Pith. sign in

REVIEW 4 major objections 4 minor 98 references

Is Visual in-Context Learning for Compositional Medical Tasks within Reach?

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Visual in-context learners can follow multi-step medical imaging pipelines defined at test time.

desk verdict Compositional in-context learning with a single model is a real and new setting, and sequence-token masking works for generative/transformation tasks, but the discriminative results sit too close to the copy baseline to support the full claim. read the letter →

arxiv 2507.00868 v2 pith:7OEOIAPQ submitted 2025-07-01 cs.CV

Simon Rei{ss} , Zdravko Marinov , Alexander Jaus , Constantin Seibold , M. Saquib Sarfraz , Erik Rodner , Rainer Stiefelhagen This is my paper
classification cs.CV
keywords visualin-contextlearningcompositionaltasksmedicalimageanalysistasksequencegenerationmaskedtokenpredictionVQ-GANcodebooksyntheticdatamulti-steppipelines
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper asks whether a single vision model can execute an entire image-processing pipeline—say, denoise, rotate, then segment—that a user defines on the spot from a few example pairs, without any retraining. The authors answer provisionally yes: they show that a transformer can be trained on synthetically generated chains of tasks, drawn from 38 medical segmentation datasets, to predict the full sequence of intermediate outputs for a new query image. The decisive training choice is to mask exactly the output sequence the model will have to produce at test time, which outperforms generic token or image masking by a wide margin. The paper also pins down where the ceiling sits: a pre-trained image-synthesis codebook cannot faithfully represent fine-grained or many-class segmentation maps, so they fine-tune the codebook on task outputs with random recoloring, which roughly doubles the upper performance bound for discriminative tasks. The net claim is that test-time-defined compositional vision pipelines are trainable today, though still far from clinical utility and confined to the task families seen in training.

What carries the argument

The central mechanism is a pipeline of three coupled components. First, a synthetic task-generation engine bootstraps diverse task sequences from 38 segmentation datasets by enriching each image–mask pair with generative tasks (super resolution, inpainting, denoising, color jitter, inversion, equalization, brightness), geometric transformations (flips, rotations), and discriminative outputs (segments, boxes, edges, skeletons, points), then sampling chains under guardrails: generative steps first, transformations second, discriminative tasks last, with class subsampling, random recoloring, parameter sampling, and sequence length capped at 30 images. Second, a task-informed codebook, a vector-quantized autoencoder (VQ-GAN) that maps each $200\times200$ image to 144 discrete tokens from a vocabulary of 16,384, is fine-tuned on the enriched data with data-balancing and random color remapping so that discrete tokens can actually represent task outputs. Third, the in-context learner, a GPT-2 Transformer, is trained with sequence-token masking: all tokens of the output sequence $O$ are replaced and must be predicted from the context–query token sequence, an objective that mirrors exactly what the model must do at test time and that clearly outperforms random token masking and image-token masking.

What would settle it

Train or evaluate on a compositional prompt whose task family or imaging modality is absent from the 38 MedSAM datasets—for example, a novel discriminative output type like cell counting or a pathology whole-slide modality—and check whether performance collapses to the copy baseline. A sharper in-distribution test: build a sequence where a class appears in the query's segmentation output but never in any context example, and verify whether the model still emits that class, which would reveal memorization rather than in-context following.

Watch

Extended reading notes

Core claim

The paper's central discovery is that visual in-context learning, previously demonstrated for single tasks, can be extended to compositional task sequences: a single GPT-style transformer, fed a context of example image-to-image step pairs, can generate the entire chain of outputs for a held-out query image across medical modalities. On a test set of 2,000 synthetic compositional tasks built from image and task outputs held out from training, the model with sequence-token masking reaches a segmentation IoU of 61.2% and box F-1 of 65.9%, far above a copy baseline (about 50.8% IoU) but below the codebook upper bound (about 86.7% IoU). The authors further establish that the representation bottleneck sits in the codebook: off-the-shelf VQ-GAN codebooks trained on natural images recover at most 29% IoU on a 142-class anatomy dataset, whereas fine-tuning the codebook on the task outputs themselves, with dataset/task balancing and color-remapping augmentation, raises the upper bound to near-90% IoU on binary and few-class tasks while generalizing to arbitrary class colors. Sequence-token masking—switching out all tokens of the output sequence during training—is identified as the objective that forces the model to learn long-range dependencies from context to output.

Load-bearing premise

The approach rests on the assumption that the synthetically generated task chains—constrained by hand-designed ordering, class subsampling, and recoloring rules—actually resemble the multi-step pipelines a clinician or user would define at test time; evaluation never leaves the 38 training datasets or the task families the generator can produce.

Editorial extensions

If this is right

  • A medical user could define a bespoke pipeline—such as artifact removal, orientation correction, then structure segmentation—by showing the model a few example input–output pairs, and run it on new images without retraining or programming.
  • The masking comparison establishes a direct recipe: to make an in-context learner solve a test-time task, train it to predict exactly that output region under masking, not generic token corruption.
  • Codebook quality is the binding constraint; any attempt to scale this approach must address representation fidelity for fine structures and many-class semantics before the learner itself improves.
  • Because intermediate outputs are generated and inspectable step by step, the approach offers a form of interpretability useful in medical imaging, where each pipeline stage can be verified.
  • The gap between the best model and the codebook upper bound quantifies the remaining headroom—roughly 25 IoU points on segmentation—indicating where training objectives and data scale must improve.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The guardrail ordering (generative before transformation before discriminative) implicitly encodes a prior on real clinical pipelines—restore, align, then label. A direct test would train on a permuted order and see whether the learners themselves impose the ordering, which would tell us whether the structure is learned or injected.
  • The class-subsampling guardrail implies a concrete failure mode worth probing: if a query output contains a class never shown in the context, the model should be unable to produce it; this yields a diagnostic for true context adherence versus dataset-memorized class priors.
  • Since the generation engine works off arbitrary segmentation datasets, the same recipe could transfer to non-medical domains; a natural extension is to test whether out-of-domain capability appears at larger data and model scale, as suggested by prior scaling results.
  • The color-remapping augmentation points toward a more general principle: in-context learners generalize to unseen output appearances only when the tokenizer itself is invariant to those appearances; training the tokenizer with output-remapping noise may be a general precondition for visual in-context learning.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper investigates whether visual in-context learning can be extended from single tasks to compositional sequences of medical imaging tasks. The authors propose a synthetic task generation engine that bootstraps compositional task sequences from 38 MedSAM segmentation datasets, covering generative, geometric transformation, and discriminative tasks. They analyze codebook reconstruction bounds, propose a task-informed codebook with color remapping augmentation and data balancing, and compare three masking-based training objectives for a GPT-2 style in-context learner. The main quantitative result (Table 1) is that sequence-token masking outperforms token and image-token masking on held-out compositional tasks, with particular gains on generative and transformation tasks; however, discriminative task scores are close to or below a trivial copy baseline. Qualitative examples show coherent multi-step predictions on some sequences, while the appendix documents failures such as plausible but mislocated structures. The paper explicitly acknowledges several limitations, including generalization beyond the training datasets.

