pith. sign in

arxiv: 2508.20845 · v2 · submitted 2025-08-28 · 🧮 math.AP

Homogenisation of phase-field functionals with linear growth

Francesco Colasanto , Matteo Focardi , Caterina Ida Zeppieri This is my paper

Pith reviewed 2026-05-18 20:39 UTC · model grok-4.3

classification 🧮 math.AP
keywords homogenizationphase-fieldfree-discontinuitylinear growthAmbrosio-TortorelliGamma-convergencestationary randomimage segmentation
0
0 comments X

The pith

Phase-field functionals with linear growth in the gradient homogenise to free-discontinuity energies whose surface term depends explicitly on jump amplitude.

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

The paper shows that Ambrosio-Tortorelli-type elliptic functionals, when the regularised volume term grows linearly rather than superlinearly in the gradient, undergo homogenisation to a free-discontinuity limit energy. In this limit the interfacial cost is no longer a fixed perimeter but varies directly with the size of the jump that the limit function makes. The result holds under mild integrand assumptions that include the important case of stationary random coefficients. A reader would care because this supplies the first rigorous upscaling procedure for image-segmentation models that use linear regularisation, which behave differently from the classical quadratic-growth case.

Core claim

The functionals homogenise to a free-discontinuity energy whose surface term explicitly depends on the jump amplitude of the limit variable. The convergence is proved under very mild assumptions on the integrands that permit stationary random coefficients.

What carries the argument

Gamma-convergence of linear-growth Ambrosio-Tortorelli functionals to a jump-amplitude-dependent free-discontinuity energy

If this is right

  • Homogenised models can now incorporate explicit jump-size dependence in the surface energy for segmentation tasks.
  • Stationary random media become admissible without requiring periodicity or strong mixing assumptions.
  • The linear-growth regime produces a qualitatively different limit from the superlinear case previously studied.
  • The procedure supplies a rigorous bridge between fine-scale oscillations and macroscopic discontinuities.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The amplitude-dependent surface term may alter optimal jump profiles in heterogeneous media, suggesting new numerical tests on random microstructures.
  • The same linear-growth setup could be examined for its effect on crack-path selection in materials with random defects.
  • Extensions to vectorial or anisotropic integrands would test whether the jump dependence survives beyond the scalar elliptic setting.

Load-bearing premise

The volume term grows linearly in the gradient variable, together with mild integrand assumptions that allow stationary random coefficients.

What would settle it

A concrete computation of the Gamma-limit for a simple periodic integrand in which the surface density is shown to be independent of jump height, or a counter-example sequence of stationary random integrands satisfying the paper's assumptions for which the claimed convergence fails.

read the original abstract

We propose a first rigorous homogenisation procedure in image-segmentation models by analysing the relative impact of (possibly random) fine-scale oscillations and phase-field regularisations for a family of elliptic functionals of Ambrosio and Tortorelli type, when the regularised volume term grows \emph{linearly} in the gradient variable. In contrast to the more classical case of superlinear growth, we show that our functionals homogenise to a free-discontinuity energy whose surface term explicitly depends on the jump amplitude of the limit variable. The convergence result as above is obtained under very mild assumptions which allow us to treat, among other, the case of \emph{stationary random integrands}.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper establishes a Gamma-convergence result for a family of Ambrosio-Tortorelli-type phase-field functionals with linear growth in the gradient term and stationary random coefficients. It shows that the functionals converge to a free-discontinuity energy whose surface term depends explicitly on the jump amplitude of the limit function, under mild integrand assumptions that permit random media.

