{"id":"1701fa81-bc3d-450a-af66-8bd5e9b6408c","arxiv_id":"2607.25994","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Richardson extrapolation of grid-based isosurface triangulation volumes yields higher-order accuracy for enclosed volumes across analytic test cases and a simplified electromagnetic heating model.","lead":"This paper presents a numerical trick for computing the volume inside a 3D surface that is known only on a grid: combine a standard triangulated-surface volume estimate at two grid resolutions to cancel the leading error. The method is tested on spheres, ellipsoids, tori, and a skin-heating model, and consistently reduces error with no extra simulation cost.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (9)'s extrapolation rests on an unproved same-C2 quadratic error expansion; in the §4.3 heating test the triangulated isosurface is not closed, so V(Δx) is not well-defined without specifying the cap or x0.","rationale":"The paper's core idea is credible: Richardson extrapolation of triangulated isosurface volumes is a natural and potentially useful technique, and the seven analytic tests consistently support a quadratic leading error on smooth closed surfaces. I do not dispute the mathematical claim for closed isosurfaces. The load-bearing weakness is that the method's domain of validity is not established, and the one non-closed example — the actual electromagnetic-heating application — is exactly where the formulation is incomplete. Eq. (6) is justified by the divergence theorem for a closed surface; for a region bounded by an isosurface and the skin plane z=0, the cap must either be included in the triangulation or the reference point x0 must be placed in the cap plane so that the cap's contribution vanishes. The paper does neither explicitly, making the reported V(Δx) and Vextrap(Δx) non-reproducible. This is a fixable gap rather than evidence that the central idea is wrong, so the reader's CONDITIONAL verdict remains appropriate. My stress test does not move the verdict.","tokens_in":13159,"tokens_out":9282,"duration_ms":97237,"concrete_test":"Reproduce the §4.3 test problem (28) on a grid with Δx=1/8. Compute the sum in Eq. (6) over the isosurface triangles only, using (i) x0=(0,0,0) and (ii) x0=(0,0,1), and (iii) with an explicit triangulated cap on z=0 appended. If (i) and (ii) differ by roughly one-third of the cap area on z=0, or if (iii) differs materially from (i), then the method as written is sensitive to an unspecified parameter and the extrapolated volume in Fig. 10 is not a well-defined approximation to volume(D).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The extrapolation identity (9) is valid only if V(Δx) and V(2Δx) share the same leading quadratic error coefficient C2 (Eqs. 7–8). The paper asserts this by analogy to polygonal curve approximation, but no derivation is given for cell-based marching-cubes-type triangulations of a level set. The assumption is not harmless: it can fail when the isosurface is not a smooth closed surface or when the coarse and fine triangulations have different topology. The paper's own electromagnetic-heating application (§4.3) is a concrete instance of a non-closed surface: the activated region D is bounded by the isosurface and the skin plane z=0. Eq. (6) is derived from the divergence theorem for a closed surface, so applied to the open isosurface triangulation alone it is not a well-defined volume; the tetrahedron sum changes with the arbitrary point x0 unless x0 is chosen in the cap plane. The manuscript never states that choice or explains how the planar cap is included in the triangulation integral. As a result, V(Δx) and V(2Δx) entering Eq. (9) for the headline heating test are not unambiguously defined by the text, and the reported extrapolated accuracy cannot be independently reproduced or verified from the paper alone.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a method for computing the volume of the region where a function defined on a rectangular grid exceeds a threshold. The boundary is approximated by a triangulated isosurface with vertices obtained by linear interpolation on grid lines, and the enclosed volume is computed as one sixth of the sum of signed scalar triple products (Eq. 6). The central device is Richardson extrapolation: from the fine-grid volume V(Δx) and the downsampled coarse-grid volume V(2Δx), Eq. 9 defines Vextrap = (4V(Δx) − V(2Δx))/3. This is justified by assuming the error expansion V(Δx) = Vtrue + C2(Δx)^2 + o((Δx)^2) with the same coefficient C2 for the coarse grid (Eqs. 7–8). The method is tested on seven analytic benchmark volumes — sphere, ellipsoids, superellipsoids, and elliptic ring tori, in aligned and rotated orientations — and on a simplified electromagnetic-heating model with an analytic activated volume. In all tests the error in V(Δx) is