Significance. If the main claim were fully supported, this would be a useful first demonstration that a single transformer-based model can be trained to execute multi-step, test-time-defined compositional vision pipelines in the medical domain, with interpretable intermediate outputs. The experimental design is a genuine strength: the sandwich of a copy baseline and a measured codebook upper bound provides a clear frame for isolating the in-context learner's contribution, and the systematic comparison of masking strategies is informative even if the interpretation requires care. The codebook analysis (Section 4 and Table 2) is also valuable, showing that off-the-shelf VQ-GANs severely under-represent fine-grained discriminative outputs and that task-informed fine-tuning substantially raises the upper bound. The paper is honest about limitations in Appendix E, which strengthens its credibility. However, the central claim about mastering discriminative compositional steps is not yet established by the evidence, and the generalization claims are narrower than the abstract suggests.

major comments (4)
  1. [Table 1; Sec. 5.2.1; Appendix E]
  2. [Sec. 5.3; Table 1]
  3. [Sec. 3.1; Sec. 6.4; Appendix E]
  4. [Sec. 5.3; Sec. 6.4]
minor comments (4)
  1. [Sec. 6.2]
  2. [Sec. 5.2.2]
  3. [Table 1]
  4. [Appendix D.2]

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: codebook bounds are measured, the training objective matches the test protocol by design but does not feed the output back as input, and the paper's self-citations are not load-bearing.

full rationale

The paper's empirical chain is self-contained. Codebook upper bounds (Table 2 and the last row of Table 1) are obtained by directly encoding and decoding task outputs with the deployed codebook; this is a measured ceiling for any token-space predictor, not a prediction derived from the model. The in-context learner is trained with a masked-token objective (Sec. 5.3) and evaluated by decoding its predicted tokens; the quantity being predicted (O) is never used as an input to the model, so the result does not reduce to its inputs by construction. The sequence-token masking variant matches the test protocol by design, but this is standard supervised training rather than circularity: the model must still generate the full output sequence from the context C and query Q. The copy baseline is an external reference, and the paper explicitly reports that discriminative tasks remain near this baseline, which is a weakness in evidence strength, not a circular step. Self-citations (e.g., ATLAS [45], interactive segmentation taxonomy [57], prior segmentation work [70,71]) are used as data sources and related work; no uniqueness theorem or load-bearing claim rests on them. The main scope caveat, that test tasks are sampled from the same synthetic generation engine as training tasks, is explicitly disclosed in Appendix E as a limitation; it affects external validity, not the internal derivation chain.

Assumptions & free parameters 7 free parameters · 4 assumptions · 1 invented entities

The central claim rests on two empirically tested assumptions: the codebook can represent task outputs (bounded in Table 2) and the transformer can learn masked-token prediction on synthetic sequences (Table 1). The least-examined input is the synthetic task generator itself: its guardrails, class subsampling, recoloring, and ordering rules are hand-designed and not individually ablated. No new physical entities are introduced; the only new constructs are the special boundary tokens.

free parameters (7)
  • Sequence length cap = 30 images (up to 15 tasks)
    Chosen in Sec 5.2.2 to keep token sequences within compute limits; not swept.
  • Masking ratio p = 15%
    Used for token and image masking in Table 1; no sweep across p reported.
  • Number of END tokens k = 2
    Ablated over k=1,2,4 in Table 1; k=2 performed best.
  • Codebook tokens per image = 144 (VQ-GAN f16, vocabulary 16,384)
    Taken from pretrained model zoo [33,74]; controls the upper bound on reconstruction.
  • Best training epochs = 52
    Longer training with sequence-token masking improved all metrics (Table 1, row 8); no convergence analysis.
  • Data balance weights = Equal task outputs vs images; dataset oversampling
    Balancing scheme in Table 2 selected the best codebook; exact per-dataset weights not given.
  • Image resolution = 200x200
    All images rescaled to 200x200 before tokenization (Sec 6.1); affects the token count q=144.
assumptions (4)
  • domain assumption The VQ-GAN token space preserves enough task-relevant structure for in-context prediction
    The paper measures this via codebook reconstruction upper bounds (Table 2), but the final model still sits far below those bounds, so the assumption holds only partially.
  • domain assumption Synthetic task sequences derived from segmentation annotations are a valid training distribution for compositional medical pipelines
    Central to training data generation (Sec 5.1-5.2); the authors admit real-world tasks may not be captured (Appendix E).
  • domain assumption Masked token prediction with a GPT2 transformer can learn the conditional structure of task sequences
    Adapted from language modeling and visual ICL [8,38]; supported empirically by Table 1 but not theoretically guaranteed.
  • ad hoc to paper The enforced {generative} to {transformation} to {discriminative} ordering and class-subsampling guardrails yield informative tasks
    Guardrails in Sec 5.2 are hand-designed to prevent trivial sequences; no ablation shows their individual contribution.
invented entities (1)
  • i-END and c-END special tokens independent evidence
    purpose: Inserted after each image and after each context element to delimit sequence boundaries for the transformer.
    Ablation over k=1,2,4 (Table 1) gives small but consistent gains; they are new discrete input constructs, not physical entities.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Is Visual in-Context Learning for Compositional Medical Tasks within Reach?." pith.science (2026). https://pith.science/paper/7OEOIAPQ

@misc{pith2026250700868,
  author       = {Pith},
  title        = {Pith review of: Is Visual in-Context Learning for Compositional Medical Tasks within Reach?},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7OEOIAPQ}},
  note         = {Machine review of arXiv:2507.00868}
}
read the original abstract

