Pith. sign in

REVIEW 3 major objections 7 minor 42 references

Contour Field based Elliptical Shape Prior for the Segment Anything Model

T0 review · 3 major / 7 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A variational ellipse prior, unrolled inside SAM, makes its segmentation outputs elliptical and beats fine-tuned SAM.

desk verdict A useful unrolled elliptical regularizer for SAM with consistent but modest gains, overclaimed as a guaranteed elliptical-output prior and undermined by an invalid proof of its key proposition. read the letter →

arxiv 2504.12556 v2 pith:A2QLMEKH submitted 2025-04-17 cs.CV

Xinyu Zhao , Jun Liu , Faqiang Wang , Li Cui , Yuping Duan This is my paper · ORCID
classification cs.CV
keywords ImagesegmentationEllipticalshapeSoftthresholddynamicsDualitySegmentAnythingModelContourfieldprior
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 claims that a segmentation network can be made to output elliptical regions by replacing the final decoding step of the Segment Anything Model with an unrolled variational solver that enforces a contour-field constraint. The constraint requires the gradient of the soft segmentation to be orthogonal to the tangent field of a parametrized ellipse, which forces every contour line of the output to be a concentric ellipse. Solving this variational problem with a dual algorithm yields a closed-form softmax update, and the ellipse parameters are re-estimated from the second moments of the current segmentation. Unrolling this solver as 100 network modules inside SAM gives SAM-ESP, which the paper reports outperforms fine-tuned SAM on the REFUGE, ACDC, CASIA.v4, and DTU/Herlev datasets and on two external datasets. The significance is that a mathematically derivable shape prior can be folded directly into a generalist segmentation model rather than being applied only through post-processing or loss modifications.

What carries the argument

The load-bearing object is the parametrized elliptical contour field $T_\Lambda(x,y)$, the tangent vector field of a family of concentric, co-oriented ellipses. Proposition 1 establishes that $\langle\nabla u, T_\Lambda\rangle = 0$ everywhere is equivalent to all contours of $u$ being ellipses from that family, turning the shape prior into a convex-set constraint. The solution mechanism is a primal-dual alternating scheme: gradient ascent on the dual variable $q$, a strictly convex softmax update for $u$ (derived via the DC algorithm with a soft-threshold-dynamics regularization term), and a second-moment computation of the ellipse center, axes, and rotation that updates $T_\Lambda$. This mechanism carries the argument because the orthogonality constraint is what guarantees ellipticity, and the unrolled primal-dual iterations are what allow the constraint to be trained with backpropagation.

What would settle it

Run SAM-ESP on a test set whose ground-truth masks are elliptical but whose images contain strong non-elliptical distractors or heavy occlusion, so that SAM's initial features mislocate the object's center; if Dice falls below that of fine-tuned SAM while the loss and iteration count are held fixed, the claim that the moment-based ellipse re-estimation reliably guides segmentation fails. A more direct check of the 'outputs are elliptical regions' assertion is to measure the intersection-over-union of each predicted mask with its own best-fit ellipse and test whether SAM-ESP outputs reach the near-perfect overlap the variational constraint should enforce.

Watch

Extended reading notes

Core claim

The central claim is that SAM's final Heaviside decoding step can be replaced by a variational problem whose feasible set is defined by the orthogonality condition $\langle\nabla u, T_\Lambda\rangle = 0$, where $T_\Lambda$ is the tangent vector field of a parametrized ellipse with parameters $\Lambda=(x_0,y_0,a,b,\theta)$. Proposition 1 in the paper states that all contours of $u$ are concentric, co-oriented ellipses if and only if this condition holds everywhere. The proposed solver alternates a dual ascent step for the Lagrange multiplier $q$, a softmax closed-form update for the segmentation $u$, and a second-moment re-estimation of the ellipse parameters from the current soft segmentation. Unrolled over 100 iterations, this becomes the ESP module that replaces SAM's final decoding operator, ensuring the network output consists of elliptical regions. The paper reports that SAM-ESP achieves higher Dice, lower boundary distance, and lower boundary-distance standard deviation than fine-tuned SAM on the four in-distribution datasets and on two external generalization datasets, and that it retains higher accuracy under salt-and-pepper noise.

