Uncertainty Gated Network for Land Cover Segmentation

TL;DR

This paper introduces an uncertainty gated network that models multiscale contexts and uses pixel-wise uncertainty to improve land cover segmentation accuracy in remote sensing images.

Contribution

It proposes a novel approach combining uncertainty estimation with gating mechanisms to enhance land cover segmentation performance.

Findings

Improved segmentation accuracy over baseline models

Effective modeling of multiscale contextual information

Utilization of heteroscedastic uncertainty for decision gating

Abstract

The production of thematic maps depicting land cover is one of the most common applications of remote sensing. To this end, several semantic segmentation approaches, based on deep learning, have been proposed in the literature, but land cover segmentation is still considered an open problem due to some specific problems related to remote sensing imaging. In this paper we propose a novel approach to deal with the problem of modelling multiscale contexts surrounding pixels of different land cover categories. The approach leverages the computation of a heteroscedastic measure of uncertainty when classifying individual pixels in an image. This classification uncertainty measure is used to define a set of memory gates between layers that allow a principled method to select the optimal decision for each pixel.