In this paper, we explore the potential of visual in-context learning to enable a single model to handle multiple tasks and adapt to new tasks during test time without re-training. Unlike previous approaches, our focus is on training in-context learners to adapt to sequences of tasks, rather than individual tasks. Our goal is to solve complex tasks that involve multiple intermediate steps using a single model, allowing users to define entire vision pipelines flexibly at test time. To achieve this, we first examine the properties and limitations of visual in-context learning architectures, with a particular focus on the role of codebooks. We then introduce a novel method for training in-context learners using a synthetic compositional task generation engine. This engine bootstraps task sequences from arbitrary segmentation datasets, enabling the training of visual in-context learners for compositional tasks. Additionally, we investigate different masking-based training objectives to gather insights into how to train models better for solving complex, compositional tasks. Our exploration not only provides important insights especially for multi-modal medical task sequences but also highlights challenges that need to be addressed.

Figures

Figures reproduced from arXiv: 2507.00868 by the authors.

Figure 1
Figure 1. Top: In green, our scenario of visual in-context learning for compositional tasks using one general model. Bottom: Task￾specific image processing with multiple specialized models. supervised learning [6, 14]. Recently, a new approach has emerged: in-context learning [8, 13, 88]. In this paradigm, a model is trained to derive the correct output O for a query image Q given a task-specific context C. This approach is a… view at source ↗
Figure 2
Figure 2. Visual in-context learning architectures and paradigms. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Investigation on how well pre-trained codebooks represent task data. Left to right: Evaluation of image reconstruction, evaluation [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: On the left, we present the sequence structure, which we enforce to construct compositional tasks, with some guardrails shown on the bottom right. The top right shows the general processing pipeline of compositional tasks through a visual in-context learner. segments. …
Figure 5
Figure 5. Figure 5: Token masking variants for training visual in-context [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: We show three different qualitative compositional medical task predictions by the best visual in-context learning variant. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Different colormaps for the ATLAS dataset with which [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 9
Figure 9. Figure 9: Additional compositional contexts, predicted task sequences and the associated ground-truth task sequences. [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]
Figure 10
Figure 10. Figure 10: Additional compositional contexts, predicted task sequences and the associated ground-truth task sequences. [PITH_FULL_IMAGE:figures/full_fig_p019_10.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

98 extracted references · 50 canonical work pages

  1. [1]

    Deep learning based automated detection of intraretinal cys- toid fluid

    Zeeshan Ahmed, Shahbaz Qamar Panhwar, Attiya Baqai, Fahim Aziz Umrani, Munawar Ahmed, and Arbaaz Khan. Deep learning based automated detection of intraretinal cys- toid fluid. International Journal of Imaging Systems and Technology, 32(3):902–917, 2022. 16

    2022

  2. [2]

    Dataset of breast ultrasound images

    Walid Al-Dhabyani, Mohammed Gomaa, Hussien Khaled, and Aly Fahmy. Dataset of breast ultrasound images. Data in brief, 28:104863, 2020. 16

    2020

  3. [3]

    P. An, S. Xu, S. A. Harmon, E. B. Turkbey, T. H. Sanford, A. Amalou, M. Kassin, N. Varble, M. Blain, V . Anderson, G. Patella, F.and Carrafiello, B. T. Turkbey, and B. J. Wood. Ct images in covid-19 [data set]., 2020. 16

    2020

  4. [4]

    The medical segmentation decathlon.Nature communications, 13(1):4128, 2022

    Michela Antonelli, Annika Reinke, Spyridon Bakas, Key- van Farahani, Annette Kopp-Schneider, Bennett A Landman, Geert Litjens, Bjoern Menze, Olaf Ronneberger, Ronald M Summers, et al. The medical segmentation decathlon.Nature communications, 13(1):4128, 2022. 16

    2022

  5. [5]

    Data from lidc-idri [data set]

    SG Armato III, G McLennan, L Bidaut, MF McNitt-Gray, CR Meyer, AP Reeves, B Zhao, DR Aberle, CI Henschke, EA Hoffman, et al. Data from lidc-idri [data set]. the cancer imaging archive, 2015. 16

    2015

  6. [6]

    Masked siamese networks for label-efficient learning

    Mahmoud Assran, Mathilde Caron, Ishan Misra, Piotr Bo- janowski, Florian Bordes, Pascal Vincent, Armand Joulin, Mike Rabbat, and Nicolas Ballas. Masked siamese networks for label-efficient learning. InEuropean Conference on Com- puter Vision, pages 456–473. Springer, 2022. 1, 6

    2022

  7. [7]

    Self-supervised learning from images with a joint-embedding predictive architecture

    Mahmoud Assran, Quentin Duval, Ishan Misra, Piotr Bo- janowski, Pascal Vincent, Michael Rabbat, Yann LeCun, and Nicolas Ballas. Self-supervised learning from images with a joint-embedding predictive architecture. arXiv preprint arXiv:2301.08243, 2023. 6

    arXiv 2023

  8. [8]

    Sequential modeling enables scal- able learning for large vision models

    Yutong Bai, Xinyang Geng, Karttikeya Mangalam, Amir Bar, Alan Yuille, Trevor Darrell, Jitendra Malik, and Alexei A Efros. Sequential modeling enables scal- able learning for large vision models. arXiv preprint arXiv:2312.00785, 2023. 1, 2, 5, 6, 18

    arXiv 2023

Show all 98 references
  1. [9]

    Segmentation labels and radiomic features for the pre-operative scans of the tcga- gbm collection (2017)

    Spyridon Bakas, Hamed Akbari, Aristeidis Sotiras, Michel Bilello, Martin Rozycki, Justin Kirby, John Freymann, Key- van Farahani, and Christos Davatzikos. Segmentation labels and radiomic features for the pre-operative scans of the tcga- gbm collection (2017). DOI: https://doi...

    2017

  2. [10]

    Segmentation labels and radiomic features for the pre-operative scans of the tcga-lgg collection [data set]

    Spyridon Bakas, Hamed Akbari, Aristeidis Sotiras, Michel Bilello, Martin Rozycki, JS Kirby, JB Freymann, Keyvan Farahani, and Christos Davatzikos. Segmentation labels and radiomic features for the pre-operative scans of the tcga-lgg collection [data set]. the cancer imaging ar...

