Semantic Segmentation of iPS Cells: Case Study on Model Complexity in Biomedical Imaging

Bisser Raytchev; Maoquan Zhang; Xiujuan Sun

REVIEW 5 major objections 5 minor 22 references

Semantic Segmentation of iPS Cells: Case Study on Model Complexity in Biomedical Imaging

T0 review · 5 major / 5 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 For iPS cell colony segmentation, a carefully configured DeepLabv3 outperforms larger foundation models like SAM2 and MedSAM2, indicating that model scale alone does not decide accuracy on subtle, low-contrast boundaries.

desk verdict A plausible but under-specified comparison: DeepLabv3 beats SAM2/MedSAM2 on iPS segmentation, but missing baselines and adaptation details undercut the claim until revised. read the letter →

arxiv 2507.21608 v1 pith:T5NBTP2H submitted 2025-07-29 cs.CV

Maoquan Zhang , Bisser Raytchev , Xiujuan Sun This is my paper

classification cs.CV

keywords semanticsegmentationiPScellsstemcellcoloniesDeepLabv3SAM2Medfoundationmodelsphase-contrastimaging

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

The paper sets out to show that for segmenting induced pluripotent stem (iPS) cell colonies in phase-contrast microscopy images, a carefully configured DeepLabv3 model achieves higher accuracy and lower computational cost than large pretrained foundation models like SAM2 and MedSAM2. Under the paper's controlled conditions, DeepLabv3 reaches $97.5 \pm 2.21\%$ mean IoU in 50 training epochs, while SAM2 and MedSAM2 plateau at $81.0 \pm 1.56\%$ and $63.5 \pm 2.17\%$ after 1,000 epochs, with higher GPU memory usage. The authors interpret this as evidence that increased model complexity does not automatically buy better segmentation, at least when the target task has faint, ambiguous boundaries and a small domain-specific dataset. The wider point is that task-adapted, mid-size networks can be a practical alternative to foundation-model-scale architectures in specialized biomedical imaging.

What carries the argument

The load-bearing mechanism is Atrous Spatial Pyramid Pooling (ASPP), a module in DeepLabv3 that runs several dilated convolutions with different dilation rates in parallel and fuses their outputs. Each atrous layer widens the effective receptive field to $(k-1)r+1$ pixels for kernel size $k$ and rate $r$, so a $3\times3$ kernel at rate 2 covers a $5\times5$ region without adding weights. This gives the network both global colony context and fine edge detail, which is exactly what faint, ambiguous phase-contrast boundaries need. The paper contrasts this with SAM2's $O(dN^2)$ multi-head self-attention, whose cost grows with the number of tokens and is motivated by temporal video memory rather than single-frame static images.

What would settle it

A re-run of the comparison on the same 590-patch test set with per-model hyperparameter sweeps for SAM2 and MedSAM2, including learning rate, weight decay, optimizer, and a prompt-to-dense-mask adapter appropriate to their design, would settle the claim: if either foundation model reaches or exceeds $97.5\%$ IoU after that tuning, the gap is a protocol artifact rather than an architectural limitation.

Watch

Extended reading notes

Core claim

The central claim is that, with the compared architectures left structurally unmodified, a properly configured DeepLabv3 achieves the best iPS colony segmentation among the models tested under the paper's fixed protocol. With a ResNet-50 backbone of about 42 million parameters, DeepLabv3 reaches a mean IoU of $97.5 \pm 2.21\%$ on a standardized subset of 590 patch images of size $1024 \times 1024$, compared with $81.0 \pm 1.56\%$ for SAM2 and $63.5 \pm 2.17\%$ for MedSAM2, while using roughly one twentieth of the training epochs and less GPU memory. The paper argues that the reason is architectural fit: DeepLabv3's atrous convolutions and Atrous Spatial Pyramid Pooling expand the receptive field without adding parameters, matching the multi-scale, low-contrast structure of colony edges, whereas SAM2's memory bank and quadratic self-attention are built for streaming video and add overhead without benefit on static images. MedSAM2's lower accuracy is attributed to fine-tuning on unrelated medical modalities, which misaligns its feature space for iPS data.

Load-bearing premise

The central claim stands on the assumption that SAM2 and MedSAM2 were not handicapped by the shared, fixed training protocol through optimizer or learning-rate choices that suit a CNN better than a Transformer, or through an awkward conversion of prompt-based models to dense masks; if that protocol favored DeepLabv3, the performance gap is not a fair test of architecture.

