Open xiapengchng opened 5 years ago
I actually find it annoying also that the sigma array is hardcoded (len and values) without any documentation about what is it. I guess this toolbox was made to evaluate a very restricted case of mscoco database which is a shame because with a little improvement it could be generalized.
I has my own keypoint dataset. Every object has 30 points (not 17). So when I evaluate the detection results. I find the dimension the metrics have something with sigmas
ious = np.zeros((len(dts), len(gts))) sigmas = np.array([.26, .25, .25, .35, .35, .79, .79, .72, .72, .62,.62, 1.07, 1.07, .87, .87, .89, .89])/10.0 vars = (sigmas * 2)**2 k = len(sigmas)
the position was cocoapi/PythonAPI/pycocotools/cocoeval.py fron line 206
So what did you do? How did you change sigmas
..i am having the same problem here for hand landmark detection..
same problem, any update?
Same problem, any update?
Same problem, any update?
Same problem, any update?
Same problem, any update?
I would also like to know how these values are determined.
I think issue #399 answers this question.
same problem
I has my own keypoint dataset. Every object has 30 points (not 17). So when I evaluate the detection results. I find the dimension the metrics have something with sigmas
ious = np.zeros((len(dts), len(gts))) sigmas = np.array([.26, .25, .25, .35, .35, .79, .79, .72, .72, .62,.62, 1.07, 1.07, .87, .87, .89, .89])/10.0 vars = (sigmas * 2)**2 k = len(sigmas)
the position was cocoapi/PythonAPI/pycocotools/cocoeval.py fron line 206