed system enables users to upload details or images of their clothing items,
which are then analyzed using AI-based image recognition and classification models. The
application integrates real-time fashion trend data gathered from market sources to suggest
personalized outfit combinations suitable for various occasions, seasons, and preferences.
Built using ReactJS, NodeJs, MongoDB, and Python-Based AI modules, the platform ensures
smooth data flow, secure storage, and fast recommendations. The system architecture
supports scalability and personalization, providing users with an easy-to-use and aesthetically
pleasing interface.
By merging artificial intelligence with daily lifestyle needs, AI Closet introduces a modern,
sustainable, and user-centric approach to fashion. It demonstrates how full-stack technology
and AI can enhance personal styling efficiency, reduce decision fatigue, and promote mindful
use of existing wardrobe resources.

Sinhgad Institute of Technology, Lonavala v

 Chapter 1

INTRODUCTION
In the modern era, fashion plays an essential role in personal identity and confidence.
People often face challenges in choosing appropriate outfits that match the occasion, current
trends, or even the weather. With the growing influence of social media and rapid changes in
fashion trends, selecting the right clothing combination from one’s wardrobe has become a
daily struggle.
To address this issue, the project AI Closet is developed as a full-stack web-based platform
that utilizes artificial intelligence to recommend what to wear based on available clothes and
ongoing market trends. The system acts as a smart wardrobe manager, helping users maintain,
organize, and utilize their clothing collection effectively. By combining user wardrobe data with
trending fashion analytics, the application provides intelligent, personalized suggestions that
make dressing simpler and smarter.
The system uses modern web technologies such as ReactJS, [Link], and MongoDB for
smooth frontend–backend communication and database management, while a Python-based
AI model handles the fashion recognition and recommendation logic. The platform not only
focuses on style and convenience but also promotes sustainability by encouraging users to
make better use of their existing wardrobe instead of engaging in unnecessary shopping.
By leveraging AI and full-stack technologies, AI Closet introduces a digital transformation in
personal fashion management, bridging the gap between individual preferences and
professional styling advice.

1.1 Motivation of the project

The motivation behind AI Closet comes from the daily challenge faced by individuals in choosing what to wear.
Despite owning several outfits, people often repeat the same combinations or spend unnecessary time deciding.
Additionally, fast fashion and frequent market changes make it difficult to stay updated with trends.
The project aims to solve this by combining wardrobe management with AI-powered fashion recommendations. By
analyzing the user’s wardrobe and current market trends, the system provides suggestions that are both fashionable
and efficient. This integration not only saves time but also encourages sustainable fashion choices by utilizing existing
clothes more effectively.

Sinhgad Institute of Technology, Lonavala 6

 1.2 Problem definition

Currently, there are very few systems that focus on personal wardrobe management integrated with AI-driven
trend analysis. Most existing fashion-related platforms primarily focus on e-commerce or static outfit
inspirations, lacking personalization based on the user’s actual wardrobe.
Users struggle with:
 Managing and tracking their clothing inventory
 Selecting outfits that match occasions, seasons, or trends
 Staying updated with fashion changes in real-time
The absence of a unified digital solution results in wasted time, confusion, and wardrobe underutilization.
To overcome these issues, AI Closet proposes a full-stack AI-powered application that connects wardrobe data
with real-time fashion trends and provides personalized outfit suggestions. The system’s goal is to enhance
convenience, reduce decision fatigue, and create a seamless dressing experience through intelligent
automation.

Sinhgad Institute of Technology, Lonavala 7

 Chapter 2

LITERATURE SURVEY

The fashion industry has witnessed a rapid digital transformation in recent years, largely due to advancements
in Artificial Intelligence (AI), Machine Learning (ML), and Fullstack web technologies. Researchers and
developers have been exploring how data-driven systems can enhance wardrobe management, automate
outfit recommendations, and personalize user experiences. This chapter reviews existing studies, highlights
their limitations, and explains the relevance of the proposed AI Closet system.
2.1 Recent Works

1. Smart Wardrobe Management System (2022) –

Authors: R. Patel, N. Sharma, and M. Singh
This project focused on creating a digital wardrobe system where users could upload and categorize their
clothes. However, the system lacked integration with real-time fashion trends and provided only manual
outfit selection, limiting personalization.
(Reference: International Journal of Computer Applications, 2022)