    2017

  3. [11]

    Advancing the cancer genome atlas glioma mri collections with expert seg- mentation labels and radiomic features

    Spyridon Bakas, Hamed Akbari, Aristeidis Sotiras, Michel Bilello, Martin Rozycki, Justin S Kirby, John B Freymann, Keyvan Farahani, and Christos Davatzikos. Advancing the cancer genome atlas glioma mri collections with expert seg- mentation labels and radiomic features. Scient...

    2017

  4. [12]

    Identifying the best machine learning algorithms for brain tumor segmentation, progression assessment, and over- all survival prediction in the brats challenge

    Spyridon Bakas, Mauricio Reyes, Andras Jakab, Stefan Bauer, Markus Rempfler, Alessandro Crimi, Russell Takeshi Shinohara, Christoph Berger, Sung Min Ha, Martin Rozycki, et al. Identifying the best machine learning algorithms for brain tumor segmentation, progression assessment...

    2018 arXiv

  5. [13]

    Visual prompting via image inpaint- ing

    Amir Bar, Yossi Gandelsman, Trevor Darrell, Amir Glober- son, and Alexei Efros. Visual prompting via image inpaint- ing. Advances in Neural Information Processing Systems , 35:25005–25017, 2022. 1, 2

    2022

  6. [14]

    Label-efficient se- mantic segmentation with diffusion models

    Dmitry Baranchuk, Andrey V oynov, Ivan Rubachev, Valentin Khrulkov, and Artem Babenko. Label-efficient se- mantic segmentation with diffusion models. In International Conference on Learning Representations, 2021. 1

    2021

  7. [15]

    xl- stm: Extended long short-term memory

    Maximilian Beck, Korbinian P ¨oppel, Markus Spanring, An- dreas Auer, Oleksandra Prudnikova, Michael Kopp, G ¨unter Klambauer, Johannes Brandstetter, and Sepp Hochreiter. xl- stm: Extended long short-term memory. arXiv preprint arXiv:2405.04517, 2024. 6

    2024 arXiv

  8. [16]

    Retrieval-augmented diffusion models, 2022

    Andreas Blattmann, Robin Rombach, Kaan Oktay, and Bj¨orn Ommer. Retrieval-augmented diffusion models, 2022. 14

    2022

  9. [17]

    Retouch: The retinal oct fluid detection and segmenta- tion benchmark and challenge

    Hrvoje Bogunovi ´c, Freerk Venhuizen, Sophie Klimscha, Ste- fanos Apostolopoulos, Alireza Bab-Hadiashar, Ulas Bagci, Mirza Faisal Beg, Loza Bekalo, Qiang Chen, Carlos Ciller, et al. Retouch: The retinal oct fluid detection and segmenta- tion benchmark and challenge. IEEE trans...

    2019

  10. [18]

    Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Sub- biah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakan- tan, Pranav Shyam, Girish Sastry, Amanda Askell, Sand- hini Agarwal, Ariel Herbert-V oss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler...

    2005 arXiv

  11. [19]

    Uni- verseg: Universal medical image segmentation

    Victor Ion Butoi, Jose Javier Gonzalez Ortiz, Tianyu Ma, Mert R Sabuncu, John Guttag, and Adrian V Dalca. Uni- verseg: Universal medical image segmentation. arXiv preprint arXiv:2304.06131, 2023. 2

    2023 arXiv

  12. [20]

    Lung segmentation in chest radiographs using anatomical atlases with nonrigid registration

    Sema Candemir, Stefan Jaeger, Kannappan Palaniappan, Jonathan P Musco, Rahul K Singh, Zhiyun Xue, Alexandros Karargyris, Sameer Antani, George Thoma, and Clement J McDonald. Lung segmentation in chest radiographs using anatomical atlases with nonrigid registration. IEEE transa...

    2013

  13. [21]

    Huiwen Chang, Han Zhang, Lu Jiang, Ce Liu, and William T. Freeman. Maskgit: Masked generative image transformer. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 2

    2022

  14. [22]

    Muse: Text-to-image generation via masked generative transform- ers

    Huiwen Chang, Han Zhang, Jarred Barber, AJ Maschinot, Jose Lezama, Lu Jiang, Ming-Hsuan Yang, Kevin Mur- phy, William T Freeman, Michael Rubinstein, et al. Muse: Text-to-image generation via masked generative transform- ers. arXiv preprint arXiv:2301.00704, 2023. 2

    2023 arXiv

  15. [23]

    Can ai help in screening viral and covid-19 pneumonia? Ieee Access, 8:132665–132676, 2020

    Muhammad EH Chowdhury, Tawsifur Rahman, Amith Khandakar, Rashid Mazhar, Muhammad Abdul Kadir, Zaid Bin Mahbub, Khandakar Reajul Islam, Muham- mad Salman Khan, Atif Iqbal, Nasser Al Emadi, et al. Can ai help in screening viral and covid-19 pneumonia? Ieee Access, 8:132665–13267...

    2020

  16. [24]

    Skin lesion analysis toward melanoma detection 2018: A challenge hosted by the interna- tional skin imaging collaboration (isic)

    Noel Codella, Veronica Rotemberg, Philipp Tschandl, M Emre Celebi, Stephen Dusza, David Gutman, Brian Helba, Aadi Kalloo, Konstantinos Liopyris, Michael Marchetti, et al. Skin lesion analysis toward melanoma detection 2018: A challenge hosted by the interna- tional skin imagin...

    2018 arXiv

  17. [25]

    Noel CF Codella, David Gutman, M Emre Celebi, Brian Helba, Michael A Marchetti, Stephen W Dusza, Aadi Kalloo, Konstantinos Liopyris, Nabin Mishra, Harald Kit- tler, et al. Skin lesion analysis toward melanoma detection: A challenge at the 2017 international symposium on biomed...

