🖼️ Duplicate Image Detection API

A FastAPI-based web service that detects and returns the most visually similar image to a user-uploaded file by comparing it with previously stored images using feature matching techniques.

Features

• Upload and store original images via REST API
• Upload a new image to compare with stored ones
• Detects visually similar images using:
  • ORB (Oriented FAST and Rotated BRIEF) for feature extraction
  • BFMatcher with Hamming distance for descriptor comparison
• Returns the most similar image using a match scoring algorithm
• Streaming response for efficient image delivery

How It Works

1. Upload Original Image

   • http://127.0.0.1:8000/docs - detect-duplicate/v1/upload
   • Saves the image in the original_images directory

2. Find Matching Image

   • http://127.0.0.1:8000/docs - detect-duplicate/v1/match
   • Accepts an uploaded image
   • Uses ORB to detect keypoints and descriptors
   • Compares descriptors with each stored image using BFMatcher
   • Calculates match score (lower score = more similar)
   • Returns the closest matching image as a response

Technologies Used

• FastAPI – for building RESTful APIs
• OpenCV – for image processing and feature detection
• Python – core programming language
• NumPy – for handling image buffers

Running the App

• Install dependencies: pip install -r requirements.txt

• Start the server: uvicorn script:app --reload