Referring Expression Segmentation on ReferIt (test) benchmark leaderboard · SOTA2 Research

Referring Expression Segmentation on ReferIt (test)

34.02Precision @ 0.5 IoUrecurrent and convolutional network model

Evaluation Results

Method
recurrent and convolutional network model2016.03	34.02	26.71	19.32	11.63	3.92	48.03
recurrent and convolutional network model2016.03	29.54	21.61	13.69	5.94	0.75	45.57
GroundeR2016.03	14.09	9.62	5.78	2.65	0.62	20.09
MCG classification2016.03	12.72	9.88	7.38	4.73	1.88	18.08
SCRC2016.03	11.91	7.71	4.33	1.78	0.36	17.84
GroundeR2016.03	11.08	6.2	2.74	0.78	0.2	20.5
per-word average2016.03	10.97	5.94	2.35	0.45	0	27.23
per-word union2016.03	10.46	5.65	2.28	0.44	0	19.37
SCRC2016.03	9.73	4.43	1.51	0.27	0.03	21.72
per-word intersection2016.03	9.58	5.35	2.2	0.43	0	26.69
whole image2016.03	5.07	2.85	1.58	0.81	0.41	15.12