医学DR图像分割处理过程仿真分析

摘要

The article put forward a method based on artificial fish algorithm of medical DR image segmentation processing method.By enhancing the DR image which is captured to increase the contrast of the DR image,SUSAN operator was used to remove the disturbed signal,calculate the boundary of the initial lesion area,and seek the optimal boundary pixels according to the artificial fish method.The experiment results show that the accuracy of DR medical image segmentation is improved effectively by the optimized algorithm,which is helpful for clinical medical diagnosis.%研究医学DR图像准确分割方法.人体组织的分布特征很难运用准确的数学模型进行描述,由于厚度不均匀,在进行CT图像采集的过程中,图像细节信息会被噪声等不利因素埋没,边缘变得不清晰,对比度降低.传统分割算法主要针对像素的某一个具体特征做出判断,在有噪的环境下,像素提纯受到干扰,很难对非可控信息进行模型控制,导致对医学DR图像的分割效果不好.为了避免上述缺陷,提出了一种人工鱼群算法的医学DR图像分割处理方法.通过对采集的DR图像进行增强处理,提高DR图像的对比度,利用SUSAN算子去除干扰信号,准确计算初始病变区域边界,对边界像素运用人工鱼群方法寻求.实验结果表明,利用改进算法能够有效提高医学DR图像分割的准确性,有利于临床医疗诊断.