[Enhance] Refine body3d_two_stage_video_demo.py

[liqikai9]

Motivation

Inrich the keypoint conversions from pose_lift_dataset to pose_det_dataset.

Now, the demo support the following conversions:

1. from Body3DH36MDataset to ['TopDownCocoDataset', 'TopDownPoseTrack18Dataset', 'TopDownPoseTrack18VideoDataset', 'TopDownAicDataset'], ['TopDownCrowdPoseDataset'];
2. from Body3DMpiInf3dhpDataset to ['TopDownCocoDataset', 'TopDownPoseTrack18Dataset', 'TopDownPoseTrack18VideoDataset', 'TopDownAicDataset'], ['TopDownCrowdPoseDataset'];

Fix the problem mentioned in #1485.
Credit to @pallgeuer.

Modification

BC-breaking (Optional)

Use cases (Optional)

A command to run the demo using the 1-frame videopose3d model trained on Body3DMpiInf3dhpDataset:

python demo/body3d_two_stage_video_demo.py \
    demo/mmdetection_cfg/faster_rcnn_r50_fpn_coco.py \
    https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth \
    configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/hrnet_w48_coco_256x192.py \
    https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w48_coco_256x192-b9e0b3ab_20200708.pth \
    configs/body/3d_kpt_sview_rgb_vid/video_pose_lift/mpi_inf_3dhp/videopose3d_mpi-inf-3dhp_1frame_fullconv_supervised_gt.py \
    https://download.openmmlab.com/mmpose/body3d/videopose/videopose_mpi-inf-3dhp_1frame_fullconv_supervised_gt-d6ed21ef_20210603.pth \
    --video-path https://user-images.githubusercontent.com/87690686/164970135-b14e424c-765a-4180-9bc8-fa8d6abc5510.mp4 \
    --out-video-root vis_results \
    --rebase-keypoint-height

Checklist

Before PR:

• I have read and followed the workflow indicated in the CONTRIBUTING.md to create this PR.
• Pre-commit or linting tools indicated in CONTRIBUTING.md are used to fix the potential lint issues.
• Bug fixes are covered by unit tests, the case that causes the bug should be added in the unit tests.
• New functionalities are covered by complete unit tests. If not, please add more unit tests to ensure correctness.
• The documentation has been modified accordingly, including docstring or example tutorials.

After PR:

• CLA has been signed and all committers have signed the CLA in this PR.

[codecov]

Codecov Report

Merging #1490 (2ad5efd) into master (afb37d4) will decrease coverage by 0.01%.
The diff coverage is 40.00%.

@@            Coverage Diff             @@
##           master    #1490      +/-   ##
==========================================
- Coverage   84.47%   84.45%   -0.02%     
==========================================
  Files         236      236              
  Lines       20036    20038       +2     
  Branches     3602     3603       +1     
==========================================
- Hits        16925    16923       -2     
- Misses       2231     2233       +2     
- Partials      880      882       +2     

Flag	Coverage Δ
unittests	84.36% <40.00%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
mmpose/apis/inference_3d.py	76.64% <40.00%> (-0.94%)	⬇️
mmpose/datasets/pipelines/shared_transform.py	83.26% <0.00%> (-0.43%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update afb37d4...2ad5efd. Read the comment docs.

[jin-s13]

Is it possible to use dataset_info in https://github.com/open-mmlab/mmpose/tree/master/configs/_base_/datasets?

[liqikai9]

I think the keypoint indexes conversion is hard to use the info in dataset_info. We may consider adding a tool to convert the keypoint orders later.