Closet ๐๐๏ธ


Project Report
Abstract
Closet is an innovative AI-powered fashion assistant designed to enhance user experience in selecting outfits, hairstyles, makeup, and accessories. The platform integrates AI-driven virtual try-on features, customizable avatars, and personalized styling recommendations. With the rise of digital fashion solutions, Closet aims to bridge the gap by offering a seamless, interactive, and inclusive styling platform. ๐๐ญ
Problem Overview
Fashion enthusiasts often struggle with selecting the perfect outfit, accessories, and hairstyles that complement their style. Existing platforms offer isolated services such as outfit recommendations or virtual try-ons, but none provide a comprehensive, all-in-one styling assistant. Closet eliminates these challenges by consolidating multiple fashion services into a single AI-driven platform. Common pain points include:
Lack of personalized outfit recommendations ๐
Difficulty in visualizing outfits before purchasing ๐
Need for an all-in-one styling solution ๐
Absence of inclusive fashion suggestions for all genders ๐
App Engineering
Prospective Tech Stack
Frontend: React.js for a dynamic user interface, React Native for mobile app development ๐ฑ.
Backend: Node.js with Express.js for server-side logic.
Database: MongoDB or PostgreSQL for structured and scalable data storage.
AI Integration: Machine learning models for outfit and style recommendations ๐ง .
Cloud Hosting: AWS โ๏ธ or Vercel for efficient and scalable deployment.
Frontend
Pages: Home ๐ , User Dashboard ๐, Virtual Try-On ๐ญ, Outfit Suggestions ๐, Custom Avatar Creator ๐ค, Hairstyle & Makeup Ideas ๐.
UX: Focus on intuitive navigation, responsive design ๐, and engaging visuals to enhance user experience.
Backend
Sort by Preferences: AI-powered sorting of outfits and accessories.
User Registration: Secure authentication ๐ and profile customization.
AI Virtual Try-On: Real-time outfit visualization using uploaded images.
Database
Types of Data:
Users: id, name, email, password ๐
Outfits: id, category, brand, color, price ๐ท๏ธ
Avatars: id, user_id, hairstyle, body type ๐งโ๐จ
Accessories: id, type, outfit_id ๐
Current Solutions
While there are various platforms addressing different aspects of fashion assistance, no single solution comprehensively covers all needs in one application. Existing options include:
- Dresso-An online fashion platform that offers AI-driven outfit recommendations and virtual styling solutions.
Dresso
- StyleSnap-An AI-powered visual search tool by Amazon that helps users find fashion items by uploading images.
StyleSnap
- Zeekit-A virtual try-on technology that allows users to see how clothes will look on different body types before purchasing.
Zeekit
Comparison
Feature | Dresso | StyleSnap | Zeekit | Closet |
Price | Varies | Varies | Free | Competitive ๐ฐ |
Core Features | Outfit matching | Visual search | Virtual try-on | All-in-one fashion solution ๐ |
Missing Features | AI outfit suggestions | Hairstyle ideas | Accessories recommendations | None ๐ฏ |
Issues | Limited AI scope | No custom avatars | No gender-inclusive styles | Comprehensive โ |
User Documentation Provided
Detailed user guide ๐
FAQs for troubleshooting common issues โ
Future Improvements or Enhancements
Expand AI capabilities for seasonal outfit suggestions ๐โ๏ธ.
Integrate voice-assisted outfit recommendations ๐๏ธ.
Potential Risks or Areas for Further Attention
Continuous AI model updates for improved styling suggestions ๐.
Scalability of the platform for a larger user base ๐.
Team Members ๐ฅ
Alisha Kaur
Ashish Yadav
Mansha Soni
Pranjal Pandey
Vaishnavi Sagariya
Acknowledgment
We extend our heartfelt gratitude to our mentor Smaranjit Ghose for their invaluable guidance throughout this project. A special thanks to our team members for their dedication, collaborative spirit, and constructive feedback, which greatly shaped Closet. We also appreciate the open-source community for providing essential tools and resources. Lastly, we thank our families and friends for their unwavering support and motivation. โค๏ธ
Subscribe to my newsletter
Read articles from Pranjal Pandey directly inside your inbox. Subscribe to the newsletter, and don't miss out.
Written by
