Learning a Discriminative Feature Network for Semantic Segmentation

TL;DR

This paper introduces a Discriminative Feature Network (DFN) with two sub-networks to improve semantic segmentation by addressing intra-class inconsistency and inter-class indistinction, achieving state-of-the-art results.

Contribution

The paper proposes a novel DFN architecture with Smooth and Border sub-networks, incorporating attention and boundary supervision for enhanced segmentation performance.

Findings

Achieved 86.2% mean IOU on PASCAL VOC 2012

Achieved 80.3% mean IOU on Cityscapes

Outperforms previous state-of-the-art methods

Abstract

Most existing methods of semantic segmentation still suffer from two aspects of challenges: intra-class inconsistency and inter-class indistinction. To tackle these two problems, we propose a Discriminative Feature Network (DFN), which contains two sub-networks: Smooth Network and Border Network. Specifically, to handle the intra-class inconsistency problem, we specially design a Smooth Network with Channel Attention Block and global average pooling to select the more discriminative features. Furthermore, we propose a Border Network to make the bilateral features of boundary distinguishable with deep semantic boundary supervision. Based on our proposed DFN, we achieve state-of-the-art performance 86.2% mean IOU on PASCAL VOC 2012 and 80.3% mean IOU on Cityscapes dataset.