多模态医学图像配准算法综述OA北大核心CSTPCD
Review of multimodal medical image registration algorithm
以多模态医学影像(多序列磁共振(MR)、计算机断层扫描(CT)和X光等模态)为研究对象,对近10 a的多模态医学影像配准相关研究工作进行归纳和分析.首先阐述面向临床应用的多模态医学配准的必要性,分析影像配准的一般流程;然后提出三种归纳角度对现有研究进行分析,重点总结多序列MR、CT和X光等模态之间的五种配准模式、五类常见配准的解剖结构及四类常用的多模态影像配准算法;接着分析七个用于多模态医学影像配准数据集和六个常用的配准评价指标;最后指出多模态医学影像配准算法面临的挑战和未来趋势.
The multimodal medical image(magnetic resonance(MR),computed tomography(CT),Xray and so on)registration was investigated,and a comprehensive analysis and summary of the research work related to multimodal medical image registration in the past decades were carried out.First,the necessity of multimodal medical image registration was introduced,and the general process of image registration was analyzed.Then,three inductive perspectives were proposed to analyze existing research.Five registration modes between multiple sequences of MR,CT and Xray modalities,five common anatomical structures for registration,and four commonly used multimodal image registration algorithms were focused on and summarized.Seven multimodal medical image registration benchmarks and six popular evaluation metrics were analyzed.Finally,challenges and potential future research directions for multimodal medical image registration algorithm were indicated.
冯筠;邓佳慧;周末;陈宝莹
西北大学信息科学与技术学院,陕西 西安 710127西安国际医学中心医院影像诊疗科,陕西 西安 710061
计算机与自动化
医学影像多模态配准优化算法监督学习无监督学习形变估计
medical imagemultimodal registrationoptimization algorithmsupervised learningunsupervised learningdeformation estimation
《华中科技大学学报(自然科学版)》 2024 (005)
29-49,157 / 22
国家自然科学基金面上资助项目(62073260).
评论