Pith. sign in

REVIEW 2 major objections 5 minor 44 references

A lightweight vision foundation model can refine semantic segmentation on the fly, without gradients or memory banks, from the first test frame.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-11 23:47 UTC pith:HXP6RK6X

load-bearing objection Practical gradient-free TTDA refinement that actually works from the first frame and plugs into existing pipelines. the 2 major comments →

arxiv 2607.03810 v1 pith:HXP6RK6X submitted 2026-07-04 cs.CV

TestMate: Test-Time Domain Adaptation Aided by Lightweight Vision Foundation Model

Dimitrios Fotiou , Vasileios Mygdalis , Ioannis Pitas This is my paper
classification cs.CV
keywords test-time domain adaptationsemantic segmentationvision foundation modelsource-free domain adaptationonline TTDAFastSAMbackpropagation-freemask fusion
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Test-time domain adaptation for semantic segmentation has been stuck between two bad options: expensive gradient updates that risk forgetting and produce noisy boundaries, or memory-bank methods that need many samples before they help and struggle when the target distribution keeps shifting. This paper argues that a small, zero-shot instance segmenter can supply multi-scale object and part masks that already generalize well, and that a simple size-ordered fusion of those masks with the main model's logits is enough to correct high-entropy regions and sharpen boundaries immediately. The fusion is parameter-free, runs in real time, and can stand alone or be dropped into existing adaptation pipelines. On standard sim-to-real and real-to-real benchmarks the method raises mean IoU from the first frame and sets new state-of-the-art numbers under three related settings: ordinary test-time adaptation, source-free adaptation, and online adaptation under continuous shift.

Core claim

A parameter-free, size-ordered soft fusion of multi-scale unlabeled masks from a lightweight zero-shot instance segmenter (FastSAM / YOLOv8-seg) with a primary segmentation model's logits yields immediate first-frame domain adaptation for semantic segmentation, improves small-object and boundary accuracy, avoids catastrophic forgetting, and raises final mIoU both alone and when plugged into existing TTDA methods.

What carries the argument

TestMate: masks are sorted ascending by size; each mask's dominant class is found by majority vote on the primary logits; small high-confidence regions refine larger, less certain ones via soft logit blending gated by local entropy, with each pixel updated at most once.

Load-bearing premise

The zero-shot foundation-model masks stay accurate and multi-scale enough under the target domain shift that majority-vote labeling of those masks reliably corrects the primary model's high-entropy regions.

What would settle it

On a target domain where FastSAM (conf=0.1, 640 px) itself produces systematically wrong or incomplete multi-scale masks, TestMate's first-frame and final mIoU gains over the unadapted source model should disappear or reverse.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Existing TTDA, SFDA and online-TTDA pipelines can gain immediate first-frame accuracy and higher final mIoU simply by refining their outputs or pseudo-labels with TestMate.
  • Gradient-free, memory-free instance-based refinement becomes a practical real-time option for safety-critical segmentation under continuous distribution shift.
  • Small and rare classes, and object boundaries, can be improved without waiting for memory-bank convergence or risking forgetting of earlier domains.
  • Lightweight foundation models can be used as external spatial priors rather than as models that themselves need test-time fine-tuning.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same size-ordered competitive fusion could be tried with other real-time zero-shot segmenters or with multi-scale proposals from classical computer vision, testing how much of the gain is specific to FastSAM.
  • If the VFM degrades under extreme weather or sensor shifts, a lightweight confidence gate on the foundation masks themselves would become necessary before fusion.
  • The approach suggests a broader pattern: treat a frozen generalist model as a cheap spatial regularizer that supplies structure the specialist model has lost under domain shift.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper introduces TestMate, a backpropagation-free, real-time TTDA method for semantic segmentation. It uses a lightweight zero-shot VFM (FastSAM/YOLOv8-seg) to produce multi-scale unlabeled object/part masks that are fused with the primary model’s logits via a size-ordered (ascending), soft-blended, entropy-filtered competitive scheme (Algorithm 1, Eqs. 1–5). The fusion is claimed to be parameter-free, to enable immediate first-frame adaptation without catastrophic forgetting, and to preserve fine boundaries and small objects. TestMate is evaluated standalone and as a plug-in to CoTTA/DIGA/DT-ST, reporting new SOTA mIoU on GTA-V o Cityscapes (SFDA/TTDA) and FMB o MVSeg (online-TTDA) under both shuffled and sequential continuous shifts (Tables 2–4), with supporting ablations (Tables 6–8) and qualitative boundary improvements (Figs. 3, 5).

