Pith. sign in

REVIEW 4 major objections 5 minor 33 references

From level set evolution to threshold optimization: A grayscale level set framework for image segmentation

T0 review · 4 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read Once a degraded image is smoothed into a threshold-separable form, the paper proves, level-set segmentation reduces to picking a single gray-level cutoff — with no length regularization or PDE evolution.

desk verdict The optimality theorem is circular and the framework is essentially thresholding after denoising, but the incremental O(NS) threshold scan is a genuine speedup and the empirical observation is worth a second look. read the letter →

arxiv 2607.22255 v1 pith:F5MYARGO submitted 2026-07-24 cs.CV math-phmath.MP

Xingkai Li , Jiebao Sun , Fanghui Song , Zhichang Guo This is my paper · ORCID
classification cs.CVmath-phmath.MP MSC 68U10
keywords imagesegmentationlevelsetmethodgrayscalethresholdoptimizationregularization-freedenoisingintensityinhomogeneityvariational
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 tries to establish that the length-regularization term pervading level-set segmentation is unnecessary once the image has been smoothed, and that the segmentation problem can be recast as a one-dimensional search over gray-level thresholds. It defines a class of smooth images on which object and background are separable by a single threshold, and builds the grayscale level set: the signed-distance functions of the threshold contours of the smoothed image. Under a separation assumption, it proves that minimizing the regularization-free fitting energy over these threshold contours equals minimizing it over all admissible partitions, so PDE evolution can be replaced by a threshold scan. The practical payoff is speed: an incremental update scheme makes the cost independent of how finely thresholds are sampled, and experiments on noisy, low-contrast, and intensity-inhomogeneous images show accurate contours without a length term. A sympathetic reader would care because this offers a theoretical justification for a much simpler, faster segmentation pipeline.

What carries the argument

Grayscale level set: the family Φ_c = {φ_c(x) = SDF(Γ_c) : c ∈ [Cmin, Cmax]}, where J is the smoothed image, Γ_c is the boundary of the superlevel set {J > c}, and SDF is the signed distance function (distance from each pixel to the contour, signed by which side it lies on). This family is a subspace of traditional level-set space and is the mechanism that converts curvature-driven PDE evolution into one-dimensional threshold optimization. Supporting it are the definition of a smooth image (bounded gradient plus threshold separability of the two regions), the co-area formula (which bounds the total arc length of candidate contours by L·|Ω|), and an incremental update algorithm that reuses co

What would settle it

Take a synthetic image whose two regions have overlapping intensity distributions even after strong denoising and bias correction, and check whether any threshold c satisfies sup_{Ω2} J < c < inf_{Ω1} J for the ground-truth partition. If no such threshold exists, or if the regularization-free fitting energy has a global minimizer that is not a threshold partition, Theorem 3.3.3 fails. Concretely, a two-Gaussian-mixture image with means separated by less than the combined standard deviations after optimal preprocessing would falsify the framework's central premise.

Watch

Extended reading notes

Core claim

The central claim, stated as Theorem 3.3.3, is that on a smooth image satisfying the separation condition of Definition 3.3.1, the global minimum of the regularization-free fitting energy over threshold partitions equals the global minimum over all admissible two-region partitions. In other words, the optimal segmentation boundary can be represented as a level contour Γ_c = ∂{x : J(x) > c*} of the smoothed image, so the search for a level-set function can be compressed to the grayscale level set without loss of optimality. The paper further shows the length term conflicts with distance regularization (driving |∇φ| toward 0 while distance regularization drives it toward 1), and that on smooth

Load-bearing premise

The argument rests on the assumption that, after preprocessing, the object and background can be separated by a single gray-level cutoff, and that the true best two-region partition is one of those cutoffs — if a real image cannot be smoothed into that form without moving the boundary, the claimed equivalence between threshold search and full variational segmentation no longer holds.