Editorial extensions

If this is right

For single-frame biomedical images with low-contrast boundaries, a mid-size CNN with dilated multi-scale pooling can match or exceed much larger foundation models, so scale alone is not the deciding factor in domain-specific segmentation.
Video-oriented components such as memory banks and temporal self-attention add training time and GPU use without improving static-image accuracy; the paper shows this cost concretely in the SAM2 comparison.
A small dataset of 60 high-resolution images, yielding 590 patches, is enough to train a high-performing specialized segmenter when the architecture and data preparation are matched to the task.
Pretraining on unrelated medical modalities can hurt rather than help: MedSAM2's fine-tuning for other imaging types is associated with a lower IoU than the general SAM2, warning that domain adaptation must be verified on the target domain.

Reading between the lines

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

The paper's Section 2.1 states identical optimizer configuration, input normalization, and evaluation metrics, but does not report learning rates, weight decay, or how the prompt-based SAM2 and MedSAM2 were converted to dense semantic segmentation; if those choices favored DeepLabv3, part of the measured gap could be implementation rather than architecture.
The same controlled comparison could be run on other phase-contrast colony datasets, such as different cell lines, magnifications, or dish densities; a persistent DeepLabv3 margin there would confirm that boundary fidelity, not dataset size, is what drives the result.
A cheap screening rule follows from this study: before spending GPU hours fine-tuning a video-oriented foundation model on a static, low-contrast task, try a tuned dilated CNN baseline with a multi-scale pooling head.

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.

Desk Editor's Note

read the letter

Quick take: this is a useful case study that reports a specific, checkable result—DeepLabv3 with 42M parameters gets 97.5 IoU on iPS colony segmentation while SAM2 gets 81.0 and MedSAM2 63.5 under the same hardware and batch size. The comparison is new and the code is promised, so the core number could be independently verified.

What's genuinely good: the paper tries to quantify compute and memory as well as accuracy, and the idea that a small specialized CNN beats generalist foundation models on low-contrast biomedical boundaries is worth taking seriously. The self-citation to Iwamoto et al. is just a baseline and doesn't bother me.

The soft spots are real. The abstract and intro say DeepLabv3 outperforms TransUNet, nnU-Net, and BioSAM2 too, but Table 1 only reports SAM2 and MedSAM2. Those results don't appear anywhere. That's not a minor omission; it's a claim without data. Second, the paper says the comparison used 'identical optimizer configuration' but never says what the optimizer was, what learning rate or weight decay, loss function, or how the prompt-based SAM2 was turned into a dense semantic segmentation predictor. That matters because SAM2 is interactive; a naive grid-prompt adaptation could handicap it. And there's a contradiction: Section 2.4.2 says SAM2 works well 'even without task-specific fine-tuning,' yet Table 1 shows 1,000 training epochs. Those can't both be true as stated.

The architecture discussion (atrous convolutions expand receptive field, SAM2's memory is unnecessary for static images) is plausible but speculative without ablations. I'd like to see an ASPP ablation or at least a DeepLabv3 variant without dilated convolutions.

On balance, the headline result might well survive a rigorous check—the margin is large—but the paper currently doesn't give the reader enough to trust that SAM2 and MedSAM2 weren't handicapped. It needs the missing baselines, full hyperparameters and adaptation details, and a rewording of the fine-tuning contradiction. As it stands I'd send it to review rather than desk reject, because the question is worthwhile and the result is verifiable. But I'd expect major revision.

Referee Report

5 major / 5 minor

Summary. The manuscript reports a comparative study of semantic segmentation models for induced pluripotent stem (iPS) cell colonies in phase-contrast images. Using a dataset of 60 high-resolution images split into 590 patches of size 1024×1024, the authors compare a DeepLabv3 model with SAM2 and MedSAM2, reporting that DeepLabv3 achieves substantially higher IoU (97.5±2.21%) than SAM2 (81.0±1.56%) and MedSAM2 (63.5±2.17%), while using less GPU memory and fewer training epochs. The paper interprets these results as evidence that a carefully configured, moderate-sized CNN can outperform large foundation models for specialized biomedical segmentation, and it argues that atrous convolutions and multi-scale pooling are responsible for the advantage. The introduction also claims superiority over TransUNet, nnU-Net, and BioSAM2, but no results for these models appear in the paper. An open-source implementation is promised.