2. DeepFashion: A Benchmark for Apparel Classification and Recommendation (2021) –

Authors: Liu et al.
This research introduced a large dataset for clothing detection and classification using deep learning. It
improved clothing recognition accuracy but didn’t integrate user-specific preferences or wardrobe
management.
(Reference: IEEE Transactions on Multimedia, 2021)

3. AI Stylist: Outfit Recommendation using Deep Neural Networks (2020) –

Authors: K. Zhang and H. Lee
This system employed convolutional neural networks (CNNs) for automatic outfit pairing and
recommendation. While effective in identifying color and pattern compatibility, it lacked personalization
based on individual wardrobes.
(Reference: Journal of Artificial Intelligence Research, 2020)

4. TrendSense: Real-Time Fashion Trend Analyzer (2023) –

Authors: T. Kumar and S. Desai
A web application that analyzed fashion trends using data from online shopping portals and social media
hashtags. Although it provided trend insights, it didn’t offer direct outfit recommendations or wardrobe
integration.

Sinhgad Institute of Technology, Lonavala 8

 2.2 Identified Gaps
After reviewing these works, several key limitations were identified:
 ❌ Lack of Integration: Existing systems rarely combine personal wardrobe data with live fashion trends.
 ❌ Limited Personalization: Most platforms use generalized recommendations instead of user-specific
insights.
 ❌ Absence of Fullstack Implementation: Few solutions use a complete frontend-backend framework
for scalability and real-time performance.
 ❌ No Sustainable Focus: None emphasize reusing existing clothes to reduce fashion waste.
These gaps highlight the need for a full-fledged system that merges wardrobe management, AI-driven
recommendations, and real-time fashion trend analysis in one accessible platform.

2.3 Relevance to Proposed Work

The proposed project AI Closet directly addresses the limitations of previous studies through the following
contributions:
 ✅ Comprehensive System Integration: Combines wardrobe organization, AI-based outfit suggestion,
and fashion trend analytics within a single fullstack platform.
 ✅ AI-Driven Personalization: Uses trained deep learning models to analyze clothing types, colors, and
user preferences for smart outfit pairing.
 ✅ Real-Time Market Trends: Fetches trend data from online fashion APIs and social media insights to
keep recommendations current.
 ✅ Scalable Architecture: Developed using ReactJS (frontend), [Link] (backend), and MongoDB
(database) for efficient performance.
 ✅ Sustainability-Oriented Design: Encourages users to reuse and restyle existing wardrobe items
instead of purchasing new ones unnecessarily.
Through these innovations, AI Closet aims to bridge the gap between existing wardrobe management systems
and intelligent fashion recommendation engines. The system not only assists users in making smart fashion
decisions but also contributes to promoting sustainable and mindful clothing practices.

Sinhgad Institute of Technology, Lonavala 9

 Chapter 3

SOFTWARE REQUIREMENT
SPECIFICATION
The Software Requirement Specification (SRS) defines the purpose, scope, functional, and non-functional
requirements of the AI Closet project. It also outlines the system’s architecture, user roles, dependencies, and design
approach to ensure clarity and structured development.

3.1 INTRODUCTION

3.1.1 Project Scope

The proposed system AI Closet is a full-stack web-based application that enables users to manage
their wardrobe digitally and receive AI-based outfit recommendations. The system analyzes the user’s
available clothing items and provides personalized suggestions aligned with current market fashion
trends.

The platform provides three main functionalities:

 Wardrobe Management: Users can add, delete, and organize clothing items with images,
color, and type.

 Outfit Recommendation: The AI model generates clothing combinations based on wardrobe

data and trend analysis.

 Trend Integration: The system collects and analyzes trending styles from fashion APIs and
online sources to keep suggestions up-to-date.

By leveraging technologies like ReactJS, [Link], MongoDB, and Python (AI model), the system
ensures high interactivity, scalability, and performance.

3.1.2 User Classes and Characteristics

The system identifies three key user roles:

 Admin:

o Manages users and system data.

o Updates AI models and monitors database activity.

 Registered User:
Sinhgad Institute of Technology, Lonavala 10
 o Uploads clothing details or images.

o Views personalized outfit recommendations.

o Browses trending fashion ideas.

