Visual Object Tracking On Got 10K

评估指标

Average Overlap

Success Rate 0.5

Success Rate 0.75

评测结果

各个模型在此基准测试上的表现结果

				Paper Title	Repository
SAMURAI-L	81.7	92.2	76.9	SAMURAI: Adapting Segment Anything Model for Zero-Shot Visual Tracking with Motion-Aware Memory
DAM4SAM	81.1	-	-	A Distractor-Aware Memory for Visual Object Tracking with SAM2
MITS	80.4	89.8	75.8	Integrating Boxes and Masks: A Multi-Object Framework for Unified Visual Tracking and Segmentation
MCITrack-L384	80.0	88.5	80.2	Exploring Enhanced Contextual Information for Video-Level Object Tracking
ARTrackV2-L	79.5	87.8	79.6	ARTrackV2: Prompting Autoregressive Tracker Where to Look and How to Describe
LoRAT-g-378	78.9	87.8	80.7	Tracking Meets LoRA: Faster Training, Larger Model, Stronger Performance
ARTrack-L	78.5	87.4	77.8	Autoregressive Visual Tracking	-
ODTrack-L	78.2	-	-	ODTrack: Online Dense Temporal Token Learning for Visual Tracking
MCITrack-B224	77.9	88.2	76.8	Exploring Enhanced Contextual Information for Video-Level Object Tracking
RTracker-L	77.9	87	76.9	RTracker: Recoverable Tracking via PN Tree Structured Memory
LoRAT-L-378	77.5	86.2	78.1	Tracking Meets LoRA: Faster Training, Larger Model, Stronger Performance
HIPTrack	77.4	88.0	74.5	HIPTrack: Visual Tracking with Historical Prompts
ODTrack-B	77.0	-	-	ODTrack: Online Dense Temporal Token Learning for Visual Tracking
TATrack-L-GOT	76.6	85.7	73.4	Target-Aware Tracking with Long-term Context Attention
DropMAE	75.9	86.8	72	DropMAE: Learning Representations via Masked Autoencoders with Spatial-Attention Dropout for Temporal Matching Tasks
NeighborTrack-OSTrack	75.7	85.72	73.3	NeighborTrack: Improving Single Object Tracking by Bipartite Matching with Neighbor Tracklets
MixViT-L(ConvMAE)	75.7	85.3	75.1	MixFormer: End-to-End Tracking with Iterative Mixed Attention
MixFormer-L	75.6	85.73	72.8	MixFormer: End-to-End Tracking with Iterative Mixed Attention
SeqTrack-L384	74.8	81.9	72.2	Unified Sequence-to-Sequence Learning for Single- and Multi-Modal Visual Object Tracking
OSTrack-384	73.7	83.2	70.8	Joint Feature Learning and Relation Modeling for Tracking: A One-Stream Framework

0 of 39 row(s) selected.