-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Feature] add a score_per_joint option in test_cfg #1169
Labels
enhancement
New feature or request
Comments
piercus
added a commit
to piercus/mmpose
that referenced
this issue
Feb 5, 2022
6 tasks
piercus
added a commit
to piercus/mmpose
that referenced
this issue
Feb 10, 2022
ly015
pushed a commit
that referenced
this issue
Feb 11, 2022
* [Fix] #1169 score_per_joint * [Fix] docstring * [Fix] #1169 coverage following review remark #1170 (comment)
ly015
pushed a commit
to ly015/mmpose
that referenced
this issue
Feb 11, 2022
* [Fix] open-mmlab#1169 score_per_joint * [Fix] docstring * [Fix] open-mmlab#1169 coverage following review remark open-mmlab#1170 (comment)
shuheilocale
pushed a commit
to shuheilocale/mmpose
that referenced
this issue
May 6, 2023
* [Fix] open-mmlab#1169 score_per_joint * [Fix] docstring * [Fix] open-mmlab#1169 coverage following review remark open-mmlab#1170 (comment)
HAOCHENYE
pushed a commit
to HAOCHENYE/mmpose
that referenced
this issue
Jun 27, 2023
ajgrafton
pushed a commit
to ajgrafton/mmpose
that referenced
this issue
Mar 6, 2024
* [Fix] open-mmlab#1169 score_per_joint * [Fix] docstring * [Fix] open-mmlab#1169 coverage following review remark open-mmlab#1170 (comment)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Related to #9 (comment)
Describe the feature
It would be great to have a 'score_per_joint' option in test_cfg in order to output one score per joint, instead of having only a global score for the pose, my use case is related to associative embedding
Motivation
In the context of bottom-up models, currently, we have 2 options :
return_heatmap=False
will only return the pose with one score per posereturn_heatmap=True
will give the full details about heatmaps as outputThose 2 choices are inconvenient if you do not want to carry the whole heatmaps, and you still want to know exaclty which joint, in the pose is present for sure, and which joint might be occluded, it is even more the case when using
refine=True
and missing joint are completed, loosing the information about which joint is confident and which one is not.I feel that all the use cases related to detecting occluded parts vs non-occlueded parts may benefit from this feature
The text was updated successfully, but these errors were encountered: