Distractor-Aware Fast Tracking via Dynamic Convolutions and MOT Philosophy

Bineng Zhong

Abstract: A practical long-term tracker typically contains three key properties, i.e. an efficient model design, an effective global re-detection strategy and a robust distractor aware- ness mechanism. However, most state-of-the-art long-term trackers (e.g., Pseudo and re-detecting based ones) do not take all three key properties into account and therefore may either be time-consuming or drift to distractors. To ad- dress the issues, we propose a two-task tracking frame- work (named DMTrack), which utilizes two core compo- nents (i.e., one-shot detection and re-identification (re-id) association) to achieve distractor-aware fast tracking via Dynamic convolutions (d-convs) and Multiple object track- ing (MOT) philosophy. To achieve precise and fast global detection, we construct a lightweight one-shot detector us- ing a novel dynamic convolutions generation method, which provides a unified and more flexible way for fusing target information into the search field. To distinguish the tar- get from distractors, we resort to the philosophy of MOT to reason distractors explicitly by maintaining all potential similarities’ tracklets. Benefited from the strength of high recall detection and explicit object association, our tracker achieves state-of-the-art performance on the LaSOT, Ox- UvA, TLP, VOT2018LT and VOT2019LT benchmarks and runs in real-time (3x faster than comparisons).