Advanced E-Commerce
VirtuFit: Cross-Platform AR Shopping Assistant
with AI-Based Fit & Style Recommender
Team Members:
C. Lebi Maria
S. Varshini
M. Sreeja
J. Renu Sri
Team mentor:
Mrs. C. Manonmani
Project Overview
Introducing VirtuFit
Cross-Platform AR Shopping Assistant with AI-Based Fit & Style Recommender
Addressing Virtual Shopping Bridging the Gap Real-time Personalized
Gaps VirtuFit bridges this gap between
Experience
Virtual shopping often lacks realism in virtual try-on and accurate style Users can try on clothes and
size, fit, and personalization, leading prediction using Augmented Reality accessories, receive accurate fit
consumers to rely on assumptions or and an AI Fit Engine. suggestions, and shop confidently—all
trial-and-error purchasing. within a standalone, cross-platform
application.
Our platform works seamlessly on both desktop and mobile devices, eliminating the need for specialized hardware or third-party plugins.
VirtuFit's Vision
Transforming the Shopping Experience
Interactive Try-On Eliminate Guesswork
To transform online and offline shopping by introducing a To eliminate guesswork, size confusion, and product returns
truly interactive, intelligent, and inclusive try-on experience. by applying AI-driven fit estimation, ensuring customers find
their perfect match.
Privacy-First Design Accessible Innovation
To offer a privacy-first, installable application that functions To prove that innovative AR can be lightweight, accessible,
independently without browser reliance or social media and scalable, extending beyond what current retail
platforms. technology offers.
Current Limitations
Critical Gaps in Today’s AR E-
Commerce Systems
No Real-Time Fit Intelligence No Human-Like Style Recommendation
Most AR try-ons simply overlay a product visually without Current systems offer suggestions based on popularity, not
calculating fit, size compatibility, or body measurements. For what truly suits the user's body type, skin tone, or occasion.
example, Lenskart shows glasses on your face but doesn’t Amazon's suggestions, for instance, are click-based, not
confirm if the frame fits your face width. personalized.
AR as a Toy, Not a Tool Privacy is a Blind Spot
Present AR experiences often function as marketing gimmicks Apps frequently access live camera feeds without user
or filters, lacking integration with the cart or persistence for awareness or control, posing privacy risks. Snapchat filters, for
real decision-making. Sephora's try-on, for example, instance, have been flagged for running the camera in the
disappears after preview without AI follow-up. background during privacy reviews.
Our Game-Changing Approach
Our Disruption: From Overlay to
Intelligent Fitting System
AI-Powered Real-Time Fit Engine Humanized Personal Style Assistant
Utilizes MediaPipe + AI to calculate fit compatibility with real- Introduces an AI stylist that understands user body shape,
time body dimensioning, warning users about incorrect sizes preferences, and goals, recommending smart choices like a
before purchase. real assistant.
Full AR-to-Purchase Flow Privacy by Design
Integrates AR try-on, fit analysis, recommendation, add to cart, All visual tracking is performed locally with no server-side
and checkout into one unified, core shopping journey. processing of camera input, ensuring user trust and control
over privacy.
Technical Foundation
Development Stack & Tools Overview
Programming Languages Frontend Tools Backend & Database
• Python: Core backend logic, AI/ML • [Link]: Desktop app framework. • [Link] + [Link]: API handling
models. • Flutter/React Native: Cross- and routing.
• JavaScript: Electron (desktop), Web platform mobile UI. • MongoDB Atlas: Cloud-based
integration. • Tailwind CSS: Fast, responsive UI design. NoSQL database.
• Dart/TypeScript: Flutter or React • • Firebase Auth: Secure login & user
Figma: UI/UX prototyping and
Native (mobile). wireframes. sessions (optional).
APIs & Services
We leverage Cloudinary or Firebase Storage for 3D model/image hosting, OAuth 2.0 for secure login, and Postman for robust API
testing and debugging.
Key Technologies & Roadmap
AI, AR Libraries & Project Timeline
AI & ML Libraries Updated Project Timeline (2 Months / 8 Weeks)
• MediaPipe / OpenPose: Real-time face/body dimension detection.
Research Week 1–2 Market study, UI/UX wireframes,
• TensorFlow Lite: Lightweight ML for fit recommendation. & Design dataset collection
• scikit-learn / pandas: For user style modeling & logic.
• OpenCV: Camera feed preprocessing (if required). Frontend Week 3–4 Desktop/mobile UI setup, base
MVP navigation, component layout
AR Technologies
AI + AR Week 5–6 Body detection, product overlay, fit
• Unity 3D + Vuforia SDK: For AR markerless product try-ons. Integratio logic development
• [Link] (fallback): Lightweight web-embedded AR alternative. n
Backend/ Week 7 User login, cart flow, product
• 8thWall / WebXR: Optional web-based AR preview integration.
API management, DB schema
Testing & Week 8 Bug fixes, UI polish, app bundling, and
Packagin presentation/demo prep
g
Thank You!!!