Find this example on our SD-image
This is a ultra fast C++ implementation of the face detector of Linzaer running on a MNN framework.
https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB.
Paper: https://arxiv.org/abs/1905.00641.pdf
Size: 320x320
Special made for a bare Raspberry Pi see https://qengineering.eu/deep-learning-examples-on-raspberry-32-64-os.html
Three frameworks are supported:
- Alibaba's MNN framework
- Tencent ncnn framework
- OpenCV dnn
The frame rate is based upon the average execution time of the single frames.
Loading frames from a file, plotting boxes, and showing the result on the screen are not taken into account.
The MNN framework has also 8 bit quantized models. These are very fast.
See the video at https://youtu.be/DERA83C9K2A
https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB
| Model | framework | model | size | mAP | Jetson Nano 2015 MHz |
RPi 4 64-OS 1950 MHz |
|---|---|---|---|---|---|---|
| Ultra-Light-Fast | ncnn | slim-320 | 320x240 | 67.1 | - FPS | 26 FPS |
| Ultra-Light-Fast | ncnn | RFB-320 | 320x240 | 69.8 | - FPS | 23 FPS |
| Ultra-Light-Fast | MNN | slim-320 | 320x240 | 67.1 | 70 FPS | 65 FPS |
| Ultra-Light-Fast | MNN | RFB-320 | 320x240 | 69.8 | 60 FPS | 56 FPS |
| Ultra-Light-Fast | OpenCV | slim-320 | 320x240 | 67.1 | 48 FPS | 40 FPS |
| Ultra-Light-Fast | OpenCV | RFB-320 | 320x240 | 69.8 | 43 FPS | 35 FPS |
| Ultra-Light-Fast + Landmarks | ncnn | slim-320 | 320x240 | 67.1 | 50 FPS | 24 FPS |
| LFFD | ncnn | 5 stage | 320x240 | 88.6 | 16.4 FPS | 4.85 FPS |
| LFFD | ncnn | 8 stage | 320x240 | 88.6 | 11.7 FPS | 3.45 FPS |
| LFFD | MNN | 5 stage | 320x240 | 88.6 | 2.6 FPS | 2.17 FPS |
| LFFD | MNN | 8 stage | 320x240 | 88.6 | 1.8 FPS | 1.49 FPS |