reported to scale roughly as (Δx)^2, and extrapolation markedly reduces the error.","tokens_in":13539,"tokens_out":6131,"duration_ms":59412,"significance":"If the extrapolation assumption holds, the method is simple, inexpensive, and practically valuable: it requires no additional PDE solves, only a downsampled coarse representation of an already computed grid function. The numerical evaluation is a genuine strength: the benchmarks are independent analytic solutions, no parameters are fitted to the target volumes, and the extrapolation formula is derived before testing rather than tuned afterward. The convergence behavior is consistent across smooth closed surfaces with nontrivial topology (tori) and across orientations. However, the central error expansion is asserted rather than proved, and the electromagnetic-heating application in §4.3 leaves the treatment of the planar boundary cap unspecified. These two issues bear directly on whether the headline claim — Vextrap = Vtrue + o((Δx)^2) — is established, and they must be addressed before the paper can be accepted.","major_comments":[{"comment":"The entire extrapolation procedure rests on the asymptotic expansion V(Δx) = Vtrue + C2(Δx)^2 + o((Δx)^2) with the same coefficient C2 for the downsampled coarse grid. This is asserted, not derived; the text says only 'Thus, we expect' and draws an analogy with polygonal curve approximation. For cell-based triangulations of a level set, the per-cell error is not obviously of the form C2 times the global mesh spacing squared summed over the surface, and the coarse-grid triangulation is not the same geometric object as the fine-grid triangulation. The equality of C2 between the two grids is especially nontrivial. Since Eq. (9) is the paper's central contribution, I ask for either a local error analysis that yields Eq. (7)–(8), or an explicit statement that extrapolation is a heuristic validated only empirically. The paper should also discuss and test failure modes, such as isosurfaces pass","section":"§2, Eqs. (7)–(9)"},{"comment":"The activated skin region D in the electromagnetic-heating test is bounded by the isosurface and the skin surface z = 0. Equation (6), however, is derived from the divergence theorem for a closed surface S = ∂D. The manuscript states that D is 'enclosed by the isosurface and the skin surface' but never specifies how the planar cap on z = 0 is included in the triangulation or in the tetrahedron sum. If only the isosurface triangles are used, the sum in Eq. (6) is not a well-defined volume and depends on the arbitrary point x0; the reported V(Δx) and V(2Δx) for the headline application then cannot be reproduced or verified from the text. The authors must specify the cap triangulation (or its equivalent) and the choice of x0, and clarify how the finite computational domain is handled.","section":"§4.3, Eq. (28)"}],"minor_comments":[{"comment":"The sentence 'the extrapolation component does not require addition data or information' contains a typo: 'addition' should be 'additional'.","section":"§3.1"},{"comment":"The convergence claims are supported only by visual inspection of log-log plots. Reporting the fitted slopes of the error curves, or a small table of errors at each Δx, would make the O((Δx)^2) statement and the improvement factor quantitative and easier to verify.","section":"Figures 2–10"},{"comment":"The paper is framed as computing the volume 'enclosed by an isosurface,' but in the physical skin-heating problem the activated region is truncated by the skin surface z = 0. This distinction is made in §4.3 but should appear earlier, since it affects the definition of the volume being approximated.","section":"Abstract and §1"},{"comment":"The analytical solution for U(z,t) has a typesetting issue: the exponentials and erfc arguments appear to have lost fraction bars and parentheses. Please check the formula carefully, since it is used to generate the benchmark volume in §4.3.","section":"Eq. (25)"},{"comment":"The summary says both curve and surface approximations 'are expected to produce second order approximations.' This is an honest hedging, but the earlier text and the abstract state the second-order behavior as a fact. Reconcile these statements or add a caveat that Eq. (7) is an assumption to be verified.","section":"§5"}],"recommendation":"major_revision","confidential_remarks":"The paper has a useful practical idea and a solid set of independent numerical tests, but the central extrapolation claim is not proven and the EM-heating test is underspecified at a load-bearing point. I do not think rejection is warranted: the numerical evidence is strong enough that the method is plausible, and the gaps are potentially fixable by adding a theoretical justification (or appropriately hedging the claim) and by specifying the boundary cap treatment. However, as written the