{"id":"f2e33f47-16d6-4e08-9b29-d7f6e457207f","arxiv_id":"2604.22518","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"NONSAC is a general, estimator-agnostic framework that improves scalability and robustness for geometric model estimation on very large noisy datasets by sampling non-minimal subsets and scoring candidate hypotheses.","lead":"The paper introduces NONSAC, a sampling framework that repeatedly draws non-minimal data subsets, runs a robust estimator on each, and picks the best model via a scoring rule. Smart readers in computer vision and robotics should care because it targets model fitting on datasets too large for standard RANSAC while claiming better outlier tolerance.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"Scoring rules' ability to reliably pick high-quality models from non-minimal-subset hypotheses in arbitrarily large, high-outlier data remains the least-secured step.","rationale":"The reader's weakest assumption is exactly the load-bearing point: the scoring rules must work on hypotheses produced from non-minimal subsets of arbitrarily contaminated data. My analysis reaches the same location after reading the abstract's description of the pipeline. Because the full manuscript was not supplied in the query, I treat the concern as still open rather than refuted, leaving the provisional UNVERDICTED verdict unchanged.","tokens_in":1626,"tokens_out":381,"duration_ms":29817,"concrete_test":"Reproduce the relative-pose experiment on a synthetic 1 M-point set with 70 % outliers; generate 500 NONSAC hypotheses using each scoring rule, then measure the fraction of trials in which the highest-scoring model has rotation error < 5° and translation error < 0.1 (normalized); if this fraction drops below 60 % for any rule, the selection step fails to support the scalability claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that, after repeatedly drawing non-minimal subsets and running a robust estimator on each, one of the proposed scoring rules will consistently surface a model whose quality is close to the global optimum. Nothing in the framework description supplies a guarantee that the scoring function remains discriminative once the input subsets are drawn from a distribution whose outlier fraction and scale are unbounded; the non-minimal size itself can amplify the effect of any residual contamination inside the sampled block. The abstract states that scoring rules were proposed and evaluated on three tasks, yet the argument supplies no analysis of how those rules behave when the underlying estimator is itself applied only to non-minimal blocks rather than to minimal samples or the full set.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper introduces NONSAC, a general framework for robust and scalable model estimation from arbitrarily large datasets contaminated with noise and outliers. NONSAC repeatedly samples non-minimal subsets, applies a robust estimator to each to generate model hypotheses, and selects the final model using one of several proposed scoring rules. The framework is presented as estimator-agnostic and integrable with existing algorithms such as RANSAC. It is evaluated on relative camera pose estimation, Perspective-n-Point, point cloud registration, and correspondence-free point cloud registration.","tokens_in":1781,"tokens_out":365,"duration_ms":41396,"significance":"If the empirical claims hold, NONSAC would offer a practical, scalable alternative to minimal-sampling consensus methods for very large contaminated datasets in geometric computer vision. The estimator-agnostic design and the extension to correspondence-free registration are potentially useful strengths. The approach could improve both runtime and outlier robustness when integrated with existing robust estimators.","major_comments":[{"comment":"Abstract: The central claim that the scoring rules reliably select a high-quality model from hypotheses generated on non-minimal subsets is load-bearing, yet the manuscript supplies no analysis or quantitative results showing how these rules behave when the underlying robust estimator is applied only to non-minimal blocks drawn from data with arbitrarily high outlier fractions. The non-minimal size can amplify residual contamination inside each block, and the evaluation on three tasks does not address this regime.","section":null}],"minor_comments":[{"comment":"The abstract lists tasks and claims but contains no quantitative results, error bars, or ablation details; these should be summarized with key metrics (e.g., success rates, runtime scaling) to allow readers to assess the reported improvements.","section":null}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for their constructive feedback on our manuscript. We address the major comment point by point below and outline the revisions we will implement.","responses":[{"response":"We acknowledge that a dedicated quantitative analysis of scoring-rule behavior under arbitrarily high outlier fractions would strengthen support for the central claim. Our current evaluations on relative camera pose estimation, PnP, and point cloud registration include challenging outlier levels and show that NONSAC with the proposed scoring rules yields more accurate and robust models than minimal-sampling baselines; however, these experiments do not systematically isolate the effect of non-minimal subset size on residual contamination at extreme outlier ratios. In the revision we will add a new experimental subsection that varies outlier fraction from 50% to 95%, applies the robust estimator to non-minimal blocks, and reports the selection accuracy of each scoring rule. This will directly quantify robustness to amplified contamination and allow us to update the abstract accordingly.","revision_made":"yes","referee_comment":"Abstract: The central claim that the scoring rules reliably select a high-quality model from hypotheses generated on non-minimal subsets is load-bearing, yet the manuscript supplies no analysis or quantitative results showing how these rules behave when the underlying robust estimator is applied only to non-minimal blocks drawn from data with arbitrarily high outlier fractions. The non-minimal size can amplify residual contamination inside each block, and the evaluation on three tasks does not address this regime."