KSNN: yolov3 hand detection and face detection demo

Why to do single classification yolo3 demo

This demo project originated from 5 months ago. Many users reported to me that the single classification model based on yolov3 could not run correctly on the NPU. So I made two single classification models based on yolov3 for user reference.

How To Use

Follow our docs to install .

Face detection demo

$ cd examples/darknet/
$ python3 yolov3-face.py --model ./models/VIM3/yolov3-face.nb --library ./libs/libnn_yolov3-face.so --video-device X

Hand detection demo

base on yolov3

$ cd examples/darknet/
$ python3 hand-cap.py --model ./models/VIM3/hand.nb --library ./libs/libnn_hand.so --video-device X

Original file

You can get the original files in my github.

  1. Face:
  1. Hand
  1. You can train on my repository

TODO

  1. Add person detect demo
  2. Optimize model parameters