Sign-Language

Why we included this project

This project walks you through building a gesture recognizer for American Sign Language from scratch, and it's best treated as a learning exercise rather than something to drop into production. The author shows the whole pipeline: capturing 44 hand gestures with OpenCV, flipping the images to double the dataset, training a small CNN with TensorFlow and Keras, then pointing the trained model at a live webcam feed. The code is short enough to read in one sitting, and the README explains what each script does, including practical touches like a hand-skin histogram for segmentation and confusion-matrix reports. If you're curious how a tiny convolutional network gets wired to real video input, this is a good place to start. Just expect to extend the model and capture scripts yourself, since the repo is a template for experiments, not a finished app.