{"id":"4a856c41-3e5a-4eec-8e81-f21629e579fb","arxiv_id":"2606.03069","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":2.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":1,"one_line_summary":"Training tweaks (augmentations, hybrid loss, scheduling, ablation flags) raise optic disc Dice from 0.939 (baseline epoch 5) to 0.956 (final epoch) on a fundus benchmark.","lead":"The paper describes four training enhancements to an existing 2024 whitening-transform segmentation method for medical images. These changes aim to boost cross-device robustness on fundus optic disc segmentation without changing the core model.","discovery_kind":"unclear","skeptic_critique":{"model":"grok-4.3","headline":"Performance gains attributed to four enhancements without ablations, variance, or statistical tests to confirm causality or control for confounders like training length.","rationale":"The reader's weakest_assumption precisely identifies the missing controls that would be required to support the causal claim. The epoch-5 versus final-epoch reporting detail adds a secondary comparability issue but does not change the primary gap. No independent verification elements (reproducible code, multiple seeds, or formal checks) are mentioned that would override this.","tokens_in":1793,"tokens_out":352,"duration_ms":22933,"concrete_test":"Using the proposed command-line ablation flags, recompute Dice/ASD for the four single-enhancement variants and the full combination on the same fundus benchmark; if any single fix alone reaches within 0.01 Dice of 0.956 or if the full pipeline gain disappears under matched epoch count and seed averaging, the causal attribution to the listed limitations weakens.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the four enhancements (domain-adaptive augmentations, hybrid BCE+Dice loss, curriculum Dice scheduling, ablation flags) produce the observed gains (final-epoch Dice 0.956 / ASD 13.31 vs. baseline epoch-5 Dice 0.939)—rests on the premise that the listed original limitations were the primary performance bottlenecks and that the fixes address them cleanly. The reported comparison provides only the full pipeline versus a single baseline point at epoch 5; no per-enhancement ablations, no run-to-run standard deviations, and no significance tests are described. This leaves the attribution vulnerable to alternative explanations such as longer training, altered optimization dynamics, or seed effects.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper identifies four limitations in the original WT-PSE framework (limited augmentations, per-pixel BCE loss, unscheduled loss weighting, and lack of ablation controls) and proposes four corresponding enhancements (domain-adaptive augmentations including random erasing/gamma/salt-and-pepper, hybrid BCE+Dice loss, curriculum Dice weight scheduling, and command-line ablation flags). On the fundus optic disc segmentation benchmark, the full enhanced pipeline is reported to reach a final-epoch Dice of 0.956 and ASD of 13.31, outperforming the baseline's epoch-5 Dice of 0.939; the central claim is that these training-level changes produce the gains without modifying the WT-PSE architecture.","tokens_in":1959,"tokens_out":478,"duration_ms":19298,"significance":"If the attribution of gains to the specific enhancements were supported by ablations, variance estimates, and fair epoch-matched comparisons, the work would demonstrate that modest training modifications can improve cross-domain robustness in medical segmentation pipelines. The manuscript provides no machine-checked proofs or parameter-free derivations; its value would rest entirely on the strength of the empirical controls.","major_comments":[{"comment":"Abstract and results: the reported comparison evaluates the proposed pipeline at its final training epoch but the baseline only at epoch 5. This epoch mismatch leaves open the possibility that the 0.017 Dice improvement is driven by additional training steps rather than the four listed enhancements.","section":"Abstract / Experiments"},{"comment":"Experiments section: despite the introduction of command-line ablation flags, no per-enhancement ablation results, incremental addition tables, or leave-one-out controls are presented. Without these, the claim that the four specific fixes address the four listed limitations cannot be isolated from confounders such as longer training or altered optimization.","section":"Experiments"},{"comment":"Results: the manuscript reports point estimates (Dice 0.956, ASD 13.31) with neither run-to-run standard deviations, error bars, nor statistical significance tests. Given the empirical nature of the work and known sensitivity of segmentation metrics to random seeds, this omission weakens the reliability of the performance claim.","section":"Results"}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive feedback