{"id":"aa84f385-4d48-4e48-b27d-e1d063819bb2","arxiv_id":"2603.28896","paper_version":2,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"An adaptation framework lets perspective-trained multi-view 3D foundation models handle fisheye inputs natively, improving pose, depth, point maps, and FoV without regressing on perspective images.","lead":"Fisheye3R adapts multi-view 3D foundation models so they work on highly distorted fisheye images without needing large fisheye ground-truth datasets. That matters for robotics, XR, and any camera system that prefers wide field of view over pinhole geometry.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified on the paper's own terms; the review cannot audit Fisheye3R because the cached full text is a different manuscript.","rationale":"The reader correctly flagged that only the abstract of Fisheye3R is present and that the full-text cache is a mismatched finance/MPC paper (2603.28898). Under that constraint, UNVERDICTED with LOW confidence is the only defensible outcome. The abstract's strongest claim (native fisheye accommodation without perspective regression via self-supervised or no-fisheye-data schemes, with consistent gains across VGGT, π³, and MapAnything) cannot be confirmed or refuted without the missing body. The reader's weakest-assumption statement is well-chosen as the place that would matter once the paper is available, but it is not yet a load-bearing attack that lands. No new technical concern can be substantiated from the abstract alone without inventing details. Therefore the verdict stays UNVERDICTED; agreement with the reader is full; no adjustment is warranted until the correct manuscript is supplied.","tokens_in":3512,"tokens_out":540,"duration_ms":5405,"concrete_test":"Replace the cached full-text block with the actual Fisheye3R PDF/source for 2603.28896; re-run the reader pass on the real method, synthetic fisheye generation protocol, and perspective non-regression tables. If those sections show measured non-regression on perspective benchmarks and gains on real fisheye that survive ablation of the projection-adaptation modules, the central claim holds; if not, re-score correctness_risk and move the verdict.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption (that nonlinear projection-induced pixel rearrangement is the dominant failure mode and can be fixed without real fisheye GT) is the right place to pressure—but it cannot be stress-tested here. The CACHEABLE PAPER SOURCE CONTEXT supplies the full text of arXiv:2603.28898 (Model Predictive Control For Trade Execution), not arXiv:2603.28896 (Fisheye3R). Only the Fisheye3R abstract is available. Without the method body, equations, synthetic-data protocol, ablations, or tables, there is no concrete internal inconsistency, missing baseline, or failed non-regression check that can be verified. Manufacturing a technical attack on unshown sections would violate good-faith reading. The load-bearing gap is therefore evidentiary availability, not a demonstrated flaw in the claimed adaptation schemes.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"Based solely on the abstract of arXiv:2603.28896, the paper claims that feed-forward multi-view 3D foundation models trained on perspective imagery degrade on fisheye inputs because nonlinear projection rearranges pixel geometry. It proposes Fisheye3R, an adaptation framework that natively handles high radial distortion without regressing on perspective images, using flexible schemes that allow self-supervised adaptation from unlabeled perspective data and supervised adaptation without any fisheye training data. Experiments across VGGT, π³, and MapAnything are said to improve camera pose, depth, point maps, and field-of-view estimation on fisheye imagery, with code released. The body text supplied under this review, however, is an unrelated manuscript on model predictive control for trade execution (arXiv:2603.28898), so the method, equations, data protocol, and results of Fisheye3R cannot be audited from the materials provided.","tokens_in":3739,"tokens_out":867,"duration_ms":12345,"significance":"If the abstract claims hold under proper evaluation, the work would be practically significant for robotics, AR/XR, and autonomous systems that rely on wide-FoV cameras: it would let existing large-scale 3D foundation models be reused on fisheye inputs without collecting scarce fisheye ground truth, while preserving perspective performance. The dual self-supervised / no-fisheye-GT supervised adaptation design and multi-model evaluation would be a useful systems contribution. Those strengths cannot be credited or stress-tested here because the full Fisheye3R manuscript is not available in this review package.","major_comments":[{"comment":"Manuscript identity mismatch: the CACHEABLE full text is “Model Predictive Control For Trade Execution” (arXiv:2603.28898, q-fin.TR), not Fisheye3R (arXiv:2603.28896, cs.CV). Sections, equations, figures, tables, and experimental protocols for Fisheye3R are absent. No load-bearing claim about adaptation, non-regression on