A Symmetric Fusion Learning Model for Detecting Visual Relations and Scene Parsing

Table 2

Comparison with state-of-the-art on the VRD data set.

Method	Relation detection				Phrase detection
	k = 1		k = 70		k = 1			k = 70
	R@50	R@100	R@50	R@100	R@50	R@100	R@50		R@100
VRD [3]	17.03	16.17	24.90	20.04	14.70	13.86	21.51		17.35
KL distillation [23]	19.17	21.34	22.68	31.89	23.14	24.03	26.32		29.43
Zoom-net [25]	18.92	21.41	21.37	27.30	24.82	28.09	29.05		37.34
CAI + SCA-M [25]	19.54	22.39	22.34	28.52	25.21	28.89	29.64		38.39
RelDN [37]	18.92	22.96	21.52	26.38	26.37	31.42	28.24		35.44
AVR [42]	22.83	25.41	27.35	32.96	29.33	33.27	34.51		41.36
GPS-Net [43]	21.50	24.30	—	—	28.90	34.00	—		—
SABRA [44]	24.47	29.16	27.27	33.99	30.57	36.80	33.39		41.79
Ours