多模态MRI脑肿瘤图像分割方法研究进展  

Research progress of multimodal MRI brain tumor image segmentation methods

作　　者：孙康康 陈伟[1,2,3] 李奇轩 孙佳伟[2,3] 焦竹青 倪昕晔 SUN Kangkang;CHEN Wei;LI Qixuan;SUN Jiawei;JIAO Zhuqing;NI Xinye(School of Computer Science and Artificial Intelligence,Changzhou University,Changzhou 213164,China;Department of Radiotherapy,theSecond People's Hospital of Changzhou Affiliated to Nanjing Medical University,Changzhou 213003,China;Central Laboratory ofMedical Physics,Nanjing Medical University,Changzhou 213003,China)

机构地区：[1]常州大学计算机与人工智能学院,常州213164 [2]南京医科大学附属常州第二人民医院放疗科,常州213003 [3]南京医科大学医学物理研究中心,常州213003

出　　处：《磁共振成像》2023年第11期164-169,176,共7页Chinese Journal of Magnetic Resonance Imaging

基　　金：江苏省重点研发计划社会发展项目(编号:BE2022720);江苏省卫健委面上项目(编号:M2020006)。

摘　　要：MRI是一种非侵入性多模态成像方法,广泛应用于脑肿瘤检测和诊断。多模态MRI脑肿瘤图像分割对脑肿瘤的诊断和治疗具有重要意义。目前大部分分割工作还是由医生手动完成,效率低且主观性强,因此寻求一种高效准确的脑肿瘤自动分割方法对临床应用至关重要。本文就基于多模态MRI图像的脑肿瘤分割的研究进展进行综述,对比分析了传统分割方法和基于深度学习的分割方法,并对现有的脑肿瘤图像分割方法存在的问题进行总结并做出展望,以便该领域的研究者更好地了解目前多模态MRI脑肿瘤图像分割方法的研究进展。MRI is a non-invasive multimodal imaging method,which is widely used in the detection and diagnosis of brain tumors.Multimodal MRI brain tumor image segmentation has important significance for the diagnosis and treatment of brain tumors.At present,most of the segmentation work is still manually completed by doctors,with low efficiency and strong subjectivity.Therefore,seeking an efficient and accurate automatic segmentation method for brain tumors is crucial for clinical applications.We reviewed the research progress of brain tumor segmentation based on multimodal MRI images,compared and analyzed traditional segmentation methods and deep learning based segmentation methods in this paper,and then summarized the problems of existing brain tumor image segmentation methods and makes prospects,so that researchers in this field could better understand the current research progress of multimodal MRI brain tumor image segmentation methods.

关 键 词：磁共振成像 脑肿瘤 多模态 图像分割 深度学习 

分 类 号：R445.2[医药卫生—影像医学与核医学] R739.41[医药卫生—诊断学]