    2017

  18. [26]

    Neuralizer: General neuroimage analysis without re-training

    Steffen Czolbe and Adrian V Dalca. Neuralizer: General neuroimage analysis without re-training. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6217–6230, 2023. 2

    2023

  19. [27]

    Vision transformers need registers

    Timoth ´ee Darcet, Maxime Oquab, Julien Mairal, and Pi- otr Bojanowski. Vision transformers need registers. arXiv preprint arXiv:2309.16588, 2023. 6

    2023 arXiv

  20. [28]

    Covid-19 in- fection map generation and detection from chest x-ray im- ages

    Aysen Degerli, Mete Ahishali, Mehmet Yamac, Serkan Ki- ranyaz, Muhammad EH Chowdhury, Khalid Hameed, Tahir Hamid, Rashid Mazhar, and Moncef Gabbouj. Covid-19 in- fection map generation and detection from chest x-ray im- ages. Health information science and systems, 9(1):15, 2021. 16

    2021

  21. [29]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 7

    2009

  22. [31]

    BERT: pre-training of deep bidirectional trans- formers for language understanding.CoRR, abs/1810.04805,

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: pre-training of deep bidirectional trans- formers for language understanding.CoRR, abs/1810.04805,

    arXiv

  23. [32]

    Cross- moda 2021 challenge: Benchmark of cross-modality do- main adaptation techniques for vestibular schwannoma and cochlea segmentation

    Reuben Dorent, Aaron Kujawa, Marina Ivory, Spyridon Bakas, Nicola Rieke, Samuel Joutard, Ben Glocker, Jorge Cardoso, Marc Modat, Kayhan Batmanghelich, et al. Cross- moda 2021 challenge: Benchmark of cross-modality do- main adaptation techniques for vestibular schwannoma and co...

    2021

  24. [33]

    Taming transformers for high-resolution image synthesis

    Patrick Esser, Robin Rombach, and Bjorn Ommer. Taming transformers for high-resolution image synthesis. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12873–12883, 2021. 2, 3, 6, 7

    2021

  25. [34]

    An annotated test-retest collection of prostate mul- tiparametric mri

    Andriy Fedorov, Michael Schwier, David Clunie, Christian Herz, Steve Pieper, Ron Kikinis, Clare Tempany, and Fiona Fennessy. An annotated test-retest collection of prostate mul- tiparametric mri. Scientific data, 5(1):1–13, 2018. 16

    2018

  26. [35]

    Mamba: Linear-time sequence modeling with selective state spaces

    Albert Gu and Tri Dao. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752, 2023. 6

    2023 arXiv

  27. [36]

    Analogist: Out-of-the-box visual in-context learning with image diffusion model

    Zheng Gu, Shiyuan Yang, Jing Liao, Jing Huo, and Yang Gao. Analogist: Out-of-the-box visual in-context learning with image diffusion model. ACM Transactions on Graphics (TOG), 43(4):1–15, 2024. 2

    2024

  28. [37]

    Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning

    Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948, 2025. 8

    2025 arXiv

  29. [38]

    Data-efficient large vision models through sequential autoregression.arXiv preprint arXiv:2402.04841, 2024

    Jianyuan Guo, Zhiwei Hao, Chengcheng Wang, Yehui Tang, Han Wu, Han Hu, Kai Han, and Chang Xu. Data-efficient large vision models through sequential autoregression.arXiv preprint arXiv:2402.04841, 2024. 2, 5, 6

    2024 arXiv

  30. [39]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll´ar, and Ross Girshick. Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 16000– 16009, 2022. 6

    2022

  31. [40]

    Whole-cell organelle segmentation in volume electron microscopy

    Larissa Heinrich, Davis Bennett, David Ackerman, Woohyun Park, John Bogovic, Nils Eckstein, Alyson Petruncio, Jody Clements, Song Pang, C Shan Xu, et al. Whole-cell organelle segmentation in volume electron microscopy. Nature, 599(7883):141–146, 2021. 3, 4, 14

    2021

  32. [41]

    Isles 2022: A multi-center magnetic resonance imag- 10 ing stroke lesion segmentation dataset

    Moritz R Hernandez Petzsche, Ezequiel de la Rosa, Uta Hanning, Roland Wiest, Waldo Valenzuela, Mauricio Reyes, Maria Meyer, Sook-Lei Liew, Florian Kofler, Ivan Ezhov, et al. Isles 2022: A multi-center magnetic resonance imag- 10 ing stroke lesion segmentation dataset. Scientif...

    2022

  33. [42]

    Automatic lung segmentation in routine imaging is primarily a data di- versity problem, not a methodology problem

    Johannes Hofmanninger, Forian Prayer, Jeanny Pan, Sebas- tian R ¨ohrich, Helmut Prosch, and Georg Langs. Automatic lung segmentation in routine imaging is primarily a data di- versity problem, not a methodology problem. European ra- diology experimental, 4:1–13, 2020. 16

    2020

  34. [43]

    Cholecseg8k: a semantic segmentation dataset for laparoscopic cholecystectomy based on cholec80

    W-Y Hong, C-L Kao, Y-H Kuo, J-R Wang, W-L Chang, and C-S Shih. Cholecseg8k: a semantic segmentation dataset for laparoscopic cholecystectomy based on cholec80. arXiv preprint arXiv:2012.12453, 2020. 16

    2012 arXiv

  35. [44]

    Automatic tuberculosis screening using chest radio- graphs

    Stefan Jaeger, Alexandros Karargyris, Sema Candemir, Les Folio, Jenifer Siegelman, Fiona Callaghan, Zhiyun Xue, Kannappan Palaniappan, Rahul K Singh, Sameer Antani, et al. Automatic tuberculosis screening using chest radio- graphs. IEEE transactions on medical imaging , 33(2):...

    2013

  36. [45]

    Towards unifying anatomy segmentation: Automated generation of a full-body ct dataset

    Alexander Jaus, Constantin Seibold, Kelsey Hermann, Ne- gar Shahamiri, Alexandra Walter, Kristina Giske, Johannes Haubold, Jens Kleesiek, and Rainer Stiefelhagen. Towards unifying anatomy segmentation: Automated generation of a full-body ct dataset. In 2024 IEEE International ...

