We Care Each Pixel: 医学分割模型的校准
We Care Each Pixel: Calibrating on Medical Segmentation Model
AI总结：
针对现有医学图像分割指标未评估校准质量的问题，提出像素级校准误差指标、形态学适配策略及符号距离校准损失，同时提升分割性能与校准可靠性。
摘要：
医学图像分割是计算机辅助诊断的基础，可精准勾勒解剖结构与病理区域。Accuracy、DSC、IoU、HD 等常用指标主要量化预测结果与真实标签的空间一致性，却未评估分割模型的校准质量，而校准质量对临床可靠性至关重要。为解决这一局限，我们提出像素级期望校准误差（pECE），这是一种全新指标，可明确衡量像素层面的校准失准，从而同时保障空间精度与置信度可靠性。我们进一步引入形态学适配策略，在计算校准损失前对真实掩码施加形态学操作，尤其对 Margin SVLS、NACL 等基于边界的损失有益。此外，我们提出符号距离校准损失（SDC），通过惩罚预测符号距离函数（SDF）与真实符号距离函数的差异，使边界几何结构与校准目标对齐。大量实验表明，我们的方法不仅提升了分割性能，还改善了校准质量，得到更可信的置信度估计。代码可在以下链接获取：https://github.com/EagleAdelaide/SDC-Loss。
英文摘要：
Medical image segmentation is fundamental for computer-aided diagnostics, providing accurate delineation of anatomical structures and pathological regions. While common metrics such as Accuracy, DSC, IoU, and HD primarily quantify spatial agreement between predictions and ground-truth labels, they do not assess the calibration quality of segmentation models, which is crucial for clinical reliability. To address this limitation, we propose pixel-wise Expected Calibration Error (pECE), a novel metric that explicitly measures miscalibration at the pixel level, thereby ensuring both spatial precision and confidence reliability. We further introduce a morphological adaptation strategy that applies morphological operations to ground-truth masks before computing calibration losses, particularly benefiting margin-based losses such as Margin SVLS and NACL. Additionally, we present the Signed Distance Calibration Loss (SDC), which aligns boundary geometry with calibration objectives by penalizing discrepancies between predicted and ground-truth signed distance functions (SDFs). Extensive experiments demonstrate that our method not only enhances segmentation performance but also improves calibration quality, yielding more trustworthy confidence estimates. Code is available at: https://github.com/EagleAdelaide/SDC-Loss.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
