Pith. sign in

REVIEW 3 major objections 5 minor 48 references

A Unified Variational Framework for Deep Weakly Supervised Image Segmentation

T0 review · 3 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 Scribble labels become a fuzzy membership map via kernel interpolation, feeding one convex Potts energy used as both iterative solver and network loss, nearing full-supervision accuracy without ground-truth masks.

desk verdict Useful empirical package, broken convexity claim: the energy is concave, not convex, and the convergence justification in Appendix C does not hold. read the letter →

arxiv 2607.19669 v1 pith:SKBCIBDT submitted 2026-07-22 cs.CV

Yin King Chu , Lingfeng Li , Sung Ha Kang , Jianping Zhang , Xue-Cheng Tai This is my paper · ORCID
classification cs.CV MSC 68U10
keywords weaklysupervisedsegmentationscribblelabelsPottsmodelRKHSfunctionextensionfuzzymembershipthresholddynamicsconvexvariationallosspartialcross-entropy
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 sparse scribble labels are enough to drive accurate image segmentation, provided they are first propagated across the whole image as a fuzzy membership function and then fed into a single convex variational energy. That energy — a simplex-constrained Potts model whose perimeter is a Gaussian-smoothed surrogate — can either be minimized directly by iterative thresholding for one image, or discretized into the training loss of a standard segmentation network. Trained this way, the network denoises the fuzzy map, sharpens object edges, and removes boundary artifacts; on the test set it beats partial cross-entropy (the usual scribble-supervision loss, which supervises only the labeled pixels) by roughly three mIoU points (mean intersection-over-union, a standard overlap metric), while staying within about three points of fully supervised training, all without seeing a ground-truth mask. If the claim holds, pixel-level annotations are not necessary to train competitive segmentation models — scribbles plus a well-chosen kernel suffice.

What carries the argument

Three components carry the argument. (1) The label-based fuzzy membership function u: sparse one-hot scribbles ψ on the labeled pixels D are extended to the whole domain by solving the RKHS linear system (K + mγI)a = ψ, giving Ψ(x) = Σ_i K(x,x_i) a_i; a Euclidean projection onto the probability simplex — implemented as a sorting-and-thresholding loop — turns Ψ into u, and the data fidelity takes the form f_k = 1 − 2u_k. The kernel is a product of a Gaussian in patch-intensity difference and a Gaussian in pixel distance, so its parameters control how far and how selectively labels propagate. (2) The smooth perimeter regularizer ('threshold dynamics', TD): λ Σ_k ∫ (1−v_k)(G_σ * v_k) dx, a Gaus

What would settle it

Retrain the same network on the same scribbles but with deliberately corrupted u — for example, inflating the spatial kernel scale σ_s until labels bleed across true boundaries. The paper's account predicts the loss, and therefore the predictions, should follow u and degrade; if the network holds its mIoU anyway, the energy is not the mechanism carrying performance. A sharper quantitative version: on the 200-image CSSD test set, the RKHS-over-PCE gap should widen as the scribbles are made sparser; if the gap stays flat across scribble densities, the fuzzy membership is not doing the work the p

Watch

Extended reading notes

Core claim

The paper's central claim is that a single convex, smooth energy — a simplex-constrained Potts model over membership functions v, with data term driven by a fuzzy membership u and perimeter approximated by Gaussian convolution — serves both classical iterative segmentation and deep-network training. Sparse one-hot scribbles ψ are extended to the whole image by solving a regularized least-squares problem in a Reproducing Kernel Hilbert Space (RKHS, a function space whose kernel makes point evaluation continuous and enables smooth interpolation); the extension is projected onto the probability simplex by a thresholding loop to give u, and the data fidelity takes the form f_k = 1 − 2u_k. The pe

Load-bearing premise

The load-bearing premise is that the fuzzy membership u, interpolated from scribbles by an RKHS whose kernel parameters are hand-chosen (the paper concedes they are empirical), faithfully represents the true class assignment on unlabeled pixels; wherever u is wrong, the data term 1−2u drives the solver and the network toward that wrong answer — a failure the paper itself exhibits with cactus glochids (Fig. 4) and an under-segmented pillar (Fig. 26).