Significance. If the comparison were fully specified and the claimed results reproduced, the paper would provide a useful case study on model complexity in biomedical imaging, with practical implications for choosing task-adapted architectures over generic foundation models on small, domain-specific datasets. The paper also makes a falsifiable empirical claim and provides an open-source toolkit, which are strengths. However, the central comparison currently rests on an underspecified adaptation protocol for SAM2/MedSAM2 and on unsupported performance claims for several named baselines, so the significance cannot be fully assessed as written.

major comments (5)

[Section 1 and Section 2] The introduction states that the proposed model 'outperforms both established CNNs (TransUNet, nnU-Net) and recent foundation models (SAM2, BioSAM2)', but Table 1 and the rest of the paper report results only for SAM2, MedSAM2, and DeepLabv3. No IoU or other metrics are given for TransUNet, nnU-Net, or BioSAM2. These claims must either be supported with experimental results or removed from the claims.
[Section 2.1 and Table 1] The evaluation protocol for adapting SAM2 and MedSAM2 to dense semantic segmentation is not specified. Since SAM2 and MedSAM2 are prompt-based, interactive segmentation models, the paper must describe how they were converted into dense multi-class predictors: what prompt-generation strategy was used, how the four classes (Good, Bad, medium/background, uncertain) were assigned, what loss function was optimized, and what learning rate, weight decay, scheduler, and optimizer settings were used. The statement 'identical training dataset size, optimizer configuration, input normalization, and evaluation metrics' is asserted but not documented. Without this information, the reported IoU gap could reflect a mismatched adaptation rather than an architectural limitation, which is load-bearing for the central claim.
[Section 2.4.2 and Table 1] There is a direct contradiction between the text and the training protocol. Section 2.4.2 says SAM2 performs well 'even without task-specific fine-tuning', while Table 1 reports that SAM2 and MedSAM2 were trained for 1,000 epochs. If the models were fine-tuned, 'without task-specific fine-tuning' is inaccurate; if they were not fine-tuned, the 1,000-epoch row in Table 1 and the comparison with DeepLabv3's 50 epochs are misleading. The manuscript must clarify whether and how each foundation model was fine-tuned on the iPS training set.
[Section 2.2 and Table 1] The dataset and evaluation description is too sparse to assess the reported differences. The paper states that 60 images were split at the image level with a 6:3:1 ratio, but it does not report the number of training, validation, and test images, the class distribution, the annotation protocol, the preprocessing or augmentation steps, or whether evaluation is performed on individual patches or reconstructed full images. In addition, the mean±std over five runs is reported without any statistical significance test or overlap analysis, so it is unclear whether the 16-point IoU gap between DeepLabv3 and SAM2 is robust across the five runs.
[Section 2.3.1] The claim that DeepLabv3's ASPP and atrous convolutions are 'particularly effective' and 'responsible for its superior performance' is not supported by any ablation study. The paper presents only an architectural rationale, not a controlled test that isolates the contribution of dilated convolutions from other differences such as backbone size, training length, or optimizer. Since this causal interpretation is part of the paper's stated contribution, it should either be backed by an ablation or softened to a hypothesis.

minor comments (5)

[Figure captions] The color legend is inconsistent between Figures 1 and 2: Figure 1 says red indicates 'Good', green 'Bad', blue medium, and pink uncertain, while Figure 2 says green indicates 'Good', pink uncertain, and blue background. Please align the captions and the visual legend.
[Table 1] The row 'Epochs' is described as 'number of epochs to reach peak accuracy', but the text says DeepLabv3 'converges in 50 epochs'. Clarify whether 50 is a stopping criterion, a convergence point, or the maximum epochs used, and state the stopping rule for SAM2 and MedSAM2.
[Section 2.2] The statement that SAM2 produced comparable IoU for both patch sizes while MedSAM2 was stable only at 1024×1024 is not accompanied by any numerical results; either provide these numbers or omit the observation.
[References] Several citations are not appropriate for the claims they support. Reference [12] is a Towards Data Science blog post used for the ResNet-50 parameter count; cite the original ResNet paper instead. References [17] and [19] are also blog posts and should be replaced with peer-reviewed sources where possible.
[Reference [22]] The author name 'Z. Maoquan' appears to be formatted incorrectly; the corresponding GitHub author should be listed as M. Zhang or as in the author list of the paper.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: this is an empirical benchmark study, and its only author-overlapping citation ([5], Iwamoto et al.) is a non-load-bearing historical baseline. The central claim rests on the directly measured Table 1 comparison, not on a fit-then-predict quantity, a self-citation chain, or an imported uniqueness theorem.

