Self-Ensembling GAN for Cross-Domain Semantic Segmentation

12/15/2021
by   Yonghao Xu, et al.
28

Deep neural networks (DNNs) have greatly contributed to the performance gains in semantic segmentation. Nevertheless, training DNNs generally requires large amounts of pixel-level labeled data, which is expensive and time-consuming to collect in practice. To mitigate the annotation burden, this paper proposes a self-ensembling generative adversarial network (SE-GAN) exploiting cross-domain data for semantic segmentation. In SE-GAN, a teacher network and a student network constitute a self-ensembling model for generating semantic segmentation maps, which together with a discriminator, forms a GAN. Despite its simplicity, we find SE-GAN can significantly boost the performance of adversarial training and enhance the stability of the model, the latter of which is a common barrier shared by most adversarial training-based methods. We theoretically analyze SE-GAN and provide an 𝒪(1/√(N)) generalization bound (N is the training sample size), which suggests controlling the discriminator's hypothesis complexity to enhance the generalizability. Accordingly, we choose a simple network as the discriminator. Extensive and systematic experiments in two standard settings demonstrate that the proposed method significantly outperforms current state-of-the-art approaches. The source code of our model will be available soon.

READ FULL TEXT

page 1

page 4

page 5

page 7

page 12

research
10/22/2021

Semi-Supervised Semantic Segmentation of Vessel Images using Leaking Perturbations

Semantic segmentation based on deep learning methods can attain appealin...
research
10/27/2019

Multi-source Domain Adaptation for Semantic Segmentation

Simulation-to-real domain adaptation for semantic segmentation has been ...
research
04/30/2020

Improving Semantic Segmentation via Self-Training

Deep learning usually achieves the best results with complete supervisio...
research
10/26/2022

SCP-GAN: Self-Correcting Discriminator Optimization for Training Consistency Preserving Metric GAN on Speech Enhancement Tasks

In recent years, Generative Adversarial Networks (GANs) have produced si...
research
01/01/2023

Self-Supervised Object Segmentation with a Cut-and-Pasting GAN

This paper proposes a novel self-supervised based Cut-and-Paste GAN to p...
research
06/14/2018

EL-GAN: Embedding Loss Driven Generative Adversarial Networks for Lane Detection

Convolutional neural networks have been successfully applied to semantic...
research
09/29/2018

Deep Adversarial Training for Multi-Organ Nuclei Segmentation in Histopathology Images

Nuclei segmentation is a fundamental task that is critical for various c...

Please sign up or login with your details

Forgot password? Click here to reset