Editorial extensions

If this is right

  • Scribble-only training becomes competitive with full supervision: the reported test gap to the fully supervised baseline is roughly 3 mIoU points, about half the gap that vanilla partial cross-entropy leaves.
  • One energy spans both regimes: the same convex problem solved by threshold dynamics for a single image becomes, after discretization, the network's training loss, so algorithmic insights transfer directly between classical and deep segmentation.
  • Because the energy is convex and smooth, training needs no alternating projections or hard label constraints, and the perimeter term is just a Gaussian convolution, keeping per-iteration cost low.
  • The TD term acts as a training regularizer, not merely a post-processing step: it penalizes over-segmentation and stabilizes the network against inventing spurious foreground regions that are absent from u.
  • The framework is generic in its regularizer — replacing the smooth perimeter with normalized-cut or compactness terms changes only one component, so the same RKHS label extension and loss derivation chain transfers to other spatial priors.

Reading between the lines

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

  • If the energy truly governs what the network learns, the three reported training effects should be visible as decreases in the corresponding terms of loss (9) over training; a direct check is to log the fidelity term and the TD term separately and verify that the TD term drops most when predictions over-segment.
  • The whole chain hinges on label propagation, and the paper concedes that kernel and weight parameters are picked empirically; this points to a testable upgrade — replacing the hand-tuned RKHS extension with a learnable propagation front end — which should raise the ceiling in exactly the cases the paper flags as weak (cactus glochids, an under-segmented pillar).
  • The paper notes that on the test set the TD term can erase true positives when predictions under-segment (mAcc drops by 1.23 points), which suggests a practical variant: make λ depend on the confidence or current perimeter of the prediction instead of fixing it globally.
  • Two load-bearing claims — convexity of the energy and equivalence between the relaxed and hard-constrained Potts problems — are inherited from previously published results cited by the paper rather than proved here; checking those proofs against the present formulation would settle whether the 'convex and smooth' guarantee is intact.
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 / 5 minor

Summary. The paper proposes a unified variational framework for weakly supervised image segmentation. Sparse scribble labels are extended to a dense fuzzy membership function via an RKHS function-extension problem; this membership is then fed into a simplex-constrained Potts model with a smooth nonlocal perimeter surrogate. The resulting energy is proposed both as an iterative single-image segmentation model, solved by threshold dynamics, and as a training loss for a UNet. Experiments on ECSSD and PASCAL VOC report improvements over partial cross-entropy and several regularized baselines, together with analyses of the training effects. The central theoretical claim, repeated in the abstract and conclusion, is that the energy is convex and smooth and that this convexity justifies the optimization algorithms.

Significance. The paper's overall direction is valuable: converting sparse labels into dense soft supervision through an RKHS extension is a principled and computationally efficient idea, and the reported experiments are extensive, including ablations, failure cases, and a candid discussion of limitations. The single-image experiments cover noise, illumination bias, and water splashes, and the paper is transparent about parameter sensitivity. If the empirical claims are reproducible, the approach is a useful addition to weakly supervised segmentation. However, the advertised convexity of the energy is false, and the convergence justification for the threshold-dynamics solver is therefore unsupported as written. The empirical comparison also contains an unlabeled table row that contradicts a stated outperformance claim. These issues are central to the paper's contribution and require substantive correction.