paper does not allow an independent reader to reproduce the headline result. Whether the editor judges this acceptable depends on whether the journal permits heuristic numerical methods; if so, a carefully qualified revision might be sufficient, but the current version needs more than cosmetic changes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does one thing and does it cleanly: it applies Richardson extrapolation to the standard triangulated-isosurface volume estimate. Given a fine-grid volume V(Δx) and a downsampled coarse-grid volume V(2Δx), the combination (4V(Δx)-V(2Δx))/3 cancels the leading O(Δx^2) error and gives a more accurate volume at no extra simulation cost. That's a genuinely useful, if modest, trick. The seven analytic test problems (spheres, ellipsoids, superellipsoids, tori, rotated versions) all show clean second-order convergence and consistent extrapolation gains. The divergence-theorem derivation is correct, and the paper is honest: it treats the error expansion as an expectation, not a theorem.\n\nThe soft spots are real but not fatal. First, the asymptotic expansion V(Δx)=V_true+C2(Δx)^2+o((Δx)^2) with a grid-independent C2 is asserted by analogy to curve approximation, not proved for marching-cubes-type triangulations. It is plausible for smooth surfaces in generic position and the numerics support it, but it can fail for surfaces passing through grid vertices or with non-manifold topology. Second—and more concretely—the heating test in §4.3 defines the activated region as bounded by the isosurface and the skin plane z=0, but the paper never says how that planar cap is included in the tetrahedron sum. For an open triangulation, the sum in Eq. (6) depends on the reference point x0 unless x0 is chosen in the cap plane. So V(Δx) for the heating test is not unambiguously defined by the text, and the headline extrapolation results there are not reproducible as written. This is an exposition gap, not a sign the method is wrong; the same issue could affect any clipped isosurface.\n\nMinor: no code or data, no comparison with alternative volume estimators, and the literature on isosurface volume errors is thin. Still, the central idea is credible and the numerical evidence is consistent.\n\nWho it's for: anyone computing volumes of level sets on gridded data—electromagnetic heating safety, medical imaging, computational geometry. It deserves serious peer review; a good referee should push for a proof or precise conditions on the error expansion and an explicit description of the cap handling. I'd send it out.","headline":"A useful, modest numerical trick—Richardson extrapolation on triangulated isosurface volumes—supported by clean experiments; main gap is the underspecified boundary cap in the heating test.","tokens_in":13912,"tokens_out":4268,"would_cite":true,"duration_ms":38482,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65D18","65B05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Combining fine- and coarse-grid triangulated volumes cancels the leading error and gives higher-order-accurate isosurface volumes at no extra simulation cost.","keywords":["isosurface","volume computation","triangulation","extrapolation","divergence theorem","electromagnetic heating","activated volume","rectangular grid"],"falsifier":"For a known-volume isosurface, compute V(Δx) and V(2Δx) on a sequence of nested grids and form the ratio (V(2Δx) − V_true)/(V(Δx) − V_true). If the claimed expansion holds, this ratio should approach 4; persistent deviations from 4 across resolutions—for example, for an isosurface aligned to pass through grid vertices—would falsify the method's accuracy claim.","tokens_in":13059,"feed_emoji":"🔥","tokens_out":5474,"duration_ms":50480,"temperature":0.7,"pith_summary":"This paper claims that the volume enclosed by an isosurface of a function known only on a rectangular grid can be computed to higher accuracy by triangulating the isosurface, computing the enclosed volume from the divergence theorem as a sum of tetrahedron volumes, and then extrapolating the fine-grid result against the volume obtained from a downsampled coarse grid. The extrapolated combination Vextrap = (4V(Δx) − V(2Δx))/3 cancels the dominant second-order error of the raw triangulation, leaving an error of order o((Δx)^2) rather than O((Δx)^2). Seven test problems spanning spheres, ellipsoids, superellipsoids, ring tori, and a model of electromagnetic skin heating support the claim. The importance is practical: the activated skin volume in millimeter-wave exposure—where temperature is numerically computed on a grid—can be assessed more accurately using information already present in the solution, without refining the simulation.","feed_headline":"Downsampling once sharpens isosurface volume accuracy","feed_subtitle":"Extrapolating a fine-grid triangulation against its