MotionCaptureSuite

Status: Experimental

Three small Tkinter/OpenCV desktop tools for recording video, reviewing it frame by frame, and measuring shoulder range of motion with MediaPipe pose tracking. I built this to track a physical-therapy exercise (arm range of motion) without needing dedicated motion-capture hardware.

flowchart LR
    A[record_video.py<br/>webcam capture] --> B[video_playback.py<br/>frame-by-frame review, pixel RGB]
    B --> C[mediapipe_detection.py<br/>pose landmarks, min/max joint angle]

Loading

Tools

• record_video.py: live camera preview with a filename field, records to .mp4
• video_playback.py: steps a recorded video forward/backward one frame at a time and reports the RGB value under the cursor
• mediapipe_detection.py: runs MediaPipe Pose on a video, computes the shoulder-elbow-hip angle for each arm per frame, and tracks the min/max angle to give a range-of-motion reading

Install

pip install opencv-python mediapipe Pillow numpy

Requires Python 3 with Tk available (standard on most desktop Python installs; on Linux you may need python3-tk from your package manager).

Usage

python record_video.py        # record a clip
python video_playback.py      # step through a saved clip, inspect pixels
python mediapipe_detection.py # load a clip, track arm angle range

Status notes

This was a personal utility, not a packaged library. There's no CLI argument handling or test suite, and the exception handling around pose detection is broad (except: pass) rather than precise. The older platform-specific install scripts referenced in earlier revisions of this README pointed at a renamed repo and have been dropped in favor of the plain pip install above.

License

No license file yet. Treat as all-rights-reserved until one is added.