Adaptive Adversarial Network for Source-free Domain Adaptation

Xia, Haifeng; Zhao, Handong; Ding, Zhengming

doi:10.1109/iccv48922.2021.00888

Cited by 153 publications

(62 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In most cases, one can only access the unlabeled target data and the model pretrained by source data. To this end, some recent works [20,23,25,26,48,52,56,58] regarding sourcefree domain adaptation emerge. These methods provide solutions to adapt the model to unseen domains without using original training data.…”

Section: Related Workmentioning

confidence: 99%

“…G-SFDA [58] forces the network to activate different channels for different domains while paying attention to the neighborhood structure of data. A 2 Net [52] introduces a new target classifier to align two domains via adversarial training manner. SoFA [59] uses a Variational Auto-Encoder to encode target distribution in latent space while reconstructing the target data in image space to constrain the latent features.…”

Section: Related Workmentioning

confidence: 99%

“…G-SFDA [58] learns different feature activations by exploiting neighborhood structure of target data. A 2 Net [52] introduces a new classifier and adopt adversarial training to align two domains. Despite the fact that these SFDA methods utilize the source domain knowledge contained by the pretrained model, none of them explicitly align the distributions between source domain and target domain to achieve adaptation.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Source-Free Domain Adaptation via Distribution Estimation

Ding¹,

Xu²,

Chen³

et al. 2022

Preprint

View full text Add to dashboard Cite

Domain Adaptation aims to transfer the knowledge learned from a labeled source domain to an unlabeled target domain whose data distributions are different. However, the training data in source domain required by most of the existing methods is usually unavailable in real-world applications due to privacy preserving policies. Recently, Source-Free Domain Adaptation (SFDA) has drawn much attention, which tries to tackle domain adaptation problem without using source data. In this work, we propose a novel framework called SFDA-DE to address SFDA task via source Distribution Estimation. Firstly, we produce robust pseudo-labels for target data with spherical k-means clustering, whose initial class centers are the weight vectors (anchors) learned by the classifier of pretrained model. Furthermore, we propose to estimate the class-conditioned feature distribution of source domain by exploiting target data and corresponding anchors. Finally, we sample surrogate features from the estimated distribution, which are then utilized to align two domains by minimizing a contrastive adaptation loss function. Extensive experiments show that the proposed method achieves state-of-the-art performance on multiple DA benchmarks, and even outperforms traditional DA methods which require plenty of source data.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Source-Free Domain Adaptation via Distribution Estimation

Ding¹,

Xu²,

Chen³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Kurmi et al [90] treat the pre-trained source model as an energy-based function, in order to learn the joint distribution, and train a GAN that generates annotated samples that are used throughout the adaptation procedure, without the need of accessing source samples. Xia et al [214] propose first to rely on a learnable target classifier that improves the recognition ability on source-dissimilar target features, and then to perform adversarial domain-level alignment and contrastive matching -at category level.…”

Section: Source-free Domain Adaptationmentioning

confidence: 99%

Unsupervised Domain Adaptation for Semantic Image Segmentation: a Comprehensive Survey

Volpi¹,

Chidlovskii²

2021

Preprint

View full text Add to dashboard Cite

Semantic segmentation plays a fundamental role in a broad variety of computer vision applications, providing key information for the global understanding of an image. Yet, the state-of-the-art models rely on large amount of annotated samples, which are more expensive to obtain than in tasks such as image classification. Since unlabelled data is instead significantly cheaper to obtain, it is not surprising that Unsupervised Domain Adaptation reached a broad success within the semantic segmentation community.This survey is an effort to summarize five years of this incredibly rapidly growing field, which embraces the importance of semantic segmentation itself and a critical need of adapting segmentation models to new environments. We present the most important semantic segmentation methods; we provide a comprehensive survey on domain adaptation techniques for semantic segmentation; we unveil newer trends such as multi-domain learning, domain generalization, test-time adaptation or source-free domain adaptation; we conclude this survey by describing datasets and benchmarks most widely used in semantic segmentation research. We hope that this survey will provide researchers across academia and industry with a comprehensive reference guide and will help them in fostering new research directions in the field.1 See as example the https://github.com/errolPereira/ Semi-Automatic-Image-Annotation-Tool 2 We refer to the dedicated page https://github.com/ heartexlabs/awesome-data-labeling for an up-to-date collection of annotation tools.

show abstract

“…Unlike the above methods, there have been a few works for classification tasks which do not use source data, but only the source model for adaptation. The methods involve -entropy minimization with divergence maximization [27], pseudo-labeling with selfreconstruction [60], generating additional target images [23] and self-supervision [58]. For source-free adaptation for semantic segmentation, recently, [28] proposed an algorithm that combines ideas from the above methods like image generation, pseudo-labeling and output space adaptation by dividing the target into easy and hard, and learning an additional discriminator to align them.…”

Section: Related Workmentioning

confidence: 99%

Unsupervised Adaptation of Semantic Segmentation Models without Source Data

Shrivastava¹,

Khurana²,

Aggarwal³

2021

Preprint

View full text Add to dashboard Cite

We consider the novel problem of unsupervised domain adaptation of source models, without access to the source data for semantic segmentation. Unsupervised domain adaptation aims to adapt a model learned on the labeled source data, to a new unlabeled target dataset. Existing methods assume that the source data is available along with the target data during adaptation. However, in practical scenarios, we may only have access to the source model and the unlabeled target data, but not the labeled source, due to reasons such as privacy, storage, etc. In this work, we propose a self-training approach to extract the knowledge from the source model. To compensate for the distribution shift from source to target, we first update only the normalization parameters of the network with the unlabeled target data. Then we employ confidence-filtered pseudo labeling and enforce consistencies against certain transformations. Despite being very simple and intuitive, our framework is able to achieve significant performance gains compared to directly applying the source model on the target data as reflected in our extensive experiments and ablation studies. In fact, the performance is just a few points away from the recent state-of-the-art methods which use source data for adaptation. We further demonstrate the generalisability of the proposed approach for fully test-time adaptation setting, where we do not need any target training data and adapt only during test-time.

show abstract

Adaptive Adversarial Network for Source-free Domain Adaptation

Cited by 153 publications

References 33 publications

Source-Free Domain Adaptation via Distribution Estimation

Source-Free Domain Adaptation via Distribution Estimation

Unsupervised Domain Adaptation for Semantic Image Segmentation: a Comprehensive Survey

Unsupervised Adaptation of Semantic Segmentation Models without Source Data

Contact Info

Product

Resources

About