    2024

  37. [46]

    Kvasir-seg: A segmented polyp dataset

    Debesh Jha, Pia H Smedsrud, Michael A Riegler, P ˚al Halvorsen, Thomas De Lange, Dag Johansen, and H˚avard D Johansen. Kvasir-seg: A segmented polyp dataset. In In- ternational conference on multimedia modeling, pages 451–

  38. [47]

    Amos: A large-scale abdominal multi-organ benchmark for versatile medical image segmentation

    Yuanfeng Ji, Haotian Bai, Jie Yang, Chongjian Ge, Ye Zhu, Ruimao Zhang, Zhen Li, Lingyan Zhang, Wanling Ma, Xi- ang Wan, et al. Amos: A large-scale abdominal multi-organ benchmark for versatile medical image segmentation. arXiv preprint arXiv:2206.08023, 2022. 16

    2022 arXiv

  39. [48]

    Categorized digital database for low en- ergy and subtracted contrast enhanced spectral mammogra- phy images

    R Khaled et al. Categorized digital database for low en- ergy and subtracted contrast enhanced spectral mammogra- phy images. The Cancer Imaging Archive, 2021. 16

    2021

  40. [49]

    Standardized assessment of automatic segmentation of white matter hyperintensities and results of the wmh segmentation challenge

    Hugo J Kuijf, J Matthijs Biesbroek, Jeroen De Bresser, Rut- ger Heinen, Simon Andermatt, Mariana Bento, Matt Berseth, Mikhail Belyaev, M Jorge Cardoso, Adria Casamitjana, et al. Standardized assessment of automatic segmentation of white matter hyperintensities and results of t...

    2019

  41. [50]

    Decoupled con- text processing for context augmented language modeling

    Zonglin Li, Ruiqi Guo, and Sanjiv Kumar. Decoupled con- text processing for context augmented language modeling. Advances in Neural Information Processing Systems , 35: 21698–21710, 2022. 1

    2022

  42. [51]

    Unified-io: A unified model for vision, language, and multi-modal tasks

    Jiasen Lu, Christopher Clark, Rowan Zellers, Roozbeh Mot- taghi, and Aniruddha Kembhavi. Unified-io: A unified model for vision, language, and multi-modal tasks. arXiv preprint arXiv:2206.08916, 2022. 2

    2022 arXiv

  43. [52]

    Toward data-efficient learning: A benchmark for covid- 19 ct lung and infection segmentation

    Jun Ma, Yixin Wang, Xingle An, Cheng Ge, Ziqi Yu, Jianan Chen, Qiongjie Zhu, Guoqiang Dong, Jian He, Zhiqiang He, et al. Toward data-efficient learning: A benchmark for covid- 19 ct lung and infection segmentation. Medical physics, 48 (3):1197–1210, 2021. 16

    2021

  44. [53]

    Fast and low-gpu-memory abdomen ct organ seg- mentation: the flare challenge

    Jun Ma, Yao Zhang, Song Gu, Xingle An, Zhihe Wang, Cheng Ge, Congcong Wang, Fan Zhang, Yu Wang, Yinan Xu, et al. Fast and low-gpu-memory abdomen ct organ seg- mentation: the flare challenge. Medical Image Analysis, 82: 102616, 2022. 16

    2022

  45. [54]

    Abdomenct-1k: Is abdominal organ segmentation a solved problem? IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(10):6695–6714, 2022

    Jun Ma, Yao Zhang, Song Gu, Cheng Zhu, Cheng Ge, Yichi Zhang, Xingle An, Congcong Wang, Qiyuan Wang, Xin Liu, Shucheng Cao, Qi Zhang, Shangqing Liu, Yunpeng Wang, Yuhui Li, Jian He, and Xiaoping Yang. Abdomenct-1k: Is abdominal organ segmentation a solved problem? IEEE Transac...

    2022

  46. [55]

    Segment anything in medical images

    Jun Ma, Yuting He, Feifei Li, Lin Han, Chenyu You, and Bo Wang. Segment anything in medical images. Nature Communications, 15(1):654, 2024. 5, 6, 14, 15

    2024

  47. [56]

    m2caiseg: Semantic segmentation of laparoscopic images using convolutional neural networks

    Salman Maqbool, Aqsa Riaz, Hasan Sajid, and Osman Hasan. m2caiseg: Semantic segmentation of laparoscopic images using convolutional neural networks. arXiv preprint arXiv:2008.10134, 2020. 16

    2008 arXiv

  48. [57]

    Deep interactive segmentation of medical images: A systematic review and taxonomy

    Zdravko Marinov, Paul F J ¨ager, Jan Egger, Jens Kleesiek, and Rainer Stiefelhagen. Deep interactive segmentation of medical images: A systematic review and taxonomy. IEEE transactions on pattern analysis and machine intelligence ,

  49. [58]

    Martin, C

    D. Martin, C. Fowlkes, D. Tal, and J. Malik. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecologi- cal statistics. In Proc. 8th Int’l Conf. Computer Vision, pages 416–423, 2001. 3, 14

    2001

  50. [59]

    Data structures for statistical comput- ing in python

    Wes McKinney et al. Data structures for statistical comput- ing in python. In Proceedings of the 9th Python in Science Conference, pages 51–56. Austin, TX, 2010. 20

    2010

  51. [60]

    Factors of influence for transfer learning across diverse appearance domains and task types

    Thomas Mensink, Jasper Uijlings, Alina Kuznetsova, Michael Gygli, and Vittorio Ferrari. Factors of influence for transfer learning across diverse appearance domains and task types. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(12):9298–9314, 2021. 1

    2021

  52. [61]

    The multimodal brain tumor image segmentation benchmark (brats)

    Bjoern H Menze, Andras Jakab, Stefan Bauer, Jayashree Kalpathy-Cramer, Keyvan Farahani, Justin Kirby, Yuliya Burren, Nicole Porz, Johannes Slotboom, Roland Wiest, et al. The multimodal brain tumor image segmentation benchmark (brats). IEEE transactions on medical imaging , 34(...

    1993

  53. [62]

    Supervised transfer learning at scale for medical imaging

    Basil Mustafa, Aaron Loh, Jan Freyberg, Patricia MacWilliams, Megan Wilson, Scott Mayer McKinney, Marcin Sieniek, Jim Winkens, Yuan Liu, Peggy Bui, et al. Supervised transfer learning at scale for medical imaging. arXiv preprint arXiv:2101.05913, 2021. 1

    2021 arXiv

  54. [63]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zem- ing Lin, Natalia Gimelshein, Luca Antiga, Alban Desmai- son, Andreas Kopf, Edward Yang, Zachary DeVito, Mar- tin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steine...