Significance. If the result holds, it provides the first rigorous homogenization analysis for linear-growth phase-field models in image segmentation, extending classical superlinear-growth results and accommodating stationary random integrands. This is a notable technical advance for heterogeneous media applications, with the explicit jump-amplitude dependence in the surface energy offering a sharper limit than standard perimeter functionals.

major comments (2)
  1. [§4.2] §4.2, lower-bound argument for the surface term: the construction separating the diffuse-interface contribution from the jump appears to rely on a cut-off function whose gradient control is justified only by linear growth; it is unclear whether this preserves the explicit dependence on |u^+ - u^-| without additional higher-integrability or convexity assumptions that the mild hypotheses omit. This is load-bearing for the central claim.
  2. [Theorem 3.1] Theorem 3.1 (Gamma-convergence statement): the surface density σ(x,ν,s) is asserted to depend on the jump size s, but the proof sketch does not exhibit the explicit formula or the test-function family that would confirm this dependence survives the random homogenization; a concrete example or explicit computation for a periodic case would strengthen the result.
minor comments (2)
  1. [§2] Notation for the random integrands (e.g., the stationary ergodic assumption) is introduced without a dedicated preliminary subsection; a short paragraph recalling the precise definition of stationarity would improve readability.
  2. Figure 1 caption refers to 'numerical illustration' but no corresponding figure or caption details appear in the provided text; please ensure all figures are present and labeled consistently.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of the manuscript and the positive evaluation of its significance. We address the two major comments point by point below, providing clarifications on the technical points raised and indicating the revisions we will make to strengthen the presentation.

read point-by-point responses

  1. Referee: [§4.2] §4.2, lower-bound argument for the surface term: the construction separating the diffuse-interface contribution from the jump appears to rely on a cut-off function whose gradient control is justified only by linear growth; it is unclear whether this preserves the explicit dependence on |u^+ - u^-| without additional higher-integrability or convexity assumptions that the mild hypotheses omit. This is load-bearing for the central claim.

    Authors: We thank the referee for highlighting this aspect of the lower-bound construction. The cut-off is supported in a tubular neighborhood of width proportional to the phase-field parameter ε around the approximate jump set. Because the integrand has linear growth in the gradient, the contribution of ∇(cut-off) to the energy is bounded by a constant (depending on the jump amplitude s) times the measure of the support; this estimate follows directly from the linear-growth assumption and the stationarity of the integrand without invoking higher integrability or convexity. The explicit dependence on s is retained because the comparison is performed against the homogenized functional evaluated on a function whose jump is exactly of size s, and the subsequent application of the ergodic theorem in the homogenization step preserves the s-dependence in the limit surface density. We have revised §4.2 to spell out this estimate in greater detail. revision: yes

  2. Referee: [Theorem 3.1] Theorem 3.1 (Gamma-convergence statement): the surface density σ(x,ν,s) is asserted to depend on the jump size s, but the proof sketch does not exhibit the explicit formula or the test-function family that would confirm this dependence survives the random homogenization; a concrete example or explicit computation for a periodic case would strengthen the result.

    Authors: The surface density σ(x,ν,s) is introduced variationally in the proof of Theorem 3.1 as the Γ-limit (after suitable rescaling and blow-up) of the phase-field energies for sequences whose limit has a jump of amplitude s across a hyperplane with normal ν. The recovery sequences are built from the standard Ambrosio-Tortorelli transition profiles, rescaled and modulated by the stationary random coefficients; the s-dependence enters through the height of these transitions and survives the homogenization by the ergodic theorem. While the main text contains a sketch, the full construction appears in the appendix. To make the dependence more transparent, we have added a new remark after Theorem 3.1 containing an explicit computation in the periodic case (a simple one-dimensional periodic integrand) that yields a closed-form expression for σ(s) and confirms the dependence on the jump amplitude. revision: yes

