Siam R-CNN: Visual Tracking by Re-Detection

TL;DR

Siam R-CNN introduces a Siamese re-detection architecture combined with dynamic programming and hard example mining, significantly improving long-term visual object tracking performance across multiple benchmarks.

Contribution

The paper presents Siam R-CNN, a novel architecture that leverages re-detection, dynamic programming, and hard example mining for enhanced long-term object tracking.

Findings

Achieves state-of-the-art results on ten tracking benchmarks.

Excels particularly in long-term tracking scenarios.

Demonstrates robustness to distractors and occlusions.

Abstract

We present Siam R-CNN, a Siamese re-detection architecture which unleashes the full power of two-stage object detection approaches for visual object tracking. We combine this with a novel tracklet-based dynamic programming algorithm, which takes advantage of re-detections of both the first-frame template and previous-frame predictions, to model the full history of both the object to be tracked and potential distractor objects. This enables our approach to make better tracking decisions, as well as to re-detect tracked objects after long occlusion. Finally, we propose a novel hard example mining strategy to improve Siam R-CNN's robustness to similar looking objects. Siam R-CNN achieves the current best performance on ten tracking benchmarks, with especially strong results for long-term tracking. We make our code and models available at www.vision.rwth-aachen.de/page/siamrcnn.