End-to-end Active Object Tracking via Reinforcement Learning

TL;DR

This paper introduces an end-to-end deep reinforcement learning approach for active object tracking, enabling the camera to autonomously follow objects with good generalization from simulation to real-world scenarios.

Contribution

It presents a novel deep RL framework with environment augmentation and custom rewards for end-to-end active object tracking, improving joint tuning and reducing human effort.

Findings

The tracker generalizes well to unseen scenarios in simulation.

It can recover from tracking failures effectively.

Simulated training shows potential for real-world application.

Abstract

We study active object tracking, where a tracker takes as input the visual observation (i.e., frame sequence) and produces the camera control signal (e.g., move forward, turn left, etc.). Conventional methods tackle the tracking and the camera control separately, which is challenging to tune jointly. It also incurs many human efforts for labeling and many expensive trial-and-errors in realworld. To address these issues, we propose, in this paper, an end-to-end solution via deep reinforcement learning, where a ConvNet-LSTM function approximator is adopted for the direct frame-toaction prediction. We further propose an environment augmentation technique and a customized reward function, which are crucial for a successful training. The tracker trained in simulators (ViZDoom, Unreal Engine) shows good generalization in the case of unseen object moving path, unseen object appearance, unseen background, and distracting object. It can restore tracking when occasionally losing the target. With the experiments over the VOT dataset, we also find that the tracking ability, obtained solely from simulators, can potentially transfer to real-world scenarios.