Unveiling the Power of Deep Tracking

TL;DR

This paper investigates the limitations of deep features in object tracking and proposes an adaptive fusion method to enhance tracking accuracy and robustness, achieving significant improvements on challenging datasets.

Contribution

It systematically analyzes deep and shallow features, identifies key challenges, and introduces a novel adaptive fusion approach to improve deep tracking performance.

Findings

Significant 17% gain in EAO on VOT2017 dataset.

Deep and shallow feature fusion improves robustness and accuracy.

Identified data scarcity and low resolution as main challenges.

Abstract

In the field of generic object tracking numerous attempts have been made to exploit deep features. Despite all expectations, deep trackers are yet to reach an outstanding level of performance compared to methods solely based on handcrafted features. In this paper, we investigate this key issue and propose an approach to unlock the true potential of deep features for tracking. We systematically study the characteristics of both deep and shallow features, and their relation to tracking accuracy and robustness. We identify the limited data and low spatial resolution as the main challenges, and propose strategies to counter these issues when integrating deep features for tracking. Furthermore, we propose a novel adaptive fusion approach that leverages the complementary properties of deep and shallow features to improve both robustness and accuracy. Extensive experiments are performed on four challenging datasets. On VOT2017, our approach significantly outperforms the top performing tracker from the challenge with a relative gain of 17% in EAO.