Editorial extensions

If this is right

  • Length regularization can be dropped from level-set segmentation after preprocessing, removing the conflicting dynamics between the length term and distance regularization.
  • Segmentation cost becomes independent of threshold discretization: for an N-pixel image with kernel support S, the full threshold sweep costs O(N·S) instead of O(M·N·S), making large images far cheaper to segment.
  • Because the optimal contour is found by a global threshold scan rather than curve evolution, results are insensitive to the initialization that trips iterative level-set methods.
  • Preprocessing and segmentation decouple cleanly: the segmentation stage is unchanged regardless of which denoising or bias-correction operator produced the smooth image, as long as the separation condition holds.
  • On images that meet the smoothness and separation assumptions, the variational segmentation problem is no longer a local-minima-prone evolution but a well-posed one-dimensional optimization with a global minimizer among threshold contours.

Reading between the lines

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

  • If the equivalence holds broadly, classical variational segmentation could regain a speed advantage over deep networks on very large images, at the price of overlap accuracy on complex textured regions; the paper's own histopathology comparison hints at this trade-off, though the paper frames the result as a geometric benefit.
  • The threshold-separability premise suggests a testable extension: the method could be viewed as histogram-mode analysis of the smoothed image, and extended to multi-phase segmentation by searching over multiple thresholds — a direction the paper lists as future work.
  • The advertised complexity gain depends on the kernel support S remaining small; for very large smoothing scales the local updates become global, so the speedup is most meaningful for localized fitting kernels.
  • One could stress-test the central assumption by measuring, after each preprocessing operator, the intensity-range gap between ground-truth regions; images with no gap should directly predict failures, making the assumption a diagnostic tool rather than only a hypothesis.
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 / 5 minor

Summary. The paper proposes a "grayscale level set" segmentation framework that, after a preprocessing step, replaces PDE-based level-set evolution by a one-dimensional threshold search. The authors claim that under a new definition of smooth images the length regularization term is unnecessary, that the optimal solution of a regularization-free fitting energy is representable in the grayscale-level-set space, and that an incremental update reduces complexity from O(M·N·S) to O(N·S). Experiments on synthetic and real images report competitive accuracy and large speedups.

Significance. If the theoretical claims were valid, the framework would offer a genuinely fast and simple alternative to variational level-set segmentation for degraded images, and the decoupling of preprocessing from segmentation would be practically attractive. The incremental threshold update in Algorithm 1 is a clear algorithmic idea, and the experimental results on several datasets (WBC, BrainWeb, DIAS, MoNuSeg) suggest that the pipeline works well on the selected examples. However, the central theoretical result, Theorem 3.3.3, is circular, and Definition 3.3.1 absorbs the desired conclusion into an assumption. Because the main contribution is the claimed optimality equivalence, the paper as it stands does not support its headline claim. The paper would need either a genuine proof that preprocessing can produce the separation condition, or a substantial reframing of the contribution as an engineering heuristic rather than a proven equivalence.

