Family Matters: A Systematic Study of Spatial vs. Frequency Masking for Continual Test-Time Adaptation
Faculty of Science and Technology (REALTEK), Norwegian University of Life Sciences (NMBU)（科学与技术学院（REALTEK）、挪威生命科学大学）；Tulane University（路易斯安那州立大学）；Gwangju Institute of Science and Technology（全州科学技术学院）；Hanyang University（翰阳大学）
Chandler Timm C. Doloriel, Yunbei Zhang, Yeonguk Yu, Taki Hasan Rafi, Muhammad salman siddiqui, Tor Kristian Stevik, Fadi Al Machot, Kristian Hovde Liland, Habib Ullah
最近的连续测试时自适应（CTTA）方法采用掩码图像建模来稳定分布偏移下的学习，但每种方法都将其掩码家族F视为固定的设计选择，并仅沿着选择策略S进行创新，从而使得家族轴未被充分探索。我们提出了一项系统的实证研究，隔离了这一轴。通过使用一个受控的CTTA实例——Mask to Adapt (M2A)——它固定S为随机和标准损失，我们仅改变F，跨越空间（补丁、像素）和频率（全频带、低频带、高频带）家族，同时保持其他所有组件相同。该研究的贡献在于为我们评估的CTTA设置提取了设计指导：（1）掩码家族决定了自适应是累积有用的结构还是累积错误——在补丁标记化架构上，空间掩码在长流中积累稳定的表示，而频率掩码则灾难性地崩溃。我们通过结构保持解释来表征这种不稳定性，其中空间相干性维持了避免与腐败的频谱特征最终重叠所需的宽谱冗余；（2）最优家族取决于架构-任务对齐——在CNN上，其重叠的感受野稀释了补丁遮挡，家族差距消失，而在具有全局线索和大容量ViT的细粒度任务上，频率掩码变得有竞争力。在混杂的系统级比较中——其中基线在损失和辅助组件上也不同——M2A的随机选择与启发式策略表现相当，尽管我们将这一观察视为提示性背景，而非对S相对重要性的受控量化。
英文摘要：
Recent continual test-time adaptation (CTTA) methods adopt masked image modeling to stabilize learning under distribution shift, yet each treats its masking family F as a fixed design choice and innovates exclusively along the selection strategy S, leaving the family axis underexplored. We present a systematic empirical study that isolates this axis. Using a controlled CTTA instantiation, Mask to Adapt (M2A), that fixes S=random and standard losses, we vary only F across spatial (patch, pixel) and frequency (all-band, low-band, high-band) families while keeping every other component identical. The study's contributions are the design guidance it extracts for the CTTA settings we evaluated: (1) the masking family determines whether adaptation compounds useful structure or compounds errors, on patch-tokenized architectures, spatial masking accumulates stable representations over long streams while frequency masking collapses catastrophically. We characterize this instability through a structural-preservation account, where spatial coherence maintains the broad-spectrum redundancy needed to avoid terminally overlapping with a corruption's spectral signature; (2) the optimal family depends on architecture-task alignment, on CNNs, whose overlapping receptive fields dilute patch occlusion, the family gap vanishes, whereas on fine-grained tasks with global cues and large-capacity ViTs, frequency masking becomes competitive. In confounded system-level comparisons, where baselines also differ in losses and auxiliary components, M2A's random selection performs comparably to heuristic strategies, though we treat this observation as suggestive context rather than a controlled quantification of S's relative importance.