Significance. If the reported gains hold, TestMate supplies a practical, low-overhead solution for real-time semantic segmentation under domain shift—especially valuable for autonomous driving and other streaming applications where back-propagation or large memory banks are undesirable. Strengths include the explicit parameter-free heuristic, the ability to operate from the first frame, seamless integration with both gradient-based and memory-bank baselines, consistent improvements across three distinct adaptation protocols, component ablations that isolate ordering/soft/entropy contributions, and public code. The reliance on FastSAM generalization under shift is a natural soft spot, yet the multi-domain results and first-frame/iteration-stability evidence already bound the risk; degradation of the VFM would mainly reduce gain magnitude rather than invert rankings.

major comments (2)
  1. [§3.3, Algorithm 1, Eq. (1)] §3.3 / Algorithm 1 / Eq. (1): The dominant-class majority vote is taken directly from the primary model’s argmax inside each VFM mask. Under severe domain shift the primary logits can be highly noisy; while soft blending (Eq. 2) and entropy filtering (Eqs. 3–5) mitigate propagation, the paper never quantifies how often VFM masks land on high-entropy primary regions or how the fusion behaves when the majority vote itself is unreliable. A short failure-case analysis or confidence-weighted vote would make the “immediate first-frame” claim more robust.
  2. [Tables 2–4] Tables 2–4: All mIoU numbers are single-run point estimates with no error bars, multi-seed averages, or batch-order sensitivity. TTDA methods can be sensitive to streaming order; given that the SOTA claim rests on relatively modest absolute gains (especially the 0.9–2.8 pp improvements when TestMate is added to DIGA/CoTTA), reporting variance would strengthen the ranking assertions.
minor comments (5)
  1. [Abstract, §1] Abstract and §1 repeatedly write “paremeter-free”; correct to “parameter-free”.
  2. [Figure 4] Figure 4 x-axis label “Itteration” should be “Iteration”.
  3. [Table 2] Table 2 (TestMate+DT-ST row) contains concatenated digits (“731.747.7161.6”); re-typeset the per-class columns for readability.
  4. [§3.2, §4.1.3] §3.2 and §4.1.3: FastSAM confidence/IoU thresholds (0.1/0.6) and 640 px resolution are fixed for speed; a one-sentence sensitivity note (or pointer to Table 8) would clarify that they are not tuned to target mIoU.
  5. [throughout] Terminology oscillates between “Visual Foundation Model” and “Vision Foundation Model”; pick one consistently.

Circularity Check

0 steps flagged

No significant circularity; TestMate is a fixed, parameter-free heuristic whose gains are measured empirically on held-out benchmarks.

full rationale

The paper's central claim is an engineering construction (Algorithm 1 in §3.3): FastSAM zero-shot masks are sorted by ascending size, labeled by majority vote on the primary model's logits, then softly blended only on high-entropy non-dominant pixels via the closed-form ratio λ = |Ri|/|Mi|. No free parameters are fitted to target mIoU; the only knobs (conf=0.1, IoU=0.6, 640 px) are chosen for real-time speed and ablated in Tables 7–8. Reported SOTA numbers (Tables 2–4) are ordinary held-out evaluations under TTDA/SFDA/online-TTDA protocols, both standalone and as a plug-in to CoTTA/DIGA. There are no self-definitional equations, no fitted quantities re-labeled as predictions, no uniqueness theorems imported from the authors, and no load-bearing self-citations that close a definitional loop. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