major comments (4)
  1. [§3.3.4, Theorem 3.3.3] The theorem assumes that the global minimizer (Ω*1, Ω*2) of the regularization-free fitting energy over the admissible class A is representable as a threshold partition: 'Ω*1 = {x : J(x) > c*}, Ω*2 = Ω \ Ω*1.' This is exactly the representability claim the theorem purports to prove. The proof then uses this assumption to conclude min_c E_gray(c) = min_A E(Ω1,Ω2). Thus the equality is not derived from the definition of smooth image or the fitting energy; it is an assumption restated as a theorem. The paper must either prove representability of the global minimizer from the properties of J (without assuming it), or explicitly label this as a conjecture. As stated, Contribution 2 is unsupported.
  2. [Definition 3.3.1 and §3.3.1] The definition of a smooth image includes condition 2: for the target partition (Ω1,Ω2), there exists c* with sup_{Ω2} J < c* < inf_{Ω1} J. This is a strong separation requirement that is precisely what the preprocessing Sσ would have to guarantee. The paper never shows that any of the proposed preprocessing operators (BM3D, PDE filtering, N4ITK, or Eq. (16) generally) can deliver this property for real degraded images while preserving boundary locations. If this condition fails, the subsequent threshold search has no optimality guarantee and the segmentation can be arbitrarily wrong. The paper needs a theorem or at least a systematic empirical validation on a test set showing that the separation condition holds, rather than assuming it per image.
  3. [§3.1, Lemma 3.1.1] Lemma 3.1.1 is stated without proof and appears to misstate the time direction. The surrounding text says 'the solutions stabilize as t→0' and 'we know from the lemma 3.1.1 and Fig. 1,' but the lemma concerns t ≥ T0, which in the context of evolution should be t → ∞, not t → 0. Moreover, the lemma claims that the solution equals the spatial mean for all t ≥ T0, a finite-time exact convergence statement for the total variation flow in N=2. This is nontrivial and the manuscript gives no citation or proof. Since this lemma is used to motivate the incompatibility of length and distance regularization, the motivation is not rigorous.
  4. [§3.1, Length-term incompatibility] The paper asserts that 'the length term constrains |∇ϕ| to approach 0 near the zero level set region, while the distance regularization term enforces |∇ϕ| to approach 1,' and concludes there is an incompatibility. No derivation or equation is provided for the claim that the length term drives |∇ϕ| to 0. The length term's gradient flow is curvature motion; it does not by itself drive |∇ϕ| to 0 in general. This assertion is load-bearing because it motivates dropping the length term entirely. Please provide a derivation or a counterexample; otherwise this remains an unsupported heuristic.
minor comments (5)
  1. [§4.5, Table 5] Table 5 appears to have a formatting error: the line 'Ours 0.6422-8 0.3357' is unclear. Probably '0.6422' with a missing metric and '−8' is not a value. Please correct.
  2. [§3.3.2, Algorithm 1 and §4.1] The notation Δc is used in the algorithm and complexity analysis but the choice of Δc is not discussed; please clarify whether it is a fixed grayscale step and how it affects accuracy.
  3. [§4.2] The text references 'Section 3.2.1' for the imaging models, but the models are presented in §3.3.1. Please correct the cross-reference.
  4. [Abstract and §1] Minor typos: 'from from O(M·N·S)' in the contributions list, and 'annoted' in the Introduction should be 'annotated.'
  5. [§3.3.2, Definitions] In Definition 3.3.2, the interface Γc is written as Γc = ∂Ω1(c), but the later use of ∂{x : J(x) > c*} in Eq. (21) is consistent; please define Ω1(c) explicitly in the definition to avoid ambiguity.

Circularity Check

2 steps flagged · score 8.0 of 10

Theorem 3.3.3 assumes the representability it purports to prove: the claimed equality between threshold search and full variational minimization is an assumption restated as a theorem.

  1. self definitional [Section 3.3.4, Theorem 3.3.3 (and its proof)]
    "Let J: Ω → [Cmin, Cmax] be a smooth image satisfying the separation condition: there exist a partition (Ω∗1, Ω∗2) and a threshold c∗ ∈ [Cmin, Cmax] such that Ω∗1 = {x ∈ Ω : J(x) > c∗}, Ω∗2 = Ω\Ω∗1. Assume that (Ω∗1, Ω∗2) is a global minimizer of the regularization-free fitting energy over the admissible partition class A. Then min c∈[Cmin,Cmax] E_gray(c) = min (Ω1,Ω2)∈A E(Ω1, Ω2)."

    The theorem assumes that the global minimizer of the fitting energy is a threshold partition (Ω∗1 = {J > c∗}), which is exactly the representability claim that Contribution 2 says the paper establishes. The proof's only step crossing the two minima is: 'by the separation assumption, the global minimizer (Ω∗1, Ω∗2) can be represented by the threshold c∗', giving E_gray(c∗) = E(Ω∗1, Ω∗2) = min_A E. No argument shows that a global minimizer over all partitions A must have threshold form; the equality is imported by the hypothesis rather than derived.

