Scene Graph Generation on Open Images V6 (test)

56.38wmAPrelRLIPv2-ParSeDA

Evaluation Results

Method
RLIPv2-ParSeDA2023.08	56.38	64.72	72.49	50.7	57.34	—	—
RLIPv2-ParSeDA2023.08	52.7	59.61	68.81	48.01	54.05	—	—
RLIPv2-ParSeDA2023.08	49.54	51.31	65.99	45.71	51.3	—	—
RLIPv2-ParSeDA2023.08	49.14	52.07	64.53	46.14	51.01	—	—
RLIPv2-ParSeDA2023.08	47.65	50.42	63.35	45.23	49.82	—	—
Dual-SGG2026.07	46	48	—	46.5	—	52.9	79.6
RLIPv2-ParSeDA2023.08	45.71	45.59	61.15	43.73	48.01	—	—
Salience-SGG2026.01	45.6	48	—	44.9	—	51.8	78.1
Salience-SGG2026.01	45.6	48	—	44.9	—	51.8	78.1
Salience-SGG2026.07	45.6	48	—	44.9	—	51.8	78.1
RLIPv2-ParSeD2023.08	44.73	44.88	60.2	43.36	47.28	—	—
RLIPv2-ParSeD2023.08	43.3	44.58	58.04	43.12	46.18	—	—
Hydra-SGG2026.01	42.8	—	—	44.3	—	50.1	76.1
Hydra-SGG2026.01	42.8	—	—	44.3	—	50.1	76.1
Hydra-SGG2026.07	42.8	—	—	44.3	—	50.1	76.1
EGTR2024.04	42	—	—	41.9	—	48.6	75
EGTR2026.01	42	—	—	41.9	—	48.6	75
EGTR2026.01	42	—	—	41.9	—	48.6	75
SSR-CNN2024.04	41.5	—	—	43.6	—	49.4	76.7
SSR-CNN2026.01	41.5	42.8	—	43.6	—	49.4	76.7
SSR-CNN2026.01	41.5	42.8	—	43.6	—	49.4	76.7
SSR-CNN2026.07	41.5	42.8	—	43.6	—	49.4	76.7
Structured Sparse R-CNN2021.06	41.47	42.84	76.66	43.64	49.38	—	—
Structured Sparse R-CNN2021.06	41.14	50.73	75.7	43.24	48.89	—	—
DRM2026.01	40.5	—	—	41.4	—	47.9	75.9
DRM2026.01	40.5	—	—	41.4	—	47.9	75.9
DRM2026.07	40.5	—	—	41.4	—	47.9	75.9
RelTR2026.01	37.2	—	—	37.5	—	43	71.7
RelTR2026.01	37.2	—	—	37.5	—	43	71.7
RelTR2026.07	37.2	—	—	37.5	—	43	71.7
SGTR2024.04	37	—	—	38.7	—	42.3	59.9
SGTR2026.01	37	42.6	—	38.7	—	42.3	59.9
SGTR2026.01	37	42.6	—	38.7	—	42.3	59.9
SGTR2026.07	37	42.6	—	38.7	—	42.3	59.9
SGTR2023.08	36.98	42.61	59.91	38.73	42.28	—	—
PE-Net2023.03	36.6	—	76.3	37.4	44.9	—	—
PE-Net2026.01	36.6	—	—	37.4	—	44.9	76.5
PE-Net2026.01	36.6	—	—	37.4	—	44.9	76.5
PE-Net2026.07	36.6	—	—	37.4	—	44.9	76.5
Mg-RMPN2026.01	35.5	45.5	—	36.4	—	43.6	74.2
Mg-RMPN2026.01	35.5	45.5	—	36.4	—	43.6	74.2
Mg-RMPN2026.07	35.5	45.5	—	36.4	—	43.6	77.8
RU-Net2023.03	35.4	—	76.9	34.9	43.5	—	—
RU-Net2023.04	35.4	—	76.9	34.9	43.5	—	—
RU-Net2026.01	35.4	—	—	34.9	—	43.5	76.9
RU-Net2026.01	35.4	—	—	34.9	—	43.5	76.9
RU-Net2026.07	35.4	—	—	34.9	—	43.5	76.9
HL-Net2023.04	35.1	—	76.5	34.7	43.2	—	—
SQUAT2023.04	34.9	—	75.8	35.9	43.5	—	—
SQUAT2026.07	34.9	—	—	35.9	—	43.5	75.8
VCTree2023.03	34.2	—	74.1	33.1	40.2	—	—
VCTree2023.04	34.2	—	76.1	33.1	42.1	—	—
VCTree2024.04	34.2	—	—	33.1	—	40.2	74.1
RelTR2024.04	34.2	—	—	37.5	—	43	71.7
RelTR2023.08	34.17	—	64.47	37.44	41.54	—	—
VCTree2021.06	34.16	33.91	74.08	33.11	40.21	—	—
VCTree2023.08	34.16	33.91	74.08	33.11	40.21	—	—
BGNN2021.06	33.51	40.45	74.98	34.15	42.06	—	—
G-RCNN2023.08	33.51	34.04	74.51	34.21	41.84	—	—
BGNN2023.03	33.5	—	75	34.2	42.1	—	—
BGNN2023.04	33.5	—	75	34.2	42.1	—	—
BGNN2024.04	33.5	—	—	34.2	—	42.1	75
BGNN2026.01	33.5	40.5	—	34.2	—	42.1	75
BGNN2026.01	33.5	40.5	—	34.2	—	42.1	75
BGNN2026.07	33.5	40.5	—	34.2	—	42.1	75
GR-CNN2023.03	33.2	—	74.5	34.2	41.8	—	—
G-RCNN2021.06	33.15	34.04	74.51	34.21	41.84	—	—
GPS-Net2023.03	32.9	—	74.8	34	41.7	—	—
GPS-Net2023.04	32.9	—	74.8	34	41.7	—	—
GPS-Net2024.04	32.9	—	—	34	—	41.7	74.8
GPS-Net2026.01	32.9	35.2	—	34	—	41.7	74.8
GPS-Net2026.01	32.9	35.2	—	34	—	41.7	74.8
GPS-Net2021.06	32.85	35.26	74.81	33.98	41.69	—	—
GPS-Net2023.08	32.85	35.26	74.81	33.98	41.69	—	—
GPS-Net2023.04	32.8	—	74.7	33.9	41.6	—	—
RelDN2023.04	32.2	—	73.1	33.4	40.9	—	—
RelDN2021.06	32.16	33.98	73.08	33.39	40.84	—	—
BGNN2023.08	31.15	39.41	74.93	31.37	40	—	—
Unbiased2023.08	30.74	35.47	69.3	32.8	39.27	—	—
VCTree+TDE2023.04	30.7	—	69.3	32.8	39.3	—	—
MOTIFS2021.06	29.91	32.68	71.63	31.59	38.93	—	—
Motifs2023.08	29.91	32.68	71.63	31.59	38.93	—	—
Motifs2023.03	29.9	—	71.6	31.6	38.9	—	—
Motifs2023.04	29.9	—	71.6	31.6	38.9	—	—
Motifs2024.04	29.9	—	—	31.6	—	38.9	71.6
RelDN2023.08	29.87	36.8	72.75	30.42	38.67	—	—