SALAD : Source-free Active Label-Agnostic Domain Adaptation for Classification, Segmentation and Detection

Kothandaraman, Divya; Shekhar, Sumit; Sancheti, Abhilasha; Manoj, Ghuhan,; Shukla, Tripti; Manocha, Dinesh

doi:10.1109/wacv56688.2023.00046

Cited by 8 publications

(7 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Self-attention is promising, especially transformerbased SFDA methods. Although the self-attention-based SFDA methods are less used in image classification [70], [124] and mostly seen in semantic segmentation [11], [112], [129], the transformer-based method TransDA [124] achieves the highest accuracy of 79.3% on the Office-Home dataset, which is 4.8% higher than the second highest method [139]. This somewhat suggests that encouraging models to turn their attention to the object region may be quite effective for reducing domain shift.…”

Section: Domain-based Reconstruction Is Commonly Used In Sfdamentioning

confidence: 99%

“…Based on this, CADX [128] divides the target domain image into supported images and query images under the sourcefree setting, and improves the original patch-to-patch operation to image-to-image in order to capture the overall representations and reduce the computational burden. In addition, some approaches [11], [129] further process the features from both spatial attention and channel attention perspectives to enrich the contextual semantics of the representations. Currently, attention-based SFDA methods are still relatively rare, especially transformer-based SFDA methods.…”

Section: Self-attentionmentioning

confidence: 99%

“…Image classification can also extend to various application scenarios such as real-world image dehazing [143], cross-scene hyperspectral image classification [144], and blind image quality assessment [114]. Semantic segmentation methods [11], [12], [50], [100], [129], [145], [146] have also emerged rapidly and have been widely used, including multi-organ segmentation [113], cross-modal segmentation [147], [148], cross-device [100], cross-central domain segmentation [85], multi-site and lifespan brain skull stripping [77], and road segmentation [98], [149]. Others applications include object detection [13], [14], [62], person re-identification [150], and video analysis [88], [144] can also benefit from SFDA.…”

Section: Computer Visionmentioning

confidence: 99%

See 2 more Smart Citations

Source-Free Domain Adaptation for Real-World Image Dehazing

Huang

Liu

et al. 2022

Proceedings of the 30th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Over the past decade, domain adaptation has become a widely studied branch of transfer learning that aims to improve performance on target domains by leveraging knowledge from the source domain. Conventional domain adaptation methods often assume access to both source and target domain data simultaneously, which may not be feasible in real-world scenarios due to privacy and confidentiality concerns. As a result, the research of Source-Free Domain Adaptation (SFDA) has drawn growing attention in recent years, which only utilizes the source-trained model and unlabeled target data to adapt to the target domain. Despite the rapid explosion of SFDA work, yet there has no timely and comprehensive survey in the field. To fill this gap, we provide a comprehensive survey of recent advances in SFDA and organize them into a unified categorization scheme based on the framework of transfer learning. Instead of presenting each approach independently, we modularize several components of each method to more clearly illustrate their relationships and mechanics in light of the composite properties of each method. Furthermore, we compare the results of more than 30 representative SFDA methods on three popular classification benchmarks, namely Office-31, Office-home, and VisDA, to explore the effectiveness of various technical routes and the combination effects among them. Additionally, we briefly introduce the applications of SFDA and related fields. Drawing from our analysis of the challenges facing SFDA, we offer some insights into future research directions and potential settings.

show abstract

Section: Domain-based Reconstruction Is Commonly Used In Sfdamentioning

confidence: 99%

Section: Self-attentionmentioning

confidence: 99%

Section: Computer Visionmentioning

confidence: 99%

See 1 more Smart Citation

Source-Free Domain Adaptation for Real-World Image Dehazing

Huang

Liu

et al. 2022

Proceedings of the 30th ACM International Conference on Multimedia

View full text Add to dashboard Cite

show abstract

“…Bohdal et al [28] partitioned target domain images into support images and query images under the unsupervised setting and improved the original patch-to-patch operation to image-to-image to capture holistic representations and reduce computational burden. Kothandaraman et al [29] further processed features from both spatial and channel perspectives, conducted feature distillation from pre-trained networks to target networks, and supplemented target samples mined based on transferability and uncertainty criteria to enrich contextual semantics. However, for some complex domains and tasks, self-attention mechanisms might not capture all critical features as they primarily focus on local relationships within input sequences while neglecting global relationships.…”

Section: Related Workmentioning

confidence: 99%

Unsupervised domain adaptation of dynamic extension networks based on class decision boundaries

Chen

Wang

Zhu

et al. 2023

Preprint

View full text Add to dashboard Cite

In response to the problems of inaccurate feature alignment, loss of source domain information, imbalanced sample distribution, and biased class decision boundaries in traditional unsupervised domain adaptation methods, this paper proposes a class decision boundary-based dynamic expansion network unsupervised domain adaptation method called CDE-Net. Specifically, our method dynamically expands the autoencoder-based network structure, which can preserve source domain feature information while gradually adapting to the target domain data distribution and learning useful feature information from the target domain. Meanwhile, by minimizing clustering loss and conditional entropy loss, CDE-Net can explore the intrinsic structure of the data and push class decision boundaries away from dense data areas. We experimentally verify our method on three medical image datasets, chest X-rays, intracranial hemorrhage, and mammography, and achieve an average AUC improvement of 25.8% or more compared to non-transfer methods. In addition, we compare our method with previous unsupervised domain adaptation methods, and the experimental results show that our method achieves better classification accuracy and generalization performance.

show abstract

“…In real-world scenarios, the source data is usually inaccessible due to data privacy, leading to the SFOD problem (Lee et al 2022;Zong et al 2022;Ding et al 2022;Kothandaraman et al 2022;Wang et al 2022b). Due to complex background and negative examples, SFOD is far more challenging than conventional source-free image classification (Agarwal et al 2022;Ambekar et al 2022;Bohdal et al 2022;Xia, Zhao, and Ding 2021).…”

Section: Related Workmentioning

confidence: 99%

Adversarial Alignment for Source Free Object Detection

Chu,

Li,

Chen

et al. 2023

AAAI

View full text Add to dashboard Cite

Source-free object detection (SFOD) aims to transfer a detector pre-trained on a label-rich source domain to an unlabeled target domain without seeing source data. While most existing SFOD methods generate pseudo labels via a source-pretrained model to guide training, these pseudo labels usually contain high noises due to heavy domain discrepancy. In order to obtain better pseudo supervisions, we divide the target domain into source-similar and source-dissimilar parts and align them in the feature space by adversarial learning.Specifically, we design a detection variance-based criterion to divide the target domain. This criterion is motivated by a finding that larger detection variances denote higher recall and larger similarity to the source domain. Then we incorporate an adversarial module into a mean teacher framework to drive the feature spaces of these two subsets indistinguishable. Extensive experiments on multiple cross-domain object detection datasets demonstrate that our proposed method consistently outperforms the compared SFOD methods. Our implementation is available at https://github.com/ChuQiaosong.

show abstract

SALAD : Source-free Active Label-Agnostic Domain Adaptation for Classification, Segmentation and Detection

Cited by 8 publications

References 38 publications

Source-Free Domain Adaptation for Real-World Image Dehazing

Source-Free Domain Adaptation for Real-World Image Dehazing

Unsupervised domain adaptation of dynamic extension networks based on class decision boundaries

Adversarial Alignment for Source Free Object Detection

Contact Info

Product

Resources

About