Jas
> I see it has supported heatmap method of face datasets now. Please support regression method of face dataset! Do you have any recommended papers/codes ?
Support 3dpw dataset #682
> It would be so helpful for better analysis if AP for each type of body joints are printed, for example 17 AP value for 17 kinds of body joints...
https://github.com/open-mmlab/mmpose/blob/8014e546df98a2cb73e08a59328d5e97bca32893/tests/test_pipelines/test_bottom_up_pipelines.py#L96 印象中，也支持矩形输入
Not supported yet. We do not support evaluation of pose tracking (MOTA), but only pose estimation (mAP)
You do not need tracking to quantify the effect of smoothing. Evaluation of video based pose estimation also works.
Choose a video pose estimation dataset, e.g. sub-jhmdb. Use smoother (/or not) to inference the whole dataset, and compare the results of using smoother or not.
demo 可以照着这个 https://github.com/open-mmlab/mmpose/blob/master/demo/docs/2d_human_pose_demo.md 改一下 onnx模型导出，可以试一下 mmdeploy
可以参考教程 https://github.com/open-mmlab/mmpose/blob/master/docs/zh_cn/tutorials/4_new_modules.md
use resume_from please