Load-bearing premise

The method depends on the second moments of the current soft segmentation being a trustworthy estimate of the true object's ellipse; if the initial SAM features are poor or the object deviates from an ellipse, the re-estimated tangent field points the constraint toward the wrong shape.

Editorial extensions

If this is right

  • SAM-ESP's outputs are structurally constrained to be elliptical regions, a hard guarantee that vanilla SAM, fine-tuned SAM, or shape-loss training do not provide.
  • On the four datasets, SAM-ESP outperforms fine-tuned SAM on Dice, boundary distance, and boundary-distance standard deviation, with the largest Dice gain on DTU/Herlev (96.45 vs 95.54).
  • On external datasets RIM-ONE DL and BinRushed, SAM-ESP generalizes better than fine-tuned SAM, while the shape-loss variant degrades generalization.
  • Under salt-and-pepper noise, SAM-ESP retains a Dice of 93.75 on DTU/Herlev versus 89.97 for fine-tuned SAM, indicating the constraint also acts as a noise regularizer.
  • The ESP module is not tied to SAM: experiments with Unet++ on CASIA.v4 show reduced boundary errors for pupil and iris segmentation, suggesting the module can be grafted onto other backbones.

Reading between the lines

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

  • Beyond the paper: the same unrolled-dual construction could be reused for other parametric shape families, such as circles, superellipses, or star shapes, by swapping the tangent field and re-deriving the moment formulas.
  • Beyond the paper: because the ellipse parameters are re-estimated at every iteration from the current soft mask, the prior is self-adjusting at inference time, so a testable extension is whether the method degrades gracefully when the target object is a deformed or partially occluded ellipse.
  • Beyond the paper: the paper fine-tunes the entire image encoder; a lighter protocol that freezes SAM's encoder and trains only the ESP decoder remains untested and would clarify how much of the gain comes from the shape constraint versus from full fine-tuning.
  • Beyond the paper: the reported gains are uneven across datasets, so the benefit likely scales with how strongly the objects satisfy the ellipse prior; applying the method to a convex but non-elliptical object class (e.g., organ shapes with rounded corners) would probe the limits of the prior.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper proposes SAM-ESP, a modification of the Segment Anything Model (SAM) in which the final decoding step is replaced by an unrolled variational algorithm that constrains the segmentation gradient to be orthogonal to the tangent field of a parameterized ellipse. The authors derive alternating subproblems for a dual variable, a softmax segmentation, and an ellipse parameter update based on second moments, and integrate the resulting iterative scheme as ESP modules into SAM. Experiments on four datasets (optic cup in REFUGE, cardiac MRI in ACDC, pupil in CASIA.v4, and cell nuclei in DTU/Herlev) and two external retinal datasets report modest improvements in Dice and boundary metrics over fine-tuned SAM and competing shape-prior baselines.

Significance. The proposed direction—unrolling a variational ellipse-prior model into a foundation model's decoder—is timely, and the closed-form updates make the module efficient and differentiable. If the elliptical constraint were rigorously enforced and the improvements were statistically robust, this would be a useful recipe for injecting shape priors into SAM-like models. However, the paper's central guarantee that outputs 'consist of elliptical regions' is not established: Proposition 1's proof is invalid, the deployed 100-iteration truncation is not shown to satisfy the constraint, and no shape-fidelity metric is reported. The improvements in Tables 3–6 are consistent but small, and the adaptively fitted ellipse makes the 'prior' partially data-driven. The paper does not ship machine-checked proofs or code, and the mathematical proof that underpins the method is logically flawed; nonetheless, the experimental results indicate a potentially useful soft regularizer, which justifies revision rather than rejection.