full rationale

The paper makes empirical comparison claims rather than analytic derivations, so the main circularity patterns do not apply. (1) No self-definitional step: 'Acc(IOU)' in Table 1 is a measured outcome on a held-out test subset; no equation defines DeepLabv3's performance in terms of the conclusion, and the reported IoU values (97.5, 81.0, 63.5) are outputs of training runs, not parameters fitted from the data and renamed as predictions. (2) No fitted-input-called-prediction: the only parameter-like choices (patch size, epochs, optimizer) are reported as configuration, and the IoU is not a transformed version of those inputs. (3) The only self-citation is reference [5] (Iwamoto et al., co-authored by B. Raytchev), used as a prior baseline ('IoU = 0.797 [5]') and as motivation for uncertainty-aware methods; the paper's conclusion does not depend on [5] being correct, so the self-citation is not load-bearing. (4) No uniqueness theorem or ansatz is imported from the authors' prior work to force the DeepLabv3 choice; the ASPP/atrous discussion in Sections 2.3-2.4 is a post-hoc interpretation of measured results, and the outcome was not imposed by that interpretation. (5) No known result is renamed as an organization. Two genuine non-circular risks remain and belong to correctness rather than circularity: (a) the paper never specifies how prompt-based SAM2/MedSAM2 were converted to dense multi-class semantic segmentation (no prompt-generation rule, class-assignment rule, loss, or learning-rate/weight-decay details), so the reported gap may partly reflect a protocol mismatch rather than an architectural limitation; and (b) Section 2.4.2 states SAM2 performs well 'even without task-specific fine-tuning' although Table 1 lists 1,000 training epochs, an internal inconsistency. Neither risk involves the derivation reducing to its own inputs. Per the rubric, this is a normal empirical study with only a minor non-load-bearing self-citation: score 1, no circular steps identified.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the experimental setup (dataset, patch size, training protocols) and on the assumption that the comparison is fair. No new theoretical entities are introduced. The architectural explanation is a post-hoc hypothesis, not an invariant.

free parameters (4)

Training epochs = 50 for DeepLabv3, 1000 for SAM2/MedSAM2
Number of epochs to reach reported peak IoU, chosen by the authors; this large difference may affect the fairness of the comparison and is not justified in depth.
Patch size = 1024x1024
All Table 1 results use this patch size; MedSAM2 was unstable at 512x512, so patch size selection is a choice that influences results.
Batch size = 2
Fixed batch size for all models, limited by GPU memory; the effect of batch size on convergence and final accuracy is not analyzed.
Optimizer configuration = Not reported
The paper claims an identical optimizer setup but does not specify learning rate, momentum, weight decay, or schedule, which can advantage or disadvantage particular architectures.

assumptions (3)

domain assumption The dataset of 60 images, split into 590 patches, is representative of iPS colony segmentation tasks.
The paper generalizes to conclusions about model complexity from this small dataset; if the images are not representative, the comparison may not transfer.
domain assumption The evaluation protocol is fair to all models
The paper states all models used identical training dataset, optimizer, normalization, and evaluation metrics, but does not account for fundamental differences in training requirements between architectures; the fairness is asserted, not demonstrated.
ad hoc to paper DeepLabv3's ASPP and atrous convolutions are responsible for its superior performance
Sections 2.3.1 and 2.4.1 attribute the performance gain to these components without ablation studies; this is an untested explanation rather than an established fact.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Semantic Segmentation of iPS Cells: Case Study on Model Complexity in Biomedical Imaging." pith.science (2026). https://pith.science/paper/T5NBTP2H