3 free parameters · 3 axioms · 1 invented entities

As an empirical systems paper the load-bearing content is the algorithmic construction and the experimental protocol; free parameters are the few FastSAM hyper-parameters chosen for real-time operation, axioms are standard CV modeling assumptions plus the claim that the VFM generalizes, and the only invented entity is the TestMate fusion procedure itself.

free parameters (3)
  • FastSAM confidence threshold = 0.1
    Set to 0.1 (Implementation Details); controls which masks enter the fusion and is not derived from first principles.
  • FastSAM NMS IoU threshold = 0.6
    Set to 0.6; directly affects the multi-scale proposal set used by Algorithm 1.
  • FastSAM input resolution = 640
    Chosen as 640 px for the main experiments after ablation (Table 8); trades accuracy for speed.
axioms (3)
  • domain assumption A lightweight zero-shot instance segmenter pretrained on large diverse data produces useful multi-scale object/part masks under the target domain shifts of interest.
    Stated in §3.2 and used as the sole source of spatial priors throughout Algorithm 1; if false the fusion has nothing clean to vote over.
  • ad hoc to paper Smaller masks are more likely to correspond to single coherent objects than larger composite masks.
    Justifies the ascending-size ordering in §3.3.1; empirically helpful (Table 7) but not proven.
  • ad hoc to paper Softmax-blended logits weighted by dominant-region fraction preserve uncertainty better than hard reassignment.
    Eq. (2) and §3.3.2; ablation shows modest gain over hard refinement.
invented entities (1)
  • TestMate size-ordered competitive soft fusion no independent evidence
    purpose: Parameter-free mechanism that lets small high-confidence VFM regions dominate and refine surrounding larger uncertain areas of the primary model.
    The entire algorithmic contribution of the paper (Algorithm 1); no independent existence outside this work.

pith-pipeline@v1.1.0-grok45 · 19700 in / 2711 out tokens · 23289 ms · 2026-07-11T23:47:39.229600+00:00 · methodology

0 comments
read the original abstract

Test-Time Domain Adaptation (TTDA) aims to adapt Deep Neural Networks to distribution shifts using only streaming, unlabeled test data in real time. Current methods for semantic segmentation tasks suffer from critical limitations. Entropy minimization techniques require costly backpropagation, risking catastrophic forgetting and producing noisy segmentation boundaries. Memory-bank methods, while backpropagation-free, exhibit slow adaptation, requiring numerous samples to converge and struggle to handle continuous domain shifts. We introduce TestMate, a novel, real-time, and backpropagation-free TTDA framework that overcomes these issues. TestMate leverages generalization capability of a lightweight Visual Foundation Model to guide the adaptation. We use a zero-shot instance segmentation YOLOv8-seg based model to generate unlabeled mask proposals for objects and their parts at multiple scales in real time. These proposals are fused with the primary model via a heuristic, size-ordered competitive scheme, where small, high-confidence regions dominate and refine predictions in surrounding larger, less certain areas. This paremeter-free mechanism enables immediate adaptation from the first frame, inherently avoids catastrophic forgetting and effectively preserves fine object details and boundaries, even for small objects. TestMate can be used as a standalone, efficient refinement module or seamlessly integrated into existing TTDA methods to significantly boost their performance. We demonstrate state-of-the-art results across two benchmark datasets, proving TestMate's effectiveness in three distinct adaptation tasks: TTDA, Source-Free Domain Adaptation (SFDA), and online-TTDA. Code is available.

Figures

Figures reproduced from arXiv: 2607.03810 by Dimitrios Fotiou, Ioannis Pitas, Vasileios Mygdalis.

