Arshad221b/Sign-Language-Recognition — reverse-engineered prompt

Build me a simple sign language recognition project that uses my webcam to read hand signs and show the predicted letter on screen in real time.

I want it to use OpenCV for the camera view and hand region selection, and a CNN model with TensorFlow or Keras for the prediction. Use the Sign Language MNIST dataset from Kaggle for training, and include an option to load an existing saved model so I don’t have to train every time. The app should open a camera window, let me place my hand inside a clear box, process that hand image, and display the predicted sign clearly.

Please make it easy to run from a Python file, with simple setup instructions and a requirements file. Keep the code understandable, add comments where helpful, and include a short notebook or training script for retraining the model if needed. Look up current docs online if you need to.