major comments (3)
  1. [§2.3 (Eq. (6)), §2.4 (Eq. (9)), Appendix C] The energy is not convex. Expanding the smooth perimeter term in (6) gives ∫(1−v_k)(Gσ∗v_k)dx = ∫v_k dx − ∫v_k(Gσ∗v_k)dx. Since Gaussian convolution is positive semidefinite, the second term is a negative quadratic form, so E is concave in v; the same holds for the loss in (9), where the Hessian with respect to Nθ,k contains −2λĜ. This contradicts the abstract's 'convex, smooth energy functional' and the conclusion. It also invalidates the Appendix C justification of Algorithm 3: the premise 'Since E is convex' is false, and the cited convergence results [30, Prop. 2] and [32, Thm. 2.2] cannot be invoked as stated. In addition, the printed update uses g_k^(t) = 1−2u_k + λGσ∗(1−v_k^(t)), but the first variation of the perimeter term is λ(1−2Gσ∗v_k), not λGσ∗(1−v_k); the missing factor of 2 changes the argmin. The authors should correct the characterization of the energy, the algorithm, an
  2. [Table 2, §5.3.1] Table 2 contains an unnamed row immediately after PCE+CV with λ=1, reporting mIoU 75.76 and mDice 85.76. These values are higher than the RKHS row (mIoU 75.39, mDice 85.52), directly contradicting the sentence 'none of them outperforms our RKHS.' The row should be labeled (presumably PCE+CV with λ=1) and the claim revised, or the row must be removed. This is not merely a typesetting issue: the comparison is central to the paper's claim of 'consistent improvements.'
  3. [§2.2 (Eqs. (4)–(5))] The equivalence between the binary problem (4) and its relaxation (5) is asserted by citation to [36] and [32] rather than proved. Given the concavity issue above, the usual convex-relaxation reasoning does not apply, so the authors should state the precise conditions and provide a self-contained argument, or at least an exact statement of the cited lemma, so the reader can verify the equivalence.
minor comments (5)
  1. [Eq. (22)] The notation x, y is overloaded: they are both points in the image domain and (2R_k+1)×(2R_k+1) patches centered at those points. Please define the patch vectors explicitly, e.g., P(x) and P(y).
  2. [§5.1] The text states that all models are trained for 8000 epochs. For 1400 training images and a 31M-parameter UNet, this is an unusually large number of full passes; please clarify whether 'epochs' means iterations/steps or true epochs, and report the actual training time.
  3. [Appendix B.2 (proof of Lemma 2)] The last line of the proof contains a typographical artifact: 'ΨT he(1)γ ≥ · · · ≥ Ψγ(K)'. This should read 'Ψ(1)^γ ≥ ... ≥ Ψ(K)^γ' and should be corrected.
  4. [Figure 2 caption] The caption says 'Smoothing effect of the perimeter term ... under different values of σ_s,0', but σ_s,0 is a kernel parameter of the RKHS membership function, not a parameter of the perimeter term. Please reword to avoid confusion.
  5. [Appendix E.3.3] The NCut and KCut formulas use a bilateral kernel with a normalization constant C, but C is never specified. Please state how C is computed for the discrete implementation.

Circularity Check

0 steps flagged · score 2.0 of 10

No construction-level circularity; the loss is an intended label-extrapolation target. Mild self-citation burden on the relaxation-equivalence and convergence claims.

full rationale

The central empirical derivation is not circular: the fuzzy membership u is computed from sparse scribbles by the RKHS extension (12)-(15), and the same precomputed u enters the data fidelity coefficients f_k = 1-2u_k in energy (6) and loss (9). The network is trained on this fixed target and tested on held-out ECSSD/CSSD ground truth; no test label enters the construction of u, and no fitted parameter is renamed as a prediction. The single-image threshold-dynamics experiments are evaluations of an optimization problem, not of a fitted target. The main self-citation-heavy link is the relaxation-equivalence step: the paper states 'the solution of the relaxed problem (5) is equivalent to the original problem (4). See explanations and proofs in [36, section 3] and [32, Lemma 2.1],' where [36] is the authors' own survey, and Appendix C delegates convergence to '[30, Proposition 2] and [32, Theorem 2.2]' with [30] also author-overlapping. Because an independent source ([32]) is cited alongside these, the claims do not reduce to a self-citation chain, but the burden is real. A separate mathematical risk, not a circularity, is that the asserted convexity of (6)/(9) is not proved and appears inconsistent with the expansion of the smooth perimeter term into a linear part minus a positive-semidefinite quadratic form; this would undermine Appendix C's 'Since E is convex' argument, but it is a correctness issue rather than an input-output equivalence.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. The main free parameters are the kernel scales and the TD weight, all hand-tuned or validation-tuned. The convexity of the energy is asserted but contradicted by the paper's own equations, and the relaxation-equivalence axiom is borrowed from the authors' prior work.

