BP神经网络算法在头部CT图像恢复中的应用研究  

Research on application of BP neural network algorithm in head CT image restoration

作　　者：陶春贵 江宇恒 王喜荣 刘晓燕 TAO Chungui;JIANG Yuheng;WANG Xirong;LIU Xiaoyan(School of Medical Imaging,Wannan Medical College,Wuhu,Anhui 241002,China)

机构地区：[1]皖南医学院医学影像学院,安徽芜湖241002

出　　处：《计算机应用文摘》2024年第11期48-51,共4页Chinese Journal of Computer Application

基　　金：校级中青年项目(WK2023ZQNZ04);国家级大学生创新创业训练计划(202210368051)。

摘　　要：CT影像检查是现代医学诊断的重要手段之一。然而,在CT设备成像过程中,电磁干扰、器件限制和人为操作等因素可能导致点扩散和高斯噪声等干扰,从而使成像质量退化,影响医生的诊断。因此,需要通过图像恢复算法来消除这些干扰,以提高影像质量。文章构建了点扩散和高斯噪声退化模型,并利用BP神经网络算法对头部CT图像的降噪恢复效果进行了对比研究。仿真结果表明,BP神经网络算法对头部CT图像的恢复效果较好,同时影响成像质量的主要因素是高斯噪声和像素移动。CT imaging examination is one of the important means of modern medical diagnosis.However,during the imaging process of CT equipment,factors such as electromagnetic interference,device limitations,and human operation may lead to interference such as point spread and Gaussian noise,resulting in degraded imaging quality and affecting the diagnosis of doctors.Therefore,it is necessary to eliminate these interferences through image restoration algorithms to improve image quality.The article constructs point diffusion and Gaussian noise degradation models,and compares the denoising and restoration effects of head CT images using the BP neural network algorithm.The simulation results indicate that,The BP neural network algorithm has a good restoration effect on head CT images,and the main factors affecting imaging quality are Gaussian noise and pixel movement.

关 键 词：神经网络 BP算法 CT图像 图像恢复 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]