Fix metric for small distances #32
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- all keypoints are in ints (and represent topleft corner of their pixel, aka zero-indexed) - heatmaps are really centered on the int pixel - metric max distances are ints and represent <= L2 distances (i.e. pixel-perfect == dmax = 0) - integration test: dummy dataset results in good scores for all mAP threholds down to 0
…ection into metric-scrutinization
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a bug in the metrics that caused the heatmap generation and hence the metric to fail for d=1 if the keypoints were annotated as floats.
This is now fixed by casting keypoints to the top-left (zero-index as in COCO) keypoint before training.