Spaces:
Sleeping
Sleeping
| """LSF_transcripter — French Sign Language (Langue des Signes Française) interpreter. | |
| The package turns a camera frame into: | |
| 1. A holistic skeleton (face / pose / both hands) via MediaPipe. | |
| 2. A normalised feature vector describing fingers, eyes, mouth, chest and arms. | |
| 3. A recognised LSF gloss using a learn-by-example (DTW) recogniser, with an | |
| optional deep-learning model hook and a few built-in heuristic signs. | |
| Optional depth refinement is applied when the camera exposes a dual-pixel / | |
| depth stream (or a monocular-depth model is installed). | |
| """ | |
| from .landmarks import HolisticTracker, LandmarkFrame | |
| from .recognizer import SignRecognizer, Recognition | |
| __all__ = [ | |
| "HolisticTracker", | |
| "LandmarkFrame", | |
| "SignRecognizer", | |
| "Recognition", | |
| ] | |
| __version__ = "0.1.0" | |