Liang Dong

@legan78 Did you fix the problem of difference in evaluation with python? I wonder if it is the scales issue. The matlab code calculate the scales https://github.com/ZheC/Realtime_Multi-Person_Pose_Estimation/blob/master/testing/src/applyModel.m#L44. Could you share...

Is there any updates for For python code evaluation? I could get 0.55 use the first 2644 images in the val2014 set. I change the original python evaluation code a...

The original paper has python demo and evaluation matlab code. They talked about the evaluation in original caffe implementation. The original evaluation code seems to be Matlab( I am trying...

If you are interested in it, we can collaborate in reimplementing the evaluation code in python. I am also interested to the performance of CPM refinement. @hufengshuo07

Hi @hufengshuo07, I finished the evaluation code. https://github.com/dragonfly90/mxnet_Realtime_Multi-Person_Pose_Estimation/blob/master/evaluation_coco.py

I am still working on pose now. I want to try some new ideas like mask rcnn for pose recognition. I may not have time to reproduce face and hand...

Good implementation. Maybe I could borrow some code from you. My training error seems to converge. Like the following 20 to 30 epochs. But My mAP performance is not improved...

Hi @insomnia250 , you could change https://github.com/dragonfly90/mxnet_Realtime_Multi-Person_Pose_Estimation/blob/master/pose_io/annotation.ipynb to generate new json file. You also need to change https://github.com/dragonfly90/mxnet_Realtime_Multi-Person_Pose_Estimation/blob/master/GenerateLabelCPM.py#L83 to make it fit 14 keypoints (Cao's original paper use 18 key...

@insomnia250 pycocotools is coco API (https://github.com/pdollar/coco). You need to install it. I think you could just change GenerateLabelCPM.py if you want to use other dataset. You only need the pixel...

@qqsh0214 I update https://github.com/dragonfly90/mxnet_Realtime_Multi-Person_Pose_Estimation/blob/master/pose_io/annotation.ipynb. You could check Zhe Cao's original code for the data format. Basically you have main person keypoints location and other persons' keypoints location in each image,...