Predicate Detection on Visual Relation Detection (VRD) (All) benchmark leaderboard · SOTA2 Research

Predicate Detection on Visual Relation Detection (VRD) (All)

86.14Recall@50U+W+SF+L: T+S

Evaluation Results

Method	Links
U+W+SF+L: T+S2017.07		86.14	—	95.19
U+W+SF+L: T+S2017.07		85.64	—	94.65
U+SF+L:T2017.07		83.84	—	87.71
U+W+L:T2017.07		83.26	—	88.98
U+W+SF+L: T2017.07		82.97	—	90.09
U+W+SF+L: T2017.07		82.54	—	89.41
Baseline: L only2017.07		80.64	—	85.34
U+W+SF+L: S2017.07		75.45	—	87.15
U+W+SF+L: S2017.07		74.98	—	86.97
U+W+SF2017.07		72.83	—	85.89
U+W+SF2017.07		72.29	—	84.89
U+W+L:S2017.07		71.83	—	84.94
U+SF+L:S2017.07		71.27	—	84.81
VRD-Full2017.07		70.97	—	84.34
U+W2017.07		70.75	—	83.78
Baseline: U2017.07		70.19	—	84.49
Baseline: U only2017.07		70.02	—	83.15
U+SF2017.07		69.87	—	83.68
U+W+SF+L: T+S2017.07		55.67	—	55.67
U+W+SF+L: T+S2017.07		55.16	—	55.16
U+W+SF+L: T2017.07		54.61	—	54.61
U+W+SF+L: T2017.07		54.13	—	54.13
U+W+L:T2017.07		52.96	—	52.96
U+SF+L:T2017.07		51.67	—	51.67
Baseline: L only2017.07		51.34	—	51.34
U+W+SF+L: S2017.07		48.61	—	48.61
VRD-Full2017.07		47.87	—	47.87
U+W+SF+L: S2017.07		47.5	—	47.5
U+W+L:S2017.07		42.98	—	42.98
U+W+SF2017.07		42.08	—	42.08
U+W+SF2017.07		41.33	—	41.33
U+SF+L:S2017.07		41.06	—	41.06
U+W2017.07		37.15	—	37.15
Baseline: U2017.07		36.97	—	36.97
U+SF2017.07		36.33	—	36.33
Baseline: U only2017.07		34.82	—	34.82
VRD-V only2017.07		28.36	—	37.2
VRD-V only2017.07		7.11	—	7.11
Joint CNN2017.07		2.03	—	2.03
Visual Phrases2017.07		1.91	—	1.91
FCL2021.03		—	58.31	—
MFURLN2021.03		—	58.2	—
MFURLN*2021.03		—	57.87	—