Open-set Object Detection on Cityscapes to BDD100k (3 novel categories) benchmark leaderboard · SOTA2 Research

Open-set Object Detection on Cityscapes to BDD100k (3 novel categories)

16.94mAPCCKM

Evaluation Results

Method	Links
CCKM2026.04		16.94	13.29	0.746	3,570
SOMA2026.04		16.81	15.67	0.869	4,220
OpenDet2026.04		16.01	14.87	0.948	4,254
PROSER2026.04		15.98	12.92	0.949	4,320
DDETR2026.04		15.91	0	0.908	7,402
OW-DETR2026.04		15.8	9.68	0.963	4,294
CCKM2026.04		14.34	0.91	0.07	360
SOMA2026.04		14.11	1.86	0.127	614
OpenDet2026.04		13.7	1.2	0.135	836
DDETR2026.04		13.48	0	0.153	1,448
PROSER2026.04		13.32	1.53	0.148	910
OW-DETR2026.04		13.15	1.27	0.129	792
CCKM2026.04		11.59	2.81	2.584	2,640
CCKM2026.04		11.55	2.26	1.467	4,122
SOMA2026.04		11.51	3.97	2.251	7,670
SOMA2026.04		11.17	4.56	2.556	7,420
OpenDet2026.04		10.5	3.26	2.308	9,760
OpenDet2026.04		10.47	1.68	3.228	13,578
DDETR2026.04		10.31	0	2.846	25,530
DDETR2026.04		10.02	0	3.054	22,108
OW-DETR2026.04		9.45	1.45	2.255	6,236
PROSER2026.04		9.17	2.38	2.525	13,200
PROSER2026.04		9.02	1.71	3.995	24,118
OW-DETR2026.04		8.11	1.75	2.785	9,602