@misc{pith2026250721608,
  author       = {Pith},
  title        = {Pith review of: Semantic Segmentation of iPS Cells: Case Study on Model Complexity in Biomedical Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T5NBTP2H}},
  note         = {Machine review of arXiv:2507.21608}
}

read the original abstract

Medical image segmentation requires not only accuracy but also robustness under challenging imaging conditions. In this study, we show that a carefully configured DeepLabv3 model can achieve high performance in segmenting induced pluripotent stem (iPS) cell colonies, and, under our experimental conditions, outperforms large-scale foundation models such as SAM2 and its medical variant MedSAM2 without structural modifications. These results suggest that, for specialized tasks characterized by subtle, low-contrast boundaries, increased model complexity does not necessarily translate to better performance. Our work revisits the assumption that ever-larger and more generalized architectures are always preferable, and provides evidence that appropriately adapted, simpler models may offer strong accuracy and practical reliability in domain-specific biomedical applications. We also offer an open-source implementation that includes strategies for small datasets and domain-specific encoding, with the aim of supporting further advances in semantic segmentation for regenerative medicine and related fields.

Figures

Figures reproduced from arXiv: 2507.21608 by the authors.

**Figure 1.** (a) A representative raw iPS image from our test set. (b) A foundation model’s segmentation mask overlaid on the same image. (c) Our specialized DeepLabv3 output, showing markedly improved boundary fidelity in circled regions. In these overlays, red indicates “Good” colonies, green “Bad” colonies, blue culture medium, and pink denotes uncertain areas. example, their Dice scores may decline from approximately 0.84 … view at source ↗

**Figure 2.** Visual comparison of (a) input iPS image, (b) ground truth segmentation, and (c) DeepLabv3 prediction. Green indicates “Good” colonies, pink marks uncertain regions, and blue denotes background. The prediction (c) shows close alignment with the ground truth (b), including regions with subtle or ambiguous boundaries. MedSAM2, and DeepLabv3 (DLV3): BatchSize (minibatch size per iteration), T(s)/Epoch (average second… view at source ↗

**Figure 3.** Core modules of (a) DeepLabv3 [9] and (b) SAM2 [7]. DeepLabv3 leverages spatial pyramid pooling and atrous convolutions for efficient feature extraction, while SAM2 incorporates a memory bank and video-specific modules optimized for temporal segmentation. same settings. These results suggest that, for specialized biomedical segmentation tasks, a carefully tuned, moderate-sized network can be both effective and ef… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

22 extracted references · 19 canonical work pages

[1]

Automated human induced pluripotent stem cell colony segmen- tation for use in cell culture automation applications,

K. A. Powell, L. R. Bohrer, N. E. Stone, B. Hittle, K. R. Anfinson, V. Luangphakdy, G. Muschler, R. F. Mullins, E. M. Stone, and B. A. Tucker, “Automated human induced pluripotent stem cell colony segmen- tation for use in cell culture automation applications,” SLAS technology, vol. 28, no. 6, pp. 416–422, 2023

work page 2023
[2]

TransUNet: Re- thinking the U-Net architecture design for medical im- age segmentation through the lens of transformers,

J. Chen, J. Mei, X. Li, Y. Lu, Q. Yu, Q. Wei, X. Luo, Y. Xie, E. Adeli, Y. Wang, et al., “TransUNet: Re- thinking the U-Net architecture design for medical im- age segmentation through the lens of transformers,” Medical Image Analysis, vol. 97, p. 103280, 2024

work page 2024
[3]

nnu-net revisited: A call for rigorous validation in 3d medical image seg- mentation,

F. Isensee, T. Wald, C. Ulrich, M. Baumgartner, S. Roy, K. Maier-Hein, and P. F. Jaeger, “nnu-net revisited: A call for rigorous validation in 3d medical image seg- mentation,” in International Conference on Medical Image Computing and Computer-Assisted Intervention, 2024, pp. 488–498

work page 2024
[4]

nnU-Net: a self-configuring method for deep learning-based biomedical image seg- mentation,

F. Isensee, P. F. Jaeger, S. A. A. Kohl, J. Petersen, and K. H. Maier-Hein, “nnU-Net: a self-configuring method for deep learning-based biomedical image seg- mentation,” Nature methods, vol. 18, no. 2, pp. 203– 211, 2021

work page 2021
[5]

Improving the reliability of semantic segmentation of medical images by uncertainty modeling with Bayesian deep networks and curriculum learning,

