胎儿、婴幼儿及儿童脑MRI的人工智能分割研究进展.pdfVIP

国际医学放射学杂志ＩｎｔＪＭｅｄＲａｄｉｏｌ２０25Nov鸦48穴6雪：666-671

·早期发育脑和脑疾病的MRI研究专题·

综述

胎儿、婴幼儿及儿童脑MRI的人工智能分割研究进展

112

，

魏铭宇，董素贞*

【摘要】人工智能（AI）技术的快速发展，使得胎儿、婴幼儿及儿童脑MRI影像得以精细分割。近年来，对胎儿

MRIAI

至儿童这一早期脑发育阶段的结构与功能影像的分割方法不断发展，如机器学习、卷积神经网络、

Transformer、通用大模型等，使得分割精度、效率、鲁棒性均取得了显著进步，且在临床应用中具有较大的潜力。现就

MRIAI

胎儿、婴幼儿及儿童脑的分割的研究进展进行综述。

【关键词】磁共振成像；人工智能；脑分割；胎儿；婴幼儿；儿童

中图分类号：R445.2文献标志码：A

Researchprogressonartificialintelligence-basedbrainMRIsegmentationinfetuses,infants,andchildren

11,2

WEIMingyu,DONGSuzhen.1DepartmentofRadiology,ShanghaiChildren’sMedicalCenter,ShanghaiJiaoTong

UniversitySchoolofMedicine,Shanghai200127,China;2ShanghaiJiaoTongUniversity.Correspondingauthor:DONG

Suzhen,E-mail:dongsuzhen@126.com

【Abstract】Therapiddevelopmentofartificialintelligence(AI)technologyhasenabledincreasinglyprecise

segmentationofbrainMRIinfetuses,infants,andchildren.Inrecentyears,AI-basedsegmentationmethodsforstructural

andfunctionalMRIduringtheearlystagesofbraindevelopment,fromfetustochildhood,haveadvancedcontinuously.

Techniquessuchasmachinelearning,convolutionalneuralnetworks(CNNs),transformers,andgeneralfoundationmodels

haveachievedremarkableimprovementsinsegmentationaccuracy,efficiency,androbustness,andhaveshowngreat

potentialforclinicalapplications.Thisreviewsummarizesrecentadvances