Circularity Check

0 steps flagged

No circularity: direct rigorous homogenization derivation

full rationale

The paper conducts a standard Gamma-convergence analysis for Ambrosio-Tortorelli-type functionals with linear growth in the gradient term, deriving the limit free-discontinuity energy (including an explicit jump-amplitude dependence in the surface density) from the original sequence of functionals. The proof relies on unfolding or two-scale techniques adapted to stationary random coefficients and linear coercivity, without any self-definitional steps, fitted parameters renamed as predictions, or load-bearing self-citations whose validity reduces to the present work. The result is self-contained against external benchmarks such as classical homogenization theory for free-discontinuity problems and does not invoke uniqueness theorems or ansatzes from the authors' prior papers in a circular manner.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The result rests on standard background from calculus of variations and Gamma-convergence theory together with the linear-growth modeling choice; no new free parameters or invented entities are introduced.

axioms (2)
  • standard math Standard ellipticity and growth conditions for Ambrosio-Tortorelli type functionals
    Invoked to ensure the functionals are well-defined and the homogenization limit exists.
  • domain assumption Very mild assumptions on the integrands permitting stationary random coefficients
    Abstract states these assumptions allow treatment of random media.

pith-pipeline@v0.9.0 · 5639 in / 1258 out tokens · 38298 ms · 2026-05-18T20:39:49.952334+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages

  1. [1]

    Ergodic theorems for superadditive processes

    Mustafa Akcoglu and Ulrich Krengel. Ergodic theorems for superadditive processes. J. Reine Angew. Math. , 323:53–67, 1981

    work page 1981

  2. [2]

    Rank one property for derivatives of functions with bounded variation

    Giovanni Alberti. Rank one property for derivatives of functions with bounded variation. Proc. Roy. Soc. Edinburgh Sect. A, 123(2):239–274, 1993

    work page 1993

  3. [3]

    Phase-field modelling of cohesive fracture

    Roberto Alessi, Francesco Colasanto, and Matteo Focardi. Phase-field modelling of cohesive fracture. Part I: Γ-convergence results. preprint, 2025

    work page 2025

  4. [4]

    Phase-field modelling of cohesive fracture

    Roberto Alessi, Francesco Colasanto, and Matteo Focardi. Phase-field modelling of cohesive fracture. Part II: Reconstruction of the cohesive law. preprint, 2025

    work page 2025

  5. [5]

    Phase-field modelling of cohesive fracture

    Roberto Alessi, Francesco Colasanto, and Matteo Focardi. Phase-field modelling of cohesive fracture. Part III: From mathematical results to engineering applications. preprint, 2025

    work page 2025

  6. [6]

    Free-discontinuity problems via functionals involving the L1-norm of the gradient and their approximations

    Roberto Alicandro, Andrea Braides, and Jayant Shah. Free-discontinuity problems via functionals involving the L1-norm of the gradient and their approximations. Interfaces Free Bound., 1(1):17–37, 1999

    work page 1999

  7. [7]

    Variational approximation of free-discontinuity energies with linear growth

    Roberto Alicandro and Matteo Focardi. Variational approximation of free-discontinuity energies with linear growth. Commun. Contemp. Math. , 4(4):685–723, 2002. 50 F. COLASANTO, M. FOCARDI, AND C. I. ZEPPIERI

    work page 2002

  8. [8]

    On the relaxation in BV (Ω; Rm) of quasi-convex integrals

    Luigi Ambrosio and Gianni Dal Maso. On the relaxation in BV (Ω; Rm) of quasi-convex integrals. Journal of functional analysis, 109(1):76–97, 1992

    work page 1992

  9. [9]

    Functions of bounded variation and free discontinuity prob- lems

    Luigi Ambrosio, Nicola Fusco, and Diego Pallara. Functions of bounded variation and free discontinuity prob- lems. Oxford Mathematical Monographs. The Clarendon Press, Oxford University Press, New York, 2000

    work page 2000

  10. [10]

    Approximation of functional depending on jumps by elliptic functional via Γ-convergence

    Luigi Ambrosio and Vincenzo Maria Tortorelli. Approximation of functional depending on jumps by elliptic functional via Γ-convergence. Communications on Pure and Applied Mathematics , 43(8):999–1036, 1990

    work page 1990

  11. [11]

