faace_recognitionj

Real-Time Face Detection using OpenCV (Python & Android)

This project implements real-time face detection using the Haar Cascade Classifier from OpenCV. It supports both:

• A Python script that performs face detection using a webcam.
• An Android app that detects faces using the phone's camera with OpenCV (no internet or ML Kit required).

Features

✅ Real-time face detection
✅ Uses OpenCV’s pre-trained Haar Cascade model
✅ Works offline (no cloud or API needed)
✅ Lightweight and fast
✅ Cross-platform: Desktop (Python) and Mobile (Android)

Model Description

This project uses the Haar Cascade Classifier, a classical machine learning-based approach for object detection developed by Viola and Jones.

How It Works

• The algorithm uses Haar-like features to detect facial structures.
• It applies integral images to compute features rapidly.
• A cascade of classifiers is trained from positive (with faces) and negative (without faces) samples.
• At runtime, the model quickly eliminates non-face regions in early stages and focuses only on potential face areas.

Model Used

• Model file: haarcascade_frontalface_default.xml
• Provided by OpenCV under BSD license
• Pre-trained for detecting frontal human faces

Android App

This Android app demonstrates real-time face detection. It uses the device camera to capture video frames and detects faces with contours, displaying bounding boxes and facial landmarks on the preview.

Features

• Real-time face detection with camera preview
• Switch between front and back camera
• Draw face contours and bounding boxes overlay
• Camera permission handling

Project Structure

• MainActivity.kt
  Handles UI, camera permission requests, and initializes the camera manager.

• CameraManager.kt
  Manages CameraX APIs, provides camera preview, and routes image frames to the face detection analyzer.

• FaceContourDetectionProcessor.kt
  Implements face detection and processes detection results to draw overlays.

• BaseImageAnalyzer.kt
  Abstract class that integrates InputImage with CameraX image analysis.

• GraphicOverlay.kt
  Custom view for rendering face bounding boxes and landmarks on top of camera preview.

• res/layout/activity_main.xml
  UI layout containing PreviewView and overlay views.

Video Demonstration

https://drive.google.com/file/d/15bpNL2LVgBU6pILCCU7wSlgLxAp3LwWk/view?usp=drive_link