Multi-stereo fusion and calibration engine for precise 3D hand and gesture key-poses.
CoreFusion consumes pre-computed 3D keypoints and ROI data from multiple EdgeTrack stereo rigs over LAN.
It performs time alignment, multi-view calibration, bundle adjustment, and low-latency fusion, and publishes a unified, high-confidence 3D key-pose stream to downstream clients such as MotionCoder.
Status: early prototype (WIP). Interfaces and schemas may evolve.
- Scale: Fuse 2–4 independent stereo rigs (4–8 cameras) to achieve robust, occlusion-resistant tracking.
Cooming soon!