期刊
ARTIFICIAL INTELLIGENCE IN MEDICINE
卷 140, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.artmed.2023.102559
关键词
Magnetic resonance images; Spine; Semantic image segmentation; Convolutional neural networks; Deep learning; Ensembles of classifiers
This study addresses challenges in medical image segmentation, including image variability, human anatomy variation, disease severity, and age and gender effects. The researchers used convolutional neural networks to automatically segment lumbar spine MRI images, assigning class labels to each pixel corresponding to structural elements. The proposed network topologies, with complementary blocks, achieved more accurate segmentation than the standard U-Net, especially when used in ensembles.
Significant difficulties in medical image segmentation include the high variability of images caused by their origin (multi-center), the acquisition protocols (multi-parametric), the variability of human anatomy, illness severity, the effect of age and gender, and notable other factors. This work addresses problems associated with the automatic semantic segmentation of lumbar spine magnetic resonance images using convolutional neural networks. We aimed to assign a class label to each pixel of an image, with classes defined by radiologists corresponding to structural elements such as vertebrae, intervertebral discs, nerves, blood vessels, and other tissues. The proposed network topologies represent variants of the U-Net architecture, and we used several complementary blocks to define the variants: three types of convolutional blocks, spatial attention models, deep supervision, and multilevel feature extractor. Here, we describe the topologies and analyze the results of the neural network designs that obtained the most accurate segmentation. Several proposed designs outperform the standard U-Net used as a baseline, primarily when used in ensembles, where the outputs of multiple neural networks are combined according to different strategies.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据