智慧医学视域下基于连续帧上下文信息与扩散模型的医学影像分割方法研究

打开文本图片集

中图分类号TP391文献标识码A文章编号：2096-7721（2025）06-0887-06

Application of medical image segmentation method based on continuous frame context information and diffusion model in the perspective of smart medicine

CHEN Lixia, LI Jun, SHI Hui,GU Jianan, WANG Yong (DepartmentofRadiology,Nantong Tumor Hospital,Nantong, China)

AbstractObjective:To propose a medical image segmentation method basedoncontinuous frame contextual information and difusionmodeltoaddresstheissesofmissed diagnosisand misdiagnosiscaused byexcessiverelianceonphysician experiencein MRIdiagnosis,aiming toimprovediagnosticaccuracyandeficiency.Methods:232rectalcancerpatients whowere treatedatthe DepartmentofGastrointestialSurgery,NantongTumorHospitalfomJanuaryO21toJanuary225wreselectedasthestudysujects. Based on different MRI image segmentation methods,they were divided into the group A( n =116) and the group B( n =116). Group A receivedtraditioalsgmetationethod,hileroupBnderwenttheproposedmedicaliagesgmentationasedoncotiuous framecontealfadelgicacsgalittitte of disputes werecomparedbetweenthetwo groups.Results:ComparedwithgroupA,shorterdiagnostictime,lowerratesofmised diagosis,desllodoi nursing quality were observed in group B with statistically significant (P

KeyWordsMagneticResonanceImaging;Continuous FrameContext Information;DifusionModel;Medical ImageSegmentation; Nursing Quality

直肠癌是全球最常见的恶性肿瘤之一，其发病率和死亡率均较高，严重威胁人类健康[。（剩余13028字）