Improving the Segmentation of Anatomical Structures in Chest Radiographs using U-Net with an ImageNet Pre-trained Encoder
This work improves segmentation accuracy for anatomical structures in chest X-rays, aiding computer-aided diagnosis, but it is incremental as it builds on existing U-Net architectures with pre-training.
The paper tackled the problem of multi-class segmentation of lungs, heart, and clavicles in chest radiographs by evaluating U-Net variants with pre-trained encoders and different loss functions, achieving Jaccard scores of 96.1%, 90.6%, and 85.5% respectively, outperforming prior state-of-the-art methods.
Accurate segmentation of anatomical structures in chest radiographs is essential for many computer-aided diagnosis tasks. In this paper we investigate the latest fully-convolutional architectures for the task of multi-class segmentation of the lungs field, heart and clavicles in a chest radiograph. In addition, we explore the influence of using different loss functions in the training process of a neural network for semantic segmentation. We evaluate all models on a common benchmark of 247 X-ray images from the JSRT database and ground-truth segmentation masks from the SCR dataset. Our best performing architecture, is a modified U-Net that benefits from pre-trained encoder weights. This model outperformed the current state-of-the-art methods tested on the same benchmark, with Jaccard overlap scores of 96.1% for lung fields, 90.6% for heart and 85.5% for clavicles.