University of Illinois at Chicago（伊利诺伊大学芝加哥分校）；ByteDance（字节跳动）
Existing continual learning techniques focus on either task incremental learning (TIL) or class incremental learning (CIL) problem, but not both. CIL and TIL differ mainly in that the task-id is provided for each test sample during testing for TIL, but not provided for CIL. Continual learning methods intended for one problem have limitations on the other problem. This paper proposes a novel unified approach based on out-of-distribution (OOD) detection and task masking, called CLOM, to solve both problems. The key novelty is that each task is trained as an OOD detection model rather than a trad…
Domain Adaptation (DA) techniques are important for overcoming the domain shift between the source domain used for training and the target domain where testing takes place. However, current DA methods assume that the entire target domain is available during adaptation, which may not hold in practice. This paper considers a more realistic scenario, where target data become available in smaller batches and adaptation on the entire target domain is not feasible. In our work, we introduce a new, data-constrained DA paradigm where unlabeled target samples are received in batches and adaptation is p…
Continual Model Merging with Test-Time Adaptation for Whole-Slide Image Analysis
University of Information Technology（信息技术大学）；Viet Nam National University Ho Chi Minh City（胡志明市越南国立大学）；ETIS (UMR 8051), CY Cergy Paris University, ENSEA, CNRS, France（法国赛尔齐巴黎大学、ENSEA、法国国家科学研究中心的ETIS（UMR 8051））
Continual-MAE: Adaptive Distribution Masked Autoencoders for Continual Test-Time Adaptation
National Key Laboratory for Multimedia Information Processing（多媒体信息处理国家重点实验室）；School of Computer Science（计算机学院）；Peking University（北京大学）；CUHK（香港中文大学）；University of Science Technology of China（中国科学技术大学）
OMUDA: Omni-level Masking for Unsupervised Domain Adaptation in Semantic Segmentation
School of Mechanical Engineering, Chengdu University（成都大学机械工程学院）；School of Information Science and Technology, Harbin Institute of Technology（哈尔滨工业大学信息科学与技术学院）；School of Computer Science and Informatics, Cardiff University（卡的夫大学计算机科学与信息学院）；School of Artificial Intelligence and Computer Science, Jiangnan University（江南大学人工智能与计算机科学学院）；College of Intelligence Science and Technology, National University of Defense Technology（国防科技大学智能科学与技术学院）
ETH Zurich（苏黎世联邦理工学院）；MPI for Informatics（马克斯·普朗克信息学研究所）；EPFL（洛桑联邦理工学院）
Test-time domain adaptation aims to adapt a source pre-trained model to a target domain without using any source data. Existing works mainly consider the case where the target domain is static. However, real-world machine perception systems are running in non-stationary and continually changing environments where the target domain distribution can change over time. Existing methods, which are mostly based on self-training and entropy regularization, can suffer from these non-stationary environments. Due to the distribution shift over time in the target domain, pseudo-labels become unreliable. …
Seoul National University（首尔大学）；KAIST（韩国科学技术院）；Kyung Hee University（庆熙大学）
Unsupervised Domain Adaptation (UDA) for semantic segmentation has been favorably applied to real-world scenarios in which pixel-level labels are hard to be obtained. In most of the existing UDA methods, all target data are assumed to be introduced simultaneously. Yet, the data are usually presented sequentially in the real world. Moreover, Continual UDA, which deals with more practical scenarios with multiple target domains in the continual learning setting, has not been actively explored. In this light, we propose Continual UDA for semantic segmentation based on a newly designed Expanding Ta…
Peking University（北京大学）；South China University of Technology（华南理工大学）；National University of Singapore（新加坡国立大学）；MIT-IBM Watson AI Lab（麻省理工学院-IBM沃森人工智能实验室）
This paper studies a new, practical but challenging problem, called Class-Incremental Unsupervised Domain Adaptation (CI-UDA), where the labeled source domain contains all classes, but the classes in the unlabeled target domain increase sequentially. This problem is challenging due to two difficulties. First, source and target label sets are inconsistent at each time step, which makes it difficult to conduct accurate domain alignment. Second, previous target classes are unavailable in the current step, resulting in the forgetting of previous knowledge. To address this problem, we propose a nov…
