Automated Sign Language Translation & Voice Synthesis via MediaPipe
Bidirectional American/Indian Sign Language translator utilizing 3D hand landmarks and Bidirectional LSTM with instant text-to-speech.
Project Overview
Captures 21 3D hand keypoints and facial landmarks per frame via standard webcams without costly sensor gloves. A temporal BiLSTM/GRU neural network classifies continuous sign sentences and speaks them aloud via Google TTS, bridging communication for the speech-impaired.
Captures 21 3D hand keypoints and facial landmarks per frame via standard webcams without costly sensor gloves. A temporal BiLSTM/GRU neural network classifies continuous sign sentences and speaks them aloud via Google TTS, bridging communication for the speech-impaired.
Bidirectional American/Indian Sign Language translator utilizing 3D hand landmarks and Bidirectional LSTM with instant text-to-speech.