  2. self definitional [Section 3.3.1, Definition 3.3.1]
    "For target partition (Ω1, Ω2), there exists a threshold c∗ such that: sup Ω2 J < c∗ < inf Ω1 J. ... The first point of the definition controls image noise, and the second point provides guarantee for subsequent segmentation."

    The class of 'smooth images' is defined so that the target (ground-truth) partition is already separable by a single intensity threshold. This is the same separation condition later assumed in Theorem 3.3.3. Thus the 'smoothness assumptions' under which the optimal solution is claimed to be representable in the grayscale-level-set space themselves contain the desired threshold-representability of the target partition. The paper never proves that the preprocessing operators (BM3D, PDE filtering, N4ITK, Eq. (16)) can produce such an image for real degraded inputs while preserving boundaries.

full rationale

The central theoretical contribution—that the optimal segmentation solution can be represented in the grayscale level set space—is not derived from independent smoothness properties. Definition 3.3.1 bakes in threshold separability of the target partition, and Theorem 3.3.3 assumes that the global minimizer of the fitting energy is exactly such a threshold partition. The theorem then concludes min over thresholds equals min over all partitions; that conclusion is the assumption itself, so the key optimality equivalence is circular. This is not a minor self-citation issue or a purely external-benchmark concern: the paper's own proof text shows the reduction. Other parts of the work, notably the incremental update complexity analysis and the experimental comparisons, have independent algorithmic content and are not circular. But since the paper's headline claim of replacing PDE level-set evolution with threshold search rests on the unproven threshold-representability of the global minimizer, the circularity is load-bearing and substantial.

Assumptions & free parameters 5 free parameters · 5 assumptions · 2 invented entities

Free parameters are per-image weights, kernel scale, threshold step, and preprocessing operator. Axioms include the co-area formula (standard), an unproved TV-flow stabilization lemma, and two ad hoc assumptions: threshold separability in the definition of smooth images and threshold-representability of the global minimizer in Theorem 3.3.3. The grayscale level set is not an independently evidenced entity; it is a relabeling of threshold contours.

free parameters (5)
  • λ1, λ2 region weight coefficients = 1,1 (image 1); 1,4 (image 2)
    Set per experiment to improve Dice; no automatic selection or sensitivity analysis (§4.2).
  • Gaussian kernel scale σ = 3
    Selected by hand for all experiments; controls local fitting scale and directly affects result (§4.2).
  • Threshold step Δc = not specified
    Discretization step for threshold search; not reported, affects complexity and accuracy (Algorithm 1).
  • Preprocessing operator Sσ and its parameters = BM3D, PDE filter, N4ITK; denoising strength varied
    The choice of denoiser and parameters determines whether the separation condition holds; no fixed rule is given (§3.3.1, §4.6).
  • Constant L in smooth image definition = not quantified
    L is assumed finite but never estimated or verified on preprocessed images (Def. 3.3.1).
assumptions (5)
  • standard math Co-area formula for H1 functions: ∫ H1({J=c}) dc = ∫ |∇J| dx
    Used in Theorem 3.3.2 to bound level-set perimeters; valid for BV/Sobolev functions on Lipschitz domains.
  • domain assumption Lemma 3.1.1: 2D total variation flow reaches the constant mean in finite time for L2 initial data
    Stated without proof or citation; used to argue length regularization reduces gradient variation; direction 'as t→0' appears to be a typo for t→∞.
  • ad hoc to paper Definition 3.3.1: there exists a threshold c* with sup_{Ω2} J < c* < inf_{Ω1} J
    This separation condition is exactly the property that makes threshold segmentation succeed; it is assumed, not derived from the preprocessing.
  • ad hoc to paper Theorem 3.3.3: the global minimizer of the regularization-free fitting energy over A can be represented by a threshold partition
    This is the conclusion restated as an assumption; A is never defined.
  • domain assumption Preprocessing operator Sσ maps BV to H1 and preserves the target boundary
    The framework depends on Sσ producing a smooth image whose level sets still match the true object boundary; no operator is proven to have this property (Eq. 16).
invented entities (2)
  • Grayscale level set Φ_c
    purpose: Restricts the search space for segmentation to contours induced by intensity thresholds of the smoothed image
    A mathematical relabeling of threshold contours; no falsifiable external prediction, and it is this entity whose optimality proof is circular (Def. 3.3.2, Thm. 3.3.3).
  • Smooth image (H1 smooth with separation)
    purpose: Defines the class of images on which regularization-free thresholding is claimed to be optimal
    The separation condition is chosen to make the method work; no independent evidence that real preprocessed images satisfy it (Def. 3.3.1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of From level set evolution to threshold optimization: A grayscale level set framework for image segmentation." pith.science (2026). https://pith.science/paper/F5MYARGO

