Classes Matter: A Fine-Grained Adversarial Approach to Cross-Domain Semantic Segmentation

Wang, Haoran; Shen, Tong; Zhang, Wei; Duan, Ling-Yu; Mei, Tao

doi:10.1007/978-3-030-58568-6_38

Cited by 210 publications

(234 citation statements)

References 27 publications

Supporting

Mentioning

200

Contrasting

Order By: Relevance

“…In this work, we focus on seeking a set of reliable target predictions as guidance to facilitate the adaptation of the diverse target data. We adopt the fine-grained adversarial learning network [17] as our baseline and develop an end-to-end self-guided adaptive framework upon this model. As shown in Fig.…”

Section: Proposed Methodsmentioning

confidence: 99%

“…Other techniques, such as pseudo-label re-training [12,13], curriculum learning [14,15], and source-data selection [16], have also been exploited to reduce the cross-domain gap. Recently, in [17], the authors proposed a fine-grained discriminator to narrow the cross-domain gap through class-level alignment and showed promising performance over previous methods.…”

Section: Introductionmentioning

confidence: 99%

“…For example, in Fig. 1 (b), even though the source and target domains are aligned through the class-level discriminator [17], not all the pixels of the "road" class are correctly predicted. Thus, focusing on aligning the cross-domain gap alone is far from enough to achieve a high-quality prediction on real-world data.…”

Section: Introductionmentioning

confidence: 99%

“…Note that, we treat each class independently and determine a set of target pixels with high predication confidence as the guide of the corresponding class. With these high-confident guide pixels, we conduct class-level adaptation on the low-confident target pixels through the fine-grained adversarial learning [17] in an endto-end framework. Our contributions are summarized below:…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Self-Guided Adversarial Learning For Domain Adaptive Semantic Segmentation

Pang¹,

Chang²,

Hsu

2021

2021 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

Unsupervised domain adaptation has been introduced to generalize semantic segmentation models from labeled synthetic images to unlabeled real-world images. Although much effort was devoted to minimize the cross-domain gap, the segmentation results on real-world data remain highly unstable. In this paper, we discuss two main issues which hinder previous methods from achieving satisfactory results and propose a novel self-guided adversarial learning to leverage the capability of domain adaptation. Firstly, to deal with the unpredictable data variation in the real-world domain, we develop a self-guided adversarial learning method by selecting reliable target pixels as guidance to lead the adaptation of the other pixels. Secondly, to address the class-imbalanced issue, we devise the selection strategy in each class independently and incorporate this idea with a class-level adversarial learning in a unified framework. Experimental results show that the proposed method significantly improves the previous methods on several benchmark datasets.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Self-Guided Adversarial Learning For Domain Adaptive Semantic Segmentation

Pang¹,

Chang²,

Hsu

2021

2021 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

show abstract

“…The other, called a discriminator network, aims to minimize the feature distribution discrepancy between two domains in adversarial learning. To make full use of class-level information, Wang et al [52] proposed fine-grained adversarial training for class-level feature alignment and preserved the internal structure of semantics across domains. Chen et al [53] developed an adversarial training method in the feature space for semantic segmentation while adopting global adaptation and category-level adaptation.…”

Section: B Domain Adaptationmentioning

confidence: 99%

Weakly-Supervised Domain Adaptation With Adversarial Entropy for Building Segmentation in Cross-Domain Aerial Imagery

Yao¹,

Wang²,

Wu³

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Building segmentation is a classical and challenging task in high-resolution remote sensing imagery. This approach has achieved remarkable performance based on a fully convolutional network (FCN) with adequate pixel-wise annotations. However, due to differences in sensor technology as well as appearance in different regions, datasets gathered from these various sources are quite distinct, and dense annotations for a particular area are not always available. Thus, directly applying a segmentation model trained on one dataset (source domain) to another unseen dataset (target domain) usually results in a drop in performance, called the domain gap. In this paper, we propose a weakly-supervised domain adaptation method using adversarial entropy for building segmentation to address this problem. First, we use an adversarial entropy strategy to decrease the entropy and improve the prediction certainty for target images, causing the distributions between the source and target domains to become closer to each other. Second, we propose a simple and effective self-training strategy for the target domain that produces high-confidence predictions using pseudo-labels. We use a series of thresholds to generate the pseudo-labels without introducing extra parameters. This strategy effectively enhances the discriminability of the target domain and further minimizes the distribution discrepancy between the two domains. Experiments on cross-domain aerial datasets have demonstrated the effectiveness and superiority of our proposed method when compared to other state-of-the-art methods.

show abstract

Unsupervised Domain Adaptation for Simultaneous Segmentation and Classification of the Retinal Arteries and Veins

Xue,

Zhang,

et al. 2024

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Automatic segmentation of the fundus retinal vessels and accurate classification of the arterial and venous vessels play an important role in clinical diagnosis. This article proposes a fundus retinal vascular segmentation and arteriovenous classification network that combines the adversarial training and attention mechanism to address the issues of fundus retinal arteriovenous classification error and ambiguous segmentation of fine blood vessels. It consists of three core components: discriminator, generator, and segmenter. In order to address the domain shift issue, U‐Net is employed as a discriminator, and data samples for arterial and venous vessels are generated with a generator using an unsupervised domain adaption (UDA) approach. The classification of retinal arterial and venous vessels (A/V) as well as the segmentation of fine vessels is improved by adding a self‐attention mechanism to improve attention to vessel edge features and the terminal fine vessels. Non‐strided convolution and non‐pooled downsampling methods are also used to avoid losing fine‐grained information and learning less effective feature representations. The performance of multi‐class blood vessel segmentation is as follows, per test results on the DRIVE dataset: F1‐score (F1) has a value of 0.7496 and an accuracy of 0.9820. The accuracy of A/V categorization has increased by 1.35% when compared to AU‐Net. The outcomes demonstrate that by enhancing the baseline U‐Net, the strategy we suggested enhances the automated classification and segmentation of blood vessels.

show abstract

Classes Matter: A Fine-Grained Adversarial Approach to Cross-Domain Semantic Segmentation

Cited by 210 publications

References 27 publications

Self-Guided Adversarial Learning For Domain Adaptive Semantic Segmentation

Self-Guided Adversarial Learning For Domain Adaptive Semantic Segmentation

Weakly-Supervised Domain Adaptation With Adversarial Entropy for Building Segmentation in Cross-Domain Aerial Imagery

Unsupervised Domain Adaptation for Simultaneous Segmentation and Classification of the Retinal Arteries and Veins

Contact Info

Product

Resources

About