Mobile Vision Transformer-based Visual Object Tracking

TL;DR

This paper introduces a lightweight, accurate, and fast visual object tracking algorithm using Mobile Vision Transformers, which outperforms existing lightweight trackers and even larger models in accuracy and speed.

Contribution

First application of MobileViT backbone in object tracking, with a novel fusion approach for improved feature encoding and performance.

Findings

Outperforms recent lightweight trackers on GOT10k and TrackingNet datasets.

Surpasses DiMP-50 in accuracy while using fewer parameters and faster inference.

Achieves high speed and accuracy balance suitable for real-time applications.

Abstract

The introduction of robust backbones, such as Vision Transformers, has improved the performance of object tracking algorithms in recent years. However, these state-of-the-art trackers are computationally expensive since they have a large number of model parameters and rely on specialized hardware (e.g., GPU) for faster inference. On the other hand, recent lightweight trackers are fast but are less accurate, especially on large-scale datasets. We propose a lightweight, accurate, and fast tracking algorithm using Mobile Vision Transformers (MobileViT) as the backbone for the first time. We also present a novel approach of fusing the template and search region representations in the MobileViT backbone, thereby generating superior feature encoding for target localization. The experimental results show that our MobileViT-based Tracker, MVT, surpasses the performance of recent lightweight trackers on the large-scale datasets GOT10k and TrackingNet, and with a high inference speed. In addition, our method outperforms the popular DiMP-50 tracker despite having 4.7 times fewer model parameters and running at 2.8 times its speed on a GPU. The tracker code and models are available at https://github.com/goutamyg/MVT