Semantic Segmentation using Adversarial Networks

TL;DR

This paper introduces an adversarial training method for semantic segmentation that improves accuracy by detecting and correcting higher-order inconsistencies between predicted and ground truth segmentation maps.

Contribution

It presents a novel adversarial training framework for semantic segmentation, leveraging a discriminator network to enhance segmentation quality.

Findings

Improved accuracy on Stanford Background dataset

Enhanced performance on PASCAL VOC 2012 dataset

Demonstrated effectiveness of adversarial training in segmentation

Abstract

Adversarial training has been shown to produce state of the art results for generative image modeling. In this paper we propose an adversarial training approach to train semantic segmentation models. We train a convolutional semantic segmentation network along with an adversarial network that discriminates segmentation maps coming either from the ground truth or from the segmentation network. The motivation for our approach is that it can detect and correct higher-order inconsistencies between ground truth segmentation maps and the ones produced by the segmentation net. Our experiments show that our adversarial training approach leads to improved accuracy on the Stanford Background and PASCAL VOC 2012 datasets.