Figure 1
Figure 1. Figure 1: TestMate is a lightweight, backpropagation-free adap￾tation module. For a test image, a Segmentation DNN produces an initial output, while a Visual Foundation Model generates high￾quality unlabeled masks. TestMate processes masks heuristically, from smallest to largest, labels them via majority voting, and fuses them with the DNN output, considering their entropy and re￾gion dominance. This parameter-free … view at source ↗
Figure 2
Figure 2. Figure 2: Example cases of VFM Masks: (a) Single-object mask capturing a person; (b) Multi-object mask capturing both person and bike. unlabeled masks {Mi} N i=1 using a VFM gϕ. Formally, {Mi} N i=1 = gϕ(Xt), Mi ∈ {0, 1} H×W , where ϕ denotes the parameters of the foundation model and N is the number of object masks generated. We then identify the dominant class within a maskMi ⊆ Yˆt, where Mi = { p | Mi(p) = 1 } de… view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative comparison on GTA-V → Cityscapes [4] and FMB [22] → MVSeg [15] Dataset of TTDA methods. benefit from improved segmentation immediately, without waiting for full adaptation. For pseudo-label-based methods, TestMate offers addi￾tional advantages. Many approaches [31, 37, 42] use a teacher–student setup, where the teacher generates pseudo￾labels for the student to learn from. The teacher is period… view at source ↗
Figure 4
Figure 4. Figure 4: Effect of TestMate (orange) when paired with (a) DIGA [39] (blue) and (b) DS-DT [42] (blue) in accuracy (mIoU) of inter￾mediate produced pseudo-labels (GTA-V [28] → Cityscapes [4]) teacher model at the right moments from the student. While conceptually related to CoTTA [37], which performs mod￾erately in TTDA, DS-DT [42] clearly outperforms it in this setting. When combined with TestMate, the DS-DT [42] + … view at source ↗
Figure 5
Figure 5. Figure 5: Visual example of TestMate (b, d) boosting boundaries and increasing coherence of semantic instances in the Cityscapes dataset compared to DIGA [39] method (a, c). aries. As shown in the first row of [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

44 extracted references · 3 linked inside Pith

  1. [1]

    Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolu- tion, and fully connected crfs.IEEE transactions on pattern analysis and machine intelligence, 40(4):834–848, 2017. 5, 6

    2017

  2. [2]

    Rethinking atrous convolution for seman- tic image segmentation.arXiv preprint arXiv:1706.05587,

    Liang-Chieh Chen, George Papandreou, Florian Schroff, and Hartwig Adam. Rethinking atrous convolution for seman- tic image segmentation.arXiv preprint arXiv:1706.05587,

    Pith/arXiv arXiv

  3. [3]

    Do- main adaptation for semantic segmentation with maximum squares loss

    Minghao Chen, Hongyang Xue, and Deng Cai. Do- main adaptation for semantic segmentation with maximum squares loss. InProceedings of the IEEE/CVF international conference on computer vision, pages 2090–2099, 2019. 5

    2090

  4. [4]

    The cityscapes dataset for semantic urban scene understanding

    Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. InProceed- ings of the IEEE conference on computer vision and pattern recognition, pages 3213–3223, 2016. 2, 5, 6, 7, 8

    2016

  5. [5]

    The mnist database of handwritten digit images for machine learning research.IEEE Signal Processing Maga- zine, 29(6):141–142, 2012

    Li Deng. The mnist database of handwritten digit images for machine learning research.IEEE Signal Processing Maga- zine, 29(6):141–142, 2012. 2

    2012

  6. [6]

    Robust mean teacher for continual and gradual test-time adaptation

    Mario D ¨obler, Robert A Marsden, and Bin Yang. Robust mean teacher for continual and gradual test-time adaptation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7704–7714, 2023. 3

    2023

  7. [7]

    Uncertainty reduction for model adaptation in semantic segmentation

    Francois Fleuret et al. Uncertainty reduction for model adaptation in semantic segmentation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9613–9623, 2021. 5, 6

