Sign Language Detection

A real-time sign language recognition system using Python, OpenCV, and a pre-trained Keras model from Teachable Machine. It captures live webcam input, identifies hand gestures, and displays the predicted label on screen with clean visuals.

---

What is it?

This project is a gesture recognition tool built to detect and classify sign language letters or commands using computer vision. It helps bridge the gap between spoken and signed communication, offering a lightweight offline prototype that runs on any laptop with a webcam.

---

Key Features

• Real-Time Gesture Recognition
  Detects hand gestures in real time using webcam and bounding boxes.

• ML-Powered Predictions
  Uses a .h5 Keras model trained with Teachable Machine to classify gestures.

• Live Visualization
  Displays cropped and white-background images alongside output with bounding boxes and label overlays.

• Clean Label UI
  Shows detected labels in a black background box with white text for clarity.

• Offline & Lightweight
  No internet required once model is downloaded — works entirely offline.

---

Why I Built This?

I built this as part of my exploration into real-world computer vision applications using Python and deep learning. Sign language detection is an impactful problem that intersects accessibility, AI, and visual recognition — and I wanted to prototype a working tool using minimal resources and simple tech.

This project also helped me:

• Practice working with OpenCV and cvzone
• Understand how to preprocess input for ML models
• Learn how to convert models trained online (Teachable Machine) into usable .h5 Keras models

---

What’s Next?

• Expand dataset for more gesture classes (A-Z or custom commands)
• Add frame smoothing to stabilize predictions
• Add guided training mode for users to record custom signs

---