Share your thoughts, 1 month free Claude Pro on usSee more

Panoptic video scene graph generation on OpenPVSG v1 (test)

11.43Recall@20

Ours - Convolution (IPS+T Baseline)

Updated 1mo ago

Evaluation Results

Method
Ours - Convolution (IPS+T Baseline) 2026.01	11.43	9.57	17.3	13.13	22.85	17.48
Motion-aware Contrastive Learning (Convolution) 2024.12	11.43	9.57	17.3	13.13	22.85	17.48
Ours - Transformer (IPS+T Baseline) 2026.01	10.59	9.56	16.98	12.39	22.33	17.47
Motion-aware Contrastive Learning (Transformer) 2024.12	10.59	9.56	16.98	12.39	22.33	17.47
IPS+T - Convolution 2026.01	10.06	8.98	14.99	12.21	18.13	15.47
IPS+T - Convolution 2024.12	10.06	8.98	14.99	12.21	18.13	15.47
IPS+T - Transformer 2026.01	9.01	6.69	14.88	11.28	17.51	13.2
IPS+T - Transformer 2024.12	9.01	6.69	14.88	11.28	17.51	13.2
IPS+T - Vanilla 2026.01	8.28	5.68	14.47	9.92	18.24	11.84
IPS+T - Vanilla 2024.12	8.28	5.68	14.47	9.92	18.24	11.84
Ours - Convolution (VPS Baseline) 2026.01	8.18	8	12.9	11.47	14.22	13.59
Motion-aware Contrastive Learning (Convolution) 2024.12	8.18	8	12.9	11.47	14.22	13.59
IPS+T - Handcrafted filter 2026.01	8.07	5.61	13.42	8.27	16.46	10.11
VPS - Convolution 2026.01	8.07	7.84	11.01	9.78	12.89	10.77
IPS+T - Handcrafted filter 2024.12	8.07	5.61	13.42	8.27	16.46	10.11
VPS - Convolution 2024.12	8.07	7.84	11.01	9.78	12.89	10.77
Ours - Transformer (VPS Baseline) 2026.01	6.71	6.94	10.27	8.68	13.42	12.09
Motion-aware Contrastive Learning (Transformer) 2024.12	6.71	6.94	10.27	8.68	13.42	12.09
VPS - Transformer 2026.01	6.5	5.75	9.64	8.25	12.26	9.51
VPS - Transformer 2024.12	6.5	5.75	9.64	8.25	12.26	9.51
VPS - Vanilla 2026.01	6.29	3.04	9.64	6.74	12.89	9.6
VPS - Vanilla 2024.12	6.29	3.04	9.64	6.74	12.89	9.6
VPS - Handcrafted filter 2026.01	5.24	2.84	7.65	7.14	9.64	8.22
VPS - Handcrafted filter 2024.12	5.24	2.84	7.65	7.14	9.64	8.22
Ours - Convolution (IPS+T Baseline) 2026.01	4.51	3.56	6.08	4.38	7.76	5.86
Motion-aware Contrastive Learning (Convolution) 2024.12	4.51	3.56	6.08	4.38	7.76	5.86
Ours - Transformer (IPS+T Baseline) 2026.01	3.98	2.98	5.97	4.2	7.44	5.15
Motion-aware Contrastive Learning (Transformer) 2024.12	3.98	2.98	5.97	4.2	7.44	5.15
IPS+T - Transformer 2026.01	3.88	2.81	5.66	4.12	6.18	4.44
IPS+T - Convolution 2026.01	3.88	2.55	5.24	3.29	6.71	5.36
IPS+T - Transformer 2024.12	3.88	2.81	5.66	4.12	6.18	4.44
IPS+T - Convolution 2024.12	3.88	2.55	5.24	3.29	6.71	5.36
IPS+T - Vanilla 2026.01	3.04	1.35	4.61	2.94	5.56	3.33
IPS+T - Vanilla 2024.12	3.04	1.35	4.61	2.94	5.56	3.33
IPS+T - Handcrafted filter 2026.01	2.52	1.72	3.77	2.36	4.72	2.79
IPS+T - Handcrafted filter 2024.12	2.52	1.72	3.77	2.36	4.72	2.79
Ours - Convolution (VPS Baseline) 2026.01	0.84	0.98	1.26	1.22	1.26	1.22
Motion-aware Contrastive Learning (Convolution) 2024.12	0.84	0.98	1.26	1.22	1.26	1.22
Ours - Transformer (VPS Baseline) 2026.01	0.63	0.83	1.05	0.76	1.05	0.76
Motion-aware Contrastive Learning (Transformer) 2024.12	0.63	0.83	1.05	0.76	1.05	0.76
VPS - Handcrafted filter 2026.01	0.42	0.13	0.52	0.5	0.94	0.92
VPS - Transformer 2026.01	0.42	0.61	0.73	0.76	1.05	0.92
VPS - Convolution 2026.01	0.42	0.25	0.63	0.67	0.63	0.67
VPS - Handcrafted filter 2024.12	0.42	0.13	0.52	0.5	0.94	0.92
VPS - Transformer 2024.12	0.42	0.61	0.73	0.76	1.05	0.92
VPS - Convolution 2024.12	0.42	0.25	0.63	0.67	0.63	0.67
VPS - Vanilla 2026.01	0.21	0.1	0.21	0.1	0.31	0.18
VPS - Vanilla 2024.12	0.21	0.1	0.21	0.1	0.31	0.18