Object Detection on CORS-ADD (test) benchmark leaderboard · SOTA2 Research

Object Detection on CORS-ADD (test)

72.06AP (bbox)LEVIRDetNet

Evaluation Results

Method	Links
LEVIRDetNet2026.06		72.06	94.88	83.56
YOLOv12x2026.06		71.7	95.46	82.69
RTMDet2026.06		69.4	92.78	80.58
DEIMv22026.06		68.26	95.28	79.41
YOLOX2026.06		67.99	94.25	79.64
DEIMv22026.06		67.39	94.27	78.14
DEIM2026.06		67.08	94.3	78.76
DynamicVis-base2026.06		65	89.4	74.2
DynamicVis-large2026.06		64.6	90.9	74.5
DINO2026.06		63.54	—	74.72
HrNet2026.06		63.19	89.61	72.68
EfficientDet2026.06		62.77	91.1	71.76
TridentNet2026.06		61.58	88.05	69.81
Fast R-CNN2026.06		60.79	89.19	69.75
Faster R-CNN2026.06		59.84	87.94	67.86
FCOS2026.06		59.19	89.66	65.47
Deformable DETR2026.06		58.16	88.78	66.06
SSD2026.06		57.58	89.13	65.84
YOLOv32026.06		52.17	85.8	56.89
Sparse R-CNN2026.06		51.79	76.9	58.21
RetinaNet2026.06		49.15	80.58	53.1
YOLOF2026.06		43.69	76.05	45.66
CenterNet2026.06		43.55	82.08	41.62