    On the approximation of free discontinuity problems

    Luigi Ambrosio and Vincenzo Maria Tortorelli. On the approximation of free discontinuity problems. Boll. Un. Mat. Ital. B (7) , 6(1):105–123, 1992

    work page 1992

  12. [12]

    Quantitative analysis of finite-difference approxima- tions of free-discontinuity problems

    Annika Bach, Andrea Braides, and Caterina Ida Zeppieri. Quantitative analysis of finite-difference approxima- tions of free-discontinuity problems. Interfaces Free Bound., 22(3):317–381, 2020

    work page 2020

  13. [13]

    Random finite-difference discretizations of the Ambrosio- Tortorelli functional with optimal mesh-size

    Annika Bach, Marco Cicalese, and Matthias Ruf. Random finite-difference discretizations of the Ambrosio- Tortorelli functional with optimal mesh-size. SIAM J. Math. Anal. , 53(2):2275–2318, 2021

    work page 2021

  14. [14]

    Interaction between oscillations and singular perturbations in a one-dimensional phase-field model

    Annika Bach, Teresa Esposito, Roberta Marziani, and Caterina Ida Zeppieri. Interaction between oscillations and singular perturbations in a one-dimensional phase-field model. In Research in mathematics of materials science, volume 31 of Assoc. Women Math. Ser. , pages 3–31. Springer, Cham, [2022] ©2022

    work page 2022

  15. [15]

    Gradient damage models for heterogeneous materials

    Annika Bach, Teresa Esposito, Roberta Marziani, and Caterina Ida Zeppieri. Gradient damage models for heterogeneous materials. SIAM J. Math. Anal. , 55(4):3567–3601, 2023

    work page 2023

  16. [16]

    Γ-convergence and stochastic homogenisation of singularly-perturbed elliptic functionals

    Annika Bach, Roberta Marziani, and Caterina Ida Zeppieri. Γ-convergence and stochastic homogenisation of singularly-perturbed elliptic functionals. Calc. Var. Partial Differential Equations , 62(7):Paper No. 199, 54, 2023

    work page 2023

  17. [17]

    A global method for relaxation

    Guy Bouchitt´ e, Irene Fonseca, and Maria Luisa Mascarenhas. A global method for relaxation. Arch. Rational Mech. Anal., 145(1):51–98, 1998

    work page 1998

  18. [18]

    Blaise Bourdin, Gilles Andre Francfort, and J.-J. Marigo. The Variational Approach to Fracture. Journal of Elasticity, 91(1):5–148, mar 2008

    work page 2008

  19. [19]

    Numerical experiments in revisited brittle fracture

    Blaise Bourdin, Gilles Andre Francfort, and Jean-Jacques Marigo. Numerical experiments in revisited brittle fracture. J. Mech. Phys. Solids , 48(4):797–826, 2000

    work page 2000

  20. [20]

    Homogenization of multiple integrals, volume 12 of Oxford Lecture Series in Mathematics and its Applications

    Andrea Braides and Anneliese Defranceschi. Homogenization of multiple integrals, volume 12 of Oxford Lecture Series in Mathematics and its Applications . The Clarendon Press, Oxford University Press, New York, 1998

    work page 1998

  21. [21]

    Second-order edge-penalization in the Ambrosio- Tortorelli functional

    Martin Burger, Teresa Esposito, and Caterina Ida Zeppieri. Second-order edge-penalization in the Ambrosio- Tortorelli functional. Multiscale Model. Simul., 13(4):1354–1389, 2015

    work page 2015

  22. [22]

    Γ-convergence of free- discontinuity problems

    Filippo Cagnetti, Gianni Dal Maso, Lucia Scardia, and Caterina Ida Zeppieri. Γ-convergence of free- discontinuity problems. Ann. Inst. H. Poincar´ e C Anal. Non Lin´ eaire, 36(4):1035–1079, 2019

    work page 2019

  23. [23]

