DomainVerse: A Benchmark Towards Real-World Distribution Shifts for Training-Free Adaptive Domain Generalization

doi:10.1109/TMM.2025.3586108

	DomainVerse: A Benchmark Towards Real-World Distribution Shifts for Training-Free Adaptive Domain Generalization
	Hou, Feng 1,2; Yuan, Jin 3,4; Yang, Ying 5; Zhang, Yao 5; Liu, Yang 1,2; Zhang, Yang 5; Zhong, Cheng 5; Shi, Zhongchao 5; Fan, Jianping 5; He, Zhiqiang 1,2,6; Rui, Yong 5
	2025
发表期刊	IEEE TRANSACTIONS ON MULTIMEDIA
ISSN	1520-9210
卷号	27 页码:6648-6660
摘要	Traditional cross-domain tasks, including unsupervised domain adaptation (UDA), domain generalization (DG) and test-time adaptation (TTA), rely heavily on the training model by source domain data whether for specific or arbitrary target domains. With the recent advance of vision-language models (VLMs), recognized as natural source models that can be transferred to various downstream tasks without any parameter training, we propose a novel cross-domain task directly combining the strengths of both UDA and DG, named Training-Free Adaptive Domain Generalization (TF-ADG). However, current cross-domain datasets have many limitations, such as unrealistic domains, unclear domain definitions, and the inability to fine-grained domain decomposition, which hinder the real-world application of current cross-domain models due to the lack of accurate and fair evaluation of fine-grained realistic domains. These insights motivate us to establish a novel realistic benchmark for TF-ADG. Benefiting from the introduced hierarchical definition of domain shifts, our proposed dataset DomainVerse addresses these issues by providing about 0.5 million images from 390 realistic, hierarchical, and balanced domains, allowing for decomposition across multiple domains within each image. With the help of the constructed DomainVerse and VLMs, we further propose two algorithms called Domain CLIP and Domain++ CLIP for training-free adaptive domain generalization. Extensive and comprehensive experiments demonstrate the significance of the dataset and the effectiveness of the proposed methods.
关键词	Adaptation models Training Benchmark testing Picture archiving and communication systems Data models Image color analysis Computational modeling Data mining Training data Painting DomainVerse training-free adaptive domain generalization vision-language models
DOI	10.1109/TMM.2025.3586108
收录类别	SCI
语种	英语
资助项目	AI Lab ; Lenovo Research
WOS研究方向	Computer Science ; Telecommunications
WOS类目	Computer Science, Information Systems ; Computer Science, Software Engineering ; Telecommunications
WOS记录号	WOS:001579069300018
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/41673
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Zhang, Yao; He, Zhiqiang
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing 100000, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100000, Peoples R China 3.Southeast Univ, Sch Comp Sci & Engn, Minist Educ, Nanjing 211189, Peoples R China 4.Southeast Univ, Key Lab Comp Network & Informat Integrat, Minist Educ, Nanjing 211189, Peoples R China 5.Lenovo Res, AI Lab, Beijing 100000, Peoples R China 6.Lenovo Ltd, Beijing 100000, Peoples R China
推荐引用方式 GB/T 7714	Hou, Feng,Yuan, Jin,Yang, Ying,et al. DomainVerse: A Benchmark Towards Real-World Distribution Shifts for Training-Free Adaptive Domain Generalization[J]. IEEE TRANSACTIONS ON MULTIMEDIA,2025,27:6648-6660.
APA	Hou, Feng.,Yuan, Jin.,Yang, Ying.,Zhang, Yao.,Liu, Yang.,...&Rui, Yong.(2025).DomainVerse: A Benchmark Towards Real-World Distribution Shifts for Training-Free Adaptive Domain Generalization.IEEE TRANSACTIONS ON MULTIMEDIA,27,6648-6660.
MLA	Hou, Feng,et al."DomainVerse: A Benchmark Towards Real-World Distribution Shifts for Training-Free Adaptive Domain Generalization".IEEE TRANSACTIONS ON MULTIMEDIA 27(2025):6648-6660.