Robust Visual Tracking by Segmentation

TL;DR

This paper introduces a segmentation-based visual tracking method that produces precise object masks instead of bounding boxes, improving accuracy and robustness in complex scenarios, and achieves state-of-the-art results on LaSOT.

Contribution

The paper presents a novel segmentation-centric tracking pipeline with an instance localization component, enabling more accurate target representation and improved tracking performance.

Findings

Achieved a success AUC score of 69.7% on LaSOT.

Validated segmentation quality on video object segmentation datasets.

Outperformed previous trackers on challenging datasets.

Abstract

Estimating the target extent poses a fundamental challenge in visual object tracking. Typically, trackers are box-centric and fully rely on a bounding box to define the target in the scene. In practice, objects often have complex shapes and are not aligned with the image axis. In these cases, bounding boxes do not provide an accurate description of the target and often contain a majority of background pixels. We propose a segmentation-centric tracking pipeline that not only produces a highly accurate segmentation mask, but also internally works with segmentation masks instead of bounding boxes. Thus, our tracker is able to better learn a target representation that clearly differentiates the target in the scene from background content. In order to achieve the necessary robustness for the challenging tracking scenario, we propose a separate instance localization component that is used to condition the segmentation decoder when producing the output mask. We infer a bounding box from the segmentation mask, validate our tracker on challenging tracking datasets and achieve the new state of the art on LaSOT with a success AUC score of 69.7%. Since most tracking datasets do not contain mask annotations, we cannot use them to evaluate predicted segmentation masks. Instead, we validate our segmentation quality on two popular video object segmentation datasets.