Peking University（北京大学）；Huawei Technologies（华为技术有限公司）；Wuhan University（武汉大学）；The Chinese University of Hong Kong（香港中文大学）；Aerospace Information Research Institute, Chinese Academy of Sciences（中国科学院空天信息创新研究院）
Continual Test-Time Adaptation (CTTA) aims to adapt the source model to continually changing unlabeled target domains without access to the source data. Existing methods mainly focus on model-based adaptation in a self-training manner, such as predicting pseudo labels for new domain datasets. Since pseudo labels are noisy and unreliable, these methods suffer from catastrophic forgetting and error accumulation when dealing with dynamic data distributions. Motivated by the prompt learning in NLP, in this paper, we propose to learn an image-level visual domain prompt for target domains while havi…
Continual Test-Time Adaptation in Computer Vision: Methods, Benchmarks, and Future Directions
The University of Texas at Dallas（德克萨斯大学达拉斯分校）；LIVIA ETS Montreal, ILLS International Laboratory on Learning Systems (ILLS)（蒙特利尔LIVIA ETS，学习系统国际实验室（ILLS））；Tulane University（路易斯安那州立大学）；Nanyang Technological University（南洋理工大学）；Hanyang University（翰阳大学）
ViDA: Homeostatic Visual Domain Adapter for Continual Test Time Adaptation
National Key Laboratory for Multimedia Information Processing（多媒体信息处理国家重点实验室）；The Chinese University of Hong Kong（香港中文大学）；Hong Kong University of Science and Technology（香港科技大学）；School of Computer Science Peking University（北京大学计算机学院）；AIRobotics
Reshaping the Online Data Buffering and Organizing Mechanism for Continual Test-Time Adaptation
Harbin Institute of Technology（哈尔滨工业大学）；Pengcheng Laboratory（鹏城实验室）；Shenzhen University of Advanced Technology（深圳先进技术学院）；Shenzhen Institute of Advanced Technology, Chinese Academy of Science（中国科学院深圳先进技术研究院）
Continual Learning for Segment Anything Model Adaptation
Southern University of Science and Technology（南方科技大学）；City University of Hong Kong（香港城市大学）；Hong Kong University of Science and Technology（香港科技大学）；Tencent YouTu Lab（腾讯优图实验室）
针对动态流式数据场景，提出持续 SAM 适应基准 CoSAM 和域适配器混合算法 MoDA，通过全局特征与辅助令牌提取分离特征，实现持续分割并保持知识，优于现有方法。
Since experiencing domain shifts during test-time is inevitable in practice, test-time adaption (TTA) continues to adapt the model after deployment. Recently, the area of continual and gradual test-time adaptation (TTA) emerged. In contrast to standard TTA, continual TTA considers not only a single domain shift, but a sequence of shifts. Gradual TTA further exploits the property that some shifts evolve gradually over time. Since in both settings long test sequences are present, error accumulation needs to be addressed for methods relying on self-training. In this work, we propose and show that…
Adaptive and Balanced Re-initialization for Long-timescale Continual Test-time Domain Adaptation
Hong Kong Polytechnic University（香港理工大学）；Eastern Institute for Advanced Study（东方高级研究所）；Ant Group（蚂蚁集团）；Australian National University（澳大利亚国立大学）；CSIRO（澳大利亚科学院）；The University of Hong Kong（香港大学）
Technische Universität Braunschweig（布伦瑞克工业大学）；Institute of Communications Technology（通信技术研究所）
Environment perception in autonomous driving vehicles often heavily relies on deep neural networks (DNNs), which are subject to domain shifts, leading to a significantly decreased performance during DNN deployment. Usually, this problem is addressed by unsupervised domain adaptation (UDA) approaches trained either simultaneously on source and target domain datasets or even source-free only on target data in an offline fashion. In this work, we further expand a source-free UDA approach to a continual and therefore online-capable UDA on a single-image basis for semantic segmentation. Accordingly…
University of Sydney（悉尼大学）；The Chinese University of Hong Kong（香港中文大学）；Sensetime Group Limited（商汤集团有限公司）
Human beings can quickly adapt to environmental changes by leveraging learning experience. However, adapting deep neural networks to dynamic environments by machine learning algorithms remains a challenge. To better understand this issue, we study the problem of continual domain adaptation, where the model is presented with a labelled source domain and a sequence of unlabelled target domains. The obstacles in this problem are both domain shift and catastrophic forgetting. We propose Gradient Regularized Contrastive Learning (GRCL) to solve the obstacles. At the core of our method, gradient reg…