major comments (3)
  1. [Appendix A, proof of Proposition 1] The converse direction of the proof of Proposition 1 is logically invalid. From the fact that the tangent of the contour and the ellipse tangent field are both orthogonal to the gradient, the authors conclude that (φ(t),ψ(t)) 'is not an ellipse', even though (φ,ψ) is by construction the parametrization of an ellipse. A correct argument should show that the contour is an integral curve of the vector field T_Λ and therefore a level set of the ellipse equation. Because Proposition 1 is the mathematical foundation of the constraint, this must be fixed before the claim is credible.
  2. [Sections 5.1, 6.1, 6.9] The claim that SAM-ESP 'ensures' elliptical outputs is not supported by the evidence. Figure 2 shows the standalone variational algorithm requiring about 500 iterations to morph a non-elliptical region into an ellipse, but the deployed model uses 100 unrolled iterations (Section 6.1) and the ablation only explores T = 75 to 150 (Section 6.9). The paper never reports a measure of ellipticity, such as the residual norm of ⟨∇u,T⟩ or the deviation of the predicted mask from its own fitted ellipse. Without such evidence, the finite-truncation output cannot be claimed to be elliptical.
  3. [Sections 3.2, 4.1, and Eq. (22)] The 'prior' is partly circular: the tangent field T is recomputed from the current segmentation mask via second-moment fitting (Eq. 22), so the constraint encourages the mask to match an ellipse derived from itself. In addition, the paper provides no convergence analysis for the alternating scheme over (q,u,T), which is nonconvex because of the T update, and no sensitivity analysis for the fixed hyperparameters (λ=ε=τ_q=1, T=100). This limits the strength of the claim that the method is a principled variational prior rather than iterative ellipse filtering.
minor comments (7)
  1. [Section 2.4] There is a typo: 'upscale opteration' should be 'upscale operation'.
  2. [Eqs. (16)–(20)] The inner product notation ⟨q, T·∇ui⟩ should explicitly indicate integration over Ω, and the sign convention for the dual variable should be stated consistently across the derivation.
  3. [Eq. (23)] The binary entropy ⟨u, ln u⟩ + ⟨1-u, ln(1-u)⟩ differs from the multi-class entropy in Eq. (15); the relationship between the two formulations should be clarified.
  4. [Section 6.1 and Figure 1] The text states that '100 ESP modules are deployed' while Figure 1 and Algorithm 1 suggest a single unrolled module with 100 iterations; please make the terminology consistent.
  5. [Table 2] The inference time of SAM-ESP is roughly twice that of SAM-fine (about 150–180 ms versus 80 ms), and this trade-off is not discussed in the main text.
  6. [References] The reference for the CASIA.v4 dataset is given as 'B. I. Test. Casia iris database v4.', which is not a proper bibliographic entry; please cite the actual dataset reference.
  7. [Proposition 1 and Section 3.1] The proposition assumes C1 contours, but binary segmentation functions are generally only Lipschitz; the regularity assumption should be stated and justified.

Circularity Check

0 steps flagged · score 2.0 of 10

No material circularity: the moment-fitted ellipse prior is self-referential but the derivation is not equivalent to its inputs; the key claim rests on an unproved convergence assumption, which is a correctness risk, not circularity.

full rationale

The derivation chain is not circular in the sense required by the review rules. The variational model (Eq. 15) constrains u to the set P={u:<grad u,T>=0}, and Proposition 1 (proved in Appendix A of this paper) independently establishes that membership in P implies elliptical level sets. The update of T from second moments of the current iterate (Eq. 22; explicitly following Leung et al. 2004) is self-referential: the 'prior' ellipse is re-estimated from the mask it constrains. However, this is a standard alternating-estimation/fixed-point heuristic, and the u-subproblem (Eq. 20) retains the image fidelity term -o and the spatial regularization R(u), so the output is not equal to the input by construction. The paper's reliance on Chen et al. 2025 (with overlapping author Jun Liu) for the contour-flow idea is not load-bearing, because the key proposition is proved in this paper and the citation supplies motivation, not the proof. The real weaknesses are non-circularity issues: (i) Section 6.1 deploys only T=100 unrolled ESP modules, while Section 5.1/Figure 2 shows the standalone algorithm needs about 500 iterations to turn a non-elliptical region into an ellipse, and Section 6.9 ablates only T=75-150; (ii) convergence of the alternating q/u/T scheme is unproven; (iii) no metric measures deviation of the output from an ellipse, so the abstract's 'ensuring' is empirically unverified. These are correctness/robustness concerns, not circular reductions. Score 2 reflects only the minor non-load-bearing self-citation.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the geometric property of the tangent field T, the concavity of the threshold-dynamics regularizer, the dual formulation, and moment-based ellipse fitting. The first three are standard or borrowed from prior work; the last is a domain assumption that makes the 'prior' adaptive to the current segmentation. No new physical or conceptual entities are postulated.