    Stochastic homogenisation of free-discontinuity problems

    Filippo Cagnetti, Gianni Dal Maso, Lucia Scardia, and Caterina Ida Zeppieri. Stochastic homogenisation of free-discontinuity problems. Arch. Ration. Mech. Anal., 233(2):935–974, 2019

    work page 2019

  24. [24]

    A global method for deterministic and stochastic homogenisation in BV

    Filippo Cagnetti, Gianni Dal Maso, Lucia Scardia, and Caterina Ida Zeppieri. A global method for deterministic and stochastic homogenisation in BV . Ann. PDE, 8(1):Paper No. 8, 89, 2022

    work page 2022

  25. [25]

    Phase-field approximation of functionals defined on piecewise-rigid maps

    Marco Cicalese, Matteo Focardi, and Caterina Ida Zeppieri. Phase-field approximation of functionals defined on piecewise-rigid maps. J. Nonlinear Sci. , 31(5):Paper No. 78, 25, 2021

    work page 2021

  26. [26]

    Phase field approximation of cohesive functionals in the vectorial case

    Francesco Colasanto. Phase field approximation of cohesive functionals in the vectorial case. in preparation, 2025

    work page 2025

  27. [27]

    Phase field approximation of cohesive fracture models

    Sergio Conti, Matteo Focardi, and Flaviana Iurlano. Phase field approximation of cohesive fracture models. Annales de l’Institut Henri Poincar´ e / Analyse non lin´ eaire, 33:1033–1067, 2016

    work page 2016

  28. [28]

    Phase-field approximation of a vectorial, geometrically nonlinear cohesive fracture energy

    Sergio Conti, Matteo Focardi, and Flaviana Iurlano. Phase-field approximation of a vectorial, geometrically nonlinear cohesive fracture energy. Arch. Ration. Mech. Anal., 248(2):Paper No. 21, 60, 2024

    work page 2024

  29. [29]

    Superlinear free-discontinuity models: relaxation and phase field approximation

    Sergio Conti, Matteo Focardi, and Flaviana Iurlano. Superlinear free-discontinuity models: relaxation and phase field approximation. preprint arXiv:2505.00852, 2025

    work page arXiv 2025

  30. [30]

    An introduction to Γ-convergence

    Gianni Dal Maso. An introduction to Γ-convergence. Progress in Nonlinear Differential Equations and their Applications, 8. Birkh¨ auser Boston Inc., Boston, MA, 1993

    work page 1993

  31. [31]

    Endowing explicit cohesive laws to the phase-field fracture theory

    Ye Feng, Jiadi Fan, and Jie Li. Endowing explicit cohesive laws to the phase-field fracture theory. Journal of the Mechanics and Physics of Solids , 152(March):104464, 2021

    work page 2021

  32. [32]

    On the variational approximation of free-discontinuity problems in the vectorial case

    Matteo Focardi. On the variational approximation of free-discontinuity problems in the vectorial case. Math. Models Methods Appl. Sci. , 11(4):663–684, 2001

    work page 2001

  33. [33]

    Revisiting brittle fracture as an energy minimization problem

    Gilles Andre Francfort and Jean-Jacques Marigo. Revisiting brittle fracture as an energy minimization problem. J. Mech. Phys. Solids , 46(8):1319–1342, 1998

    work page 1998

  34. [34]

    Lammen, S

    H. Lammen, S. Conti, and J. Mosler. Approximating arbitrary traction–separation-laws by means of phase-field theory — Mathematical foundation and numerical implementation. Journal of the Mechanics and Physics of Solids, 197:106038, apr 2025. HOMOGENISATION OF PHASE-FIELD FUNCTIONALS WITH LINEAR GROWTH 51

    work page 2025

  35. [35]

    Quasiconvexification in W 1,1 and optimal jump microstructure in BV relaxation

    Christopher Larsen. Quasiconvexification in W 1,1 and optimal jump microstructure in BV relaxation. SIAM J. Math. Anal. , 29(4):823–848, 1998

    work page 1998

  36. [36]