free parameters (6)
  • kernel intensity scale σI,k = 0.05 (PASCAL), 0.01 (ECSSD), 0.01–0.05 in single-image examples
    Controls patch-intensity similarity in kernel (22); chosen per dataset/example by hand.
  • kernel spatial scale σs,k = 3 (PASCAL), 1 (ECSSD), 100 (easy subset), 0.01–60 in single images
    Controls spatial closeness in kernel (22); hand-tuned; crucial for label propagation.
  • patch radius Rk = 3 for training; 2–4 in examples
    Patch window size in kernel (22); chosen by hand.
  • TD weight λ = 5e-5 for network; 10–35 for single-image
    Weight of perimeter term; selected on validation (Table 3) or manually for images.
  • Gaussian scale σ for TD = 0.1 for training; 3 for single-image
    Scale of Gaussian in smooth perimeter approximation (3); set by hand.
  • RKHS regularization γ = 0
    Set to 0 in Section 4.1.1, making the extension an exact interpolant; not tuned.
assumptions (5)
  • domain assumption The smooth perimeter surrogate (1-v)(Gσ*v) approximates the perimeter of each phase for small σ (Eq. 3).
    The model regularizer is built on this approximation; the paper cites [29] and [30] but does not prove convergence for this energy.
  • domain assumption The relaxed simplex problem (5) has the same minimizers as the binary problem (4).
    Section 2.2 asserts this citing [36, Sec. 3] and [32, Lemma 2.1], both by the same group; no proof is given in the paper.
  • ad hoc to paper The energy E in (6) is convex.
    Asserted in the abstract and Appendix C to justify threshold dynamics, but from (6) the perimeter term is concave; this is an incorrect assumption.
  • standard math The kernel matrix in (23) is symmetric positive definite and solvable via conjugate gradient.
    Gaussian kernels are positive definite; standard and not central to the contribution.
  • domain assumption Neural network optimization of (10) with SGD finds a usable segmentation operator.
    All deep learning segmentation methods rely on this; no convergence guarantees are given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Unified Variational Framework for Deep Weakly Supervised Image Segmentation." pith.science (2026). https://pith.science/paper/SKBCIBDT

@misc{pith2026260719669,
  author       = {Pith},
  title        = {Pith review of: A Unified Variational Framework for Deep Weakly Supervised Image Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SKBCIBDT}},
  note         = {Machine review of arXiv:2607.19669}
}
read the original abstract

We propose a unified variational framework for image segmentation under sparse pixel-level supervision. Our method is based on a simplex-constrained Potts model with a smooth perimeter regularizer, yielding a convex, smooth energy functional that can be used as a training loss in weakly supervised deep learning paradigms or optimized efficiently using iterative methods. Sparse labels are incorporated into the data fidelity term by constructing a fuzzy membership function via a function extension problem in a Reproducing Kernel Hilbert Space (RKHS), which can effectively capture inhomogeneous intensity statistics. The derived discrete loss for training standard networks demonstrates robustness and consistent improvements over non-training and partial cross-entropy (PCE) baselines in experiments, achieving comparable performance without requiring ground-truth segmentation images.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 2 linked inside Pith

  1. [36]

    In: Handbook of Mathematical Models and Algorithms in Computer Vision and Imaging: Mathematical Imaging and Vision, pp

    Tai, X., Li, L., Bae, E.: The potts model with different piecewise constant repre- sentations and fast algorithms: a survey. In: Handbook of Mathematical Models and Algorithms in Computer Vision and Imaging: Mathematical Imaging and Vision, pp. 1–41 (2023)

    2023

  2. [32]

    Journal of Computational Physics350, 657–667 (2017)

    Wang, D., Li, H., Wei, X., Wang, X.-P.: An efficient iterative thresholding method for image segmentation. Journal of Computational Physics350, 657–667 (2017)

    2017

  3. [1]

    In: 2017 IEEE International Conference on Computer Vision (ICCV), pp

    Máttyus, G., Luo, W., Urtasun, R.: Deeproadmapper: Extracting road topology 28 from aerial images. In: 2017 IEEE International Conference on Computer Vision (ICCV), pp. 3458–3466 (2017). https://doi.org/10.1109/ICCV.2017.372

    doi:10.1109/iccv.2017.372 2017

  4. [2]

    In: 2016 Fourth International Conference on 3D Vision (3DV), pp

    Milletari, F., Navab, N., Ahmadi, S.-A.: V-Net: Fully convolutional neural net- works for volumetric medical image segmentation. In: 2016 Fourth International Conference on 3D Vision (3DV), pp. 565–571 (2016). IEEE

    2016

  5. [3]