perspective images, or gains on pose/depth/point-map/FoV can be verified. This is a blocking evidentiary failure for peer review.","section":null},{"comment":"Abstract-only central claim cannot be checked: the abstract asserts that nonlinear projection-induced pixel rearrangement is the dominant failure mode and that self-supervised adaptation on unlabeled perspective images (and supervised adaptation without fisheye training data) suffice to fix it without real fisheye GT. Without the method body, synthetic-data protocol, ablations, baselines, error bars, or train/test leakage controls, this assumption remains untested. A correct resubmission must supply those sections so the claim can be audited.","section":null},{"comment":"Non-regression and multi-model consistency are load-bearing but unshown: the abstract promises no performance regression on perspective images and consistent gains across VGGT, π³, and MapAnything on four metrics. Tables and held-out protocols are not present in the supplied text, so the central empirical contribution cannot be assessed.","section":null}],"minor_comments":[{"comment":"Only the Fisheye3R abstract is available; figure/table numbering, notation, and related-work placement cannot be reviewed until the correct PDF is provided.","section":null},{"comment":"Code URL is given in the abstract (https://github.com/android-xr/fisheye3r); once the correct manuscript is attached, reproducibility of the adaptation schemes should be checked against that release.","section":null}],"recommendation":"uncertain","confidential_remarks":"The review package pairs the Fisheye3R abstract/title with the full body of an unrelated MPC trade-execution paper (2603.28898). This looks like a cache/ID mix-up rather than author misconduct, but it makes a substantive technical referee report on Fisheye3R impossible. Please re-queue with the correct full text of 2603.28896. I have not manufactured critiques of unshown sections."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The one thing you need to know: we only have a clean abstract for Fisheye3R. The body that came with the cache is an unrelated MPC trade-execution paper (2603.28898). So this is not a verdict on the method; it is a note on what the abstract actually claims and what is missing.\n\nWhat is new on the face of it is practical and well-scoped. Feed-forward multi-view 3D models (VGGT, π³, MapAnything) trained on perspective data degrade under strong radial distortion. Fisheye3R is framed as an adaptation layer that lets those models take fisheye natively, with two learning schemes that avoid the usual data bottleneck: self-supervised adaptation from unlabeled perspective images only, and supervised adaptation that does not require fisheye training data. The abstract also claims no regression on perspective inputs and consistent gains on pose, depth, point maps, and FoV. Code is promised. That combination—if real—is exactly what robotics and XR people need when they already have a strong perspective foundation model and a fisheye sensor.\n\nThe soft spot is not a flaw in the argument; it is evidentiary absence. We cannot check the projection handling, the synthetic or pseudo-label protocol, the non-regression numbers on perspective, ablations, or whether residual photometry/calibration gaps dominate. The abstract’s load-bearing assumption—that the main failure is nonlinear pixel rearrangement and that it can be fixed without real fisheye GT—is plausible but untested here. Circularity risk (train/test contamination, synthetic fisheye that is too easy) is the usual one for this genre and cannot be ruled out or confirmed.\n\nWho it is for: people shipping multi-view 3D on wide-FoV cameras who want to keep existing foundation models. A serious editor should send the real manuscript to referees once the correct PDF is attached; the problem is real, the framing is honest, and three-model evaluation plus open code is enough to deserve referee time. Until then I would not cite it or put it in reading group. Get the actual paper and re-read.","headline":"Abstract promises a useful fisheye adaptation for 3D foundation models without GT or perspective regression, but the supplied full text is a different paper, so the claims cannot be audited.","tokens_in":4340,"tokens_out":542,"would_cite":false,"duration_ms":79946,"reading_group":"no","serious_thinker":"unclear","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Fisheye3R adapts multi-view 3D foundation models to fisheye inputs without fisheye ground truth and without hurting perspective performance.","keywords":["fisheye cameras","multi-view 3D reconstruction","feed-forward foundation models","radial distortion","model adaptation","self-supervised learning","camera pose estimation","depth estimation"],"falsifier":"Take the same three foundation models, apply Fisheye3R, and evaluate on held-out real fisheye multi-view sequences with accurate pose and depth ground truth: if camera pose, depth, point-map, and FoV errors do not improve over the unadapted baselines (or if perspective performance drops), the central