free parameters (5)
  • epsilon (entropy regularization) = 1
    Set to 1 in all experiments. Controls smoothness of the soft segmentation; smaller values would produce more binary outputs. No sensitivity analysis is provided.
  • lambda (regularization weight) = 1
    Weight of the threshold dynamics regularization R(u). Set to 1 throughout; no ablation is reported.
  • tau_q (dual ascent step) = 1
    Step size in the q update (Eq. 19). Set to 1; no convergence guarantee or sensitivity study is given.
  • Gaussian kernel standard deviation sigma = 5 (support 5x5)
    Used in the kernel k for R(u) in the experiments; the paper does not state which kernel is used in the trained SAM-ESP model.
  • ESP module depth T = 100
    Number of unrolled iterations. The ablation on REFUGE shows 100 is best among 75, 100, 125, and 150; selected on validation.
assumptions (5)
  • standard math The vector field T in (14) is tangent to a family of concentric ellipses with common center, orientation, and semiaxis ratio, and its integral curves are these ellipses.
    Underlies Proposition 1 and the entire constraint u_i in P. Not proved in detail, but can be verified by direct differentiation.
  • standard math For a symmetric positive semi-definite kernel k, the regularization R(u)=lambda<u,k*(1-u)> is concave, so the DC decomposition with subgradient p=lambda k*(1-2u) is valid.
    Inherited from Liu et al. 2022 (Eq. 6); required for the u-subproblem to be convex.
  • standard math The divergence-adjoint relation <q, T dot grad u_i> = -<div(qT), u_i> holds under the Neumann boundary condition partial u_i / partial n = 0.
    Used to derive the dual formulation (17); assumes sufficient smoothness and boundary conditions on the image domain.
  • domain assumption The second moments of a (possibly soft) segmentation u_i determine a meaningful best-fit ellipse via (22).
    The moment formulas are exact only for binary indicators of a single ellipse; for general soft masks they define an ellipse with matched moments. If the segmentation is poor, the fitted ellipse may not correspond to the true object.
  • ad hoc to paper The alternating algorithm over q, u, T converges to a segmentation that satisfies the elliptical constraint and improves the SAM objective.
    The overall problem is nonconvex because R is concave and T is re-estimated. No convergence proof is given for the full alternation; the paper relies on empirical evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Contour Field based Elliptical Shape Prior for the Segment Anything Model." pith.science (2026). https://pith.science/paper/A2QLMEKH

@misc{pith2026250412556,
  author       = {Pith},
  title        = {Pith review of: Contour Field based Elliptical Shape Prior for the Segment Anything Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A2QLMEKH}},
  note         = {Machine review of arXiv:2504.12556}
}
read the original abstract

The elliptical shape prior information plays a vital role in improving the accuracy of image segmentation for specific tasks in medical and natural images. Existing deep learning-based segmentation methods, including the Segment Anything Model (SAM), often struggle to produce segmentation results with elliptical shapes efficiently. This paper proposes a new approach to integrate the prior of elliptical shapes into the deep learning-based SAM image segmentation techniques using variational methods. The proposed method establishes a parameterized elliptical contour field, which constrains the segmentation results to align with predefined elliptical contours. Utilizing the dual algorithm, the model seamlessly integrates image features with elliptical priors and spatial regularization priors, thereby greatly enhancing segmentation accuracy. By decomposing SAM into four mathematical sub-problems, we integrate the variational ellipse prior to design a new SAM network structure, ensuring that the segmentation output of SAM consists of elliptical regions. Experimental results on some specific image datasets demonstrate an improvement over the original SAM.

