Multi-Target Adversarial Frameworks for Domain Adaptation in Semantic Segmentation

Saporta, Antoine; Vu, Tuan-Hung; Cord, Matthieu; Pérez, Patrick

doi:10.1109/iccv48922.2021.00894

Cited by 37 publications

(46 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Oracle means the model is trained and tested on the specific target domain. Our method achieves Dice scores that are about 55% and 13% higher than those from [26], [24] and [27] when using OCTA (OCTA-500) and fundus image (DRIVE) as the source domain. For the dual teachers, T sim performs well on D sim and T dis achieves descent performance on D dis .…”

Section: Resultsmentioning

confidence: 92%

See 1 more Smart Citation

Student Becomes Decathlon Master in Retinal Vessel Segmentation via Dual-teacher Multi-target Domain Adaptation

Peng¹,

Li²,

Cheng³

et al. 2022

Preprint

View full text Add to dashboard Cite

Unsupervised domain adaptation has been proposed recently to tackle the so-called domain shift between training data and test data with different distributions. However, most of them only focus on singletarget domain adaptation and cannot be applied to the scenario with multiple target domains. In this paper, we propose RVms, a novel unsupervised multi-target domain adaptation approach to segment retinal vessels (RVs) from multimodal and multicenter retinal images. RVms mainly consists of a style augmentation and transfer (SAT) module and a dual-teacher knowledge distillation (DTKD) module. SAT augments and clusters images into source-similar domains and source-dissimilar domains via Bézier and Fourier transformations. DTKD utilizes the augmented and transformed data to train two teachers, one for source-similar domains and the other for source-dissimilar domains. Afterwards, knowledge distillation is performed to iteratively distill different domain knowledge from teachers to a generic student. The local relative intensity transformation is employed to characterize RVs in a domain invariant manner and promote the generalizability of teachers and student models. Moreover, we construct a new multimodal and multicenter vascular segmentation dataset from existing publicly-available datasets, which can be used to benchmark various domain adaptation and domain generalization methods. Through extensive experiments, RVms is found to be very close to the target-trained Oracle in terms of segmenting the RVs, largely outperforming other state-of-the-art methods.

show abstract

Section: Resultsmentioning

confidence: 92%

“…All methods are evaluated using two metrics, i.e., Dice[%] and 95% Hausdorff Distance (HD[px]), the results of which are tabulated in Table 1. We compare RVms with two recently-developed SOTA DA/MTDA models, namely ADVENT [26] and Multi-Dis [24]. Note that [26] is trained with mixed target domains.…”

Section: Resultsmentioning

confidence: 99%

Student Becomes Decathlon Master in Retinal Vessel Segmentation via Dual-teacher Multi-target Domain Adaptation

Peng¹,

Li²,

Cheng³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Within the DASiS context, different approaches have been proposed [83,166]. Isobe et al [83] propose a method that trains an expert model for every target domain where the models are encouraged to collaborate via style transfer.…”

Section: Multi-target Dasismentioning

confidence: 99%

“…Such expert models are further exploited as teachers for a student model that learns to imitate their output and serves as regularizer to bring the different experts closer to each other in the learned feature space. Instead, Saporta et al [166] propose to combine for each target domain T i two adversarial pipelines: one that learn to discriminate between the domain T i and the source, and one between T i and the union of other target domains. Then, to reduce the instability that the multi-discriminator model training might cause, they propose a multi-target knowledge transfer by adopting a multi-teacher/single-student distillation mechanism, which leads to a model that is agnostic to the target domains.…”

Section: Multi-target Dasismentioning

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

“…Recently, multi-target domain adaptation (MTDA) methods [11,18,40,43,48,58] have been proposed, which enables a single model to adapt a labeled source domain to multiple unlabeled target domains. Most of works train multiple STDA models and then distill the knowledge into a single multi-target domain adaptation network.…”

Section: Introductionmentioning

confidence: 99%

ADAS: A Direct Adaptation Strategy for Multi-Target Domain Adaptive Semantic Segmentation

Lee¹,

Choi²,

Changjae³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

Multi-Target Adversarial Frameworks for Domain Adaptation in Semantic Segmentation

Cited by 37 publications

References 17 publications

Student Becomes Decathlon Master in Retinal Vessel Segmentation via Dual-teacher Multi-target Domain Adaptation

Student Becomes Decathlon Master in Retinal Vessel Segmentation via Dual-teacher Multi-target Domain Adaptation

Unsupervised Domain Adaptation for Semantic Image Segmentation: a Comprehensive Survey

ADAS: A Direct Adaptation Strategy for Multi-Target Domain Adaptive Semantic Segmentation

Contact Info

Product

Resources

About