SOTA Pedestrian Detection benchmarks and papers with code

Benchmarks

Dataset Name	SOTA Method	Metric
CityPersons (val)	InterNet	AP (Reasonable)17.9	85	1mo ago
CrowdHuman (val)	Hulk	MR^-236.5	61	1mo ago
CrowdHuman	InternImage-H	mAP97.2	38	1mo ago
Caltech (test)	RPN+BF	MR9.6	36	1mo ago
KAIST (test)	AR-CNN	MR (Near)0	22	1mo ago
MultiviewX	TrackTacular	MODA96.5	21	1mo ago
Wildtrack		MODA94.1	21	1mo ago
CityPersons 1.0 (val)		Miss Rate (Reasonable)9.4	21	1mo ago
LLVIP (test)	AUNet	mAP@5097.2	20	1mo ago
Caltech	NMS-Ped	MR5.92	17	1mo ago
CrowdHuman (test)	Faster R-CNN (ours)	MR40.7	16	1mo ago
CityPersons highly occluded (HO)	PedesFormer	Miss Rate36.9	16	1mo ago
KAIST multispectral pedestrian dataset (All-dataset)		Detection Rate (Near)28.74	14	1mo ago
Waymo Open Dataset (val)	RSN	AP L1 3D80.2	12	1mo ago
KITTI Hard (val)	MS-CNN	AP64.08	12	1mo ago
KITTI Easy (val)	HyperLearner + Segmentation	AP79.43	12	1mo ago
KITTI Moderate (val)	HyperLearner + Edge	AP72.51	12	1mo ago
KITTI (test)		AP (Easy)73.14	12	1mo ago
Caltech Reasonable	PedHunter	Miss Rate2.3	11	1mo ago
CityPersons original image size (1024x2048 pixels) (test)	CornerNet	AP (Reasonable)21	11	1mo ago
CVC-14 (Night)	MBNet	MR^-213.5	11	1mo ago
CVC-14 (Day)	AR-CNN	MR^-224.7	11	1mo ago
Caltech standard (test)		Detection Rate (Reasonable)11.52	11	1mo ago
City Persons 6 (test)	F2DNet	MR-2 (Reasonable)8.7	10	1mo ago
KAIST reasonable subset (test)	MBNet	MR^-2 (IoU=0.5, All)8.13	10	1mo ago

Showing 25 of 68 rows