Figures

Figures reproduced from arXiv: 2504.12556 by the authors.

Figure 1
Figure 1. Architecture of each elliptical shape prior (ESP) module. [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. The proposed method for a non-elliptical region. [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. The comparison of different methods on a natural image. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The network architecture of SAM-ESP. architecture of the entire network is illustrated in [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Sample images from the datasets used. From left to right, columns one to four are from REFUGE, [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Visual effects of the segmentation results on four datasets. The blue line in the figure represents the [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: The generalization ability for different noise [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 8
Figure 8. Figure 8: Visualization of model segmentation results under noise influence. The first row displays segmentation results [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]
Figure 9
Figure 9. Figure 9: Visualization of segmentation results on external datasets. Blue: ground truth contours; green: model [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Segmentation results comparison. Blue: ground truth contours; green: model predictions. Red boxes [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]
Figure 11
Figure 11. Figure 11: Visualization of segmentation results on CASIA.v4 datasets [PITH_FULL_IMAGE:figures/full_fig_p019_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 32 canonical work pages

  1. [1]

    Some generalized order-disorder transformations

    Renfrey Burnard Potts. Some generalized order-disorder transformations. In Math. Proc. Cambridge Philos. Soc., volume 48, pages 106--109. Cambridge University Press, 1952

    work page 1952

  2. [2]

    Snakes: Active contour models

    Michael Kass, Andrew Witkin, and Demetri Terzopoulos. Snakes: Active contour models. Int. J. Comput. Vis., 1 0 (4): 0 321--331, January 1987. ISSN 1573-1405

    work page 1987

  3. [3]

    Active contours without edges

    Tony F Chan and Luminita A Vese. Active contours without edges. IEEE Trans. Image Process., 10 0 (2): 0 266--277, February 2001. ISSN 1941-0042

    work page 2001

  4. [4]

    Interactive graph cuts for optimal boundary & region segmentation of objects in N - D images

    Yuri Y Boykov and M-P Jolly. Interactive graph cuts for optimal boundary & region segmentation of objects in N - D images. In Proc. 8th IEEE Int. Conf. Comput. Vision. (ICCV), volume 1, pages 105--112, July 2001

    work page 2001

  5. [5]

    Semiautomatic segmentation with compact shape prior

    Piali Das, Olga Veksler, Vyacheslav Zavadsky, and Yuri Boykov. Semiautomatic segmentation with compact shape prior. Image Vis. Comput., 27 0 (1-2): 0 206--219, 2009

    work page 2009

  6. [6]

    Graph cut based image segmentation with connectivity priors

    Sara Vicente, Vladimir Kolmogorov, and Carsten Rother. Graph cut based image segmentation with connectivity priors. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit., pages 1--8, June 2008

    work page 2008

  7. [7]

    Star shape prior for graph - cut image segmentation

    Olga Veksler. Star shape prior for graph - cut image segmentation . In Proc. 10th Eur. Conf. Comput. Vis., pages 454--467, Oct 2008

    work page 2008

  8. [8]

    Convexity shape prior for binary segmentation

    Lena Gorelick, Olga Veksler, Yuri Boykov, and Claudia Nieuwenhuis. Convexity shape prior for binary segmentation. IEEE Trans. Pattern Anal. Mach. Intell., 39 0 (2): 0 258--271, February 2017

    work page 2017

Show all 42 references
  1. [9]

    Convexity shape prior for level set - based image segmentation method

    Shi Yan, Xue-Cheng Tai, Jun Liu, and Hai-Yang Huang. Convexity shape prior for level set - based image segmentation method . IEEE Trans. Image Process., 29: 0 7141--7152, June 2020

    2020

  2. [10]

