{"id":"339ac627-2b79-4c60-a50d-57e10f50cee5","arxiv_id":"2501.10474","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"Poxel is a voxel-based reconstruction pipeline that outputs printable CMYKWCl voxel grids, with no quantitative validation in the paper.","lead":"Poxel converts digital 3D reconstructions into voxel grids that use a printer-friendly CMYKWCl color model for photopolymer jetting printers. The paper claims better printed-model fidelity than an adapted 3D Gaussian Splatting pipeline, but provides only qualitative evidence.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper never specifies or validates the RGBA-to-CMYKWCl discretization that the fidelity claim rests on, so the central result is currently unfalsifiable.","rationale":"The reader's weakest assumption already identified the color-averaging and gamut-mapping step as untested. My concern sharpens that point: the paper does not even specify the discretization mapping, so it cannot be tested. This is the single most load-bearing issue because the entire contribution is the direct, printable CMYKWCl output. If that mapping is arbitrary or incorrect, the claimed fidelity advantage collapses; if it is omitted, the work is not reproducible. The paper's own text in Section III-C only promises \"efficient, parallelized processing\" without revealing the conversion, and Section IV offers only qualitative visual comparison. No code, no supplementary material, and no measured color data are provided. My proposed test is deliberately simple: use the authors' mapping to print a standard color target and measure the color error. That directly settles whether the central fidelity claim has any basis. I do not see a need to change the reader's REJECT verdict; the missing specification and validation are sufficient grounds for rejection, and my concern reinforces rather than redirects the reader's reasoning.","tokens_in":3985,"tokens_out":2229,"duration_ms":24165,"concrete_test":"Obtain the authors' exact RGBA-to-CMYKWCl mapping (function or lookup table). Print a standardized color target (e.g., a ColorChecker or IT8.7/4 chart) on the same Stratasys J850 using Poxel's pipeline, a baseline 3DGS-converted pipeline, and the printer's native color management. Measure all printed patches with a spectrophotometer and compute CIEDE2000 errors against the digital target colors. If Poxel's mean or 95th-percentile ΔE is not significantly lower than the baseline, the fidelity claim fails. Also verify whether Equation (1) with N=12 is actually used in the mapping; if not, the method description is internally inconsistent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that Poxel achieves superior printed color fidelity by directly outputting CMYKWCl voxel grids. The load-bearing link in that chain is the color discretization step described in Section III-C. The paper states that RGB colors are \"discretized into CMYKWCl\" but gives no algorithm, lookup table, or mathematical relation between the averaged RGBA values from Equation (1) and the printer's CMYKWCl ink mixture. Without that mapping, the method cannot be reproduced, and the claimed \"superior alignment with the printer's color gamut\" is an assertion, not a demonstrated result. The only evidence offered is two printed models (Fig. 2) compared subjectively; no measured color patches, no colorimetric error metric, and no comparison against a standard color-management baseline appear anywhere in Sections IV.A or IV.B. This is not merely a missing quantitative evaluation: the core mechanism of the paper is under-specified. Even if the 1x6x2 regional averaging in Section III-D were correct, the subsequent discretization is a black box. A reader cannot tell whether the claimed fidelity comes from the voxel representation, from the printer's native color management, or from an arbitrary mapping that happened to work on two test models. Until the mapping is specified and validated against measured print output, the central claim is unfalsifiable.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes 'Poxel,' a voxel-based reconstruction and color-conversion pipeline for full-color multi-material jetting 3D printing. The method replaces view-dependent RGB radiance fields with view-independent CMYKWCl voxels, uses anisotropic voxels and regional color averaging, and directly outputs printable voxel grids. The evaluation consists of two printed models on a Stratasys J850 and a qualitative comparison against a 3D Gaussian Splatting baseline. The core technical steps, particularly the RGB-to-CMYKWCl discretization and the optimization losses, are not specified, and the reported fidelity improvement is supported only by subjective statements.","tokens_in":4269,"tokens_out":4726,"duration_ms":45093,"significance":"If fully specified