Documentation

Pipeline

How a frame becomes structured intelligence, in thirteen stages. This is the summary; the full breakdown is on the Pipeline page.

Every analysed frame flows through the same stages:

  1. Input  ·  acquire a source (camera, stream, video, image).
  2. Frame Capture  ·  bounded, drop-oldest buffering.
  3. Preprocessing  ·  normalise, low-light enhance, background plate.
  4. Detection  ·  YOLO11 + ByteTrack with class gating.
  5. Segmentation  ·  mask movers vs the static plate.
  6. Depth Estimation  ·  Depth Anything V2, temporally fused.
  7. Pose Estimation  ·  keypoints for intent and gait.
  8. Geometry Fusion  ·  fuse depth, masks, background completion.
  9. 3D Reconstruction  ·  back-project to a point cloud.
  10. Mesh Optimization  ·  triangulate + edge culling.
  11. Texture Mapping  ·  project RGB and inpainted texture.
  12. Spatial Analysis  ·  foresight, identity, alerts.
  13. Output  ·  stream, export, persist.
Full detail

Expand every stage with its inputs, outputs and internals on the Pipeline page.