A robust, generalized MediaPipe 3D landmark extractor and normalizer for any video dataset, featuring seamless integration with the AUTSL Sign Language dataset.
computer-vision pytorch pose-estimation sign-language-recognition mediapipe dataset-processing autsl landmark-extraction skeleton-normalization
-
Updated
Jul 18, 2026 - Python