coarse-grid twin removes the leading error with zero extra simulations.","key_machinery":"The central mechanism is extrapolation applied to a volume computed from a triangulated isosurface. The volume is expressed via the divergence theorem as one third of a surface integral of (x − x0)·dS; over each planar triangle this integral reduces to one sixth of a triple scalar product, so the enclosed volume is a summed set of tetrahedron volumes. The key identity is the assumed second-order error expansion V(Δx) = V_true + C2(Δx)^2 + o((Δx)^2) for both the fine grid and the downsampled coarse grid, which makes the combination (4V(Δx) − V(2Δx))/3 cancel C2(Δx)^2.","core_discovery":"On the paper's own terms, the discovery is the identity Vextrap(Δx) ≡ (4V(Δx) − V(2Δx))/3 = V_true + o((Δx)^2), where V(Δx) is the volume enclosed by the linear-interpolation triangulation of the isosurface on a grid of spacing Δx and V(2Δx) is the same quantity computed after downsampling the grid. Because V(Δx) has the asymptotic form V_true + C2(Δx)^2 + o((Δx)^2) and V(2Δx) has the same leading coefficient C2, the combination eliminates the leading error term. The paper's seven analytic test problems all show raw triangulation error following a C2(Δx)^2 trend and extrapolated error far below it, including for a non-symmetric activated-skin-volume problem.","pith_inferences":["(Editorial inference) Because extrapolation can be applied recursively, combining V(Δx), V(2Δx), and V(4Δx) could push the error to higher order than the paper's single-step formula, at the cost of requiring an additional downsampling.","(Editorial inference) The assumed expansion may fail when the isosurface passes through grid vertices, has sharp creases, or is clipped by a boundary—conditions near the skin surface z = 0 in the application, where the paper does not specify how the boundary cap is included in the triangulation integral.","(Editorial inference) The same triangulation-plus-extrapolation recipe could improve other surface integrals over isosurfaces, such as integrated heat flux or surface area, but the paper does not test those."],"forward_implications":["A single fine-grid temperature solution yields both V(Δx) and, by downsampling, V(2Δx); no extra PDE solve is needed for the accuracy gain.","In every test case the extrapolated volume error is substantially smaller than the raw triangulation error, so the method can replace much finer grids in practice.","The procedure requires no symmetry and applies to any discrete 3D function, including the non-axially-symmetric skin temperature distributions that arise with moving or oblique beams.","More accurate activated volumes imply better estimates of the thermal nociceptor activation region and potential thermal injury in millimeter-wave exposure assessments.","The observed second-order trend in all seven problems supports using the extrapolation as a cheaper accuracy boost for isosurface volume computation."],"fun_headline_variants":["Two-grid extrapolation sharpens isosurface volume","Extrapolated triangulation cuts volume error","Grid extrapolation boosts isosurface volume accuracy","Volume computation refined via two-grid extrapolation","Extrapolation halves error in activated volume"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper assumes, without proof, that the triangulated-volume error has the form V(Δx) = V_true + C2(Δx)^2 + o((Δx)^2) and that the coefficient C2 is identical for the fine grid and the downsampled coarse grid; the expansion is inferred from the analogy with line-segment curve approximation and can break down for surfaces passing through grid vertices, non-manifold topology, or boundary-clipped domains.","fun_headline_variants_meta":{"raw":{"variants":["Two-grid extrapolation sharpens isosurface volume","Extrapolated triangulation cuts volume error","Grid extrapolation boosts isosurface volume accuracy","Volume computation refined via two-grid extrapolation","Extrapolation halves error in activated volume"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000134,"raw_usage":{"total_tokens":985,"prompt_tokens":759,"completion_tokens":226,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":503,"completion_tokens_details":{"reasoning_tokens":167}},"tokens_in":503,"tokens_out":226,"duration_ms":3112,"temperature":1.0,"reasoning_tokens":167,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T00:52:59.318770+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a known-volume isosurface, compute V(Δx) and V(2Δx) on a sequence of nested grids and form the ratio (V(2Δx) − V_true)/(V(Δx) − V_true). If the claimed expansion holds, this ratio should approach 4; persistent deviations from 4 across resolutions—for example, for an isosurface aligned to pass through grid vertices—would falsify the method's accuracy claim.","supporting_citations":[],"review_version":1}