claim fails.","tokens_in":4404,"feed_emoji":"📷","tokens_out":930,"duration_ms":13011,"temperature":0.7,"pith_summary":"Feed-forward multi-view 3D reconstruction models trained on ordinary perspective photos fail when the camera uses a fisheye lens, because the nonlinear projection rearranges pixels in ways those models never saw. Collecting enough labeled fisheye data to retrain them from scratch is impractical. Fisheye3R is an adaptation framework that teaches the same models to accept high radial distortion natively, while preserving their original accuracy on perspective images. It does this with flexible training recipes: one that needs only unlabeled perspective images (self-supervised) and another that needs no fisheye training data at all (supervised). Across three existing foundation models the adapted systems improve camera pose, depth, point maps, and field-of-view estimates on fisheye imagery, giving a practical path to wide-angle 3D reconstruction without a large new labeled corpus.","feed_headline":"3D models learn fisheye views without fisheye labels","feed_subtitle":"Adaptation keeps perspective accuracy while lifting pose, depth, and FoV on wide-angle cameras","key_machinery":"Fisheye3R: a modular adaptation framework plus flexible learning schemes (self-supervised on unlabeled perspective images only, and supervised without any fisheye training data) that retarget existing multi-view 3D foundation models so they accept high radial distortion natively.","core_discovery":"The authors claim that the performance drop of perspective-trained multi-view 3D foundation models on fisheye images is driven mainly by the changed spatial arrangement of pixels under nonlinear projection, and that this can be corrected by an adaptation framework (Fisheye3R) that never requires real fisheye images with ground truth—using either self-supervision on unlabeled perspective data or supervised adaptation without fisheye training data—while leaving perspective performance intact and improving pose, depth, point-map, and field-of-view estimates on fisheye inputs.","pith_inferences":["The same pixel-arrangement adaptation idea may transfer to other non-pinhole projections (catadioptric, ultra-wide rectilinear) if the dominant error remains geometric rather than photometric.","If residual failures appear under extreme distortion, they likely come from multi-view consistency or calibration error rather than pure spatial layout, suggesting a next diagnostic split of error sources.","Pairing Fisheye3R with lightweight online calibration could make consumer AR/VR headsets with fisheye optics reuse existing 3D foundation weights with minimal extra data."],"forward_implications":["Existing perspective-trained multi-view 3D models can be deployed on fisheye cameras without collecting large labeled fisheye datasets.","Self-supervised adaptation from unlabeled perspective images alone is enough to recover useful fisheye performance.","Camera pose, depth, point maps, and field-of-view estimates on wide-FOV imagery improve consistently across multiple foundation architectures.","Production pipelines can keep one adapted model that handles both perspective and high-distortion inputs without separate branches."],"fun_headline_variants":["Fisheye3R adapts 3D models to fisheye without labels","3D foundation models handle fisheye via adaptation sans GT","Adapt perspective 3D models to fisheye using only perspective data","Fisheye3R lifts pose and depth on fisheye without fisheye training","Multi-view 3D models gain fisheye support without regression"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The main reason perspective-trained 3D models fail on fisheye is the nonlinear remapping of pixels, and that gap can be closed by adaptation that never trains on real fisheye images with ground truth.","fun_headline_variants_meta":{"raw":{"variants":["Fisheye3R adapts 3D models to fisheye without labels","3D foundation models handle fisheye via adaptation sans GT","Adapt perspective 3D models to fisheye using only perspective data","Fisheye3R lifts pose and depth on fisheye without fisheye training","Multi-view 3D models gain fisheye support without regression"]},"model":"grok-4.5","effort":"low","cost_usd":0.00426,"raw_usage":{"total_tokens":1346,"prompt_tokens":858,"num_sources_used":0,"completion_tokens":105,"cost_in_usd_ticks":42600000,"prompt_tokens_details":{"text_tokens":858,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":383,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":858,"tokens_out":105,"duration_ms":3540,"temperature":1.0,"reasoning_tokens":383,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T16:04:06.990496+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Take the same three foundation models, apply Fisheye3R, and evaluate on held-out real fisheye multi-view sequences with accurate pose and depth ground truth: if camera pose, depth, point-map, and FoV errors do not improve over the unadapted baselines (or if perspective performance drops), the central claim fails.","supporting_citations":[],"review_version":1}