Model Adaptation: Unsupervised Domain Adaptation Without Source Data

Li, Rui; Jiao, Qianfen; Cao, Wenming; Wong, Hau-San; Wu, Si

doi:10.1109/cvpr42600.2020.00966

Cited by 392 publications

(245 citation statements)

References 28 publications

Supporting

Mentioning

230

Contrasting

Order By: Relevance

“…Based on hypothesis transfer learning, [21] proposes a self-training framework with mutual information maximization and pseudo-labeling strategy. In [20], a collaborative class conditional generative adversarial network is employed to avoid the usage of source data with target data generation and model adaptation. In this paper, instead of using entropy minimization, we leverage contrastive learning for cross-domain alignment.…”

Section: Related Workmentioning

confidence: 99%

Contrastive Adaptation Network for Unsupervised Domain Adaptation

Kang¹,

Jiang

Yang³

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

832

457

View full text Add to dashboard Cite

Unsupervised Domain Adaptation (UDA) makes predictions for the target domain data while manual annotations are only available in the source domain. Previous methods minimize the domain discrepancy neglecting the class information, which may lead to misalignment and poor generalization performance. To address this issue, this paper proposes Contrastive Adaptation Network (CAN) optimizing a new metric which explicitly models the intra-class domain discrepancy and the inter-class domain discrepancy. We design an alternating update strategy for training CAN in an end-to-end manner. Experiments on two real-world benchmarks Office-31 and VisDA-2017 demonstrate that CAN performs favorably against the state-of-the-art methods and produces more discriminative features.

show abstract

Section: Related Workmentioning

confidence: 99%

Contrastive Adaptation Network for Unsupervised Domain Adaptation

Kang¹,

Jiang

Yang³

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

832

457

View full text Add to dashboard Cite

show abstract

“…Model Adaptation Most of the previously mentioned methods need the explicit availability of source domain data during adaptation too, and have made tremendous strides in improving the segmentation performance in that case. A few recent papers tackle model adaptation for classification problems [44,42,16]. [39] proposes source-free domain adaptation in the case where label knowledge of the target domain is not available, and show their efficiency on a set of classification problems with varying levels of label overlap.…”

Section: Related Workmentioning

confidence: 99%

Uncertainty Reduction for Model Adaptation in Semantic Segmentation

Fleuret

2021

2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

“…Different from the above methods, 3C‐GAN 28 first generates images of each category as the style of target domain to adjust the source network and then also minimizes entropy. DAAS 29 trains a transformation network from the target domain to the source domain to make the entropy of the transformed target domain input to the classifier as small as possible than before.…”

Section: Related Workmentioning

confidence: 99%

“…Only few recent works studied source data-free domain adaptation in the following three categories: The first one adjusts the source model by self-supervised learning based on pseudo labels [25][26][27] ; the second one is to generate some target domain style samples which can be correctly classified by the source model 28 ; and the final is to transform the target domain samples to the source domain style samples. 29 First of all, without source domain data, it is not easy to do satisfied image translation between the target and source domains.…”

Section: Introductionmentioning

confidence: 99%

Source data‐free domain adaptation of object detector through domain‐specific perturbation

Xiong

Zhang

et al. 2021

Int J Intell Syst

View full text Add to dashboard Cite

The current unsupervised cross‐domain detection methods need source domain data to retrain the detection model in target domain. However, the source domain data may be unavailable due to privacy, decentralization, or computation resource restrictions. A natural idea is to optimize the parameters of the source domain model by self‐supervised learning based on pseudo labels. We propose another approach from the viewpoint of noise perturbation without pseudo‐labeling. It can be assumed that the source and target domains are actually derived from a domain invariant space through domain‐specific perturbations, respectively. A super target domain can be constructed by augmenting more target domain perturbations to the target domain images. The optimal direction of the target domain to the domain invariant space can be approximated as the alignment direction from the super target domain to the target domain. Based on this idea, we propose a novel method called SOAP (SOurce data‐free domain Adaptation through domain Perturbation) which can remove domain perturbation from the target domain. The image‐level, instance‐level, and category consistency regularizations based on Mean Teacher structure are proposed to learn the correct alignment direction. Specifically, the category consistency can also further improve the classification accuracy. Extensive experiments on multiple domain adaptation scenarios demonstrate that SOAP achieves better performance surpassing the baseline (Faster R‐CNN) and multiple state‐of‐the‐art domain adaptation methods which need to access source domain data.

show abstract

Model Adaptation: Unsupervised Domain Adaptation Without Source Data

Cited by 392 publications

References 28 publications

Contrastive Adaptation Network for Unsupervised Domain Adaptation

Contrastive Adaptation Network for Unsupervised Domain Adaptation

Uncertainty Reduction for Model Adaptation in Semantic Segmentation

Source data‐free domain adaptation of object detector through domain‐specific perturbation

Contact Info

Product

Resources

About