A bug when testing voxelpose

[Taylorminer]

Thanks for your work! When I test voxelpose in Campus dataset. I run the following command:

I meet the following error:

[liqikai9]

@Taylorminer Hi, thanks for using mmpose. This demo script body3d_multiview_detect_and_regress_img_demo.py only support CMU Panoptic data format currently, you can find more description about this demo script in this doc.

You may need to modify the demo script to run this demo on campus and shelf dataset. And if you are interested in supporting this feature in mmpose, please let us know. We are glad to help.

[Taylorminer]

@Taylorminer Hi, thanks for using mmpose. This demo script body3d_multiview_detect_and_regress_img_demo.py only support CMU Panoptic data format currently, you can find more description about this demo script in this doc.

You may need to modify the demo script to run this demo on campus and shelf dataset. And if you are interested in supporting this feature in mmpose, please let us know. We are glad to help.

Hi, I prepare some images and camera parameters like your default demo data format and modify the input data in this demo script body3d_multiview_detect_and_regress_img_demo.py . But I meet this error: Is there necessary to modify models/detectors/multiview_pose.py , or can you tell me how to fix it ? By the way, I am interested in supporting this feature in mmpose. Thanks! @liqikai9 @wusize

[ly015]

@wusize Could you please help?

[wusize]

@Taylorminer Hi, thanks for using mmpose. This demo script body3d_multiview_detect_and_regress_img_demo.py only support CMU Panoptic data format currently, you can find more description about this demo script in this doc. You may need to modify the demo script to run this demo on campus and shelf dataset. And if you are interested in supporting this feature in mmpose, please let us know. We are glad to help.

Hi, I prepare some images and camera parameters like your default demo data format and modify the input data in this demo script body3d_multiview_detect_and_regress_img_demo.py . But I meet this error: Is there necessary to modify models/detectors/multiview_pose.py , or can you tell me how to fix it ? By the way, I am interested in supporting this feature in mmpose. Thanks! @liqikai9 @wusize

Hi, Taylorminer! The demo for campus/shelf is not supported by now as pre-computed 2D heatmaps are used for the inference. For the demo, we can either add a 2D model to generate heatmaps or use pre-computed ones. We'll do it soon. And also welcome your kind suggestion/contribution.

[Taylorminer]

@Taylorminer Hi, thanks for using mmpose. This demo script body3d_multiview_detect_and_regress_img_demo.py only support CMU Panoptic data format currently, you can find more description about this demo script in this doc. You may need to modify the demo script to run this demo on campus and shelf dataset. And if you are interested in supporting this feature in mmpose, please let us know. We are glad to help.

Hi, I prepare some images and camera parameters like your default demo data format and modify the input data in this demo script body3d_multiview_detect_and_regress_img_demo.py . But I meet this error: Is there necessary to modify models/detectors/multiview_pose.py , or can you tell me how to fix it ? By the way, I am interested in supporting this feature in mmpose. Thanks! @liqikai9 @wusize

Hi, Taylorminer! The demo for campus/shelf is not supported by now as pre-computed 2D heatmaps are used for the inference. For the demo, we can either add a 2D model to generate heatmaps or use pre-computed ones. We'll do it soon. And also welcome your kind suggestion/contribution.

Thanks for your work ! Do mmpose support to output the 2D heatmap ? If I want to test voxelpose on a custom scene with 3 cameras, how can I do it with mmpose ?