and validated, a direct pipeline from digital reconstruction to printer-native CMYKWCl voxel grids would be a useful contribution to color 3D printing, potentially avoiding conversion artifacts. The use of a real multi-material printer and physical test prints is a concrete strength. However, in its current form the contribution cannot be reproduced or quantitatively assessed, and the central claim of improved fidelity is not supported by the evidence presented.","major_comments":[{"comment":"The RGB-to-CMYKWCl discretization is not specified. The text states that colors are 'discretized into CMYKWCl' but gives no algorithm, lookup table, or mathematical relationship between the averaged RGBA value from Eq. (1) and the printer's ink mixture. This mapping is the core of the claimed fidelity improvement, so its absence makes the pipeline unreproducible and the central result unfalsifiable.","section":"III-C"},{"comment":"The results are based on subjective observations of two printed models. There are no colorimetric measurements, no ΔE or other error metrics, no error bars, and no controlled baseline with a standard color-management workflow. The statement in Section IV-B that 'the Poxel model provided superior alignment with the printer's color gamut' is therefore an assertion, not a demonstrated finding. Quantitative evaluation against measured print patches is required before any fidelity claim can be accepted.","section":"IV / IV-B"},{"comment":"The optimization is under-specified. Equation (1) defines a regional averaging operation, but it is not a loss function, and the 'structural loss' mentioned in the same section is never defined. There is no description of the training procedure, hyperparameters, voxel grid resolution, or how the averaged colors are used in the optimization. Without these details, the method cannot be reproduced or compared.","section":"III-D"},{"comment":"The anisotropic voxel configuration is not described quantitatively. The paper claims that anisotropic voxels enable finer control over layer resolution and color blending, but it never states the actual voxel dimensions or how they map to the printer's 600x300 dpi planarity and 0.014 mm layer thickness. The only numeric parameter introduced later, the 1x6x2 averaging window in Section III-D, is presented without justification or sensitivity analysis.","section":"III-B"}],"minor_comments":[{"comment":"The headings and text contain spacing artifacts ('V olume', 'V oxel', 'T raining'); these should be fixed for readability.","section":"III-A, III-B, II-A"},{"comment":"The experimental protocol description is inconsistent: the text mentions 'two batches' of printed models, then 'first test' and 'second test', while Fig. 3 refers to a 'Third attempt'. Please clarify how many models were printed and which attempts correspond to which results.","section":"IV"},{"comment":"Equation (1) averages RGBA values but does not state whether the values are in linear or gamma-encoded space; this distinction is important for color mixing and should be specified.","section":"III-D, Eq. (1)"},{"comment":"The 3DGS baseline adaptation is described only qualitatively ('converting 2D splats to 3D ellipsoidal volumes, slicing along the height axis, and mapping RGB values to a CMYKWCl palette'). More detail is needed to interpret the comparison and to assess whether the baseline is a fair one.","section":"IV-A"},{"comment":"References [4]-[6] are vendor web pages; archival citations or DOIs, if available, would be more durable and useful to readers.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript reads as an extended abstract rather than a complete research paper. The missing discretization algorithm, undefined loss functions, and lack of quantitative evaluation are substantial but addressable omissions. If the authors can supply full methodological detail, measured colorimetric results, and a controlled comparison, the paper could become a valuable contribution. I recommend major revision rather than rejection because the identified problems are omissions that can be fixed within the scope of the work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nWhat you should know: this is a short paper that grafts Plenoxel-style voxel reconstruction onto a CMYKWCl multi-material jetting workflow, and actually prints two models on a Stratasys J850. That is the whole contribution: a direct printable-voxel pipeline, removing view-dependent RGB in favor of view-independent CMYKWCl voxels. That framing is sensible, and the authors clearly understand why NeRF/3DGS outputs are not directly printable.\n\nCredit where due: they did real fabrication experiments, not just simulations. The anisotropic voxel idea is reasonable, though not new; the 1x6x2 regional averaging is a simple heuristic but at least stated. The paper is honest about being a first attempt.