摘要
医学图像主要包括CT、MRI、X线、超声等,其中超声检查价格低,对软组织成像效果好,对人体基本无伤害,目前临床已广泛应用。超声图像分割对后期图像分析有很大作用,可为临床诊断及放疗摆位等提供参考。本文就超声图像分割的传统方法、基于形变模型的分割方法及结合深度学习方法的研究进展进行综述。
Medical images mainly include CT,MRI,X-ray and ultrasound.Compared with other imaging methods,ultrasound is cheaper and has better imaging effect on soft tissue with little damage to human body,so it is widely used in clinic.Ultrasound image segmentation plays an important role in later image analysis,and it can provide reference for clinical diagnosisand radiotherapy placement.This study reviews the traditional methods,deformation model methods and deep learning methodsfor ultrasonic image segmentation.
作者
张钒
陆正大
李春迎
焦竹青
倪昕晔
ZHANG Fan;LU Zhengda;LI Chunying;JIAO Zhuqing;NI Xinye(School of Microelectronics and Control Engineering,Changzhou University,Jiangsu 213164,China)
出处
《临床超声医学杂志》
CSCD
2022年第6期453-456,共4页
Journal of Clinical Ultrasound in Medicine
基金
常州市医学物理重点实验室项目(CM20193005)
江苏省卫健委面上项目(M2020006)。
关键词
超声
图像分割
深度学习
Ultrasound
Image segmentation
Deeping learning