✋ Real-Time Sign Language Recognition

This project implements a real-time recognition system for sign language alphabets and numbers using a Convolutional Neural Network (CNN) and MediaPipe for hand landmark detection. The system captures live video input, processes hand gestures, and classifies them into corresponding sign language alphabets or numbers.

✋ Hand Sign Recognition System

This project is a complete hand sign recognition pipeline using MediaPipe and TensorFlow. It supports:

• Automatic landmark extraction from hand gesture images
• Training a Multi-Layer Perceptron (MLP) classifier
• Real-time hand sign detection from webcam
• GUI interface using tkinter for ease of use

🔧 Requirements

• Python 3.7+
• OpenCV
• MediaPipe
• TensorFlow
• scikit-learn
• Pillow
• tqdm
• matplotlib
• seaborn

📁 Project Structure

├── create_landmark_dataset.py # Extracts and normalizes images hand landmarks into CSV ├── train_evaluate_model.py # Trains an MLP model on the landmark dataset ├── main.py # Real-time hand detection using webcam ├── App.py # GUI application for live prediction ├── hand_landmarks_normalized.csv # (Generated) CSV dataset of landmarks ├── landmark_model_best.keras # (Generated) Best trained model

📸 Step-by-Step Usage

1. Dataset Preparation Prepare a dataset of hand gesture images grouped by class in subfolders, e.g. Gesture Image Data/ ├── A/ ├── B/ ├── C/ └── ...

2. Extract Landmarks Run: python create_landmark_dataset.py This will create hand_landmarks_normalized.csv.

3. Train the Model Train an MLP classifier using: python train_evaluate_model.py This will output landmark_model_best.keras and show training plots and metrics.

4. Real-Time Detection (CLI Mode) To test the model using your webcam: python main.py

5. GUI Mode Launch the graphical interface using: python App.py This provides a simple App to start/stop detection and view predictions in real-time.