4D Scene Segmentation on Neural3DV benchmark leaderboard · SOTA2 Research

4D Scene Segmentation on Neural3DV

96.67mIoU (coffee_martini)Inst4DGS

Evaluation Results

Method	Links
Inst4DGS2026.03		96.67	—	—	—	—	—	99.8	94.2	99.54	97.6	99.64	95.91	99.09	95.19	99.68
TRASE2026.03		91.2	—	—	—	—	—	99.48	91.29	99.51	91.03	99.47	87.16	99.3	90.44	99.47
SA4D2026.03		85.83	—	—	—	—	—	99.1	89.87	99.41	86.45	99.14	88.98	99.4	90.47	99.48
CGC2026.03		83.59	—	—	—	—	—	98.95	90.34	99.45	90.52	99.43	85.3	99.2	87.94	99.34
GaussianGrouping2026.03		82.95	—	—	—	—	—	98.9	89.74	99.41	95.12	99.72	82.79	99	92.61	99.59
OpenGaussian2026.03		82.54	—	—	—	—	—	98.96	63.36	97.98	91.15	99.51	81.99	99.07	89.86	99.43
DGD2026.03		78.75	—	—	—	—	—	98.65	81.5	98.83	81.7	98.77	67.71	97.76	76.38	98.54
SAGA2026.03		22.01	—	—	—	—	—	80.81	81.25	98.81	69.82	97.67	84.39	99.11	89.59	99.41
DGD2025.12		—	46	98.2	77.3	59.5	62.7	—	—	—	—	—	—	—	—	—
OmniSeg3D2025.12		—	70	99.6	94.8	80.7	79.7	—	—	—	—	—	—	—	—	—
SA4D2025.12		—	66.8	98.7	86.1	75.9	81.2	—	—	—	—	—	—	—	—	—
SADG2025.12		—	64.9	98.5	82.7	74.4	81.3	—	—	—	—	—	—	—	—	—
Split4D2025.12		—	80.1	99.8	90.3	87	83.3	—	—	—	—	—	—	—	—	—