Dance Across Shifts: Forward-Facilitation Continual Test-Time Adaptation through Dynamic Style Bridging
Harbin Institute of Technology（哈尔滨工业大学）；Pengcheng Laboratory（鹏城实验室）；Shenzhen University of Advanced Technology（深圳先进技术大学）；Guangdong Provincial Key Laboratory of Computility Microelectronics（广东省计算微电子重点实验室）
Multi-Scale Global-Instance Prompt Tuning for Continual Test-time Adaptation in Medical Image Segmentation
School of Computer Science, University of Nottingham, UK（诺丁汉大学计算机科学学院）；School of Artificial Intelligence, Shenzhen University, China（深圳大学人工智能学院）；Department of Information Engineering and Computer Science, University of Trento, Italy（特伦托大学信息工程与计算机科学系）；School of Computer Science and Information Engineering, Hefei University of Technology, China（合肥工业大学计算机科学与信息工程学院）
Unsupervised Domain Adaptation (UDA) is a transfer learning task which aims at training on an unlabeled target domain by leveraging a labeled source domain. Beyond the traditional scope of UDA with a single source domain and a single target domain, real-world perception systems face a variety of scenarios to handle, from varying lighting conditions to many cities around the world. In this context, UDAs with several domains increase the challenges with the addition of distribution shifts within the different target domains. This work focuses on a novel framework for learning UDA, continuous UDA…
Continual Test-Time Adaptation (CTTA) generalizes conventional Test-Time Adaptation (TTA) by assuming that the target domain is dynamic over time rather than stationary. In this paper, we explore Multi-Modal Continual Test-Time Adaptation (MM-CTTA) as a new extension of CTTA for 3D semantic segmentation. The key to MM-CTTA is to adaptively attend to the reliable modality while avoiding catastrophic forgetting during continual domain shifts, which is out of the capability of previous TTA or CTTA methods. To fulfill this gap, we propose an MM-CTTA method called Continual Cross-Modal Adaptive Clu…
IBM Research Europe（IBM欧洲研究院）；ETH Zurich（苏黎世联邦理工学院）；Uppsala University（乌普萨拉大学）
Deep learning models can achieve high accuracy when trained on large amounts of labeled data. However, real-world scenarios often involve several challenges: Training data may become available in installments, may originate from multiple different domains, and may not contain labels for training. Certain settings, for instance medical applications, often involve further restrictions that prohibit retention of previously seen data due to privacy regulations. In this work, to address such challenges, we study unsupervised segmentation in continual learning scenarios that involve domain shift. To…
In this paper, our goal is to adapt a pre-trained convolutional neural network to domain shifts at test time. We do so continually with the incoming stream of test batches, without labels. The existing literature mostly operates on artificial shifts obtained via adversarial perturbations of a test image. Motivated by this, we evaluate the state of the art on two realistic and challenging sources of domain shifts, namely contextual and semantic shifts. Contextual shifts correspond to the environment types, for example, a model pre-trained on indoor context has to adapt to the outdoor context on…
Peking University（北京大学）；The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））；Imperial College London（帝国理工学院）
When facing changing environments in the real world, the lightweight model on client devices suffers from severe performance drops under distribution shifts. The main limitations of the existing device model lie in (1) unable to update due to the computation limit of the device, (2) the limited generalization ability of the lightweight model. Meanwhile, recent large models have shown strong generalization capability on the cloud while they can not be deployed on client devices due to poor computation constraints. To enable the device model to deal with changing environments, we propose a new l…