    2019

  55. [64]

    Pedregosa, G

    F. Pedregosa, G. Varoquaux, A. Gramfort, V . Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V . Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duchesnay. Scikit-learn: Machine learning in Python. Journal of Machine Learni...

    2011

  56. [65]

    Kvasir: A multi- class image dataset for computer aided gastrointestinal dis- ease detection

    Konstantin Pogorelov, Kristin Ranheim Randel, Carsten Gri- wodz, Sigrun Losada Eskeland, Thomas de Lange, Dag Johansen, Concetto Spampinato, Duc-Tien Dang-Nguyen, Mathias Lux, Peter Thelin Schmidt, et al. Kvasir: A multi- class image dataset for computer aided gastrointestinal...

    2017

  57. [66]

    Language models are unsu- pervised multitask learners

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsu- pervised multitask learners. OpenAI blog, 1(8):9, 2019. 2, 6, 7

    2019

  58. [67]

    Exploring the effect of image enhancement techniques on covid-19 detection using chest x-ray images.Computers in biology and medicine, 132: 104319, 2021

    Tawsifur Rahman, Amith Khandakar, Yazan Qiblawey, Anas Tahir, Serkan Kiranyaz, Saad Bin Abul Kashem, Moham- mad Tariqul Islam, Somaya Al Maadeed, Susu M Zughaier, Muhammad Salman Khan, et al. Exploring the effect of image enhancement techniques on covid-19 detection using ches...

    2021

  59. [68]

    Tyche: Stochastic in-context learning for medical image segmenta- tion

    Marianne Rakic, Hallee E Wong, Jose Javier Gonzalez Ortiz, Beth A Cimini, John V Guttag, and Adrian V Dalca. Tyche: Stochastic in-context learning for medical image segmenta- tion. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 1...

  60. [69]

    Every annotation counts: Multi-label deep supervision for medical image segmen- tation

    Simon Reiß, Constantin Seibold, Alexander Freytag, Erik Rodner, and Rainer Stiefelhagen. Every annotation counts: Multi-label deep supervision for medical image segmen- tation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9532–9542,

  61. [70]

    Graph-constrained con- trastive regularization for semi-weakly volumetric segmen- tation

    Simon Reiß, Constantin Seibold, Alexander Freytag, Erik Rodner, and Rainer Stiefelhagen. Graph-constrained con- trastive regularization for semi-weakly volumetric segmen- tation. In European Conference on Computer Vision, pages 401–419. Springer, 2022. 1

    2022

  62. [71]

    Decoupled semantic proto- types enable learning from diverse annotation types for semi- weakly segmentation in expert-driven domains

    Simon Reiß, Constantin Seibold, Alexander Freytag, Erik Rodner, and Rainer Stiefelhagen. Decoupled semantic proto- types enable learning from diverse annotation types for semi- weakly segmentation in expert-driven domains. In Proceed- ings of the IEEE/CVF Conference on Compute...

    2023

  63. [72]

    Medi- cal vision generalist: Unifying medical imaging tasks in con- text

    Sucheng Ren, Xiaoke Huang, Xianhang Li, Junfei Xiao, Jieru Mei, Zeyu Wang, Alan Yuille, and Yuyin Zhou. Medi- cal vision generalist: Unifying medical imaging tasks in con- text. arXiv preprint arXiv:2406.05565, 2024. 2

    2024 arXiv

  64. [73]

    Ct-org, a new dataset for mul- tiple organ segmentation in computed tomography.Scientific Data, 7(1):381, 2020

    Blaine Rister, Darvin Yi, Kaushik Shivakumar, Tomomi Nobashi, and Daniel L Rubin. Ct-org, a new dataset for mul- tiple organ segmentation in computed tomography.Scientific Data, 7(1):381, 2020. 16

    2020

  65. [74]

    High-resolution image syn- thesis with latent diffusion models, 2021

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨orn Ommer. High-resolution image syn- thesis with latent diffusion models, 2021. 3, 14

    2021

  66. [75]

    Rapid artificial intelli- gence solutions in a pandemic—the covid-19-20 lung ct le- sion segmentation challenge

    Holger R Roth, Ziyue Xu, Carlos Tor-D ´ıez, Ramon Sanchez Jacob, Jonathan Zember, Jose Molto, Wenqi Li, Sheng Xu, Baris Turkbey, Evrim Turkbey, et al. Rapid artificial intelli- gence solutions in a pandemic—the covid-19-20 lung ct le- sion segmentation challenge. Medical image...

    2022

  67. [76]

    Distilbert, a distilled version of BERT: smaller, faster, cheaper and lighter

    Victor Sanh, Lysandre Debut, Julien Chaumond, and Thomas Wolf. Distilbert, a distilled version of BERT: smaller, faster, cheaper and lighter. CoRR, abs/1910.01108,

    1910 arXiv

  68. [77]

    Amp: Adaptive masked proxies for few-shot segmentation

    Mennatullah Siam, Boris N Oreshkin, and Martin Jagersand. Amp: Adaptive masked proxies for few-shot segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 5249–5258, 2019. 1

    2019

  69. [78]

    A large annotated medical image dataset for the development and evaluation of segmentation algo- rithms

    Amber L Simpson, Michela Antonelli, Spyridon Bakas, Michel Bilello, Keyvan Farahani, Bram Van Ginneken, An- nette Kopp-Schneider, Bennett A Landman, Geert Litjens, Bjoern Menze, et al. A large annotated medical image dataset for the development and evaluation of segmentation a...

    1902 arXiv

  70. [79]

    Exploring effective factors for improving visual in-context learning

    Yanpeng Sun, Qiang Chen, Jian Wang, Jingdong Wang, and Zechao Li. Exploring effective factors for improving visual in-context learning. arXiv preprint arXiv:2304.04748, 2023. 2, 4

    2023

  71. [80]

    Learning to compare: Re- lation network for few-shot learning

    Flood Sung, Yongxin Yang, Li Zhang, Tao Xiang, Philip HS Torr, and Timothy M Hospedales. Learning to compare: Re- lation network for few-shot learning. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 1199–1208, 2018. 1

    2018

  72. [81]

    Covid-19 infection localization and severity grading from chest x-ray images

    Anas M Tahir, Muhammad EH Chowdhury, Amith Khan- dakar, Tawsifur Rahman, Yazan Qiblawey, Uzair Khurshid, Serkan Kiranyaz, Nabil Ibtehaz, M Sohel Rahman, Somaya Al-Maadeed, et al. Covid-19 infection localization and severity grading from chest x-ray images. Computers in bi- olo...

