Spaces:
Running
Running
| """Shared vision utilities for Reachy Mini apps. | |
| Two layers: | |
| - **Attention** (present + engaged): `Detector` protocol + `AttentionResult`, | |
| with `MotionPresenceDetector` (zero-dep default) and `VLMAttentionDetector` | |
| (real engagement via a vision LLM). `build_detector(spec)` picks one. | |
| - **Objects / pose** (optional, heavier deps): `ObjectDetector` (YOLO phones), | |
| `HeadPoseEstimator` (MediaPipe). Import these directly when needed. | |
| """ | |
| from .base import AttentionResult, Detector, BaseDetector | |
| from .motion import MotionPresenceDetector | |
| from .vlm import VLMAttentionDetector | |
| from .factory import build_detector | |
| from . import frames | |
| __all__ = [ | |
| "AttentionResult", | |
| "Detector", | |
| "BaseDetector", | |
| "MotionPresenceDetector", | |
| "VLMAttentionDetector", | |
| "build_detector", | |
| "frames", | |
| ] | |