Multi Object Tracking On Dancetrack

평가 지표

AssA

HOTA

IDF1

MOTA

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	AssA	HOTA	IDF1	MOTA	Paper Title	Repository
OC-SORT	38.0	55.1	54.2	89.4	Observation-Centric SORT: Rethinking SORT for Robust Multi-Object Tracking
MOTIP (DAB-Deformable DETR)	60.8	70.0	75.1	91.0	Multiple Object Tracking as ID Prediction
FairMOT	23.8	39.7	40.8	82.2	FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Tracking
MoveSORT	38.7	56.1	56.0	91.8	Beyond Kalman Filters: Deep Learning-Based Filters for Improved Object Tracking
C-TWiX	47.2	62.1	63.6	91.4	Learning Data Association for Multi-Object Tracking using Only Coordinates
MOTRv2	64.4	73.4	76.0	92.1	MOTRv2: Bootstrapping End-to-End Multi-Object Tracking by Pretrained Object Detectors
FCG	29.9	48.7	46.5	89.9	Multiple Object Tracking from appearance by hierarchically clustering tracklets
Hybrid-SORT	47.4	62.2	63.0	91.6	Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking
CO-MOT	58.9	69.4	71.9	91.2	Bridging the Gap Between End-to-end and Non-End-to-end Multi-Object Tracking
MotionTrack	41.7	58.2	58.6	91.3	MotionTrack: Learning Motion Predictor for Multiple Object Tracking	-
ByteTrack	31.5	47.1	51.9	88.2	ByteTrack: Multi-Object Tracking by Associating Every Detection Box
TransTrack	27.5	45.7	44.8	83.0	TransTrack: Multiple Object Tracking with Transformer
MeMOTR (Deformable DETR)	52.3	63.4	65.5	85.4	MeMOTR: Long-Term Memory-Augmented Transformer for Multi-Object Tracking
MOTIP (Deformable DETR, with CrowdHuman)	62.8	71.4	76.3	91.6	Multiple Object Tracking as ID Prediction
UCMCTrack	-	63.6	65.0	-	UCMCTrack: Multi-Object Tracking with Uniform Camera Motion Compensation
MeMOTR	58.4	68.5	71.2	89.9	MeMOTR: Long-Term Memory-Augmented Transformer for Multi-Object Tracking
MambaMOT	39.0	56.1	54.9	90.3	MambaMOT: State-Space Model as Motion Predictor for Multi-Object Tracking	-
CenterTrack	22.6	41.8	35.7	86.8	Tracking Objects as Points
DeepMoveSORT	48.6	63.0	65.0	92.6	Engineering an Efficient Object Tracker for Non-Linear Motion	-
Deep OC-SORT	45.8	61.3	61.5	92.3	Deep OC-SORT: Multi-Pedestrian Tracking by Adaptive Re-Identification

0 of 34 row(s) selected.