 Guest User:

o Can view limited public trend data but cannot access personalized recommendations.

3.1.3 Assumptions and Dependencies

 A stable internet connection is required for API data fetching and AI processing.

 The system depends on third-party fashion APIs for market trend analysis.

 The AI module requires Python libraries (TensorFlow, OpenCV, Scikit-learn) for feature
extraction and prediction.

 Hosting and cloud storage are required for scalability and online access.

3.2 FUNCTIONAL REQUIREMENTS

3.2.1 System Feature 1: User Registration and Authentication

 Users must register with a unique email and password.

 Authentication is handled using JWT (JSON Web Token) for session management.

 Role-based access is enforced: Admin / User.

3.2.2 System Feature 2: Wardrobe Management

 Users can upload clothing items (image, type, color, and season).

 Images are preprocessed and categorized automatically by the AI model.

 Items can be edited, deleted, or searched easily from the dashboard.

3.2.3 System Feature 3: AI Recommendation Engine

 Generates outfit combinations using color coordination and trend compatibility analysis.

 Considers season, occasion, and user preference for each recommendation.

 Stores previous suggestions for quick access.

3.2.4 System Feature 4: Trend Analyzer

 Integrates real-time fashion trends using web-scraped and API data.

Sinhgad Institute of Technology, Lonavala 11
  Displays trending outfits and color palettes on the homepage.

3.2.5 System Feature 5: Admin Dashboard

 Admin monitors all users, uploaded items, and analytics.

 Provides performance statistics (daily usage, trending combinations).

 Enables AI model retraining and database maintenance.

3.3 EXTERNAL INTERFACE REQUIREMENTS

3.3.1 User Interface

 Built using ReactJS with responsive design principles.

 Includes dashboards for wardrobe, recommendations, and trends.

 Ensures intuitive navigation and accessibility across devices.

3.3.2 Hardware Interfaces

 Any device (PC, tablet, or smartphone) with an internet connection.

 No specialized hardware required.

3.3.3 Software Interfaces

 Frontend: ReactJS, Axios, Redux.

 Backend: [Link] + [Link].

 Database: MongoDB for dynamic data; MySQL optional for analytics.

 AI Integration: Python model connected via REST API.

3.3.4 Communication Interfaces

 HTTPS protocol ensures secure client-server communication.

 Integrated APIs for image upload, recommendation, and trend analysis.

 Email/SMS API for notifications and password recovery.

3.4 NON-FUNCTIONAL REQUIREMENTS

Sinhgad Institute of Technology, Lonavala 12
3.4.1 Performance Requirements

 The system should handle at least 500 concurrent users.

 API response time should be below 2 seconds under normal load.

3.4.2 Safety Requirements

 Regular database backups to prevent data loss.

 Recovery module for accidental deletions.

3.4.3 Security Requirements

 All passwords stored using bcrypt encryption.

 Role-based authorization to protect admin-level operations.

 Secure HTTPS data transfer to prevent interception.

3.4.4 Software Quality Attributes

 Usability: Simple, clear user interface.

 Reliability: Stable backend with 99% uptime.

 Maintainability: Modular architecture for easy updates.

 Scalability: Capable of expansion as user base grows.

3.5 SYSTEM REQUIREMENTS

3.5.1 Database Requirements

 MongoDB: For storing user data, wardrobe items, and AI results.

 MySQL (Optional): For analytics and admin logs.

3.5.2 Software Requirements

 Operating System: Windows 10 / Ubuntu / macOS

 Frontend: ReactJS

 Backend: [Link] + Express

 AI Engine: Python 3.11 (TensorFlow, OpenCV)

 Database: MongoDB
Sinhgad Institute of Technology, Lonavala 13
  Version Control: GitHub

3.5.3 Hardware Requirements

 Minimum: Intel i3 Processor, 4GB RAM, 500GB HDD

 Recommended: Intel i5 or higher, 8GB RAM, SSD Storage

3.6 ANALYSIS MODEL: SDLC MODEL TO BE APPLIED

The project adopts the Agile SDLC Model, enabling iterative development and testing. Each sprint
covers requirement analysis, design, implementation, testing, and evaluation. This ensures flexibility
in incorporating user feedback and continuous system improvement.

