| --- |
| 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. |
|
|