}],"tokens_in":1225,"tokens_out":312,"duration_ms":38656,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The core of this paper is NONSAC, which samples non-minimal subsets from big noisy datasets, fits robust estimators to them to create several hypotheses, and picks the best one with a scoring rule. This is meant to let people run geometric fitting on datasets that are too large for ordinary RANSAC. It does a good job of keeping the approach general so it works with any robust estimator and showing it on standard problems like camera pose estimation, PnP, and point cloud registration. The part about doing registration without correspondences is a practical bonus. Laying out multiple scoring rules and testing them is also useful for seeing what works. The soft spot is exactly what the stress-test note points out. The scoring rules have to work well even when the non-minimal samples come from data with lots of outliers, and nothing guarantees that they will stay good at ranking the hypotheses. The paper says it evaluates them on the tasks, but I would want to see the actual numbers on how much it improves over baselines as the data gets larger and dirtier. If the gains are small or inconsistent, the whole thing loses its edge. This is for vision researchers who handle massive 3D data and need robust methods that scale. Someone building large reconstruction pipelines could pick up the framework and try the scoring ideas. It is worth sending to peer review so the experiments can be checked in detail. I would recommend accepting it for review.","headline":"NONSAC is a sensible framework for scaling RANSAC to huge datasets through non-minimal samples and hypothesis scoring, with the main question being whether the scoring holds up under heavy contamination.","tokens_in":2265,"tokens_out":366,"would_cite":false,"duration_ms":34818,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"NONSAC scales robust geometric model fitting to arbitrarily large noisy datasets by using non-minimal samples and hypothesis scoring.","keywords":["non-minimal sampling","consensus","robust estimation","outlier rejection","RANSAC","camera pose estimation","point cloud registration","model fitting"],"falsifier":"A large dataset with high outlier ratio on which every scoring rule selects a model whose error is markedly higher than the ground-truth model or the best hypothesis generated.","tokens_in":2531,"feed_emoji":"📊","tokens_out":626,"duration_ms":46999,"temperature":0.7,"pith_summary":"The paper introduces NONSAC as a framework that draws non-minimal subsets from very large datasets containing noise and outliers, runs a robust estimator on each subset to produce candidate models, and then selects the final model with a scoring rule. This setup is designed to work with any existing estimator and to plug into algorithms such as RANSAC. A sympathetic reader would care because standard minimal-sampling methods become impractical or inaccurate once data sizes grow without bound, while NONSAC aims to keep both speed and accuracy. The authors test the approach on relative camera pose estimation, Perspective-n-Point, and point cloud registration, including a version that hypothesizes all-to-all correspondences without prior matches.","feed_headline":"NONSAC scales model fitting to huge noisy datasets","feed_subtitle":"Non-minimal sampling and hypothesis scoring let existing estimators handle arbitrarily large contaminated data.","key_machinery":"Non-minimal subset sampling followed by robust hypothesis generation and quality scoring to select the best model.","core_discovery":"NONSAC repeatedly samples non-minimal subsets of data and generates model hypotheses using a robust estimator, producing multiple candidate models. The final model is selected based on a predefined scoring rule that evaluates hypothesis quality. The framework is estimator-agnostic and can be integrated with existing geometric fitting algorithms such as RANSAC to improve both scalability and robustness to outliers on arbitrarily large contaminated datasets.","pith_inferences":["The method may allow vision pipelines to avoid aggressive data downsampling or outlier pre-filtering steps.","Scoring rules tuned to particular sensor modalities or noise statistics could further improve results on specific tasks.","The non-minimal sampling idea might transfer to incremental or streaming settings where data arrives continuously."],"forward_implications":["Robust estimation becomes practical for datasets too large for traditional minimal-sampling techniques.","Existing algorithms such as RANSAC gain improved scalability and outlier tolerance when wrapped inside the framework.","The same procedure applies to relative camera pose estimation, Perspective-n-Point, and point cloud registration.","Correspondence-free registration is enabled by generating hypotheses over all-to-all possible matches."],"fun_headline_variants":["NONSAC fits models to huge noisy datasets with non-minimal sampling","Non-minimal sampling and scoring enables robust fitting on large data","NONSAC improves scalability of geometric estimators on outlier-rich data","NONSAC selects models using non-minimal samples and quality scoring"],"cache_read_input_tokens":64,"weakest_assumption_plain":"The scoring rules can reliably identify a high-quality model from the hypotheses produced by applying the robust estimator to non-minimal subsets drawn from arbitrarily large contaminated data.","fun_headline_variants_meta":{"raw":{"variants":["NONSAC fits models to huge noisy datasets with non-minimal sampling","Non-minimal sampling and scoring enables robust fitting on large data","NONSAC improves scalability of geometric estimators on outlier-rich data","NONSAC selects models using non-minimal samples and quality scoring"]},"model":"grok-4.3","cost_usd":0.013104,"raw_usage":{"total_tokens":5550,"prompt_tokens":563,"num_sources_used":0,"completion_tokens":69,"cost_in_usd_ticks":131040500,"prompt_tokens_details":{"text_tokens":563,"audio_tokens":0,"image_tokens":0,"cached_tokens":64},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":4918,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":563,"tokens_out":69,"duration_ms":32975,"temperature":1.0,"reasoning_tokens":4918,"cache_read_input_tokens":64,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-05-08T12:24:27.028831+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A large dataset with high outlier ratio on which every scoring rule selects a model whose error is markedly higher than the ground-truth model or the best hypothesis generated.","supporting_citations":[],"review_version":1}