Visual Tracking by means of Deep Reinforcement Learning and an Expert Demonstrator

TL;DR

This paper introduces two deep reinforcement learning-based visual trackers, A3CT and A3CTD, which leverage expert demonstrations to improve tracking accuracy and efficiency, achieving state-of-the-art results in real-time on multiple benchmarks.

Contribution

The paper presents novel trackers that incorporate expert demonstrations into reinforcement learning for visual tracking, enhancing performance and real-time capability.

Findings

Achieve state-of-the-art performance on multiple benchmarks.

Operate in real-time during tracking.

Effectively utilize expert demonstrations to improve tracking accuracy.

Abstract

In the last decade many different algorithms have been proposed to track a generic object in videos. Their execution on recent large-scale video datasets can produce a great amount of various tracking behaviours. New trends in Reinforcement Learning showed that demonstrations of an expert agent can be efficiently used to speed-up the process of policy learning. Taking inspiration from such works and from the recent applications of Reinforcement Learning to visual tracking, we propose two novel trackers, A3CT, which exploits demonstrations of a state-of-the-art tracker to learn an effective tracking policy, and A3CTD, that takes advantage of the same expert tracker to correct its behaviour during tracking. Through an extensive experimental validation on the GOT-10k, OTB-100, LaSOT, UAV123 and VOT benchmarks, we show that the proposed trackers achieve state-of-the-art performance while running in real-time.