Learning to Update for Object Tracking with Recurrent Meta-learner

TL;DR

This paper introduces a meta-learning approach to model update in object tracking, training a recurrent neural network to learn an online updater that improves tracker performance and speed.

Contribution

It formulates the model update as a meta-learning problem and learns an online updater using offline videos, which is a novel approach in object tracking.

Findings

The learned updater improves base trackers consistently.

It runs faster than real-time on GPU.

It outperforms traditional update methods on benchmarks.

Abstract

Model update lies at the heart of object tracking. Generally, model update is formulated as an online learning problem where a target model is learned over the online training set. Our key innovation is to \emph{formulate the model update problem in the meta-learning framework and learn the online learning algorithm itself using large numbers of offline videos}, i.e., \emph{learning to update}. The learned updater takes as input the online training set and outputs an updated target model. As a first attempt, we design the learned updater based on recurrent neural networks (RNNs) and demonstrate its application in a template-based tracker and a correlation filter-based tracker. Our learned updater consistently improves the base trackers and runs faster than realtime on GPU while requiring small memory footprint during testing. Experiments on standard benchmarks demonstrate that our learned updater outperforms commonly used update baselines including the efficient exponential moving average (EMA)-based update and the well-designed stochastic gradient descent (SGD)-based update. Equipped with our learned updater, the template-based tracker achieves state-of-the-art performance among realtime trackers on GPU.