    In: 2016 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp

    Kampffmeyer, M., Salberg, A.-B., Jenssen, R.: Semantic segmentation of small objects and modeling of uncertainty in urban remote sensing images using deep convolutional neural networks. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp. 680–688 (2016). https://doi. org/10.1109/CVPRW.2016.90

    doi:10.1109/cvprw.2016.90 2016

  6. [4]

    Annual Review of Biomedical Engineering2(1), 315–337 (2000)

    Pham, D.L., Xu, C., Prince, J.L.: Current methods in medical image segmenta- tion. Annual Review of Biomedical Engineering2(1), 315–337 (2000)

    2000

  7. [5]

    Communications on Pure and Applied Mathematics42(5), 577–685 (1989)

    Mumford, D.B., Shah, J.: Optimal approximations by piecewise smooth func- tions and associated variational problems. Communications on Pure and Applied Mathematics42(5), 577–685 (1989)

    1989

  8. [6]

    IEEE Transactions on Image Processing10(2), 266–277 (2001)

    Chan, T.F., Vese, L.A.: Active contours without edges. IEEE Transactions on Image Processing10(2), 266–277 (2001)

    2001

Show all 48 references
  1. [7]

    Mathematical Pro- ceedings of the Cambridge Philosophical Society48(1), 106–109 (1952) https: //doi.org/10.1017/S0305004100027419

    Potts, R.B.: Some generalized order-disorder transformations. Mathematical Pro- ceedings of the Cambridge Philosophical Society48(1), 106–109 (1952) https: //doi.org/10.1017/S0305004100027419 . Cambridge University Press

    1952 doi

  2. [8]

    Applied and Computational Harmonic Analysis48(2), 787–794 (2020)

    Zhou, D.-X.: Universality of deep convolutional neural networks. Applied and Computational Harmonic Analysis48(2), 787–794 (2020)

    2020

  3. [9]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention, pp

    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedical image segmentation. In: International Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 234–241 (2015). Springer

    2015

  4. [10]

    IEEE Transactions on Pattern Analysis and Machine Intelligence40(4), 834–848 (2017)

    Chen, L.-C., Papandreou, G., Kokkinos, I., Murphy, K., Yuille, A.L.: Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. IEEE Transactions on Pattern Analysis and Machine Intelligence40(4), 834–848 (2017)

    2017

  5. [11]

    In: Proceedings of the IEEE/CVF International Confer- ence on Computer Vision, pp

    Strudel, R., Garcia, R., Laptev, I., Schmid, C.: Segmenter: Transformer for semantic segmentation. In: Proceedings of the IEEE/CVF International Confer- ence on Computer Vision, pp. 7262–7272 (2021)

    2021

  6. [12]

    In: Proceedings of the IEEE International Conference on Computer Vision (ICCV), pp

    Pathak, D., Krahenbuhl, P., Darrell, T.: Constrained convolutional neural networks for weakly supervised segmentation. In: Proceedings of the IEEE International Conference on Computer Vision (ICCV), pp. 1796–1804 (2015) 29

    2015

  7. [13]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Ahn, J., Kwak, S.: Learning pixel-level semantic affinity with image-level super- vision for weakly supervised semantic segmentation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4981–4990 (2018)

    2018

  8. [14]

    IEEE Transactions on Medical Imaging38(9), 2151–2164 (2019)

    Duan, J., Bello, G., Schlemper, J., Bai, W., Dawes, T.J., Biffi, C., Marvao, A., Doumoud, G., O’Regan, D.P., Rueckert, D.: Automatic 3d bi-ventricular segmen- tation of cardiac images by a shape-refined multi-task deep learning approach. IEEE Transactions on Medical Imaging38(...

    2019

  9. [15]

    In: European Conference on Computer Vision, pp

    Lu, W., Jia, X., Xie, W., Shen, L., Zhou, Y., Duan, J.: Geometry constrained weakly supervised object localization. In: European Conference on Computer Vision, pp. 481–496 (2020). Springer

    2020

  10. [16]

    IEEE Transactions on Circuits and Systems for Video Technology33(8), 3935–3946 (2023)

    Chen, J., Lu, W., Li, Y., Shen, L., Duan, J.: Adversarial learning of object-aware activation map for weakly-supervised semantic segmentation. IEEE Transactions on Circuits and Systems for Video Technology33(8), 3935–3946 (2023)

    2023

  11. [17]

    IEEE Transactions on Image Processing29, 1856–1866 (2019)