@misc{pith2026260722255,
  author       = {Pith},
  title        = {Pith review of: From level set evolution to threshold optimization: A grayscale level set framework for image segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F5MYARGO}},
  note         = {Machine review of arXiv:2607.22255}
}
abstract

The segmentation of multiple degradations has been a challenging problem in the field of image segmentation. Existing level set approaches commonly adopt a length regularization term to constrain the geometric shape of the segmentation contour. However, the introduction of the length term often results in numerical instability and high computational cost. In this paper, we show that the length term is not essential under certain smoothness constraints, and theoretically prove that the presence of the length term affects the property of $|\nabla \phi|=1$. Based on the finding, we define a class of smooth images, construct the grayscale level set, and propose a fast segmentation framework for degraded images, such as heavily noisy images and intensity inhomogeneous images. The framework transforms PDE evolution into one-dimensional threshold search, which has significant advantages in computational speed, especially on large-scale images. Experiments validate the segmentation performance of the proposed framework on various degraded images.

Figures

Figures reproduced from arXiv: 2607.22255 by the authors.

Figure 1
Figure 1. Evolution of level set function corresponding to length regularization term with different iterations. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The results of variation segmentation model. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Visualization of Mixture Model (15) The first step of our framework is to construct a generalized preprocessing mapping operator Sσ : BV (Ω) → H1 (Ω) parameterized by a scale σ, which maps the observed image I(x) to a smooth image J(x) with good properties. J(x) = Sσ(I(x)) (16) Preprocessing mapping operator Sσ has a high degree of modularity and scalability. Common choices include total variation denoising, Gaussia… view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: The Gray Contour of Smooth Images Proposition 3.3.1. The grayscale level set is a subspace of the traditional level set, i.e Φc(x) ⊂ Φ(x) Proof. For any threshold c, we have ϕc(x) = SDF(Γc) ∈ Φ. Hence, Φc ⊂ Φ. Besides, the grayscale level set has the above relationship…
Figure 5
Figure 5. Figure 5: Results of Image 1 10 [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Results of Image 2 [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: (a-d) show the bias correction result and its segmentation on image 3. For comparison, the classical RSF model is applied to the same inhomogeneous image under different initial contour settings. As shown in fig. 7 (e-h), the RSF model exhibits strong sensitivity to in…
Figure 8
Figure 8. Figure 8: Results of WBC dataset [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Results of BrainWeb dataset (a) Degraded (b) GT (c) Processed (d) Result [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: Results of DIAS dataset 13 [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Comparison results. 14 [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]
Figure 12
Figure 12. Figure 12: Comparison with Unet Due to the powerful semantic feature extraction ability of deep network, U-Net has a good performance in the traditional indicators based on regional overlap such as Dice. Although U-Net achieves better overlop-based metrics, our method geometrica…
Figure 13
Figure 13. Figure 13: Additive Gaussian Noise (a) Zhou+U (b) U (c) Zhou+N (d) N (e) Zhou+O (f) O [PITH_FULL_IMAGE:figures/full_fig_p016_13.png]
Figure 14
Figure 14. Figure 14: Multiplicative Gamma Noise was further developed, reducing the overall computational complexity from O(MNS) to O(NS). The proposed framework provides a flexible and modular segmentation paradigm by decoupling image preprocessing from the segmentation stage. Under suit…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 4 canonical work pages