on our manuscript. We address each major comment below and will revise the paper to improve the experimental rigor and clarity of our claims.","responses":[{"response":"We agree that the epoch mismatch is a valid concern that could confound attribution of the observed gains. The original WT-PSE baseline was reported at epoch 5 per its published protocol, while our enhanced pipeline was evaluated at convergence. In the revised manuscript, we will add epoch-matched comparisons by reporting the baseline performance at its final epoch as well and include learning curves for both pipelines to demonstrate that improvements are not solely attributable to additional training steps.","revision_made":"yes","referee_comment":"[Abstract / Experiments] Abstract and results: the reported comparison evaluates the proposed pipeline at its final training epoch but the baseline only at epoch 5. This epoch mismatch leaves open the possibility that the 0.017 Dice improvement is driven by additional training steps rather than the four listed enhancements."},{"response":"We acknowledge that although command-line ablation flags were added to support controlled experiments, the manuscript does not present the corresponding ablation results. This limits isolation of each enhancement's contribution. In the revision, we will include incremental addition tables and leave-one-out ablation studies showing the effect of each component (augmentations, hybrid loss, scheduling) when added sequentially to the baseline.","revision_made":"yes","referee_comment":"[Experiments] Experiments section: despite the introduction of command-line ablation flags, no per-enhancement ablation results, incremental addition tables, or leave-one-out controls are presented. Without these, the claim that the four specific fixes address the four listed limitations cannot be isolated from confounders such as longer training or altered optimization."},{"response":"We agree that the absence of variability measures weakens the empirical claims. In the revised manuscript, we will rerun the experiments with multiple random seeds, report mean and standard deviation for Dice and ASD, and include error bars in the results tables. Statistical significance testing will be added where appropriate to support the comparisons.","revision_made":"yes","referee_comment":"[Results] Results: the manuscript reports point estimates (Dice 0.956, ASD 13.31) with neither run-to-run standard deviations, error bars, nor statistical significance tests. Given the empirical nature of the work and known sensitivity of segmentation metrics to random seeds, this omission weakens the reliability of the performance claim."}],"tokens_in":1522,"tokens_out":533,"duration_ms":30276,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main thing to know is that the claimed improvements rest on a comparison that pits the enhanced pipeline after full training against the original method stopped at epoch 5. That setup makes it difficult to credit the specific changes rather than just extra training time or different optimization paths.\n\nWhat is new is the application of four practical fixes to the 2024 WT-PSE method: domain-adaptive augmentations (random erasing, gamma correction, salt-and-pepper), a hybrid BCE plus Dice loss, a curriculum schedule on the Dice weight, and command-line flags meant to support ablations. The curriculum scheduling is a reasonable step for stabilizing early training when Dice loss can be noisy on medical edges. The authors also clearly list the original limitations they targeted.\n\nThe soft spots are in the evidence. The abstract gives only the aggregate final scores (0.956 Dice, 13.31 ASD) without per-enhancement breakdowns, run-to-run standard deviations, or statistical tests. Even though ablation flags were added, no actual ablation results appear. This leaves open whether the gains come from the listed changes or from longer training and altered dynamics. The work stays purely empirical with no new derivation.\n\nThis is for someone already running WT-PSE on fundus optic disc data who wants to try these tweaks in their own code. A reader seeking a new framework or tightly controlled validation will not get much. The citation pattern is narrow and focused on the prior WT-PSE paper.