    Kim, B., Ye, J.C.: Mumford–shah loss functional for image segmentation with deep learning. IEEE Transactions on Image Processing29, 1856–1866 (2019)

    2019

  12. [18]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Liang, Z., Wang, T., Zhang, X., Sun, J., Shen, J.: Tree energy loss: Towards sparsely annotated semantic segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 16907–16916 (2022)

    2022

  13. [19]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Lin, D., Dai, J., Jia, J., He, K., Sun, J.: Scribblesup: Scribble-supervised con- volutional networks for semantic segmentation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3159–3167 (2016)

    2016

  14. [20]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Tang, M., Djelouah, A., Perazzi, F., Boykov, Y., Schroers, C.: Normalized cut loss for weakly-supervised cnn segmentation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1818–1827 (2018)

    2018

  15. [21]

    In: Proceedings of the European Conference on Computer Vision (ECCV), pp

    Tang, M., Perazzi, F., Djelouah, A., Ben Ayed, I., Schroers, C., Boykov, Y.: On regularized losses for weakly-supervised cnn segmentation. In: Proceedings of the European Conference on Computer Vision (ECCV), pp. 507–522 (2018)

    2018

  16. [22]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Wu, L., Zhong, Z., Fang, L., He, X., Liu, Q., Ma, J., Chen, H.: Sparsely annotated semantic segmentation with adaptive gaussian mixtures. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 15454– 15464 (2023)

    2023

  17. [23]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Zhang,H.,Burrows,L.,Meng,Y.,Sculthorpe,D.,Mukherjee,A.,Coupland,S.E., Chen, K., Zheng, Y.: Weakly supervised segmentation with point annotations for histopathology images via contrast-based variational model. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pa...

    2023

  18. [24]

    Journal of Mathematical Imaging and Vision37(1), 49–65 (2010)

    Ha Quang, M., Kang, S.H., Le, T.M.: Image and video colorization using vector- valued reproducing kernel hilbert spaces. Journal of Mathematical Imaging and Vision37(1), 49–65 (2010)

    2010

  19. [25]

    Journal of Visual Communi- cation and Image Representation25(5), 1136–1148 (2014)

    Kang, S.H., Shafei, B., Steidl, G.: Supervised and transductive multi-class segmentation using p-laplacians and rkhs methods. Journal of Visual Communi- cation and Image Representation25(5), 1136–1148 (2014)

    2014

  20. [26]

    Journal of Mathematical Imaging and Vision49(1), 191–201 (2014)

    Bresson, X., Tai, X.-C., Chan, T.F., Szlam, A.: Multi-class transductive learning based on l1 relaxations of cheeger cut and mumford-shah-potts model. Journal of Mathematical Imaging and Vision49(1), 191–201 (2014)

    2014

  21. [27]

    Physica Medica42, 162–173 (2017)

    Gui, L., Li, C., Yang, X.: Medical image segmentation based on level set and isoperimetric constraint. Physica Medica42, 162–173 (2017)

    2017

  22. [28]

    Neurocomputing623, 129395 (2025)

    Zhang, K., Li, L., Liu, H., Yuan, J., Tai, X.-C.: Deep convolutional neural networks meet variational shape compactness priors for image segmentation. Neurocomputing623, 129395 (2025)

    2025

  23. [29]

    Annales de la Faculté des sciences de Toulouse : Mathématiques6, 16 (2007)

    Pallara-F, M.M.J.-D., Preunkert, P.-M.: Short-time heat flow and functions of bounded variation inR N. Annales de la Faculté des sciences de Toulouse : Mathématiques6, 16 (2007)

    2007

  24. [30]

    Pattern Recognition44(9), 2093–2108 (2011) https://doi.org/10

    Liu, J., Tai, X.-c., Huang, H., Huan, Z.: A fast segmentation method based on constraint optimization and its applications: Intensity inhomogeneity and texture segmentation. Pattern Recognition44(9), 2093–2108 (2011) https://doi.org/10. 1016/j.patcog.2011.02.022

  25. [31]

    Communications on Pure and Applied Mathematics68(5), 808–864 (2015)

    Esedo¯ g Lu, S., Otto, F.: Threshold dynamics for networks with arbitrary surface tensions. Communications on Pure and Applied Mathematics68(5), 808–864 (2015)

    2015

  26. [33]

