2018/01/17 by Chen Shen, Holger R. Roth, Shen, Chen +11 · 2 citations
Computer Science · Engineering · Medicine · #Advanced Neural Network Applications #Advanced X-ray and CT Imaging #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Medical Imaging Techniques and Applications #Radiomics and Machine Learning in Medical Imaging
paper · pdf · doi:10.48550/arxiv.1801.05912
openalex publication_date 2018/01/17 · openalex created_date 2019/07/30 · openalex updated_date 2026/07/28
Deep learning-based methods achieved impressive results for the segmentation\nof medical images. With the development of 3D fully convolutional networks\n(FCNs), it has become feasible to produce improved results for multi-organ\nsegmentation of 3D computed tomography (CT) images. The results of multi-organ\nsegmentation using deep learning-based methods not only depend on the choice of\nnetworks architecture, but also strongly rely on the choice of loss function.\nIn this paper, we present a discussion on the influence of Dice-based loss\nfunctions for multi-class organ segmentation using a dataset of abdominal CT\nvolumes. We investigated three different types of weighting the Dice loss\nfunctions based on class label frequencies (uniform, simple and square) and\nevaluate their influence on segmentation accuracies. Furthermore, we compared\nthe influence of different initial learning rates. We achieved average Dice\nscores of 81.3%, 59.5% and 31.7% for uniform, simple and square types of\nweighting when the learning rate is 0.001, and 78.2%, 81.0% and 58.5% for each\nweighting when the learning rate is 0.01. Our experiments indicated a strong\nrelationship between class balancing weights and initial learning rate in\ntraining.\n