    Barba, and J

    Hai-Shan Wu, J. Barba, and J. Gil. A parametric fitting algorithm for segmentation of cell images. IEEE Trans. Biomed. Eng., 45 0 (3): 0 400--407, March 1998. doi:10.1109/10.661165

    1998 doi

  3. [11]

    Lip image segmentation using fuzzy clustering incorporating an elliptic shape function

    Shu-Hung Leung, Shi-Lin Wang, and Wing-Hong Lau. Lip image segmentation using fuzzy clustering incorporating an elliptic shape function. IEEE Trans. Image Process., 13 0 (1): 0 51--62, January 2004. doi:10.1109/TIP.2003.818116

    2004

  4. [12]

    Graph cuts segmentation using an elliptical shape prior

    Greg Slabaugh and Gozde Unal. Graph cuts segmentation using an elliptical shape prior. In Proc. IEEE Int. Conf. Image Process. (ICIP), volume 2, pages 1222--1225, September 2005

    2005

  5. [13]

    Shape-based segmentation of tomatoes for agriculture monitoring

    Ujjwal Verma, Florence Rossant, Isabelle Bloch, Julien Orensanz, and Denis Boisgontier. Shape-based segmentation of tomatoes for agriculture monitoring . In Proc. 3rd Int. Conf. Pattern Recognit. Appl. Methods (ICPRAM), pages 402--411, March 2014

    2014

  6. [14]

    Ukpai, Satnam S

    Charles O. Ukpai, Satnam S. Dlay, and Wai L. Woo. Pupil segmentation using active contour with shape prior . In Sixth International Conference on Graphic and Image Processing (ICGIP 2014), volume 9443, pages 566--570, October 2014. doi:10.1117/12.2180065

    2014 doi

  7. [15]

    Segmentation of the left ventricle in cardiac cine MRI using a shape-constrained snake model

    Yuwei Wu, Yuanquan Wang, and Yunde Jia. Segmentation of the left ventricle in cardiac cine MRI using a shape-constrained snake model. Comput. Vis. Image Underst., 117 0 (9): 0 990--1003, September 2013. ISSN 1077-3142. doi:https://doi.org/10.1016/j.cviu.2012.12.008

    2013 doi

  8. [16]

    Fully convolutional networks for semantic segmentation

    Evan Shelhamer, Jonathan Long, and Trevor Darrell. Fully convolutional networks for semantic segmentation. IEEE Trans. Pattern Anal. Mach. Intell., 39 0 (4): 0 640--651, April 2017. doi:10.1109/TPAMI.2016.2572683

    2017

  9. [17]

    U- Net : Convolutional networks for biomedical image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- Net : Convolutional networks for biomedical image segmentation . In Proc. 18th Int. Conf. Med. Image Comput. Comput.-Assist. Intervent. (MICCAI)., pages 234--241, May 2015

    2015

  10. [18]

    Segment anything

    Alexander Kirillov et al. Segment anything. In Proc. IEEE/CVF Conf. Comput. Vis., pages 4015--4026, Oct 2023

    2023

  11. [19]

    Segment anything in medical images

    Jun Ma, Yuting He, Feifei Li, Lin Han, Chenyu You, and Bo Wang. Segment anything in medical images. Nat. Commun., 15 0 (1): 0 654, January 2024. ISSN 2041-1723. doi:10.1038/s41467-024-44824-z

    2024 doi

  12. [20]

    How to efficiently adapt large segmentation model ( SAM ) to medical images , June 2023

    Xinrong Hu, Xiaowei Xu, and Yiyu Shi. How to efficiently adapt large segmentation model ( SAM ) to medical images , June 2023. URL http://arxiv.org/abs/2306.13731. arXiv:2306.13731

    2023 arXiv

  13. [21]

    SAM - Med2D , August 2023

    Junlong Cheng et al. SAM - Med2D , August 2023. URL https://arxiv.org/abs/2308.16184v1. arXiv:2308.16184

    2023 arXiv

  14. [22]

