Hierarchical Recurrent Filtering for Fully Convolutional DenseNets

TL;DR

This paper introduces a hierarchical recurrent filtering architecture that extends single-frame segmentation models to process multiple frames, improving robustness in perception tasks by effectively handling data perturbations.

Contribution

The work presents a novel, parameter-efficient hierarchical recurrent filtering method that enhances existing dense convolutional networks for multi-frame perception tasks.

Findings

Model effectively copes with data perturbations

Hierarchical recurrent filtering improves robustness

Decouples temporal dependencies from scene representation

Abstract

Generating a robust representation of the environment is a crucial ability of learning agents. Deep learning based methods have greatly improved perception systems but still fail in challenging situations. These failures are often not solvable on the basis of a single image. In this work, we present a parameter-efficient temporal filtering concept which extends an existing single-frame segmentation model to work with multiple frames. The resulting recurrent architecture temporally filters representations on all abstraction levels in a hierarchical manner, while decoupling temporal dependencies from scene representation. Using a synthetic dataset, we show the ability of our model to cope with data perturbations and highlight the importance of recurrent and hierarchical filtering.