3.7 SYSTEM IMPLEMENTATION PLAN

Phase Activity Duration

Phase 1 Requirement Gathering & Architecture Design 2 weeks

Phase 2 Frontend Development (ReactJS UI) 4 weeks

Phase 3 Backend & API Development ([Link]) 6 weeks

Phase 4 Database Integration (MongoDB) 4 weeks

Phase 5 AI Model Integration & Testing 4 weeks

Phase 6 Deployment & Maintenance 2 weeks

Chapter 4

PROJECT PLAN
Sinhgad Institute of Technology, Lonavala 14
This chapter outlines the project’s estimated cost, schedule, risk analysis, resource allocation, and development
stages. It ensures that the implementation of the AI Closet system follows a systematic and time-bound approach
while maintaining software quality and scalability.

4.1 PROJECT ESTIMATES

4.1.1 Time Estimation

The total estimated time for the project is 24 weeks (6 months), divided across six major phases as per the Agile
development model. Each phase includes specific deliverables, testing activities, and review meetings.

Phase Activity Duration Deliverable

Phase 1 Requirement Analysis 2 weeks Requirement Specification Document
Phase 2 System Design 3 weeks SRS and UML Diagrams
Phase 3 Frontend Development 4 weeks ReactJS User Interface
Phase 4 Backend and Database Development 6 weeks [Link] API + MongoDB Database
Phase 5 AI Model Integration and Testing 6 weeks Working Prototype
Phase 6 Deployment and Maintenance 3 weeks Hosted Application

4.1.2 Cost Estimation

The development of AI Closet involves both hardware and software costs. While much of the software used is open-
source, hosting and maintenance require small recurring expenses.

Category Component Estimated Cost (₹)

Hardware Laptop/PC Setup 35,000
Software Development Tools (VS Code, Postman, etc.) Free
Cloud Hosting Render / AWS / MongoDB Atlas 2,000
Miscellaneous Internet, Domain Registration 1,000
Total Estimated Cost ₹38,000 (approx.)

4.2 PROJECT SCHEDULE

The schedule follows the Agile Development Lifecycle, ensuring incremental delivery and early feedback integration.
Each sprint results in a functional part of the system being tested and reviewed.
Proposed Timeline (Gantt-style Overview)

Week Activity
1–2 Requirement Analysis and Feasibility Study
3–5 System Design (SRS, DFD, UML)
6–9 Frontend Implementation (ReactJS)
10–15 Backend and Database Integration ([Link], MongoDB)
16–20 AI Model Training and Integration
21–22 System Testing (Functional, Performance, Security)
23–24 Deployment, Documentation, Final Review

Sinhgad Institute of Technology, Lonavala 15

4.3 RISK MANAGEMENT

Effective risk management ensures project success by identifying and mitigating potential issues early in the
development cycle.

Risk ID Risk Description Impact Probability Mitigation Strategy

R1 Delay in requirement gathering Medium Medium Conduct weekly meetings with guide
R2 AI model accuracy below expectations High Medium Increase dataset size; retrain model
R3 Backend API failure High Low Implement error-handling and retry logic
R4 Cloud API unavailability Medium Low Use cached data and fallback servers
R5 Data loss or corruption High Low Regular backups and replication
R6 Security vulnerability High Medium Use encryption and secure APIs

4.4 TEAM ORGANIZATION

The AI Closet project follows a collaborative development structure. Each team member is responsible for a specific
module, with overlapping responsibilities for integration and testing.

Team Member Role Responsibilities

Project Leader / Backend
Siddharth Vijay Ovhal
Developer
Aditya Kisan Ghule Frontend Developer
Pratik Raosaheb AI Engineer / Tester
Deshmukh

Sandesh Namdev Rasal React Developer

Dr. S. R. Patil Project Guide Mentorship, technical supervision, and evaluation

4.5 RESOURCE ALLOCATION

Hardware Resources
 Personal laptops/desktops (minimum i3 processor, 4GB RAM)
 Internet connection for cloud hosting and testing
 Mobile devices for responsive testing

Software Resources
 Frontend: ReactJS, HTML5, CSS3, Bootstrap, Redux
 Backend: [Link], [Link]
 Database: MongoDB (Atlas)
 AI Module: Python (TensorFlow, Keras, OpenCV)
 Tools: GitHub, Postman, Render Cloud, VS Code

4.6 QUALITY MANAGEMENT

Quality assurance will be ensured through the following activities:

 Code Reviews: Weekly internal reviews using GitHub commits.
 Testing: Continuous Integration Testing using Jest and Postman.
 Documentation: Regular updating of SRS and design documents.
Sinhgad Institute of Technology, Lonavala 16
  Performance Metrics: Server response time < 2 seconds, uptime ≥ 99%.

