Face-Detector with landmark

Features

• Retinaface-mobile0.25 model converted into ncnn python/ opencv onnx/ pytorch python
• Face-Detector-1MB slim
• 5 key points of face detection
• Support onnx export
• Network parameter and flop calculation

lightweight face detector with keypoint detection

Provides a series of face detectors suitable for mobile deployment including key face detectors: Modified the anchor size of Retinaface-mobile0.25 to make it more suitable for edge computing; Reimplemented Face-Detector-1MB and added key point detection and ncnn C++ The deployment function, in most cases, the accuracy is better than the original version.

Requirments

• Ubuntu18.04
• Python3.7
• opencv
• numpy

Accuracy

Widerface test

• Accuracy in wider face val (single-scale input resolution: 320*240）

method	Easy	Medium	Hard
libfacedetection v1（caffe）	0.65	0.5	0.233
libfacedetection v2（caffe）	0.714	0.585	0.306
version-slim(original)	0.765	0.662	0.385
version-RFB(original)	0.784	0.688	0.418
version-slim(our)	0.795	0.683	0.34.5
version-RFB(our)	0.814	0.710	0.363
Retinaface-Mobilenet-0.25(our)	0.811	0.697	0.376

• Accuracy in wider face val (single-scale input resolution:：640*480）

method	Easy	Medium	Hard
libfacedetection v1（caffe）	0.741	0.683	0.421
libfacedetection v2（caffe）	0.773	0.718	0.485
version-slim(original)	0.757	0.721	0.511
version-RFB(original)	0.851	0.81	0.541
version-slim(our)	0.850	0.808	0.595
version-RFB(our)	0.865	0.828	0.622
Retinaface-Mobilenet-0.25(our)	0.873	0.836	0.638

ps: When testing, the long side is 320 or 640, and the image is scaled in equal proportions.

Parameter and flop

method	parameter(M)	flop(M)
version-slim(our)	0.343	98.793
version-RFB(our)	0.359	118.435
Retinaface-Mobilenet-0.25(our)	0.426	193.921

ps: 320*240 as input

Python inference

coming soon

References

• FaceBoxes
• Retinaface (mxnet)
• Retinaface (pytorch)
• Ultra-Light-Fast-Generic-Face-Detector-1MB

@inproceedings{deng2019retinaface,
title={RetinaFace: Single-stage Dense Face Localisation in the Wild},
author={Deng, Jiankang and Guo, Jia and Yuxiang, Zhou and Jinke Yu and Irene Kotsia and Zafeiriou, Stefanos},
booktitle={arxiv},
year={2019}