dice coefficient computation

In biomedical image segmentation validation, metrics such as the Hausdorff distance suffer from implementation inconsistencies across open-source toolkits, compromising benchmark reliability, introducing biomarker bias, and posing clinical deployment risks. To address this, we systematically evaluate 11 widely used toolkits and introduce, for the first time, a reference implementation based on high-fidelity 3D surface meshes. Our framework integrates real-world clinical data and a cross-platform consistency analysis. Statistical analysis reveals significant inter-tool variation in Hausdorff distance computations (p < 0.001), with interpolation strategy, boundary handling, and sampling density identified as primary sources of discrepancy. Based on these findings, we propose a reproducible and verifiable paradigm for distance-based evaluation, accompanied by standardized computational guidelines. This work substantially enhances the reliability, comparability, and clinical translatability of segmentation assessment.

In clinical practice, medical image segmentation models often face the challenge of unlabeled test data, hindering reliable performance assessment. To address this, we propose Segmentation Performance Evaluator (SPE), the first lightweight, plug-and-play unsupervised performance estimation method that accurately predicts six key metrics—including Dice and HD95—without ground-truth annotations. SPE integrates universal meta-learning representations from UniverSeg, uncertainty-aware regression, and cross-domain consistency constraints, enabling generalizable evaluation across diverse segmentation architectures and metrics with zero training overhead and seamless integration into existing pipelines. Evaluated on six public benchmarks, SPE achieves an average Pearson correlation coefficient of 0.956 ± 0.046 and a mean absolute error of 0.025 ± 0.019—significantly outperforming prior approaches. The source code is publicly available.

To address the lack of clinical trustworthiness in deploying prostate cancer histopathological image segmentation models, this paper proposes a novel paradigm that directly quantifies model uncertainty as the predicted Dice score. We first discover a statistically significant Spearman correlation (p < 0.05) between sub-regional uncertainty estimates and actual Dice scores for prostate anatomy. Leveraging Monte Carlo Dropout combined with multi-initialization to robustly estimate uncertainty, we construct an interpretable linear regression model that accurately predicts Dice scores—achieving low RMSE—without compromising the original segmentation performance. This approach bridges uncertainty estimation and segmentation quality assessment, establishing a new, clinically meaningful standard for evaluating trustworthiness in pathology AI systems.

Conventional binary classifier evaluation relies on single metrics, failing to holistically address diverse application scenarios. Method: This paper proposes a unified multi-perspective analysis framework based on Tile visualization, which innovatively maps infinite-dimensional ranking scores onto a two-dimensional Tile plot and geometrically models classifier behavior in ROC space—enabling performance comparison under arbitrary metric combinations. The framework supports four user categories (theoretical analysis, algorithm design, benchmarking, and application development) via customizable preference modeling and role-adapted “flavor” interpretations. Results: Empirical evaluation across 74 state-of-the-art semantic segmentation models demonstrates that a single Tile plot comprehensively captures model performance differences, significantly improving cross-task evaluation consistency, interpretability, and practical utility.

This study addresses the critical need for quantifying dataset similarity in model generalization, transfer learning, simulation calibration, and two-sample testing. We systematically survey 118 similarity quantification methods and propose the first ten-dimensional classification framework, organizing approaches into seven technical categories: statistical distances (e.g., Wasserstein distance, Maximum Mean Discrepancy), kernel-based methods, information-theoretic measures, dimensionality-reduction embeddings, permutation tests, generative-model-based discriminators, and Gaussian process likelihood ratios. We develop a multi-dimensional evaluation system balancing theoretical guarantees, interpretability, and practical applicability, yielding a structured recommendation matrix aligned with task requirements and data characteristics. Furthermore, we introduce the first open-source, interactive tool for method selection—enabling real-time filtering and parameter configuration—to significantly enhance both selection efficiency and deployment suitability.

Existing segmentation evaluation metrics often lack transparency and modularity, making them ill-suited for diverse tasks such as transparent object, specular surface, or lesion segmentation. This work proposes a unified evaluation framework that decomposes metrics into five modular components: prediction representation, target extraction, target matching, score computation, and metric reporting. For the first time, it systematically analyzes the implicit assumptions and design limitations of mainstream binary segmentation metrics through this modular lens. The framework enables task-aware customization of evaluation protocols, reveals evolutionary trajectories among existing metrics, and is accompanied by an open-source toolkit. By offering a principled and interpretable foundation, this approach paves the way for developing more rational and adaptable segmentation evaluation methodologies.

This study addresses the lack of systematic and neutral comparisons among similarity measures for categorical datasets. It presents the first comprehensive evaluation of several prominent methods—including edge-count tests, constrained minimum distance, graph-based tests, Classifier Two-Sample Tests (C2ST), and the Maximum Mean Discrepancy with Categorical Metrics (MMCM)—assessing their ability to detect distributional differences and their computational costs in both two-sample and multi-sample settings. The results demonstrate that the Friedman–Rafsky test achieves the best overall performance in two-sample tasks, while MMCM excels in multi-sample scenarios by offering both high statistical power and computational efficiency. This work provides empirical evidence and practical guidance for selecting appropriate similarity measures when analyzing categorical data.

This study addresses the frequent oversight in existing medical image segmentation methods regarding the fundamental influence of dataset characteristics on model design, which often leads to a mismatch between architectural choices and task requirements. To bridge this gap, the authors propose the Medical Segmentation Dataset Knowledge Card (MS-DKC) framework—a novel paradigm that systematically constructs a five-dimensional knowledge representation encompassing imaging/acquisition properties, morphological features, supervision schemes, contextual dependencies, and deployment risks. This framework explicitly links dataset attributes to failure modes, design priors, and risk-alignment criteria, enabling traceable, data-driven model customization. Experiments on DRIVE, ISIC2018, and ACDC demonstrate that models designed under the MS-DKC paradigm—such as MS-DKC-AttNextTopo-VCSF-NoAug, achieving a Dice score of 0.8872—significantly outperform generic architectures, thereby validating the efficacy and superiority of the proposed approach.

This work addresses the challenge of selecting high-quality data subsets from noisy labels to achieve performance approaching that of noise-free training. The authors observe that conventional k-nearest neighbors (k-NN) suffer degraded performance in high-dimensional, label-noisy settings and propose a novel approach that integrates symmetry and invariance priors into subset selection. Specifically, they introduce symmetry into the cutstats framework for the first time and theoretically demonstrate that leveraging invariance enables k-NN to asymptotically approach the Bayes optimal classifier. Moreover, they show that even with only partial knowledge of symmetries, effective modeling is achievable through learned symmetry-aware representations. Empirical results confirm that the proposed method substantially improves subset selection quality under high-dimensional label noise, yielding downstream model performance close to that attainable with clean labels.

研究通过对比专家和模型在rPCI评分上的差异，评估了深度学习分割模型的临床决策影响，发现rPCI评分对分割变异性具有鲁棒性，但在临界值附近仍需专家审查。