\n\nI would not send this to peer review as written. The central attribution to the four enhancements needs matched-epoch comparisons and actual ablations before it is worth referee time.","headline":"The paper adds four standard training tweaks to WT-PSE and reports a Dice gain, but compares the new final-epoch result only to the baseline at epoch 5 with no ablations or variance shown.","tokens_in":2413,"tokens_out":417,"would_cite":false,"duration_ms":29243,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Four training enhancements to the WT-PSE framework raise optic-disc Dice score from 0.939 to 0.956","keywords":["medical image segmentation","whitening transform","domain adaptation","data augmentation","loss scheduling","optic disc segmentation","cross-domain segmentation"],"falsifier":"A controlled re-run of the fundus optic-disc experiments in which the four enhancements are applied yet the Dice score remains at or below the baseline value of 0.939 would falsify the central claim.","tokens_in":2717,"feed_emoji":"","tokens_out":705,"duration_ms":21988,"temperature":0.7,"pith_summary":"The paper identifies four limitations in the original WT-PSE implementation for cross-domain medical image segmentation: insufficient augmentations to mimic scanner differences, per-pixel binary cross-entropy loss vulnerable to edge noise, missing scheduled loss weighting that can destabilize training, and no built-in ablation controls. It proposes four direct fixes consisting of domain-adaptive augmentations, a hybrid BCE-Dice loss, curriculum-based Dice weight scheduling, and command-line flags for systematic ablations. Experiments on the fundus optic disc benchmark show the updated pipeline reaching a final Dice score of 0.956 and ASD of 13.31. A sympathetic reader would care because these changes target practical training bottlenecks that affect reliable segmentation when imaging devices and protocols vary.","feed_headline":"Four training tweaks lift optic-disc Dice to 0.956","feed_subtitle":"Domain-adaptive augmentations, hybrid loss, scheduled weighting and ablation flags improve the whitening framework on fundus images","key_machinery":"The four training enhancements (domain-adaptive augmentation, hybrid BCE-Dice loss, curriculum Dice-weight scheduling, and command-line ablation flags) that directly target the four listed limitations in the original WT-PSE learning framework.","core_discovery":"The improved pipeline that applies domain-adaptive augmentation including random erasing, gamma correction and salt-and-pepper noise, replaces per-pixel BCE with a hybrid BCE-Dice loss, introduces curriculum-based Dice weight scheduling, and adds command-line ablation flags achieves a final-epoch optic-disc Dice score of 0.956 and an ASD score of 13.31 on the fundus benchmark, outperforming the baseline epoch-5 Dice score of 0.939, while leaving the underlying WT-PSE architecture unchanged.","pith_inferences":["The same four enhancements could be tested on segmentation tasks outside fundus imaging to check whether the gains generalize.","The approach suggests that many existing whitening-based frameworks might benefit from similar training-level adjustments rather than architectural redesign.","Further runs with varied random seeds or additional clinical datasets would test whether the reported Dice and ASD improvements remain stable."],"forward_implications":["Performance gains are obtainable solely through training-level changes without any modification to the WT-PSE architecture.","Domain-adaptive augmentations can simulate real scanner variations more effectively than the original limited set.","The hybrid loss improves edge-aware segmentation under noisy conditions compared with per-pixel BCE alone.","Curriculum scheduling of the Dice term stabilizes early training phases that would otherwise be destabilized.","Command-line ablation flags enable systematic scientific comparison of each enhancement."],"fun_headline_variants":["Training augmentations achieve 0.956 optic-disc Dice","Hybrid BCE Dice loss reaches 0.956 on fundus images","Scheduled weighting and ablation flags achieve 0.956 Dice","Enhanced training without architecture change reaches 0.956 Dice"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The four listed limitations are the primary causes of any performance gap and the proposed fixes address them without introducing new uncontrolled variables.","fun_headline_variants_meta":{"raw":{"variants":["Training augmentations achieve 0.956 optic-disc Dice","Hybrid BCE Dice loss reaches 0.956 on fundus images","Scheduled weighting and ablation flags achieve 0.956 Dice","Enhanced training without architecture change reaches 0.956 Dice"]},"model":"grok-4.3","cost_usd":0.010326,"raw_usage":{"total_tokens":4612,"prompt_tokens":748,"num_sources_used":0,"completion_tokens":66,"cost_in_usd_ticks":103262000,"prompt_tokens_details":{"text_tokens":748,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":3798,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":748,"tokens_out":66,"duration_ms":40931,"temperature":1.0,"reasoning_tokens":3798,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-28T11:05:48.298013+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A controlled re-run of the fundus optic-disc experiments in which the four enhancements are applied yet the Dice score remains at or below the baseline value of 0.939 would falsify the central claim.","supporting_citations":[],"review_version":1}