3D Point Cloud Recognition on ScanObjectNN Sim-to-Real variants benchmark leaderboard · SOTA2 Research

3D Point Cloud Recognition on ScanObjectNN Sim-to-Real variants

76.21Accuracy (MN.11 → SONN.11 OBJ)Uni3D + Point-Cache (Global)

Evaluation Results

Method
Uni3D + Point-Cache (Global)2026.03	76.21	77.26	59.1	74.5	76.5	62.47	71.01
Uni3D + BayesMM2026.03	74.31	77.05	62.48	72.5	76	63.32	70.94
Uni3D + Point-Cache (Hierarchical)2026.03	74.11	76	57.92	77.5	78	58.89	69.07
Uni3D2026.03	72.63	74.53	55.76	67.5	68.5	57.98	66.18
O-Shape + BayesMM2026.03	69.05	71.76	56.94	71.75	74.5	64	68
O-Shape + Point-Cache (Hierarchical)2026.03	66.11	69.68	47.5	71.5	71	56.57	63.78
ULIP + BayesMM2026.03	65.41	60.42	45.34	67.25	68.5	55.36	60.38
O-Shape + Point-Cache (Global)2026.03	65.07	68.67	46.23	71	71.5	55.67	63.02
ULIP + Point-Cache (Hierarchical)2026.03	64.42	56.63	35.77	67.25	64.5	47.61	56.03
ULIP + Point-Cache (Global)2026.03	62.32	52.63	34.97	65.5	62.5	47.36	54.21
O-Shape2026.03	59.78	62.53	45.51	64	70.25	53.55	59.27
ULIP-2 + BayesMM2026.03	57.89	58.97	50.25	61.25	64.25	56.22	58.14
ULIP-2 + Point-Cache (Hierarchical)2026.03	57.26	58.95	47.71	58	70.25	52.7	57.48
ULIP2026.03	57.05	50.32	32.6	61	61	44.38	51.06
ULIP-2 + Point-Cache (Global)2026.03	55.1	58.52	47.38	56.75	65	50.68	55.57
ULIP-22026.03	50.94	52.42	39.12	51.5	59.25	46.35	49.93