Face Similarity Search System

An ethical face similarity search system built with Flask and OpenCV that allows users to upload photos and find visually similar faces from a sample dataset.

Features

• Face Detection: Uses OpenCV's Haar Cascade classifiers for robust face detection
• Feature Extraction: Generates embeddings using multiple computer vision techniques:
  • Local Binary Patterns (LBP)
  • Intensity Histograms
  • Edge Features
  • Geometric Properties
• Similarity Search: Combines cosine similarity and Euclidean distance for accurate matching
• Web Interface: Clean, responsive Bootstrap UI with drag-and-drop upload
• Ethical Design: No web crawling - uses only sample datasets provided by users

Ethical Considerations

This system is designed with privacy and ethics in mind:

• ✅ No Web Crawling: Uses only sample datasets, no unauthorized data collection
• ✅ Privacy-Focused: Local processing, no permanent storage of uploaded images
• ✅ Transparent: Open-source algorithms, no black-box AI models
• ✅ Educational: Designed for research and learning purposes
• ✅ User Control: Users provide their own dataset images

Installation & Setup

Prerequisites

• Python 3.8+
• OpenCV with face recognition module
• Flask and other Python dependencies

Quick Start

1. Clone/Download the application files

2. Install dependencies:

   pip install flask opencv-python opencv-contrib-python pillow numpy

3. Add sample images:

   • Place sample face images in the static/dataset/ directory
   • Supported formats: PNG, JPG, JPEG, GIF, BMP
   • Use clear, well-lit photos with visible faces

4. Run the application:

   python main.py