Situation Prediction on imSitu (test) benchmark leaderboard · SOTA2 Research

Situation Prediction on imSitu (test)

36.72Top-1 Verb AccFully-connected Graph

Evaluation Results

Method
Fully-connected Graph2017.08	36.72	27.52	19.25	61.9	45.39	29.96	—	—	41.4	69.16	41.36
Fusion, VGG+RNN2017.03	35.9	27.45	16.36	63.08	46.88	26.06	—	—	40.16	70.27	35.25
Chain RNN2017.08	34.63	24.65	17.89	61.06	41.73	28.15	—	—	38.54	62.94	37.32
Tensor composition + reg + self train2016.12	34.12	26.45	15.51	62.59	46.88	25.46	70.44	34.38	39.48	—	—
Tensor Composition + DataAug2017.08	34.12	26.45	15.51	62.59	46.88	25.46	—	—	39.48	70.44	34.38
Tensor Comp. + Image Reg. CRF + Extra 5M Images2017.03	34.12	26.45	15.51	62.59	46.88	25.46	—	—	39.48	70.44	34.38
Tensor composition + reg2016.12	32.96	25.32	14.57	60.12	44.64	24	69.2	32.97	37.97	—	—
Tensor Composition2017.08	32.96	25.32	14.57	60.12	44.64	24	—	—	37.97	69.2	32.97
Tensor Comp. + Image Reg. CRF2017.03	32.96	25.32	14.57	60.12	44.64	24	—	—	37.97	69.2	32.97
Baseline: Image Regression2016.12	32.34	24.64	14.19	58.88	42.76	22.55	65.66	28.96	36.25	—	—
CNN+CRF2017.08	32.34	24.64	14.19	58.88	42.76	22.55	—	—	36.25	65.66	28.96
Image Regression CRF2017.03	32.34	24.64	14.19	58.88	42.76	22.55	—	—	36.25	65.66	28.96
Baseline: Image Regression2016.12	32.3	24.95	14.77	59.52	44.08	23.99	67.82	31.46	37.36	—	—