Sign Language Detection Project

Overview

This project detects American Sign Language (ASL) gestures using computer vision and machine learning. It includes:

1. Dataset Preparation: Extracts hand landmarks from images of ASL gestures (A to Z).
2. Model Training: Trains a Random Forest classifier on the preprocessed data.
3. Real-time Inference: Detects and recognizes ASL gestures using a webcam.
4. Web Interface: Displays the predicted ASL gesture on a webpage.

Dataset

The dataset comprises images representing ASL letters from A to Z.

• Dataset: ASL Alphabet Dataset

Prerequisites

Ensure you have the following installed:

• Python 3.7 or higher
• OpenCV
• MediaPipe
• NumPy
• Scikit-learn
• Pickle

You can install the required libraries using pip install opencv-python mediapipe numpy scikit-learn pickle-mixin.

Project Structure

• create_dataset.py: Extracts hand landmarks from ASL images.
• train_classifier.py: Trains a Random Forest classifier.
• inference_classifier.py: Detects and recognizes ASL gestures using a webcam.
• templates: Contains files for the web interface.

Scripts

create_dataset.py

This script processes the ASL image dataset to extract hand landmarks using MediaPipe and saves the landmarks to a file.

train_classifier.py

This script trains a Random Forest classifier using the extracted hand landmarks and saves the trained model to a file.

inference_classifier.py

This script uses a webcam to capture real-time ASL gestures, processes the frames to extract hand landmarks, and uses the trained model to predict the gesture.

templates

This folder contains the files for the web interface that displays the predicted ASL gesture. The web interface shows the detected gesture in real-time using a terminal-style loader animation.

Running the Project

Step 1: Create Dataset

Run create_dataset.py to extract hand landmarks from the ASL images.

Step 2: Train Classifier

Run train_classifier.py to train the Random Forest classifier.

Step 3: Real-time Inference

Run inference_classifier.py to start the webcam and recognize ASL gestures in real-time.

Step 4: Web Interface

Open home.html in your browser to view the web interface. Ensure the webcam is enabled, and the predicted ASL gesture will be displayed in the terminal-style loader.

Acknowledgments

• Dataset: ASL Alphabet Dataset
• Hand Landmark Detection: MediaPipe

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.