README.md

---
license: apache-2.0
tags: [onnx, mediapipe, pose-detection, blazepose, qtmesheditor]
---

# BlazePose person detector (ONNX)

Google MediaPipe's BlazePose detector, from the Pose Landmarker `.task`
bundle, converted to ONNX.

- **Input** `[1,224,224,3]` RGB in `[-1,1]`, letterboxed.
- **Outputs** `[1,2254,12]` + `[1,2254,1]`. Anchors: strides
  `[8,16,32,32,32]` → 2254. The ROI is centred on the mid-hip keypoint,
  sized from the hip→scale keypoint distance, target angle 90°, 1.25× square.

The detector ships fp16 **block-sparse** weights behind DENSIFY ops that
tf2onnx can't parse; the export densifies them via the TFLite interpreter
first (see the export script).

## License

Apache-2.0 — this graph is a direct ONNX conversion of a Google
[MediaPipe](https://developers.google.com/mediapipe) model (Apache-2.0 code
AND weights). Conversion + numerical-parity proof (vs the Python `mediapipe`
reference): [`scripts/export-facecap-onnx.py`](https://github.com/fernandotonon/QtMeshEditor/blob/master/scripts/export-facecap-onnx.py),
contract in [`docs/MOCAP_SPIKE.md`](https://github.com/fernandotonon/QtMeshEditor/blob/master/docs/MOCAP_SPIKE.md).

## How it is used

Mirror of one graph from [`fernandotonon/QtMeshEditor-models`](https://huggingface.co/fernandotonon/QtMeshEditor-models)
(`mocap/…`), which [QtMeshEditor](https://github.com/fernandotonon/QtMeshEditor)
downloads on first use for its **Performance Capture** feature (video/webcam →
facial morph + head + full-body skeletal animation, epic #869). This standalone
repo is for discoverability; the app fetches from the aggregate repo.