    Learnable ophthalmology SAM , April 2023

    Zhongxi Qiu, Yan Hu, Heng Li, and Jiang Liu. Learnable ophthalmology SAM , April 2023. URL http://arxiv.org/abs/2304.13425. arXiv:2304.13425

    2023 arXiv

  15. [23]

    RSAM - Seg : A SAM -based approach with prior knowledge integration for remote sensing image semantic segmentation , February 2024

    Jie Zhang, Xubing Yang, Rui Jiang, Wei Shao, and Li Zhang. RSAM - Seg : A SAM -based approach with prior knowledge integration for remote sensing image semantic segmentation , February 2024. URL http://arxiv.org/abs/2402.19004. arXiv:2402.19004

    2024 arXiv

  16. [24]

    SAM fails to segment anything ? -- SAM - Adapter : Adapting SAM in underperformed scenes : camouflage , shadow , medical image segmentation , and more , May 2023

    Tianrun Chen et al. SAM fails to segment anything ? -- SAM - Adapter : Adapting SAM in underperformed scenes : camouflage , shadow , medical image segmentation , and more , May 2023. URL http://arxiv.org/abs/2304.09148. arXiv:2304.09148

    2023 arXiv

  17. [25]

    Star shape prior in fully convolutional networks for skin lesion segmentation

    Zahra Mirikharaji and Ghassan Hamarneh. Star shape prior in fully convolutional networks for skin lesion segmentation. In Proc. Int. Conf. Med. Image Comput. Comput.-Assist. Intervent., pages 737--745, June 2018

    2018

  18. [26]

    Noise-robust pupil center detection through cnn-based segmentation with shape-prior loss

    Sang Yoon Han, Hyuk Jin Kwon, Yoonsik Kim, and Nam Ik Cho. Noise-robust pupil center detection through cnn-based segmentation with shape-prior loss. IEEE Access, 8: 0 64739--64749, April 2020. doi:10.1109/ACCESS.2020.2985095

    2020

  19. [27]

    Accurate cnn-based pupil segmentation with an ellipse fit error regularization term

    Cuneyt Akinlar, Hatice Kubra Kucukkartal, and Cihan Topal. Accurate cnn-based pupil segmentation with an ellipse fit error regularization term. Expert Syst. Appl., 188: 0 116004, February 2022. ISSN 0957-4174. doi:https://doi.org/10.1016/j.eswa.2021.116004

    2022

  20. [28]

    Deep convolutional neural networks with spatial regularization, volume and star-shape priors for image segmentation

    Jun Liu, Xiangyue Wang, and Xue-Cheng Tai. Deep convolutional neural networks with spatial regularization, volume and star-shape priors for image segmentation. J. Math. Imaging Vis., 64 0 (6): 0 625--645, Jul 2022. ISSN 0924-9907. doi:10.1007/s10851-022-01087-x

    2022 doi

  21. [29]

    Contour flow constraint: Preserving global shape similarity for deep learning based image segmentation, 2025

    Shengzhe Chen, Zhaoxuan Dong, and Jun Liu. Contour flow constraint: Preserving global shape similarity for deep learning based image segmentation, 2025. URL https://arxiv.org/abs/2504.09384. arXiv:2504.09384

    2025

  22. [30]

    A continuous max-flow approach to potts model

    Jing Yuan, Egil Bae, Xue-Cheng Tai, and Yuri Boykov. A continuous max-flow approach to potts model. In Kostas Daniilidis, Petros Maragos, and Nikos Paragios, editors, Computer Vision -- ECCV 2010, pages 379--392, Berlin, Heidelberg, 2010. Springer Berlin Heidelberg. ISBN 978-3...