    Journal of Mathematical Imaging and Vision64(6), 625–645 (2022)

    Liu, J., Wang, X., Tai, X.-C.: Deep convolutional neural networks with spatial regularization, volume and star-shape priors for image segmentation. Journal of Mathematical Imaging and Vision64(6), 625–645 (2022)

    2022

  27. [34]

    Journal of Mathematical Imaging and Vision67(6), 61 (2025)

    Liu, J., Zhang, K., Tai, X.-C., Luo, S.: Convex shape prior for deep convolution neural network-based image segmentation. Journal of Mathematical Imaging and Vision67(6), 61 (2025)

    2025

  28. [35]

    SIAM Journal on Imaging Sciences 17(1), 540–594 (2024) 31

    Tai, X.-C., Liu, H., Chan, R.: Pottsmgnet: A mathematical explanation of encoder-decoder based neural networks. SIAM Journal on Imaging Sciences 17(1), 540–594 (2024) 31

    2024

  29. [37]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Luo, S., Tai, X.-C., Huo, L., Wang, Y., Glowinski, R.: Convex shape prior for multi-object segmentation using a single level set function. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 613–621 (2019)

    2019

  30. [38]

    arXiv preprint arXiv:2001.00617 (2020)

    Clason, C.: Regularization of inverse problems. arXiv preprint arXiv:2001.00617 (2020)

    2001 arXiv

  31. [39]

    In: Proceedings of the 25th International Conference on Machine Learning, pp

    Duchi, J., Shalev-Shwartz, S., Singer, Y., Chandra, T.: Efficient projections onto the l 1-ball for learning in high dimensions. In: Proceedings of the 25th International Conference on Machine Learning, pp. 272–279 (2008)

    2008

  32. [40]

    Bernard, O., Lalande, A., Zotti, C., Cervenansky, F., Yang, X., Heng, P.-A., Cetin, I., Lekadir, K., Camara, O., Ballester, M.A.G.,et al.: Deep learning tech- niques for automatic mri cardiac multi-structures segmentation and diagnosis: is the problem solved? IEEE Transactions...

    2018

  33. [41]

    Advances in Neural Information Processing Systems32(2019)

    Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al.: Pytorch: An imperative style, high- performance deep learning library. Advances in Neural Information Processing Systems32(2019)

    2019

  34. [42]

    Remote Sensing of Environment210, 113–143 (2018)

    Román, M.O., Wang, Z., Sun, Q., Kalb, V., Miller, S.D., Molthan, A., Schultz, L., Bell, J., Stokes, E.C., Pandey, B.,et al.: Nasa’s black marble nighttime lights product suite. Remote Sensing of Environment210, 113–143 (2018)

    2018

  35. [43]

    IEEE Transactions on Pattern Analysis and Machine Intelligence38(4), 717–729 (2015)

    Shi, J., Yan, Q., Xu, L., Jia, J.: Hierarchical image saliency detection on extended cssd. IEEE Transactions on Pattern Analysis and Machine Intelligence38(4), 717–729 (2015)

    2015

  36. [44]

    http://www.pascal- network.org/challenges/VOC/voc2012/workshop/index.html

    Everingham, M., Van Gool, L., Williams, C.K.I., Winn, J., Zisserman, A.: The PASCAL Visual Object Classes Challenge 2012 (VOC2012) Results. http://www.pascal- network.org/challenges/VOC/voc2012/workshop/index.html

    2012

  37. [45]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Yan, Q., Xu, L., Shi, J., Jia, J.: Hierarchical saliency detection. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1155– 1162 (2013)

    2013

  38. [46]

    arXiv preprint arXiv:1412.6980 (2014) 32

    Kingma, D.P.: Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014) 32

    2014 arXiv

  39. [47]

    Transactions of the American Mathematical Society68(3), 337–404 (1950)

    Aronszajn, N.: Theory of reproducing kernels. Transactions of the American Mathematical Society68(3), 337–404 (1950)

    1950

  40. [48]

    reproducing kernel

    Shi, J., Malik, J.: Normalized cuts and image segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence22(8), 888–905 (2000) Appendix A Reproducing Kernel Hilbert Space LetDbe a nonempty set, letVbe a real Hilbert space with inner product⟨·,·⟩V, and letV D d...

    2000

Pith tools

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