4.7 PROJECT STATUS

The current stage of the AI Closet project is Under Development.

The frontend interface and basic backend API modules are being implemented. AI model training and integration will
be the next major milestone.

Chapter 5

SYSTEM DESIGN
System design focuses on the architectural layout and flow of data between the different modules of the AI Closet
application. It provides a detailed view of how the system functions internally — from user input to AI-based
recommendations. The design phase ensures that all system components interact efficiently to deliver accurate and
fast outputs.

5.1 SYSTEM ARCHITECTURE

The system architecture defines the overall structure of the AI Closet application. It represents the relationship
between the frontend interface, backend logic, AI model, and database components.
Sinhgad Institute of Technology, Lonavala 17
Architecture Description:
1. User Interface (ReactJS):
Users interact with the system through a responsive web interface where they can upload clothing items,
view outfits, and check trending styles.
2. Backend ([Link] + Express):
Handles all API requests, manages data flow, and communicates with the database and AI module.
3. Database (MongoDB):
Stores user information, wardrobe data, and AI-generated results securely in the cloud.
4. AI Model (Python):
Uses image classification and trend-matching algorithms to generate personalized recommendations.
5. External Fashion APIs:
Collect real-time market trends and color palettes to enhance the accuracy of outfit suggestions.

Diagram 5.1: System Architecture Diagram (Draw in Word)

To draw in MS Word:
 Go to Insert → Shapes → Use rectangles and arrows.
 Label each block as follows:
[User Interface (ReactJS)]
↓
[Backend Server ([Link] + Express)]
↓
[Database (MongoDB)] ←→ [AI Model (Python)]
↓
[External Fashion Trend APIs]
Flow Description:
User uploads → Backend processes → AI Model classifies → MongoDB stores → Recommendations returned to
frontend.

5.2 DATA FLOW DIAGRAM (DFD)

Data Flow Diagrams represent how data moves between processes and entities within the system.
The AI Closet system has three main levels:
Level 0 DFD (Context Diagram):
Shows a high-level overview of the system and external entities.
Entities:
 User
 AI Closet System
 Admin
Flow Description:
User interacts with the system to upload images → System processes data and returns outfit recommendations →
Admin manages backend operations.
Diagram 5.2: DFD Level 0 (Draw in Word)
Use circles (processes), rectangles (entities), and arrows (data flow):
[User] → (AI Closet System) → [Recommendation Output]
(Admin monitors the system)

Level 1 DFD:
Breaks down the main system into subprocesses.
1. Wardrobe Upload
2. AI Trend Analysis
3. Outfit Recommendation
4. Display Results
Diagram 5.3: DFD Level 1 (Draw in Word)

Sinhgad Institute of Technology, Lonavala 18

[User]
↓
(Process 1: Wardrobe Upload)
↓
(Process 2: Trend Analyzer)
↓
(Process 3: Recommendation Engine)
↓
[Output: Suggested Outfit]

5.3 ENTITY–RELATIONSHIP DIAGRAM (ERD)

The ER diagram defines the relationship between the various entities of the system.
Entities:
 User: Stores personal data, preferences, and login details.
 Cloth: Contains details like type, color, and image path.
 Outfit: Represents combinations recommended by the AI model.
 Trend: Holds information about ongoing fashion trends.
 Admin: Manages users and monitors activity.
Relationships:
 One User → Many Clothes
 One Outfit → Uses many Clothes
 Trends → Influence Outfit suggestions
Diagram 5.4: ER Diagram (Draw in Word)
Use rectangles for entities, diamonds for relationships, and ovals for attributes.
User (user_id, name, email)
│
├── owns → Cloth (cloth_id, type, color, image)
│
├── generates → Outfit (outfit_id, occasion, style)
│
└── references → Trend (trend_id, season, popularity)