S. Iwamoto, B. Raytchev, T. Tamaki, and K. Kaneda, “Improving the reliability of semantic segmentation of medical images by uncertainty modeling with Bayesian deep networks and curriculum learning,” in Uncer- tainty for Safe Utilization of Machine Learning in Med- ical Imaging, and Perinatal Imaging, Placental and Preterm Image Analysis: 3rd International...

work page 2021
[6]

Segment anything,

A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.- Y. Lo, et al., “Segment anything,” in Proceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 4015–4026

work page 2023
[7]

SAM 2: Seg- ment Anything in Images and Videos,

N. Ravi, V. Gabeur, Y.-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R¨ adle, C. Rolland, L. Gustafson, E. Mintun, J. Pan, K. V. A. Alwala, N. Carion, C.-Y. Wu, R. Gir- shick, P. Doll´ ar, and C. Feichtenhofer, “SAM 2: Seg- ment Anything in Images and Videos,” 2024. [Online]. Available: https://arxiv.org/abs/2408.00714

arXiv 2024
[8]

Segment anything in medical images,

J. Ma, Y. He, F. Li, L. Han, C. You, and B. Wang, “Segment anything in medical images,” Nature Com- munications, vol. 15, no. 1, p. 654, 2024

work page 2024

Show all 22 references

[9]

Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation,

L.-C. Chen, Y. Zhu, G. Papandreou, F. Schroff, and H. Adam, “Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation,” inPro- ceedings of the European Conference on Computer Vi- sion (ECCV), September 2018

2018
[10]

Biomedical sam 2: Segment anything in biomedical images and videos,

Z. Yan, W. Sun, R. Zhou, Z. Yuan, K. Zhang, Y. Li, T. Liu, Q. Li, X. Li, L. He, et al., “Biomedical sam 2: Segment anything in biomedical images and videos,” arXiv preprint arXiv:2408.03286, 2024

2024 arXiv
[11]

Deeplab: Semantic image segmenta- tion with deep convolutional nets, atrous convolution, and fully connected crfs,

L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “Deeplab: Semantic image segmenta- tion with deep convolutional nets, atrous convolution, and fully connected crfs,” IEEE transactions on pat- tern analysis and machine intelligence, vol. 40, no. 4, pp. 834–848, 2017

2017
[12]

Train Neural Net for Semantic Seg- mentation with PyTorch in 50 Lines of Code,

V. Blagojevic, “Train Neural Net for Semantic Seg- mentation with PyTorch in 50 Lines of Code,” 2021. [Online]. Available: https://towardsdatascience.com/ train-neural-net-for-semantic-segmentation-with-pytorch-in-50-lines-of-code-830c71a6544f/

2021
[13]

Video ob- ject segmentation using space-time memory networks,

S. W. Oh, J.-Y. Lee, N. Xu, and S. J. Kim, “Video ob- ject segmentation using space-time memory networks,” in Proceedings of the IEEE/CVF international confer- ence on computer vision, 2019, pp. 9226–9235

2019
[14]

Segmentation of Carotid Plaques Based on Improved DeepLabV3,

D. Wang, Z. Liu, and Y. Zhang, “Segmentation of Carotid Plaques Based on Improved DeepLabV3,” in Proceeding of the 2024 5th International Conference on Computer Science and Management Technology, 2024, pp. 201–207

2024
[15]

Towards Segment Anything Model (SAM) for Medical Image Segmentation: A Sur- vey,

Y. Zhang and R. Jiao, “Towards Segment Anything Model (SAM) for Medical Image Segmentation: A Sur- vey,” 2023. [Online]. Available: https://arxiv.org/ abs/2305.03678

2023 arXiv
[16]

Adapting segment anything models to medical imaging via fine-tuning without do- main pretraining,

K. Li and P. Rajpurkar, “Adapting segment anything models to medical imaging via fine-tuning without do- main pretraining,” in AAAI 2024 Spring Symposium on Clinical Foundation Models, 2024

2024
[17]

DeepLabV3 and Medical Imaging,

IMAIOS, “DeepLabV3 and Medical Imaging,” 2022. [Online]. Available: https://www.imaios.com/en/resources/ blog/deeplabv3-and-medical-imaging

2022
[18]

On the computational complexity of self-attention,

F. D. Keles, P. M. Wijewardena, and C. Hegde, “On the computational complexity of self-attention,” in In- ternational conference on algorithmic learning theory, 2023, pp. 597–619

2023
[19]