    2010

  23. [31]

    Threshold dynamics for networks with arbitrary surface tensions

    Selim Esedoḡ Lu and Felix Otto. Threshold dynamics for networks with arbitrary surface tensions. Commun. Pure Appl. Math., 68 0 (5): 0 808--864, May 2015

    2015

  24. [32]

    An efficient iterative thresholding method for image segmentation

    Dong Wang, Haohan Li, Xiaoyu Wei, and Xiao-Ping Wang. An efficient iterative thresholding method for image segmentation. J. Comput. Phys., 350: 0 657--667, Dec 2017. ISSN 0021-9991

    2017

  25. [33]

    An image is worth 16x16 words : transformers for image recognition at scale , June 2021

    Alexey Dosovitskiy et al. An image is worth 16x16 words : transformers for image recognition at scale , June 2021. URL http://arxiv.org/abs/2010.11929. arXiv:2010.11929

    2021 arXiv

  26. [34]

    Understanding segment anything model : SAM is biased towards texture rather than shape , June 2023

    Chaoning Zhang et al. Understanding segment anything model : SAM is biased towards texture rather than shape , June 2023. URL http://arxiv.org/abs/2311.11465. arXiv:2311.11465

    2023 arXiv

  27. [35]

    Refuge challenge: A unified framework for evaluating automated methods for glaucoma assessment from fundus photographs

    Jos \'e Ignacio Orlando et al. Refuge challenge: A unified framework for evaluating automated methods for glaucoma assessment from fundus photographs. Med. Image Anal., 59: 0 101570, January 2020. ISSN 1361-8415. doi:10.1016/j.media.2019.101570

    2020

  28. [36]

    Deep learning techniques for automatic MRI cardiac multi-structures segmentation and diagnosis: Is the problem solved? IEEE Trans

    Olivier Bernard et al. Deep learning techniques for automatic MRI cardiac multi-structures segmentation and diagnosis: Is the problem solved? IEEE Trans. Med. Imag., 37 0 (11): 0 2514--2525, November 2018. doi:10.1109/TMI.2018.2837502

    2018

  29. [37]

    B. I. Test. Casia iris database v4. [Online]. Available: http://www.idealtest.org/dbDetailForUser.do?id=4#/

  30. [38]

    Towards complete and accurate iris segmentation using deep multi - task attention network for non - cooperative iris recognition

    Caiyong Wang, Jawad Muhammad, Yunlong Wang, Zhaofeng He, and Zhenan Sun. Towards complete and accurate iris segmentation using deep multi - task attention network for non - cooperative iris recognition . IEEE Trans. Inf. Forensics Secur., 15: 0 2944--2959, March 2020

    2020

  31. [39]

    Pap-smear benchmark data for pattern classification

    Jan Jantzen, Jonas Norup, Georgios Dounias, and Beth Bjerregaard. Pap-smear benchmark data for pattern classification. In Proc. Nature Inspired Smart Inf. Syst. (NiSIS), pages 1--9, January 2005

    2005

  32. [40]

    Rim-one dl: A unified retinal image database for assessing glaucoma using deep learning

    Francisco José Fumero Batista, Tinguaro Diaz-Aleman, Jose Sigut, Silvia Alayon, Rafael Arnay, and Denisse Angel-Pereira. Rim-one dl: A unified retinal image database for assessing glaucoma using deep learning. Image Anal. Stereol., 39 0 (3): 0 161--167, November 2020. ISSN 1854-5165

    2020

  33. [41]

    Retinal fundus images for glaucoma analysis: the riga dataset

    Ahmed Almazroa, Sami Alodhayb, Essameldin Osman, Eslam Ramadan, Mohammed Hummadi, Mohammed Dlaim, Muhannad Alkatee, Kaamran Raahemifar, and Vasudevan Lakshminarayanan. Retinal fundus images for glaucoma analysis: the riga dataset. In Proc. SPIE, Med. Imag., Imag. Informat. Hea...

    2018 doi

  34. [42]

    Deep learning based adversarial disturbances in fundus image analysis

    Mohammad Eslami, Lakshmi Sritan Motati, Rohan Kalahasty, Saber Kazeminasab Hashemabad, Min Shi, Yan Luo, Yu Tian, Nazlee Zebardast, Mengyu Wang, and Tobias Elze. Deep learning based adversarial disturbances in fundus image analysis. Investigative Ophthalmology & Visual Science...

    2023

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.