\n\nThe soft spots are large. The core color discretization step—RGB to CMYKWCl—is described as \"discretizes\" with no mapping, algorithm, lookup table, or measured patch data. That is the load-bearing link for the fidelity claim, and it is a black box. The optimization section mentions \"loss functions\" but never writes them down; only Eq. (1) for averaging. The evaluation is two printed models compared subjectively. No colorimetric measurements, no error bars, no baseline beyond a hand-rolled 3DGS conversion. There are also missing references to existing color 3D printing pipelines (e.g., Cuttlefish or other gamut-mapping work) that should have been compared.\n\nThe stress-test note is right: with the discretization unspecified, the central claim is currently unfalsifiable. A reader cannot tell whether the improved fidelity comes from the voxel representation, the printer's built-in color management, or an arbitrary mapping that worked on two models.\n\nIs there enough here for a referee? I think yes, with caution. The problem is real, the pipeline is plausible, and the missing pieces are fixable in principle. But as it stands this is a demo, not a validated method. If it went to review, it would need major revision: specify the discretization, report measured color errors, compare against a standard color-managed pipeline, and release code or data.\n\nI would not cite it yet, but I'd keep it in mind. Worth a look for anyone working on fabrication from radiance fields, not for a general graphics audience.\n\nRecommendation: send to peer review rather than desk reject, on the understanding that it will likely need heavy revision before it can be accepted.","headline":"A plausible pipeline idea that is under-specified and renders its central fidelity claim unverifiable.","tokens_in":4755,"tokens_out":1870,"would_cite":false,"duration_ms":17686,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proposes Poxel, a voxel-based reconstruction framework that outputs print-ready CMYKWCl voxel grids directly from photographs, avoiding the intermediate conversions that degrade color fidelity in physical 3D prints.","keywords":["Poxel","printable voxel","CMYKWCl color model","photopolymer jetting","3D reconstruction","radiance fields","voxel rendering","color discretization"],"falsifier":"Print a standardized color target reconstructed by Poxel on a CMYKWCl jetting printer and measure the patches with a spectrophotometer; if the measured colors depart from the source colors by more than the printer's normal gamut error, the claimed fidelity advantage of direct CMYKWCl voxel output fails.","tokens_in":3834,"feed_emoji":"🖨️","tokens_out":8925,"duration_ms":78596,"temperature":0.7,"pith_summary":"Poxel is a proposal for making 3D reconstructions that are ready to print, not just ready to view. Standard radiance-field methods encode scenes in view-dependent RGB colors optimized for a virtual camera, which cannot be sent directly to a multi-material printer. Poxel instead learns a grid of voxels, volume cells, where each voxel carries layered CMYKWCl ink amounts, removes view dependence, and uses anisotropic voxels so layer height and color blending match the printer's mechanics. The paper demonstrates end-to-end prints from phone and camera footage on a commercial jetting printer, reporting that printed colors stay closer to their digital counterparts than an adapted 3D-Gaussian-splatting pipeline. If the approach holds, it would let a single reconstruction feed a printer without a separate color-managed conversion stage.","feed_headline":"Voxel method prints 3D scans without extra color steps","feed_subtitle":"By encoding color as print-ready CMYKWCl inks, the digital model keeps its colors on the printed object.","key_machinery":"The central object is the Poxel, an anisotropic, printable voxel whose color is stored as layered CMYKWCl, cyan, magenta, yellow, black, white, clear, ink amounts rather than RGB. The framework learns this grid by volume rendering with view-independent colors, applies a regional color average, $\\bar{C} = \\frac{1}{N}\\sum_{i=1}^{N} C_i$ over $N=12$ voxels in a $1 \\times 6 \\times 2$ block, to smooth transitions, and discretizes to the printer's ink palette during optimization. That combination is what lets the output be sliced and printed without a separate color-space conversion.","core_discovery":"The central claim is that a voxel representation designed around the printer's actual ink set, rather than around virtual rendering, removes the main source of print error: converting a view-dependent RGB model into printable ink layers after the fact. By optimizing voxel colors directly as CMYKWCl values and averaging colors over a fixed $1 \\times 