  1. [1]

    Gore, and Zhaohua Ding

    Chunming Li, Chiu-Yen Kao, John C. Gore, and Zhaohua Ding. Minimization of region-scalable fit- ting energy for image segmentation.IEEE Transactions on Image Processing, 17(10):1940–1949, 2008. doi:10.1109/TIP.2008.2002304

    arXiv 1940

  2. [2]

    Chris Gatenby, Dimitris N

    Chunming Li, Rui Huang, Zhaohua Ding, J. Chris Gatenby, Dimitris N. Metaxas, and John C. Gore. A level set method for image segmentation in the presence of intensity inhomogeneities with application to mri.IEEE Transactions on Image Processing, 20(7):2007–2016, 2011. doi:10.1109/TIP.2011.2146190

    arXiv 2007

  3. [3]

    Deep convolutional neural networks with spatial regularization, volume and star-shape priors for image segmentation.Journal of Mathematical Imaging and Vision, 64(6): 625–645, 2022

    Jun Liu, Xiangyue Wang, and Xue-Cheng Tai. Deep convolutional neural networks with spatial regularization, volume and star-shape priors for image segmentation.Journal of Mathematical Imaging and Vision, 64(6): 625–645, 2022

    2022

  4. [4]

    Convex shape prior for deep convolution neural network-based image segmentation.Journal of Mathematical Imaging and Vision, 67(6):61, 2025

    Jun Liu, Kehui Zhang, Xue-Cheng Tai, and Shousheng Luo. Convex shape prior for deep convolution neural network-based image segmentation.Journal of Mathematical Imaging and Vision, 67(6):61, 2025

    2025

  5. [5]

    Additive-bias-correction variational model for noisy and intensity-inhomogeneous image segmentation.SIAM Journal on Imaging Sciences, 18(2):1235–1259, 2025

    Po-Wen Hsieh, Chung-Lin Tseng, and Suh-Yuh Yang. Additive-bias-correction variational model for noisy and intensity-inhomogeneous image segmentation.SIAM Journal on Imaging Sciences, 18(2):1235–1259, 2025. doi:10.1137/24M1676612. URLhttps://doi.org/10.1137/24M1676612. 16 A PREPRINT

    work page doi:10.1137/24m1676612 2025

  6. [6]

    Fronts propagating with curvature-dependent speed: Algorithms based on hamilton-jacobi formulations.Journal of Computational Physics, 79(1):12–49, 1988

    Stanley Osher and James A Sethian. Fronts propagating with curvature-dependent speed: Algorithms based on hamilton-jacobi formulations.Journal of Computational Physics, 79(1):12–49, 1988. ISSN 0021-9991. doi:https://doi.org/10.1016/0021-9991(88)90002-2. URL https://www.sciencedirect.com/ science/article/pii/0021999188900022

    arXiv 1988

  7. [7]

    Snakes: Active contour models.International Journal of Computer Vision, 1(4):321–331, 1988

    Michael Kass, Andrew Witkin, and Demetri Terzopoulos. Snakes: Active contour models.International Journal of Computer Vision, 1(4):321–331, 1988

    1988

  8. [8]

    Caselles, R

    V . Caselles, R. Kimmel, and G. Sapiro. Geodesic active contours. InProceedings of IEEE International Conference on Computer Vision, pages 694–699, 1995. doi:10.1109/ICCV .1995.466871

    arXiv 1995

Show all 33 references
  1. [9]

    Optimal approximations by piecewise smooth functions and associated variational problems.Communications on Pure & Applied Mathematics, 42(5):577–685, 1989

    David Mumford and Jayant Shah. Optimal approximations by piecewise smooth functions and associated variational problems.Communications on Pure & Applied Mathematics, 42(5):577–685, 1989