    2021

  8. [8]

    Back to the source: Diffusion- driven adaptation to test-time corruption

    Jin Gao, Jialing Zhang, Xihui Liu, Trevor Darrell, Evan Shel- hamer, and Dequan Wang. Back to the source: Diffusion- driven adaptation to test-time corruption. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11786–11796, 2023. 2

    2023

  9. [9]

    Sim- ple copy-paste is a strong data augmentation method for in- stance segmentation

    Golnaz Ghiasi, Yin Cui, Aravind Srinivas, Rui Qian, Tsung- Yi Lin, Ekin D Cubuk, Quoc V Le, and Barret Zoph. Sim- ple copy-paste is a strong data augmentation method for in- stance segmentation. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 2918–2928, 2021. 6

    2021

  10. [10]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. InProceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 5, 6, 7

    2016

  11. [11]

    Model adaptation: Historical contrastive learning for unsu- pervised domain adaptation without source data.Advances in neural information processing systems, 34:3635–3649,

    Jiaxing Huang, Dayan Guan, Aoran Xiao, and Shijian Lu. Model adaptation: Historical contrastive learning for unsu- pervised domain adaptation without source data.Advances in neural information processing systems, 34:3635–3649,

  12. [12]

    Multispectral pedestrian detection: Benchmark dataset and baseline

    Soonmin Hwang, Jaesik Park, Namil Kim, Yukyung Choi, and In So Kweon. Multispectral pedestrian detection: Benchmark dataset and baseline. InProceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 1037–1045, 2015. 5, 7, 8

    2015

  13. [13]

    Video analytics dataset.https://www.ino.ca/ en/technologies/video- analyticsdataset/,

    INO. Video analytics dataset.https://www.ino.ca/ en/technologies/video- analyticsdataset/,

  14. [14]

    Accessed: 2025-08-01. 5, 7, 8

    2025

  15. [15]

    Deflating dataset bias us- ing synthetic data augmentation

    Nikita Jaipuria, Xianling Zhang, Rohan Bhasin, Mayar Arafa, Punarjay Chakravarty, Shubham Shrivastava, Sagar Manglani, and Vidya N Murali. Deflating dataset bias us- ing synthetic data augmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 772–773, 2020. 1

    2020

  16. [16]

    Yuille, and Li Cheng

    Wei Ji, Jingjing Li, Cheng Bian, Zongwei Zhou, Jiaying Zhao, Alan L. Yuille, and Li Cheng. Multispectral video semantic segmentation: A benchmark dataset and baseline. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 1094–1104,

  17. [17]

    YOLO by ultralytics, 2023

    Glenn Jocher, Ayush Chaurasia, and Jing Qiu. YOLO by ultralytics, 2023. Available:https://github.com/ ultralytics/ultralytics, accessed 2023. 2, 3, 6

    2023

  18. [18]

    Ev-tta: Test-time adaptation for event-based object recognition

    Junho Kim, Inwoo Hwang, and Young Min Kim. Ev-tta: Test-time adaptation for event-based object recognition. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 17745–17754, 2022. 3

    2022

  19. [19]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. InProceedings of the IEEE/CVF international confer- ence on computer vision, pages 4015–4026, 2023. 3

    2023

  20. [20]

    A review of domain adap- tation without target labels.IEEE transactions on pattern analysis and machine intelligence, 43(3):766–785, 2019

    Wouter M Kouw and Marco Loog. A review of domain adap- tation without target labels.IEEE transactions on pattern analysis and machine intelligence, 43(3):766–785, 2019. 2

    2019

  21. [21]

    Rgb-t object tracking: Benchmark and baseline.Pat- tern Recognition, 96:106977, 2019

    Chenglong Li, Xinyan Liang, Yijuan Lu, Nan Zhao, and Jin Tang. Rgb-t object tracking: Benchmark and baseline.Pat- tern Recognition, 96:106977, 2019. 5, 7, 8

