太赫兹科学与电子信息学报, 2018, 16 (2): 307, 网络出版: 2018-06-09
基于自组织神经网络的多模态MRI图像分割
Multimodal MRI image segmentation based on SOM network
摘要
磁共振成像 (MRI)是一项重要的医学成像技术, 在人体组织器官的诊断治疗方面被广泛应用。在脑肿瘤的临床诊断应用中, 如何实现脑肿瘤图像的有效自动分割是一个研究的难点和重点。利用多个自组织神经网络 (SOM)构造一个并行自组织神经网络 (CSOM), 将肿瘤图像的分割问题转化为并行自组织神经网络的分类问题。实验表明, 并行自组织神经网络的应用, 有效提高了分割精确度, 有利于自动分割的实现。
Abstract
Magnetic Resonance Imaging(MRI) is a distinctly important technology of medical imaging, which is widely used in the diagnosis and treatment of tissues and organs of the human body. In the clinical diagnosis of brain tumor, it is a challenge for how to achieve effective automatic brain image segmentation. Multiple Self-Organizing feature Maps(SOM) are utilized to create a Concurrent Self-Organizing Map(CSOM) for the whole segmentation process to realize the brain tumor image segmentation. The results show that CSOM model used in the brain tumor image segmentation is effective and successful in this design, improving the precision and reducing the segmentation duration triumphantly, makes a progressive step to the automated segmentation.
王磊. 基于自组织神经网络的多模态MRI图像分割[J]. 太赫兹科学与电子信息学报, 2018, 16(2): 307. WANG Lei. Multimodal MRI image segmentation based on SOM network[J]. Journal of terahertz science and electronic information technology, 2018, 16(2): 307.