Domain Adaptation Guide,

V7 Labs, “Domain Adaptation Guide,” 2023. [Online]. Available: https://www.v7labs.com/blog/domain-adaptation-guide

2023
[20]

SAMCL: Empowering SAM to Continually Learn from Dynamic Domains,

Z. Wang, K. Ji, D. Wang, and F. Cheng, “SAMCL: Empowering SAM to Continually Learn from Dynamic Domains,” arXiv preprint arXiv:2412.05012, 2024

2024
[21]

Ad- vancing medical imaging informatics by deep learning- based domain adaptation,

A. Choudhary, L. Tong, Y. Zhu, and M. D. Wang, “Ad- vancing medical imaging informatics by deep learning- based domain adaptation,” Yearbook of medical infor- matics, vol. 29, no. 01, pp. 129–138, 2020

2020
[22]

iPS-Semantic-Segmentation,

Z. Maoquan, “iPS-Semantic-Segmentation,” 2025. [On- line]. Available: https://github.com/afountain/iPS-Semantic-Segmentation

2025

Pith tools

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

[1] [1]

Automated human induced pluripotent stem cell colony segmen- tation for use in cell culture automation applications,

K. A. Powell, L. R. Bohrer, N. E. Stone, B. Hittle, K. R. Anfinson, V. Luangphakdy, G. Muschler, R. F. Mullins, E. M. Stone, and B. A. Tucker, “Automated human induced pluripotent stem cell colony segmen- tation for use in cell culture automation applications,” SLAS technology, vol. 28, no. 6, pp. 416–422, 2023

work page 2023

[2] [2]

TransUNet: Re- thinking the U-Net architecture design for medical im- age segmentation through the lens of transformers,

J. Chen, J. Mei, X. Li, Y. Lu, Q. Yu, Q. Wei, X. Luo, Y. Xie, E. Adeli, Y. Wang, et al., “TransUNet: Re- thinking the U-Net architecture design for medical im- age segmentation through the lens of transformers,” Medical Image Analysis, vol. 97, p. 103280, 2024

work page 2024

[3] [3]

nnu-net revisited: A call for rigorous validation in 3d medical image seg- mentation,

F. Isensee, T. Wald, C. Ulrich, M. Baumgartner, S. Roy, K. Maier-Hein, and P. F. Jaeger, “nnu-net revisited: A call for rigorous validation in 3d medical image seg- mentation,” in International Conference on Medical Image Computing and Computer-Assisted Intervention, 2024, pp. 488–498

work page 2024

[4] [4]

nnU-Net: a self-configuring method for deep learning-based biomedical image seg- mentation,

F. Isensee, P. F. Jaeger, S. A. A. Kohl, J. Petersen, and K. H. Maier-Hein, “nnU-Net: a self-configuring method for deep learning-based biomedical image seg- mentation,” Nature methods, vol. 18, no. 2, pp. 203– 211, 2021

work page 2021

[5] [5]

Improving the reliability of semantic segmentation of medical images by uncertainty modeling with Bayesian deep networks and curriculum learning,

S. Iwamoto, B. Raytchev, T. Tamaki, and K. Kaneda, “Improving the reliability of semantic segmentation of medical images by uncertainty modeling with Bayesian deep networks and curriculum learning,” in Uncer- tainty for Safe Utilization of Machine Learning in Med- ical Imaging, and Perinatal Imaging, Placental and Preterm Image Analysis: 3rd International...

work page 2021

[6] [6]

Segment anything,

A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.- Y. Lo, et al., “Segment anything,” in Proceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 4015–4026

work page 2023

[7] [7]

SAM 2: Seg- ment Anything in Images and Videos,

N. Ravi, V. Gabeur, Y.-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R¨ adle, C. Rolland, L. Gustafson, E. Mintun, J. Pan, K. V. A. Alwala, N. Carion, C.-Y. Wu, R. Gir- shick, P. Doll´ ar, and C. Feichtenhofer, “SAM 2: Seg- ment Anything in Images and Videos,” 2024. [Online]. Available: https://arxiv.org/abs/2408.00714

arXiv 2024

[8] [8]

Segment anything in medical images,

J. Ma, Y. He, F. Li, L. Han, C. You, and B. Wang, “Segment anything in medical images,” Nature Com- munications, vol. 15, no. 1, p. 654, 2024

work page 2024