Radiologists have different training and clinical experiences, which may result in various segmentation annotations for lung nodules, causing segmentation uncertainty. Conventional methods usually select a single annotation as the learning target or try to learn a latent space of various annotations, but these approaches waste the valuable information of consensus or disagreements ingrained in the multiple annotations. In this paper, we propose an Uncertainty-Aware Attention Mechanism (UAAM) that utilizes consensus and disagreements among multiple annotations to facilitate better segmentation. To achieve this, we introduce the Multi-Confidence Mask (MCM), which is a combination of a Low-Confidence (LC) Mask and a High-Confidence (HC) Mask. The LC mask indicates regions with a low segmentation confidence, which may cause different segmentation options among radiologists. Following UAAM, we further design an Uncertainty-Guide Segmentation Network (UGS-Net), which contains three modules: a Feature Extracting Module that captures a general feature of a lung nodule, an Uncertainty-Aware Module that produces three features for the annotations' union, intersection, and annotation set, and an Intersection-Union Constraining Module that uses distances between the three features to balance the predictions of final segmentation, LC mask, and HC mask. To fully demonstrate the performance of our method, we propose a Complex Nodule Validation on LIDC-IDRI, which tests UGS-Net's segmentation performance on lung nodules that are difficult to segment using U-Net. Experimental results demonstrate that our method can significantly improve the segmentation performance on nodules with poor segmentation by U-Net.

翻译：放射科医生具有不同的训练背景和临床经验，可能导致对肺结节的分割标注存在差异，进而产生分割不确定性。传统方法通常选择单一标注作为学习目标，或尝试学习多种标注的潜在空间，但这些方法浪费了多标注中蕴含的一致性或分歧信息。本文提出一种基于不确定性感知的注意力机制（UAAM），该机制利用多标注间的一致性和分歧信息来优化分割效果。为此，我们引入多置信度掩膜（MCM），它由低置信度掩膜（LC Mask）和高置信度掩膜（HC Mask）组成。其中，低置信度掩膜标识分割置信度较低的区域，此类区域可能导致放射科医生间产生不同分割方案。基于UAAM，我们进一步设计了不确定性引导分割网络（UGS-Net），该网络包含三个模块：特征提取模块用于捕获肺结节的通用特征，不确定性感知模块用于生成标注并集、交集及标注集对应的三种特征，以及交并约束模块利用三种特征间的距离来平衡最终分割、低置信度掩膜和高置信度掩膜的预测结果。为全面验证方法性能，我们提出LIDC-IDRI复杂结节验证方案，该方案测试UGS-Net对U-Net难以分割的肺结节的分割性能。实验结果表明，本方法可显著提升U-Net分割效果较差的结节的分割性能。