TransTrack: Multiple Object Tracking with Transformer
read the original abstract
In this work, we propose TransTrack, a simple but efficient scheme to solve the multiple object tracking problems. TransTrack leverages the transformer architecture, which is an attention-based query-key mechanism. It applies object features from the previous frame as a query of the current frame and introduces a set of learned object queries to enable detecting new-coming objects. It builds up a novel joint-detection-and-tracking paradigm by accomplishing object detection and object association in a single shot, simplifying complicated multi-step settings in tracking-by-detection methods. On MOT17 and MOT20 benchmark, TransTrack achieves 74.5\% and 64.5\% MOTA, respectively, competitive to the state-of-the-art methods. We expect TransTrack to provide a novel perspective for multiple object tracking. The code is available at: \url{https://github.com/PeizeSun/TransTrack}.
This paper has not been read by Pith yet.
Forward citations
Cited by 10 Pith papers
-
SAM 3: Segment Anything with Concepts
SAM 3 introduces promptable concept segmentation that doubles accuracy of prior systems on images and videos while improving standard SAM segmentation performance.
-
CalibFree: Self-Supervised View Feature Separation for Calibration-Free Multi-Camera Multi-Object Tracking
CalibFree enables calibration-free multi-camera tracking via self-supervised feature separation through single-view distillation and cross-view reconstruction, reporting 3% higher accuracy and 7.5% better F1 on tested...
-
GateMOT: Q-Gated Attention for Dense Object Tracking
GateMOT proposes Q-Gated Attention to enable linear-complexity, spatially aware attention for state-of-the-art dense object tracking on benchmarks like BEE24.
-
Radar-Informed 3D Multi-Object Tracking under Adverse Conditions
RadarMOT improves 3D multi-object tracking accuracy by using radar point clouds as direct observations to refine states and recover missed objects, achieving 12.7% higher AMOTA at long range and up to 10.3% in adverse...
-
OmniTrack++: Omnidirectional Multi-Object Tracking by Learning Large-FoV Trajectory Feedback
OmniTrack++ improves omnidirectional multi-object tracking with trajectory feedback through DynamicSSM stabilization, FlexiTrack instances, ExpertTrack Memory with Mixture-of-Experts, and adaptive Tracklet Management,...
-
PS-MOT: Cultivating Instance Awareness from Point Seeds for Multi-Object Tracking
PS-Track sets a new state-of-the-art for point-supervised multi-object tracking by converting point seeds into temporally consistent pseudo-labels via Temporal-Feedback Prompting, Point-Excited Wavelet Attention, and ...
-
SMAC: Spatial-Modal Joint Modeling and Adaptive Representation Collapse for Multimodal Object Tracking
SMAC introduces a spatial-modal fusion backbone and adaptive collapse network for multimodal MOT, reporting 63.31 HOTA and 79.21 MOTA on UniRTL RNT modality.
-
Time-series Meets Complex Motion Modeling: Robust and Computational-effective Motion Predictor for Multi-object Tracking
TCMP achieves SOTA MOT metrics (HOTA 63.4%, IDF1 65.0%, AssA 49.1%) with 0.014x parameters and 0.05x FLOPs of the previous best method by using a simple dilated TCN regressor.
-
Hypergraph-State Collaborative Reasoning for Multi-Object Tracking
HyperSSM integrates hypergraphs and state space models to let correlated objects mutually refine motion estimates, stabilizing trajectories under noise and occlusion for state-of-the-art multi-object tracking.
-
HieDG: A Hierarchical Discrete Geometry-Guided Framework for Multi-Animal Tracking
HieDG discretizes geometric cues with a hierarchical residual codebook and integrates the resulting tokens into query-based trackers to boost identity consistency on animal and generic MOT benchmarks.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.