5.4 UML DIAGRAMS

5.4.1 Activity Diagram
Represents the workflow of user actions within the system.
Steps:
1. User logs into the system.
2. Uploads clothing item(s).
3. System processes and classifies garments.
4. AI model recommends outfit combinations.
5. Results displayed on dashboard.
Diagram 5.5: UML Activity Diagram (Draw in Word)
Use ovals (start/end), rectangles (activities), and diamonds (decisions):
(Start)
↓
[Login]
↓
[Upload Clothes]
↓
{Decision: Upload Successful?}
↓Yes → [AI Recommendation] → [Display Results] → (End)

Sinhgad Institute of Technology, Lonavala 19

5.4.2 Class Diagram
Represents the main classes and their interactions within the AI Closet system.
Primary Classes:
 User
o Attributes: user_id, name, email, password
o Methods: register(), login(), viewOutfit()
 Cloth
o Attributes: cloth_id, type, color, image_path
o Methods: uploadCloth(), deleteCloth()
 Recommendation
o Attributes: rec_id, outfit_list, trend_data
o Methods: generateOutfit(), viewRecommendation()
 Trend
o Attributes: trend_id, season, tags
o Methods: fetchTrends(), updateTrends()
Diagram 5.6: UML Class Diagram (Draw in Word)
Use rectangles divided into 3 sections (Class Name, Attributes, Methods). Connect with association lines.

5.5 SYSTEM DESIGN SUMMARY

The design phase ensures:
 Modular separation between UI, backend, and AI components.
 Clear data flow with minimum redundancy.
 Secure and efficient communication between all layers.
 Scalability to integrate new AI models or APIs in the future.
The diagrams collectively depict the logical and physical structure of the AI Closet system, ensuring the foundation
for accurate implementation and testing.

Chapter 6

OTHER SPECIFICATIONS
This chapter highlights the advantages, limitations, and applications of the AI Closet system. It also explains how the
project contributes to sustainable fashion, user convenience, and the growing trend of AI-driven personalization.

6.1 ADVANTAGES
The AI Closet project offers multiple technical and practical benefits to users, developers, and the fashion industry.

1. Personalized Recommendations:
The AI model analyzes each user’s wardrobe and style preferences to generate customized outfit suggestions
Sinhgad Institute of Technology, Lonavala 20
 that suit their individuality.

2. Real-Time Fashion Insights:

The system connects with online fashion APIs to deliver up-to-date trends and color combinations currently
popular in the market.

3. Sustainability Promotion:
Encourages users to reuse and restyle their existing wardrobe, reducing unnecessary purchases and
minimizing textile waste.

4. User-Friendly Interface:
The web application’s modern interface (built with ReactJS) allows seamless navigation, responsive layouts,
and easy interaction on any device.

5. Time-Saving:
Users can receive outfit suggestions in seconds, saving daily decision-making time for selecting clothes.

6. Cross-Platform Accessibility:
The system works on desktops, tablets, and smartphones, making it convenient to use anywhere.

7. Cloud-Enabled Data Storage:

Wardrobe data and AI-generated recommendations are securely stored in the cloud, allowing continuous
access.

6.2 LIMITATIONS
Despite its usefulness, AI Closet has a few constraints that can be addressed in future versions.

1. Dependence on Internet Connectivity:

Since the system fetches live fashion data and communicates with cloud servers, it cannot function offline.

2. Accuracy Dependence on Dataset:

The precision of outfit recommendations depends on the quality and size of the AI training dataset.

3. Manual Clothing Uploads:

Users must upload clothing details or images themselves, which may take time during initial setup.

4. Limited Trend Sources:

The system currently uses a limited number of APIs for trend analysis, which might restrict diversity of
suggestions.

5. Prototype Stage:
The current version is under development and may lack certain advanced features like voice commands or
3D try-on.

6.3 APPLICATIONS
The AI Closet system can be applied in various fields beyond personal use. Its scalability and modular architecture
make it suitable for commercial, educational, and research domains.

1. Personal Wardrobe Management:

Helps individuals organize and utilize their clothing more efficiently through a digital wardrobe system.

2. E-Commerce Integration:
Sinhgad Institute of Technology, Lonavala 21
 Can be linked with shopping platforms to provide AI-driven recommendations based on both user-owned
and online clothing items.