    Global-local subadditive ergodic theorems and application to homoge- nization in elasticity

    Christian Licht and G´ erard Michaille. Global-local subadditive ergodic theorems and application to homoge- nization in elasticity. Ann. Math. Blaise Pascal , 9(1):21–62, 2002

    work page 2002

  37. [37]

    Optimal approximations by piecewise smooth functions and associated variational problems

    David Mumford and Jayant Shah. Optimal approximations by piecewise smooth functions and associated variational problems. Comm. Pure Appl. Math. , 42(5):577–685, 1989

    work page 1989

  38. [38]

    The variational approach to damage: I

    Kim Pham and Jean-Jacques Marigo. The variational approach to damage: I. The foundations.Comptes Rendus Mecanique, 338(4):191–198, 2010

    work page 2010

  39. [39]

    The variational approach to damage: II

    Kim Pham and Jean-Jacques Marigo. The variational approach to damage: II. The gradient damage models. Comptes Rendus Mecanique, 338(4):199–206, apr 2010

    work page 2010

  40. [40]

    Stochastic homogenization of degenerate integral functionals and their Euler- Lagrange equations

    Matthias Ruf and Thomas Ruf. Stochastic homogenization of degenerate integral functionals and their Euler- Lagrange equations. J. ´Ec. polytech. Math., 10:253–303, 2023

    work page 2023

  41. [41]

    Stochastic homogenization of degenerate integral functionals with linear growth

    Matthias Ruf and Caterina Ida Zeppieri. Stochastic homogenization of degenerate integral functionals with linear growth. Calc. Var. Partial Differential Equations , 62(4):Paper No. 138, 36, 2023

    work page 2023

  42. [42]

    Shape recovery from noisy images by curve evolution

    Jayant Shah. Shape recovery from noisy images by curve evolution. IASTED International Conference on Signal and Image Processing (SIP-95) , 1:461–464, 1995

    work page 1995

  43. [43]

    Common framework for curve evolution, segmentation and anisotropic diffusion

    Jayant Shah. Common framework for curve evolution, segmentation and anisotropic diffusion. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition , pages 136–142, 1996

    work page 1996

  44. [44]

    Curve evolution and segmentation functionals: Application to color images

    Jayant Shah. Curve evolution and segmentation functionals: Application to color images. IEEE International Conference on Image Processing, 1:461–464, 1996

    work page 1996

  45. [45]

    Uses of elliptic approximation in Computer Vision , volume 25 of Progr

    Jayant Shah. Uses of elliptic approximation in Computer Vision , volume 25 of Progr. Nonlinear Differential Equations Appl. Birkh¨ auser, Basel, 1996

    work page 1996

  46. [46]

    Riemannian drums, anisotropic curve evolution, and segmentation

    Jayant Shah. Riemannian drums, anisotropic curve evolution, and segmentation. Journal of Visual Communi- cation and Image Representation , pages 142–153, 2000

    work page 2000

  47. [47]

    A unified phase-field theory for the mechanics of damage and quasi-brittle failure

    Jian-Ying Wu. A unified phase-field theory for the mechanics of damage and quasi-brittle failure. J. Mech. Phys. Solids, 103(Supplement C):72–99, 2017

    work page 2017

  48. [48]

    A length scale insensitive phase-field damage model for brittle fracture

    Jian-Ying Wu and Vinh Phu Nguyen. A length scale insensitive phase-field damage model for brittle fracture. Journal of the Mechanics and Physics of Solids , 119:20–42, 2018. (Francesco Colasanto) DiMaI U. Dini, Universit `a di Firenze, V.le G.B. Morgagni 67/A, 50134 Firenze, Italy Email address, Francesco Colasanto: francesco.colasanto@unifi.it (Matteo Fo...

    work page 2018