Metrics or Mirage? An Audit of Evaluation Inconsistencies in Colonoscopy Polyp Segmentation Benchmarks

arXiv:2607.08203v1 Announce Type: new Abstract: Progress in colonoscopy polyp segmentation is routinely reported through leaderboard comparisons on a small set of public benchmarks. We argue that this apparent progress is difficult to verify: a systematic audit of extbf{27 papers} published between 2015 and 2026 reveals three structural problems in how the community evaluates models. extbf{First}, 25 of 27 papers extit{omit the Hausdorff distance}. Hausdorff distance is a boundary-accuracy metric with direct clinical relevance for detecting flat or small polyps, and is a standard in radiotherapy segmentation. extbf{Second}, at least five extit{incompatible train/test split protocols} co-exist across papers reporting results on the same two datasets (Kvasir-SEG and CVC-ClinicDB), making published Dice scores non-comparable even when they appear in the same leaderboard column. extbf{Third}, 26 of 27 papers make extit{performance claims without any statistical significance test}. Strikingly, four papers published after the Metrics Reloaded frameworkite{metricsreloaded2024} (Maier-Hein et al., extit{Nature Methods} 2024) perpetuate these same problems, suggesting that general-purpose metric guidance has not yet reached the colonoscopy sub-community. To show these problems are not merely cosmetic, we re-evaluate five representative models under three controlled protocols with a single uniform scorer, and find that the reported metric conceals large boundary and recall failures, that the ``best'' model changes with the metric, and that near-tied rankings reverse across random splits. We propose a five-point extbf{Polyp Segmentation Reporting Checklist}(PSRC) as a lightweight, domain-adapted corrective.

Source: arXiv cs.CV | 2026-07-10