What it is
A web app that reads hand gestures from a camera and classifies them in real time, using MediaPipe hand landmarks and an LSTM.
What I built
I trained the model and built the web app around it.
- MediaPipe landmarks fed to an LSTM with dense layers for sequence classification
- Trained with TensorFlow and Keras; tuned for low-latency inference
Stack
- Python
- TensorFlow
- Keras
- MediaPipe
- LSTM