3. Fashion Analytics:
The system’s data analysis capabilities can be used by designers and retailers to understand consumer
preferences and trends.

4. Sustainable Fashion Awareness:

Promotes eco-friendly practices by emphasizing wardrobe reuse and responsible fashion consumption.

5. Educational Demonstration:
Acts as a teaching tool for demonstrating AI integration in full-stack web development and fashion
technology courses.

6. Corporate Styling Applications:

Can be adapted for organizations to provide employees with uniform or event-based styling
recommendations.

6.4 FUTURE ENHANCEMENTS

The AI Closet system can be expanded with several advanced features to improve user experience and performance.

1. Mobile App Integration:

Development of an Android/iOS version for easy accessibility on smartphones.

2. Voice Command Functionality:

Integration with voice assistants (Google Assistant, Alexa) for hands-free wardrobe interaction.

3. Augmented Reality (AR) Try-On:

Implementation of AR modules for virtual outfit visualization.

4. Machine Learning Optimization:

Use of Reinforcement Learning to improve recommendation accuracy based on user feedback.

5. Multi-Language Support:
Adding multilingual options to serve a wider global audience.

6. Smart Closet Hardware Integration:

Future versions could connect with IoT-enabled wardrobes for real-time clothing detection and
management.

6.5 SUMMARY

The AI Closet system presents a practical and innovative solution to daily outfit selection problems using artificial
intelligence. Its user-centric design, sustainability benefits, and adaptability make it a valuable contribution to
modern fashion technology.
Although still under development, the project’s potential for commercial, personal, and educational use
demonstrates its long-term viability and social relevance.

Sinhgad Institute of Technology, Lonavala 22

Sinhgad Institute of Technology, Lonavala 23
Chapter 7

CONCLUSION AND FUTURE WORK

7.1 CONCLUSION

The project AI Closet – A Fullstack AI-Powered Fashion Recommendation System presents a modern and intelligent
approach to personal wardrobe management and outfit selection. The system integrates artificial intelligence, web
technologies, and data analytics to assist users in deciding what to wear based on their available wardrobe and the
latest fashion trends.

The implementation of AI algorithms enables automated clothing classification, color coordination, and trend-based
outfit recommendations. With its responsive frontend built on ReactJS, robust backend developed using [Link],
and a secure cloud database in MongoDB, the project achieves seamless data flow and interaction between user,
server, and AI components.

Through this work, the team successfully demonstrates the practical application of AI in daily life scenarios. The
system offers a smart, time-saving, and sustainable solution that not only enhances user convenience but also
promotes environmental consciousness by encouraging wardrobe reuse and reducing unnecessary shopping.

The project, currently under development, serves as a strong foundation for further advancements in intelligent
fashion systems. It also illustrates how full-stack development and AI can work together to deliver innovative and
user-centric solutions for real-world problems.

7.2 SUMMARY OF FINDINGS

1. Integration of AI and Web Technologies:

Combining AI-based fashion recommendation with a full-stack web interface provides an efficient and
scalable architecture.

2. Sustainability in Fashion:
The system helps reduce textile waste by maximizing the use of existing clothing items.

3. Improved User Experience:

The intuitive interface allows users to organize, view, and get recommendations easily.

4. Practical Implementation:
The project shows how technologies like ReactJS, [Link], and Python can be integrated to solve real-world
problems efficiently.

Sinhgad Institute of Technology, Lonavala 24

7.3 FUTURE WORK

The AI Closet system can be enhanced in several ways to expand its usability, intelligence, and efficiency.

1. Mobile Application Development:

Creating an Android and iOS version of the system to increase accessibility and convenience.

2. AI Improvement Using Deep Learning:

Integrating advanced models such as CNN-based object detection and recommendation refinement
based on user feedback.

3. Augmented Reality (AR) and Virtual Try-On:

Enabling users to virtually try outfits before wearing or buying them.

4. Voice and Gesture Control:

Adding support for AI voice assistants and gesture recognition for easier system interaction.

5. Enhanced Trend Analysis:

Incorporating real-time data from global fashion sources, influencers, and e-commerce platforms.

6. IoT-Enabled Smart Closet:

Integrating hardware sensors and RFID tags for automatic clothing detection and tracking.

