Batch incremental 3D object detection on ScanNet V2 (val) benchmark leaderboard · SOTA2 Research

Batch incremental 3D object detection on ScanNet V2 (val)

72.85mAP@0.25 (Base)FI3Det

Evaluation Results

Method	Links
FI3Det2026.04		72.85	35.58	70.78
FI3Det2026.04		72.84	38.48	70.94
Baseline2026.04		72.77	—	—
Baseline2026.04		72.77	—	—
Imprinting2026.04		72.77	6.52	39.64
Imprinting2026.04		72.77	7.1	39.94
FI3Det2026.04		72.28	30.23	51.26
FI3Det2026.04		72.27	30.81	51.54
VLM-vanilla2026.04		71.81	7.5	68.24
VLM-vanilla2026.04		71.81	14.09	68.6
VLM-vanilla2026.04		71.79	17.12	44.45
VLM-vanilla2026.04		71.78	16.72	44.25
AIC3DOD2026.04		71.66	8.94	40.3
Baseline2026.04		71.47	—	—
Baseline2026.04		71.47	—	—
Imprinting2026.04		71.47	1.81	67.62
Imprinting2026.04		71.47	0.23	67.72
AIC3DOD2026.04		70.54	4.59	66.88
AIC3DOD2026.04		69.97	15.43	42.7
IL-DETR2026.04		69.78	0.03	65.91
SDCOT++2026.04		67.75	0.05	63.99
AIC3DOD2026.04		67.44	0.07	63.69
IL-DETR2026.04		67.05	13.82	40.43
IL-DETR2026.04		65.77	6.02	35.9
IL-DETR2026.04		65.63	0.35	62
SDCOT++2026.04		62.12	0.09	58.68
SDCOT++2026.04		35.87	1.35	18.61
SDCOT++2026.04		28.3	7.77	18.03