System Requirements Document (SRD)

Project Name: raven-photo

1. Introduction

The raven-photo project is an innovative platform designed to revolutionize the online shopping experience by combining AI-based body structure and skin tone analysis with a virtual try-on feature. Users can upload their photos to receive personalized clothing color recommendations and visualize garments on their own body using advanced AI techniques like GANs and image warping.

This document outlines the system requirements for developing both a mobile application and a website to deliver this functionality. The goal is to create a seamless, premium experience for users across platforms while also enabling integration with e-commerce businesses.

2. System Overview

The raven-photo system will consist of two primary components:

1. Mobile Application: A cross-platform app (iOS and Android) built using Flutter to provide users with an intuitive and interactive experience for photo analysis, recommendations, and virtual try-on.
2. Website: A responsive web application built using ReactJS to offer similar functionality for users who prefer desktop or browser-based interactions.

Both platforms will leverage a shared backend powered by Python and FastAPI, with AI models for skin tone detection, body structure analysis, and garment overlay. The system will also include APIs for integration with e-commerce platforms and cloud storage for secure handling of user-uploaded images.

3. Functional Requirements

Mobile Application

• As a User, I should be able to upload a photo for analysis.
• As a User, I should receive personalized clothing color recommendations based on my skin tone and body structure.
• As a User, I should be able to select garments and see how they look on my uploaded photo using the virtual try-on feature.
• As a User, I should be able to browse a catalog of recommended garments.
• As a User, I should be able to save my analysis results and try-on images for future reference.
• As a User, I should be able to share my try-on images on social media.
• As a User, I should be able to purchase recommended garments directly through affiliate links.

Website

• As a User, I should be able to upload a photo for analysis via the website.
• As a User, I should receive personalized clothing color recommendations on the website.
• As a User, I should be able to use the virtual try-on feature on the website.
• As a User, I should be able to browse and filter garments by categories and recommendations.
• As a User, I should be able to log in and access my saved analysis and try-on results.
• As a User, I should be able to access a help section or FAQs for guidance.

4. User Personas

1. End User (B2C)

• Description: Individual shoppers looking for personalized clothing recommendations and virtual try-on experiences.
• Goals: Find clothing that suits their skin tone and body structure, visualize garments before purchase, and shop confidently online.

2. E-commerce Partner (B2B)

• Description: Online retailers and fashion brands integrating raven-photo’s API for enhanced user experiences on their platforms.
• Goals: Increase customer engagement and sales by offering personalized recommendations and virtual try-on features.

3. Admin

• Description: Internal team managing the platform, monitoring performance, and handling user queries.
• Goals: Ensure smooth operation of the system, manage user data securely, and resolve technical issues.

5. Visuals Colors and Theme

Unique Color Palette for raven-photo

• Background: #F7F8FA (Soft Mist White)
• Surface: #FFFFFF (Pure White)
• Text: #2C2C2C (Charcoal Black)
• Accent: #FF6F61 (Coral Red)
• Muted Tones: #A3A3A3 (Soft Gray)

This palette reflects a premium, minimalist aesthetic with a focus on clarity and elegance, aligning with the high-end fashion-tech identity of raven-photo.

6. Signature Design Concept

Interactive Virtual Dressing Room

The homepage of both the app and website will feature an Interactive Virtual Dressing Room.

• Visuals: A 3D-rendered dressing room with a mirror at the center. The mirror will display the user’s uploaded photo, dynamically updated with their selected garments.
• Animations: Smooth transitions as users swipe through clothing options, with garments appearing to "float" onto their body in the mirror.
• Interactions: Users can drag and drop garments onto their photo, rotate the view to see different angles, and zoom in for details.
• Micro-interactions: Subtle haptic feedback on mobile when selecting garments, and hover effects on the website for interactive elements.
• Color Shifts: The background lighting of the dressing room will subtly change based on the user’s skin tone, creating a personalized ambiance.

This bold and immersive concept will make the raven-photo experience unforgettable and set it apart from competitors.

7. Non-Functional Requirements

• Performance: The system should process photo uploads and analyses within 5 seconds.
• Scalability: The backend must handle up to 10,000 concurrent users during peak times.
• Security: User photos and data must be encrypted in transit and at rest.
• Accessibility: The app and website must comply with WCAG 2.1 AA standards.
• Localization: Support for multiple languages, starting with English and Hindi.

8. Tech Stack

Frontend

• Mobile App: Flutter
• Website: ReactJS

Backend

• Language: Python
• Framework: FastAPI

Database

• RDBMS: MySQL (for structured data)
• NoSQL: MongoDB (for unstructured data)

AI Models

• Skin Tone Detection: Pretrained CNN with OpenCV
• Virtual Try-On: Pix2Pix GAN and DensePose

DevOps

• Containerization: Docker
• Orchestration: Kubernetes

9. Assumptions and Constraints

• Users will upload high-quality images for accurate analysis.
• The system will initially target users in India (IN) and support INR for transactions.
• The virtual try-on feature will work best with front-facing body images.

10. Glossary

• GAN: Generative Adversarial Network, used for generating realistic garment overlays.
• DensePose: A tool for mapping 2D images to 3D body models.
• WCAG: Web Content Accessibility Guidelines, ensuring accessibility for users with disabilities.