Hierarchical Attentive Recurrent Tracking

TL;DR

This paper introduces a hierarchical attentive recurrent model inspired by human visual attention for class-agnostic object tracking in cluttered videos, improving focus on relevant features through multiple attention layers.

Contribution

It presents a novel fully differentiable hierarchical attention framework for object tracking, trained with auxiliary tasks to enhance convergence.

Findings

Effective in cluttered environments

Performs well on KTH and KITTI datasets

Outperforms baseline methods

Abstract

Class-agnostic object tracking is particularly difficult in cluttered environments as target specific discriminative models cannot be learned a priori. Inspired by how the human visual cortex employs spatial attention and separate "where" and "what" processing pathways to actively suppress irrelevant visual features, this work develops a hierarchical attentive recurrent model for single object tracking in videos. The first layer of attention discards the majority of background by selecting a region containing the object of interest, while the subsequent layers tune in on visual features particular to the tracked object. This framework is fully differentiable and can be trained in a purely data driven fashion by gradient methods. To improve training convergence, we augment the loss function with terms for a number of auxiliary tasks relevant for tracking. Evaluation of the proposed model is performed on two datasets: pedestrian tracking on the KTH activity recognition dataset and the more difficult KITTI object tracking dataset.