    2019

  22. [22]

    A comprehensive survey on source-free domain adap- tation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    Jingjing Li, Zhiqi Yu, Zhekai Du, Lei Zhu, and Heng Tao Shen. A comprehensive survey on source-free domain adap- tation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. 2

    2024

  23. [23]

    Multi- interactive feature learning and a full-time multi-modality benchmark for image fusion and segmentation

    Jinyuan Liu, Zhu Liu, Guanyao Wu, Long Ma, Risheng Liu, Wei Zhong, Zhongxuan Luo, and Xin Fan. Multi- interactive feature learning and a full-time multi-modality benchmark for image fusion and segmentation. InProceed- ings of the IEEE/CVF international conference on computer vision, pages 8115–8124, 2023. 5, 7

    2023

  24. [24]

    Source-free domain adaptation for semantic segmentation

    Yuang Liu, Wei Zhang, and Jun Wang. Source-free domain adaptation for semantic segmentation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1215–1224, 2021. 5, 6

    2021

  25. [25]

    Improved self-training for test-time adaptation

    Jing Ma. Improved self-training for test-time adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 23701–23710, 2024. 3

    2024

  26. [26]

    Reading digits in natural images with unsupervised feature learning

    Yuval Netzer, Tao Wang, Adam Coates, Alessandro Bis- sacco, Baolin Wu, Andrew Y Ng, et al. Reading digits in natural images with unsupervised feature learning. InNIPS workshop on deep learning and unsupervised feature learn- ing, page 4. Granada, 2011. 2

    2011

  27. [27]

    Efficient test-time model adaptation without forgetting

    Shuaicheng Niu, Jiaxiang Wu, Yifan Zhang, Yaofo Chen, Shijian Zheng, Peilin Zhao, and Mingkui Tan. Efficient test-time model adaptation without forgetting. InInterna- tional conference on machine learning, pages 16888–16905. PMLR, 2022. 4, 5, 6, 7, 8

    2022

  28. [28]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. InInternational conference on machine learning, pages 8748–8763. PmLR, 2021. 3

    2021

  29. [29]

    Playing for data: Ground truth from computer games

    Stephan R Richter, Vibhav Vineet, Stefan Roth, and Vladlen Koltun. Playing for data: Ground truth from computer games. InComputer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part II 14, pages 102–118. Springer,

    2016

  30. [30]

    Generalization in neural networks: A broad survey.Neurocomputing, 611:128701, 2025

    Chris Rohlfs. Generalization in neural networks: A broad survey.Neurocomputing, 611:128701, 2025. 1

    2025

  31. [31]

    Acdc: The adverse conditions dataset with correspondences for se- mantic driving scene understanding

    Christos Sakaridis, Dengxin Dai, and Luc Van Gool. Acdc: The adverse conditions dataset with correspondences for se- mantic driving scene understanding. InProceedings of the IEEE/ICVF international conference on computer vision, pages 10765–10775, 2021. 1

    2021

  32. [32]

    Ecotta: Memory-efficient continual test-time adaptation via self-distilled regularization

    Junha Song, Jungsoo Lee, In So Kweon, and Sungha Choi. Ecotta: Memory-efficient continual test-time adaptation via self-distilled regularization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11920–11929, 2023. 2, 3, 5

    2023

  33. [33]

    Source- free domain adaptation with frozen multimodal foundation model

    Song Tang, Wenxin Su, Mao Ye, and Xiatian Zhu. Source- free domain adaptation with frozen multimodal foundation model. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23711– 23720, 2024. 3

    2024

  34. [34]

    Vdm-da: Virtual domain modeling for source data-free domain adap- tation.IEEE Transactions on Circuits and Systems for Video Technology, 32(6):3749–3760, 2021