    2021

  73. [82]

    Tahir, Muhammad E

    Anas M. Tahir, Muhammad E. H. Chowdhury, Yazan Qi- blawey, Amith Khandakar, Tawsifur Rahman, Serkan Ki- ranyaz, Uzair Khurshid, Nabil Ibtehaz, Sakib Mahmud, and Maymouna Ezeddin. Covid-qu-ex dataset, 2022. 16

    2022

  74. [83]

    The ham10000 dataset, a large collection of multi-source der- matoscopic images of common pigmented skin lesions

    Philipp Tschandl, Cliff Rosendahl, and Harald Kittler. The ham10000 dataset, a large collection of multi-source der- matoscopic images of common pigmented skin lesions. Sci- entific data, 5(1):1–9, 2018. 3, 4, 14, 16

    2018

  75. [84]

    En- donet: a deep architecture for recognition tasks on laparo- scopic videos

    Andru P Twinanda, Sherif Shehata, Didier Mutter, Jacques Marescaux, Michel De Mathelin, and Nicolas Padoy. En- donet: a deep architecture for recognition tasks on laparo- scopic videos. IEEE transactions on medical imaging , 36 (1):86–97, 2016. 16

    2016

  76. [85]

    Automated measurement of fetal head circumference using 2d ultrasound images

    Thomas LA van den Heuvel, Dagmar de Bruijn, Chris L de Korte, and Bram van Ginneken. Automated measurement of fetal head circumference using 2d ultrasound images. PloS one, 13(8):e0200412, 2018. 16

    2018

  77. [86]

    Neural discrete representation learning

    Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. Advances in neural information pro- cessing systems, 30, 2017. 2, 3

    2017

  78. [87]

    Gomez, Lukasz Kaiser, and Illia 12 Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia 12 Polosukhin. Attention is all you need. In Advances in Neu- ral Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2...

    2017

  79. [88]

    Images speak in images: A generalist painter for in-context visual learning

    Xinlong Wang, Wen Wang, Yue Cao, Chunhua Shen, and Tiejun Huang. Images speak in images: A generalist painter for in-context visual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6830–6839, 2023. 1, 2

    2023

  80. [89]

    Seggpt: Segmenting ev- erything in context

    Xinlong Wang, Xiaosong Zhang, Yue Cao, Wen Wang, Chunhua Shen, and Tiejun Huang. Seggpt: Segmenting ev- erything in context. arXiv preprint arXiv:2304.03284, 2023. 2

    2023 arXiv

  81. [90]

    To- talsegmentator: robust segmentation of 104 anatomic struc- tures in ct images

    Jakob Wasserthal, Hanns-Christian Breit, Manfred T Meyer, Maurice Pradella, Daniel Hinck, Alexander W Sauter, Tobias Heye, Daniel T Boll, Joshy Cyriac, Shan Yang, et al. To- talsegmentator: robust segmentation of 104 anatomic struc- tures in ct images. Radiology: Artificial In...

    2023

  82. [91]

    Thinking llms: General instruction following with thought generation, 2024a

    Tianhao Wu, Janice Lan, Weizhe Yuan, Jiantao Jiao, Jason Weston, and Sainbayar Sukhbaatar. Thinking llms: General instruction following with thought generation, 2024a. URL https://arxiv. org/abs/2410.10630, 2023. 8

    2023 arXiv

  83. [92]

    Feature generating networks for zero-shot learning

    Yongqin Xian, Tobias Lorenz, Bernt Schiele, and Zeynep Akata. Feature generating networks for zero-shot learning. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5542–5551, 2018. 1

    2018

  84. [93]

    J. Yang, G. Sharp, H. Veeraraghavan, W. Van Elmpt, A. Dekker, T. Lustberg, and M. Gooding. Data from lung ct segmentation challenge (lctsc) (version 3) [data set]., 2017. 16

    2017

  85. [94]

    Imagebrush: Learning visual in-context instructions for exemplar-based image ma- nipulation

    Yifan Yang, Houwen Peng, Yifei Shen, Yuqing Yang, Han Hu, Lili Qiu, Hideki Koike, et al. Imagebrush: Learning visual in-context instructions for exemplar-based image ma- nipulation. Advances in Neural Information Processing Sys- tems, 36, 2024. 2

    2024

  86. [95]

    An image is worth 32 tokens for reconstruction and generation

    Qihang Yu, Mark Weber, Xueqing Deng, Xiaohui Shen, Daniel Cremers, and Liang-Chieh Chen. An image is worth 32 tokens for reconstruction and generation. arXiv preprint arXiv:2406.07550, 2024. 2

    2024 arXiv

  87. [96]

    Siim-acr pneumothorax segmen- tation

    Anna Zawacki, Carol Wu, George Shih, Julia Elliott, Mikhail Fomitchev, Mohannad Hussain, ParasLakhani, Phil Culli- ton, and Shunxing Bao. Siim-acr pneumothorax segmen- tation. https : / / kaggle . com / competitions / siim- acr- pneumothorax- segmentation, 2019. Kaggle. 16

    2019

  88. [97]

    Video in-context learning

    Wentao Zhang, Junliang Guo, Tianyu He, Li Zhao, Linli Xu, and Jiang Bian. Video in-context learning. arXiv preprint arXiv:2407.07356, 2024. 2

    2024 arXiv

  89. [98]

    What makes good examples for visual in-context learning? Advances in Neural Information Processing Systems, 36:17773–17794,

    Yuanhan Zhang, Kaiyang Zhou, and Ziwei Liu. What makes good examples for visual in-context learning? Advances in Neural Information Processing Systems, 36:17773–17794,

  90. [99]

    Robust detection and segmentation for diagnosis of vertebral diseases using routine mr images

    D ˇzenan Zuki ´c, Ale ˇs Vlas ´ak, Jan Egger, Daniel Ho ˇr´ınek, Christopher Nimsky, and Andreas Kolb. Robust detection and segmentation for diagnosis of vertebral diseases using routine mr images. In Computer Graphics Forum , pages 190–204. Wiley Online Library, 2014. 16 13 g...

    2014

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.