Detectron: evaluation of kps

[jinxuan777]

I trained a model of keypoints detection with own datasets. reval.py is used to evaluate the mAP of results. I found compute_oks set
sigmas = np.array([.26, .25, .25, .35, .35, .79, .79, .72, .72, .62,.62, 1.07, 1.07, .87, .87, .89, .89])/10.0
How could I recompute the sigmas ? @rbgirshick

[gadcam]

@jinxuan777
keypoints.py reads
https://github.com/facebookresearch/Detectron/blob/e5bb3a8ff0b9caf59c76037726f49465d6b9678b/detectron/utils/keypoints.py#L16
so the OKS(Object Keypoint Similarity) is defined here http://cocodataset.org/#keypoints-eval

The sigmas used to compute it are defined in the page

As discussed, we set κ_i=2σ_i for each keypoint type i. For people, the σ's are .026, .025, .035, .079, .072, .062, .107, .087, & .089 for the nose, eyes, ears, shoulders, elbows, wrists, hips, knees, & ankles, respectively.

and the procedure used to obtain them is detailed in 1.3. Tuning OKS.