computer-vision-2

Multi-target Domain Adaptation

3260 papers • 126 benchmarks • 313 datasets

The idea of Multi-target Domain Adaptation is to adapt a model from a single labelled source domain to multiple unlabelled target domains.

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in multi-target-domain-adaptation-2

Trend

Dataset

Best Model

Actions

Office-31

Office-Home

DomainNet

Libraries

i

Use these libraries to find multi-target-domain-adaptation-2 models and implementations

thuml/Transfer-Learning-Library

2 papers 3,118

Datasets

Subtasks

No subtasks available.

Most implemented papers

Unsupervised Domain Adaptation by Backpropagation

V. Lempitsky, Yaroslav Ganin•Thu Sep 25 2014

The method performs very well in a series of image classification experiments, achieving adaptation effect in the presence of big domain shifts and outperforming previous state-of-the-art on Office datasets.

6524

Content

OBJ-MDA

kevinmusgrave/pytorch-adapt

2 papers 319

0

Paper Graph

Minimum Class Confusion for Versatile Domain Adaptation

Mingsheng Long, Jianmin Wang, Ying Jin, Ximei Wang•Sat Dec 07 2019

It is uncovered that reducing such pairwise class confusion leads to significant transfer gains, and a general loss function is proposed: Minimum Class Confusion (MCC), which can be characterized as a non-adversarial DA method without explicitly deploying domain alignment, enjoying faster convergence speed.

447 0

Paper Graph

Domain Agnostic Learning with Disentangled Representations

Kate Saenko, Xingchao Peng, Zijun Huang, Ximeng Sun•Sat Apr 27 2019

This paper devise a novel Deep Adversarial Disentangled Autoencoder (DADA) capable of disentangling domain-specific features from class identity and demonstrates experimentally that when the target domain labels are unknown, DADA leads to state-of-the-art performance on several image classification datasets.

296 0

Paper Graph

Unsupervised Multi-Target Domain Adaptation Through Knowledge Distillation

Eric Granger, J. Dolz, Le Thanh Nguyen-Meidine, M. Kiran, Louis-Antoine Blais-Morin, Atif Bela•Mon Jul 13 2020

This paper proposes a novel unsupervised MTDA approach to train a CNN that can generalize well across multiple target domains and relies on multi-teacher knowledge distillation (KD) to iteratively distill target domain knowledge from multiple teachers to a common student.

96 0

Paper Graph

Blending-Target Domain Adaptation by Adversarial Meta-Adaptation Networks

Liang Lin, Xiaodan Liang, Ziliang Chen, Jingyu Zhuang•Fri May 31 2019

Empirical results show that BTDA is a quite challenging transfer setup for most existing DA algorithms, yet AMEAN significantly outperforms these state-of-the-art baselines and effectively restrains the negative transfer effects in BTDA.

93 0

Paper Graph

Curriculum Graph Co-Teaching for Multi-Target Domain Adaptation

N. Sebe, E. Ricci, Zhun Zhong, Subhankar Roy, E. Krivosheev•Wed Mar 31 2021

This paper develops a co-teaching strategy with the dual classifier head that is assisted by curriculum learning to obtain more reliable pseudo-labels and proposes Domain-aware Curriculum Learning (DCL), a sequential adaptation strategy that first adapts on the easier target domains, followed by the harder ones.

73 0

Paper Graph

Incremental Multi-Target Domain Adaptation for Object Detection with Efficient Domain Transfer

Eric Granger, Marco Pedersoli, J. Dolz, Le Thanh Nguyen-Meidine, M. Kiran, Louis-Antoine Blais-Morin•Mon Apr 12 2021

This work proposes a new Incremental MTDA technique for object detection that can adapt a detector to multiple target domains, one at a time, without having to retain data of previously-learned target domains.

32 0

Paper Graph

Foreground-Aware Stylization and Consensus Pseudo-Labeling for Domain Adaptation of First-Person Hand Segmentation

Y. Ushiku, Takuma Yagi, Yoichi Sato, Takehiko Ohkawa, Atsushi Hashimoto•Mon Jul 05 2021

This work proposes foreground-aware image stylization and consensus pseudo-labeling for domain adaptation of hand segmentation and demonstrates promising results in challenging multi-target domain adaptation and domain generalization settings.

9 0

Paper Graph

See Eye to Eye: A Lidar-Agnostic 3D Detection Framework for Unsupervised Multi-Target Domain Adaptation

Mao Shan, Stewart Worrall, E. Nebot, Darren Tsai, J. S. Berrio•Tue Nov 16 2021

This work proposes a novel unsupervised multi-target domain adaptation framework, SEE, for transferring the performance of state-of-the-art 3D detectors across both fixed and flexible scan pattern lidars without requiring fine-tuning of models by end-users.

18 0

Paper Graph

Adding a benchmark result helps the community track progress.