Learning a Discriminative Feature Network for Semantic Segmentation
Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, Nong Sang
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/ycszen/TorchSegpytorch★ 0
- github.com/YuhuiMa/DFN-tensorflowtf★ 0
- github.com/akinoriosamura/TorchSeg-mirrorpytorch★ 0
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.
Tasks
Benchmark Results
| Dataset | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| Cityscapes test | Smooth Network with Channel Attention Block | Mean IoU (class) | 80.3 | — | Unverified |
| Cityscapes test | DFN (ResNet-101) | Mean IoU (class) | 79.3 | — | Unverified |
| PASCAL VOC 2012 test | Smooth Network with Channel Attention Block | Mean IoU | 86.2 | — | Unverified |
| PASCAL VOC 2012 test | DFN (ResNet-101) | Mean IoU | 82.7 | — | Unverified |
| PASCAL VOC 2012 val | DFN (ResNet-101) | mIoU | 80.6 | — | Unverified |