Metadata-Version: 2.5
Name: seetapsych-face-hub
Version: 0.0.2
Summary: Face modules for SeetaPsych
Project-URL: Homepage, https://github.com/seetapsych/seetapsych-face-hub
Project-URL: Repository, https://github.com/seetapsych/seetapsych-face-hub
Project-URL: Issues, https://github.com/seetapsych/seetapsych-face-hub/issues
License: Copyright (c) 2026, Visual Information Processing and Learning (VIPL) group,
        Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China;
        Southeast University, China;
        Beijing Seetatech Co., Ltd.
        All rights reserved.
        
        Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
        
        1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
        
        2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
        
        3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
        
        THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
License-File: LICENSE
Requires-Python: >=3.10
Requires-Dist: seetapsych-lib
Provides-Extra: all
Requires-Dist: mediapipe>=0.10.33; extra == 'all'
Requires-Dist: numpy>=2.2.6; extra == 'all'
Requires-Dist: onnx>=1.20.1; extra == 'all'
Requires-Dist: onnxruntime-gpu<1.24; (sys_platform != 'darwin' and python_version <= '3.10') and extra == 'all'
Requires-Dist: onnxruntime-gpu>=1.24; (sys_platform != 'darwin' and python_version > '3.10') and extra == 'all'
Requires-Dist: onnxruntime<1.24; (sys_platform == 'darwin' and python_version <= '3.10') and extra == 'all'
Requires-Dist: onnxruntime>=1.24; (sys_platform == 'darwin' and python_version > '3.10') and extra == 'all'
Requires-Dist: opencv-python>=4.11.0.86; extra == 'all'
Requires-Dist: safetensors>=0.7.0; extra == 'all'
Requires-Dist: scikit-image>=0.25.2; extra == 'all'
Requires-Dist: torch>=2.10.0; extra == 'all'
Requires-Dist: torchvision>=0.25.0; extra == 'all'
Provides-Extra: dev
Requires-Dist: bandit>=1.7; extra == 'dev'
Requires-Dist: build>=1.4.4; extra == 'dev'
Requires-Dist: mediapipe>=0.10.33; extra == 'dev'
Requires-Dist: mypy>=1.11; extra == 'dev'
Requires-Dist: numpy>=2.2.6; extra == 'dev'
Requires-Dist: onnx>=1.20.1; extra == 'dev'
Requires-Dist: onnxruntime-gpu<1.24; (sys_platform != 'darwin' and python_version <= '3.10') and extra == 'dev'
Requires-Dist: onnxruntime-gpu>=1.24; (sys_platform != 'darwin' and python_version > '3.10') and extra == 'dev'
Requires-Dist: onnxruntime<1.24; (sys_platform == 'darwin' and python_version <= '3.10') and extra == 'dev'
Requires-Dist: onnxruntime>=1.24; (sys_platform == 'darwin' and python_version > '3.10') and extra == 'dev'
Requires-Dist: opencv-python>=4.11.0.86; extra == 'dev'
Requires-Dist: pytest>=8.0; extra == 'dev'
Requires-Dist: ruff>=0.6; extra == 'dev'
Requires-Dist: safetensors>=0.7.0; extra == 'dev'
Requires-Dist: scikit-image>=0.25.2; extra == 'dev'
Requires-Dist: torch>=2.10.0; extra == 'dev'
Requires-Dist: torchvision>=0.25.0; extra == 'dev'
Requires-Dist: twine>=6.2.0; extra == 'dev'
Description-Content-Type: text/markdown

# SeetaPsych Face Hub

> Community open-source face modules for SeetaPsych

## Usage

This project is already included in the seetapsych-lib default configuration. Download and use it via `seetapsych-manager download`.

For usage, refer to [SeetaPsych](https://github.com/seetapsych/seetapsych-lib).

You can additionally add this algorithm module using the following methods.

### WebUI

Run `seetapsych-webui` with the `--dirs` argument to use it.

```
seetapsych-webui --files seetapsych_face_hub/modules/insightface/retinaface.yml
```

### Programmatic Usage

Add the following code in your program to use this algorithm module.
```python
from seetapsych_lib.runtime.factory import Factory
from seetapsych_lib.runtime.pipeline import Pipeline

factory = Factory()
factory.load_file_modules("seetapsych_face_hub/modules/insightface/retinaface.yml")

pipeline = Pipeline(factory, ...)

pipeline.add_attributes("face/detection")
```

## Introduction

### Retinaface(InsightFace)

A face detection model from InsightFace's Buffalo release, capable of detecting faces and 5-point facial landmarks simultaneously. ONNX-based with configurable input size (default 640x640) for speed/accuracy tradeoff.

Module config: [retinaface.yml](seetapsych_face_hub/modules/insightface/retinaface.yml).
Provide Attributes: `face/detection` `face/landmarks`.

Available models: `det_500m.onnx` (recommended, 500M FLOPs, lightweight), `det_2.5g.onnx` (2.5G FLOPs, balanced), `det_10g.onnx` (10G FLOPs, high accuracy).

### ArcFace(InsightFace)

A face feature extraction (recognition) model from InsightFace's Buffalo release, capable of extracting L2-normalized 512-dim face embedding vectors. Requires 5-point face landmarks for similarity transform alignment before feature extraction.

Module config: [arcface.yml](seetapsych_face_hub/modules/insightface/arcface.yml).
Provide Attributes: `face/feature`.

Available models: `w600k_r50.onnx` (recommended, ResNet-50 backbone, higher accuracy), `w600k_mbf.onnx` (MobileFaceNet backbone, lightweight, faster).

### FaceMesh(MediaPipe)

A suite of MediaPipe Tasks vision solutions including BlazeFace face detector and Face Landmarker. Face Mesh extracts 468 normalized 3D face landmarks covering the entire facial geometry. Detection and mesh are provided as two separate packages:

- **Face Detector**: Provides `face/detection`. Supports BlazeFace models in short-range, full-range, and sparse variants.
- **Face Mesh**: Provides `face/mesh`, requires `face/detection`. Outputs normalized 3D landmarks with optional face blendshapes and facial transformation matrices.

Module config: [mediapipe.yml](seetapsych_face_hub/modules/mediapipe.yml).
Provide Attributes: `face/detection` `face/mesh`.

Detector models: `blaze_face_full_range.tflite` (recommended), `blaze_face_full_range_sparse.tflite`, `blaze_face_short_range.tflite`.
Mesh model: `face_landmarker.task` (recommended).

### Retinaface(Ver. PyTorch)

PyTorch implementation of RetinaFace (Biubug6), a single-stage dense face localisation model. Detects faces and 5-point landmarks with configurable confidence threshold (default 0.6). Models are distributed in safetensors format.

Module config: [retinaface.yml](seetapsych_face_hub/modules/retinaface.yml).
Provide Attributes: `face/detection` `face/landmarks`.

Available models: `mobilenet0.25_Final.safetensors` (recommended, MobileNet-0.25 backbone, lightweight), `Resnet50_Final.safetensors` (ResNet-50 backbone, higher accuracy).

## References

- [MediaPipe](https://ai.google.dev/edge/mediapipe/solutions/guide)
- [InsightFace](https://github.com/deepinsight/insightface)
- [Pytorch_Retinaface](https://github.com/biubug6/Pytorch_Retinaface)