    Jiayi Tian, Jing Zhang, Wen Li, and Dong Xu. Vdm-da: Virtual domain modeling for source data-free domain adap- tation.IEEE Transactions on Circuits and Systems for Video Technology, 32(6):3749–3760, 2021. 2

    2021

  35. [35]

    Gda: Generalized diffusion for robust test-time adaptation

    Yun-Yun Tsai, Fu-Chen Chen, Albert YC Chen, Junfeng Yang, Che-Chun Su, Min Sun, and Cheng-Hao Kuo. Gda: Generalized diffusion for robust test-time adaptation. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23242–23251, 2024. 2

    2024

  36. [36]

    Yolov8: A novel object detection algorithm with enhanced performance and robust- ness

    Rejin Varghese and Sambath M. Yolov8: A novel object detection algorithm with enhanced performance and robust- ness. In2024 International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS), pages 1–6, 2024. 6

    2024

  37. [37]

    Tent: Fully test-time adaptation by entropy minimization.arXiv preprint arXiv:2006.10726,

    Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno Ol- shausen, and Trevor Darrell. Tent: Fully test-time adaptation by entropy minimization.arXiv preprint arXiv:2006.10726,

    Pith/arXiv arXiv 2006

  38. [38]

    Continual test-time domain adaptation

    Qin Wang, Olga Fink, Luc Van Gool, and Dengxin Dai. Continual test-time domain adaptation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7201–7211, 2022. 3, 5, 6, 7, 8

    2022

  39. [39]

    Feature alignment and uniformity for test time adap- tation

    Shuai Wang, Daoan Zhang, Zipei Yan, Jianguo Zhang, and Rui Li. Feature alignment and uniformity for test time adap- tation. InProceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, pages 20050–20060,

  40. [40]

    Dynamically instance- guided adaptation: A backward-free approach for test-time domain adaptive semantic segmentation

    Wei Wang, Zhun Zhong, Weijie Wang, Xi Chen, Charles Ling, Boyu Wang, and Nicu Sebe. Dynamically instance- guided adaptation: A backward-free approach for test-time domain adaptive semantic segmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24090–24099, 2023. 3, 5, 6, 7, 8

    2023

  41. [41]

    Sam4udass: When sam meets un- supervised domain adaptive semantic segmentation in intel- ligent vehicles.IEEE Transactions on Intelligent Vehicles, 9 (2):3396–3408, 2023

    Weihao Yan, Yeqiang Qian, Hanyang Zhuang, Chunxiang Wang, and Ming Yang. Sam4udass: When sam meets un- supervised domain adaptive semantic segmentation in intel- ligent vehicles.IEEE Transactions on Intelligent Vehicles, 9 (2):3396–3408, 2023. 3

    2023

  42. [42]

    Source data-free unsupervised domain adaptation for seman- tic segmentation

    Mucong Ye, Jing Zhang, Jinpeng Ouyang, and Ding Yuan. Source data-free unsupervised domain adaptation for seman- tic segmentation. InProceedings of the 29th ACM interna- tional conference on multimedia, pages 2233–2242, 2021. 5, 6

    2021

  43. [43]

    Towards better stability and adaptabil- ity: Improve online self-training for model adaptation in se- mantic segmentation

    Dong Zhao, Shuang Wang, Qi Zang, Dou Quan, Xiutiao Ye, and Licheng Jiao. Towards better stability and adaptabil- ity: Improve online self-training for model adaptation in se- mantic segmentation. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 11733–11743, 2023. 3, 5, 6

    2023

  44. [44]

    Fast segment any- thing.arXiv preprint arXiv:2306.12156, 2023

    Xu Zhao, Wenchao Ding, Yongqi An, Yinglong Du, Tao Yu, Min Li, Ming Tang, and Jinqiao Wang. Fast segment any- thing.arXiv preprint arXiv:2306.12156, 2023. 2, 3, 6, 8

    Pith/arXiv arXiv 2023