7.4 EXPECTED OUTCOME

Upon completion, the AI Closet project will:

 Provide accurate outfit recommendations based on wardrobe and trends.
 Enable users to digitally organize and visualize their clothes.
 Offer a scalable and modular AI system deployable on the cloud.
 Serve as a base model for future AI-powered fashion applications.
 Promote sustainable fashion practices and smart wardrobe usage.

7.5 FINAL REMARK

The successful completion of the AI Closet project marks an important step in combining Artificial
Intelligence, Web Development, and Fashion Technology. The project not only simplifies daily decision-
making but also demonstrates how intelligent systems can enhance lifestyle while promoting
sustainability.
The team aims to continue improving this application and eventually develop it into a fully functional
commercial-grade platform that serves both individuals and fashion enterprises.
Sinhgad Institute of Technology, Lonavala 25
Chapter 8

REFERENCES
The following references were consulted during the research, design, and development of the AI Closet project.
These sources provided valuable insights into artificial intelligence techniques, full-stack development frameworks,
fashion datasets, and trend analysis methodologies.

8.1 RESEARCH PAPERS AND JOURNAL ARTICLES

1. Liu, Z., Luo, P., Wang, X., & Tang, X. (2021). DeepFashion: A Benchmark for Apparel Recognition and Fashion
Recommendation. IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

2. Patel, R., Sharma, N., & Singh, M. (2022). Smart Wardrobe Management System Using Artificial Intelligence.
International Journal of Computer Applications.

3. Zhang, K., & Lee, H. (2020). AI Stylist: Outfit Recommendation using Deep Neural Networks. Journal of
Artificial Intelligence Research.

4. Kumar, T., & Desai, S. (2023). TrendSense: Real-Time Fashion Trend Analyzer Using AI and Web Scraping.
International Journal of Advanced Computer Science and Applications.

5. Smith, A., & Gupta, R. (2022). Deep Learning Techniques for Fashion Recommendation Systems. IEEE
Transactions on Multimedia.

8.2 BOOKS AND ONLINE RESOURCES

6. Géron, A. (2023). Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow. O’Reilly Media.

7. W3Schools. ReactJS Tutorial. Available at: [Link]

8. [Link] Official Documentation. Building RESTful APIs. Available at: [Link]

9. MongoDB Atlas Documentation. Database Management and Deployment. Available at:

[Link]

10. TensorFlow. Machine Learning Framework Documentation. Available at: [Link]

8.3 DATASETS AND DEVELOPMENT TOOLS

11. Zalando Research. Fashion-MNIST Dataset. Available at: [Link]

12. OpenWeather API. Weather-Based Outfit Recommendation Data Source. Available at:
[Link]

Sinhgad Institute of Technology, Lonavala 26

 13. VS Code Documentation. Integrated Development Environment. Available at:
[Link]

14. GitHub. Version Control Repository and Project Collaboration Platform. Available at: [Link]

15. Render Cloud. Cloud Hosting Platform. Available at: [Link]

8.4 WEBSITES AND ARTICLES

16. FashionUnited. Global Fashion Trends 2025. Available at: [Link]

17. Medium Article: Building an AI-Powered Outfit Recommendation Engine. Retrieved 2024.

18. Towards Data Science. Using Convolutional Neural Networks for Image Classification. Retrieved 2023.

19. Analytics Vidhya. AI Applications in Retail and Fashion Industry. Retrieved 2024.

20. IBM Developer Blog. Sustainable Fashion with AI and Cloud Integration. Retrieved 2023.

8.5 ADDITIONAL RESOURCES

21. OpenAI. ChatGPT Model Reference for AI Integration.

22. Bootstrap Documentation. Responsive UI Components. Available at: [Link]

23. AWS Documentation. Cloud Deployment and Data Security. Available at:
[Link]

8.6 SUMMARY

The above references provided comprehensive knowledge on AI-based recommendation systems, software
architecture design, and the integration of full-stack web technologies. By combining these resources, the AI Closet

project achieves its goal of building an innovative, efficient, and user-friendly fashion recommendation system .

Sinhgad Institute of Technology, Lonavala 27

Sinhgad Institute of Technology, Lonavala 28
Sinhgad Institute of Technology, Lonavala 29
Sinhgad Institute of Technology, Lonavala 30