SelfMOTR: Revisiting MOTR with Self-Generating Detection Priors
This addresses tracking challenges for computer vision applications, but it is incremental as it builds on existing MOTR-like models.
The paper tackles the problem of poor detection performance and conflict between detection and association in end-to-end tracking transformers by introducing SelfMOTR, which uses self-generated detection priors, achieving strong performance on DanceTrack that competes with recent state-of-the-art methods.
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.