.README

### Prosopagnosia Assistance Project

**Overview**

This project aims to assist individuals suffering from prosopagnosia (face blindness) by providing a real-time facial recognition system. The solution includes visual and auditory feedback mechanisms to identify people and announce their names. The long-term goal is to integrate this technology into wearable devices, such as smart glasses, for seamless assistance.

**Features Implemented**

1. Face Recognition System

A face recognition model using the Local Binary Patterns Histogram (LBPH) algorithm.
Trained with labeled images stored in the training_images directory.
Detects faces in real-time from webcam video streams and matches them with trained labels.

2. Feedback Mechanisms

Visual Feedback
The recognized person's name and confidence score are displayed on the video stream in real time.
Unrecognized faces are labeled as "Unknown."

Audio Feedback
A Text-to-Speech (TTS) system announces the name of the recognized individual.
The system ensures each name is announced only once until no faces are detected for a certain period.
For unknown individuals, the system announces: "I'm sorry, do I know you?"

3. Real-Time Detection

Integrates OpenCV for real-time webcam capture and facial detection.
Uses Haar cascades for face detection.

**Project Structure**

Prosopagnosia_project/
|-- codebase/
| |-- main.py # Main script for training and real-time recognition
| |-- training_images/ # Directory containing labeled subfolders of training images
|-- requirements.txt # Dependencies for the project
|-- README.md # Project documentation

**How It Works**

Training:
The model is trained using images stored in the training_images directory.
Each subdirectory in training_images represents a labeled class (e.g., a person's name).

Real-Time Recognition:
The system captures frames from the webcam and detects faces using Haar cascades.
Each detected face is matched against the trained model.

Feedback:
Names of recognized individuals are displayed on the screen and announced audibly.
Unknown faces trigger a generic audio message.

**Installation and Setup**

Install Dependencies: pip install -r requirements.txt
Prepare Training Data:
Create subfolders in the training_images directory for each individual.
Add grayscale images (100x100 pixels recommended) of each person to their respective folder.
Run the Application: python main.py

**Dependencies**
Python 3.x
OpenCV
NumPy
Pyttsx3

**Acknowledgments**
OpenCV for face detection and recognition.
Pyttsx3 for Text-to-Speech functionality.

**Author**
This project is designed and developed as a step toward helping individuals with prosopagnosia through innovative AI solutions.