Uncertainty-Guided Lung Nodule Segmentation with Feature-Aware Attention
This work addresses segmentation variability in medical imaging for lung nodule analysis, representing an incremental improvement by incorporating uncertainty from multiple annotations.
The paper tackles the problem of inconsistent lung nodule segmentations from radiologists by proposing an Uncertainty-Guided Segmentation Network (UGS-Net) that learns from multiple annotations to predict regions with different uncertainty levels, achieving superior performance on the LIDC-IDRI dataset.
Since radiologists have different training and clinical experiences, they may provide various segmentation annotations for a lung nodule. Conventional studies choose a single annotation as the learning target by default, but they waste valuable information of consensus or disagreements ingrained in the multiple annotations. This paper proposes an Uncertainty-Guided Segmentation Network (UGS-Net), which learns the rich visual features from the regions that may cause segmentation uncertainty and contributes to a better segmentation result. With an Uncertainty-Aware Module, this network can provide a Multi-Confidence Mask (MCM), pointing out regions with different segmentation uncertainty levels. Moreover, this paper introduces a Feature-Aware Attention Module to enhance the learning of the nodule boundary and density differences. Experimental results show that our method can predict the nodule regions with different uncertainty levels and achieve superior performance in LIDC-IDRI dataset.