Openpose implementation using darknet framework
[Openpose]
https://github.com/CMU-Perceptual-Computing-Lab/openpose
[Darknet]
https://github.com/pjreddie/darknet
[Result]
[Benchmark]
51.634ms for 200x200x3 net input size running on GTX 1060, which is 3x faster than original caffe implementation
[Weight file] (darknet version openpose.weight)
https://drive.google.com/open?id=1BfY0Hx2d2nm3I4JFh0W1cK2aHD1FSGea
[Usage]
./openpose-darknet [image file] [cfg file] [weight file]
#example
./openpose-darknet person.jpg openpose.cfg openpose.weight
[note]
- Darknet version openpose.cfg and openpose.weight are ported from COCO version pose_deploy_linevec.prototxt and pose_iter_440000.caffemodel.
- You could change net input width, height in openpose.cfg.