SelfMOTR: Revisiting MOTR with Self-Generating Detection Priors
By: Fabian Gülhan , Emil Mededovic , Yuli Wu and more
Potential Business Impact:
Tracks moving objects better by using its own smart guesses.
Despite progress toward end-to-end tracking with transformer architectures, poor detection performance and the conflict between detection and association in a joint architecture remain critical concerns. Recent approaches aim to mitigate these issues by (i) employing advanced denoising or label assignment strategies, or (ii) incorporating detection priors from external object detectors via distillation or anchor proposal techniques. Inspired by the success of integrating detection priors and by the key insight that MOTR-like models are secretly strong detection models, we introduce SelfMOTR, a novel tracking transformer that relies on self-generated detection priors. Through extensive analysis and ablation studies, we uncover and demonstrate the hidden detection capabilities of MOTR-like models, and present a practical set of tools for leveraging them effectively. On DanceTrack, SelfMOTR achieves strong performance, competing with recent state-of-the-art end-to-end tracking methods.
Similar Papers
MUT3R: Motion-aware Updating Transformer for Dynamic 3D Reconstruction
CV and Pattern Recognition
Fixes wobbly 3D pictures from moving cameras.
DepTR-MOT: Unveiling the Potential of Depth-Informed Trajectory Refinement for Multi-Object Tracking
CV and Pattern Recognition
Helps robots see through crowds and obstacles.
Attention-Bayesian Hybrid Approach to Modular Multiple Particle Tracking
Machine Learning (Stat)
Tracks many moving things even in messy scenes.