FitScan AI: An Intelligent Outfit Recognition System Based on Laravel and FastAPI

FitScan AI is an intelligent outfit recognition system combining Laravel backend and FastAPI AI services, supporting image upload, intelligent cropping, outfit classification, and fashion recommendations. The project adopts a front-end and back-end separation architecture, demonstrating the integration of traditional web development and machine learning services, which is of reference value for learners of full-stack AI application development.

Original Author and Source

• Original Author/Maintainer: Melawati (@melawatii)
• Source Platform: GitHub
• Original Project Name: smart-outfit-recognition / FitScan AI
• Original Link: https://github.com/melawatii/smart-outfit-recognition
• Release Date: May 29, 2026

Project Overview

FitScan AI is a complete intelligent outfit recognition web application. Laravel handles the web interface and user interaction, while FastAPI hosts the AI inference service, retaining the advantages of the PHP web ecosystem while leveraging Python's machine learning capabilities. The core function is that users upload clothing photos, which are cropped and then analyzed by AI to provide outfit classification and suggestions, forming a complete process loop.

Technical Architecture Analysis

• Frontend Layer: Laravel Blade template engine + Tailwind CSS to build responsive interfaces; JavaScript handles image upload and cropping
• Backend Layer: Laravel 10 (PHP8.1+) is responsible for request processing, session management, file upload, and communication with AI services
• AI Service Layer: FastAPI framework builds the inference service (deployed on Hugging Face Spaces), using TensorFlow/Keras pre-trained models to recognize clothing categories at the bottom

Core Functional Flow

1. Image Upload and Preprocessing: Users upload photos and interactively crop the target area
2. AI Inference and Classification: FastAPI receives the cropped image and outputs clothing type and style classification via the TensorFlow model
3. Intelligent Recommendation: Generate outfit suggestions (matching, occasion, style, etc.) based on classification results

Free Hosting Solution

Supports free hosting like InfinityFree, requiring special handling of file upload paths (using $_SERVER['DOCUMENT_ROOT'] instead of Laravel's default storage directory)

AI Service Cloud Deployment

The AI model is containerized with Docker and deployed on Hugging Face Spaces. The Laravel application calls it via HTTP API, which has good scalability

FitScan AI has a clear structure and reasonable tech stack, fully demonstrating the process from image upload to AI inference and result display. The Laravel+FastAPI architecture is of reference value for similar projects and can serve as a starting point for fashion-related applications or a learning case for full-stack AI development

Inspirations for full-stack AI developers:

1. Tech stack combination example: PHP handles the web layer (request management, page rendering) + Python focuses on AI inference, communicating via REST API
2. Model serviceization mode: Trained models run as independent services, facilitating updates and expansion
3. Suitable as a reference case for entry-level full-stack AI development to learn how to integrate web and AI