    1989

  2. [10]

    Chan and L.A

    T.F. Chan and L.A. Vese. Active contours without edges.IEEE Transactions on Image Processing, 10(2):266–277,

  3. [11]

    Chunming Li, Chenyang Xu, Changfeng Gui, and Martin D. Fox. Distance regularized level set evolution and its application to image segmentation.IEEE Transactions on Image Processing, 19(12):3243–3254, 2010. doi:10.1109/TIP.2010.2069690

    2010

  4. [12]

    A modified level set algorithm based on point dis- tance shape constraint for lesion and organ segmentation.Physica Medica, 57:123–136, 2019

    Xu Li, Chunming Li, Hairong Liu, and Xiaoping Yang. A modified level set algorithm based on point dis- tance shape constraint for lesion and organ segmentation.Physica Medica, 57:123–136, 2019. ISSN 1120-

    2019

  5. [13]

    Convexity shape prior for level set-based image segmen- tation method.IEEE Transactions on Image Processing, 29:7141–7152, 2020

    Shi Yan, Xue-Cheng Tai, Jun Liu, and Hai-Yang Huang. Convexity shape prior for level set-based image segmen- tation method.IEEE Transactions on Image Processing, 29:7141–7152, 2020. doi:10.1109/TIP.2020.2998981

    2020

  6. [14]

    Image segmentation for intensity inhomogeneity in presence of high noise.IEEE Transactions on Image Processing, 27(8):3729–3738, 2018

    Haider Ali, Lavdie Rada, and Noor Badshah. Image segmentation for intensity inhomogeneity in presence of high noise.IEEE Transactions on Image Processing, 27(8):3729–3738, 2018. doi:10.1109/TIP.2018.2825101

    2018

  7. [15]

    A variational image segmentation model with intensity correction in the presence of high level multiplicative noise.Inverse Problems and Imaging, 19(5):877–902, 2025

    Yamei Zhou, Zhichang Guo, Yao Li, and Boying Wu. A variational image segmentation model with intensity correction in the presence of high level multiplicative noise.Inverse Problems and Imaging, 19(5):877–902, 2025

    2025

  8. [16]

    A three-stage variational image segmentation framework incor- porating intensity inhomogeneity information.SIAM Journal on Imaging Sciences, 13(3):1692–1715, 2020

    Xu Li, Xiaoping Yang, and Tieyong Zeng. A three-stage variational image segmentation framework incor- porating intensity inhomogeneity information.SIAM Journal on Imaging Sciences, 13(3):1692–1715, 2020. doi:10.1137/20M1310618. URLhttps://doi.org/10.1137/20M1310618

    2020 doi

  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. InInternational Conference on Medical image computing and computer-assisted intervention, pages 234–241. Springer, 2015

    2015

  10. [18]

    Swin-unet: Unet-like pure transformer for medical image segmentation

    Hu Cao, Yueyue Wang, Joy Chen, Dongsheng Jiang, Xiaopeng Zhang, Qi Tian, and Manning Wang. Swin-unet: Unet-like pure transformer for medical image segmentation. In Leonid Karlinsky, Tomer Michaeli, and Ko Nishino, editors,Computer Vision – ECCV 2022 Workshops, pages 205–218, C...

    2022

  11. [19]

    Lungren, Shaoting Zhang, Lei Xing, Le Lu, Alan Yuille, and Yuyin

    Jieneng Chen, Jieru Mei, Xianhang Li, Yongyi Lu, Qihang Yu, Qingyue Wei, Xiangde Luo, Yutong Xie, Ehsan Adeli, Yan Wang, Matthew P. Lungren, Shaoting Zhang, Lei Xing, Le Lu, Alan Yuille, and Yuyin. Transunet: Rethinking the u-net architecture design for medical image segmentat...

    2024

  12. [20]

    An image is worth 16x16 words: Transformers for image recognition at scale.CoRR, abs/2010.11929, 2020

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at...