6 \\times 2$ voxel region to smooth gradients, Poxel produces a grid that can be sliced into printer layers. In the paper's comparison, the Poxel print had smoother color transitions and better alignment with the printer's color gamut than a baseline that rendered with 3D Gaussian splatting, converted the splats to ellipsoidal volumes, and then mapped RGB to CMYKWCl, because Poxel never leaves the print color space.","pith_inferences":["A natural next test, not reported in the paper, is a quantitative colorimetric evaluation between the digital reconstruction and the prints, such as CIEDE2000 color differences; the paper's fidelity improvement is currently supported by visual inspection.","The fixed $1 \\times 6 \\times 2$ averaging window is a tunable hyperparameter, and per-printer adjustment of its shape and size could further reduce color banding on devices with different native resolutions.","Because the ink palette is the only printer-specific input to the discretization, the same framework should transfer to other CMYKWCl jetting printers by swapping the palette lookup, which would broaden the evidence for the central claim.","If direct voxel output becomes accepted, printer drivers could consume reconstructed voxel grids directly, collapsing the separate slicing and color-management stages that current printing pipelines perform."],"forward_implications":["Printed Poxel models should retain the smooth color gradients of the source scene because color averaging happens in the print color space rather than in RGB.","A single Poxel reconstruction can be sliced directly into layers for a multi-material jetting printer, eliminating the RGB-to-CMYKWCl conversion step that introduced artifacts in the comparison baseline.","Removing view dependence makes the representation invariant to camera angle, which is necessary for an object that is viewed in the round as a physical print.","Anisotropic voxel spacing lets the reconstruction allocate more resolution along the print height axis, where layer thickness limits achievable detail."],"supporting_citations":[{"why":"Establishes the NeRF volume-rendering baseline whose view-dependent RGB output Poxel replaces with printable voxel colors.","marker":"[1]"},{"why":"Provides the explicit voxel grid representation that Poxel inherits and adapts from continuous radiance fields to discrete printable voxels.","marker":"[2]"},{"why":"Serves as the main comparison baseline; its 2D-splat to 3D-volume conversion and RGB-to-CMYKWCl mapping are what Poxel claims to outperform.","marker":"[3]"},{"why":"Defines the photopolymer jetting technology and CMYKWCl ink set whose layer thickness and resolution constraints motivate Poxel's anisotropic voxel design.","marker":"[4]"}],"fun_headline_variants":["Voxel method prints 3D scans in native ink colors","Printer-native voxels skip RGB conversion in 3D prints","New voxel pipeline outputs print-ready CMYKWCl grids","Direct voxel-to-printer colors fix 3D print fidelity","Poxel voxels map directly to printable inks, no conversion"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The method assumes that averaging RGBA colors over a fixed $1 \\times 6 \\times 2$ voxel block and then quantizing the result to CMYKWCl inks produces colors the printer reproduces faithfully; this was checked by visual comparison, not by measuring printed color patches against a colorimetric standard.","fun_headline_variants_meta":{"raw":{"variants":["Voxel method prints 3D scans in native ink colors","Printer-native voxels skip RGB conversion in 3D prints","New voxel pipeline outputs print-ready CMYKWCl grids","Direct voxel-to-printer colors fix 3D print fidelity","Poxel voxels map directly to printable inks, no conversion"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000367,"raw_usage":{"total_tokens":1932,"prompt_tokens":868,"completion_tokens":1064,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":484,"completion_tokens_details":{"reasoning_tokens":975}},"tokens_in":484,"tokens_out":1064,"duration_ms":10600,"temperature":1.0,"reasoning_tokens":975,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:57:09.173437+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Print a standardized color target reconstructed by Poxel on a CMYKWCl jetting printer and measure the patches with a spectrophotometer; if the measured colors depart from the source colors by more than the printer's normal gamut error, the claimed fidelity advantage of direct CMYKWCl voxel output fails.","supporting_citations":[{"cited_title":"4\" FUNCTION default.is.dash.repeated.names #1 FUNCTION default.name.format.string","cited_arxiv_id":null,"evidence_quote":"Establishes the NeRF volume-rendering baseline whose view-dependent RGB output Poxel replaces with printable voxel colors."}],"review_version":1}