    2010 arXiv

  13. [21]

    Berg, and Wan Yen Lo

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C. Berg, and Wan Yen Lo. Segment anything.IEEE, 2023

    2023

  14. [22]

    Rudin, Stanley Osher, and Emad Fatemi

    Leonid I. Rudin, Stanley Osher, and Emad Fatemi. Nonlinear total variation based noise removal algorithms. Physica D: Nonlinear Phenomena, 60(1):259–268, 1992. ISSN 0167-2789. doi:https://doi.org/10.1016/0167- 2789(92)90242-F. URLhttps://www.sciencedirect.com/science/article/p...

    1992

  15. [23]

    He and Hao Liu

    Roy Y . He and Hao Liu. Euler’s elastica-based cartoon-smooth-texture image decomposition.SIAM Journal on Imaging Sciences, 18(1):526–569, 2025. doi:10.1137/24M167411X. URL https://doi.org/10.1137/ 24M167411X. 17 A PREPRINT

    2025 doi

  16. [24]

    David Adalsteinsson and James A. Sethian. A fast level set method for propagating interfaces.Journal of Computational Physics, 118(2):269–277, 1995. ISSN 0021-9991. doi:https://doi.org/10.1006/jcph.1995.1098. URLhttps://www.sciencedirect.com/science/article/pii/S0021999185710984

    1995

  17. [25]

    Image denoising by sparse 3-d transform-domain collaborative filtering.IEEE Transactions on Image Processing, 16(8):2080–2095, 2007

    Kostadin Dabov, Alessandro Foi, Vladimir Katkovnik, and Karen Egiazarian. Image denoising by sparse 3-d transform-domain collaborative filtering.IEEE Transactions on Image Processing, 16(8):2080–2095, 2007. doi:10.1109/TIP.2007.901238

  18. [26]

    Hybrid bm3d and pde filter- ing for non-parametric single image denoising.Signal Processing, 184:108049, 2021

    Ying Wen, Zhichang Guo, Wenjuan Yao, Dong Yan, and Jiebao Sun. Hybrid bm3d and pde filter- ing for non-parametric single image denoising.Signal Processing, 184:108049, 2021. ISSN 0165-

    2021

  19. [27]

    Tustison, Brian B

    Nicholas J. Tustison, Brian B. Avants, Philip A. Cook, Yuanjie Zheng, Alexander Egan, Paul A. Yushkevich, and James C. Gee. N4itk: Improved n3 bias correction.IEEE Transactions on Medical Imaging, 29(6):1310–1320,

  20. [28]

    Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising.IEEE Transactions on Image Processing, 26(7):3142–3155, 2016

    Kai Zhang, Wangmeng Zuo, Yunjin Chen, Deyu Meng, and Lei Zhang. Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising.IEEE Transactions on Image Processing, 26(7):3142–3155, 2016

    2016

  21. [29]

    A variational model to remove multiplicative noise based on sar image feature preservation.Inverse Problems and Imaging, 19(2):253–281, 2025

    Yamei Zhou, Zhichang Guo, Yao Li, Wenjuan Yao, and Boying Wu. A variational model to remove multiplicative noise based on sar image feature preservation.Inverse Problems and Imaging, 19(2):253–281, 2025. 18

    2025

  22. [1684]

    URL https://www.sciencedirect.com/science/ article/pii/S0165168421000888

    doi:https://doi.org/10.1016/j.sigpro.2021.108049. URL https://www.sciencedirect.com/science/ article/pii/S0165168421000888

    2021

  23. [1797]

    URL https://www.sciencedirect.com/science/ article/pii/S1120179718313681

    doi:https://doi.org/10.1016/j.ejmp.2018.12.032. URL https://www.sciencedirect.com/science/ article/pii/S1120179718313681

    2018 doi

  24. [2001]

    doi:10.1109/83.902291

    doi

  25. [2010]

    doi:10.1109/TMI.2010.2046908

    2010

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.