图像分割算法在胃溃疡图像上的应用  

Image Division Algorithm in the Application of Gastric Ulcer Image

作　　者：周学友[1] 傅忠谦[1] 何力[1] 

机构地区：[1]中国科学技术大学电子科学技术系,安徽合肥230026

出　　处：《计算机仿真》2010年第5期197-200,共4页Computer Simulation

摘　　要：主要研究了对胃溃疡图像中溃疡轮廓进行精确提取的问题。首先要解决对胃溃疡图像进行图像分割时,复杂多变的背景噪音对分割效果的影响,利用小波变换的多尺度特性对胃溃疡图像进行边缘提取,获得不同尺度下的边缘提取图。通过不同尺度下的边缘效果和溃疡边界的连通性,将不同尺度下获取的边缘再次进行对比和提取,把降低噪音的影响和保持边界连续性有效的统一起来,得到效果更佳的边缘提取图。同时利用梯度矢量流模型将轮廓收敛至溃疡的边界上,并对轮廓内的面积进行计算得到溃疡的面积。计算结果显示,利用小波变换和梯度矢量流模型后得到的溃疡面积,已经能够达到临床医学的要求。This paper studied the images of the gastric ulcer accurate contour extraction problem. The wavelet multi - scale edge detection algorithm, regional growth and gradient vector flow algorithm were used to solve the image contour extraction ulcer encountered background noise and computational complexity of the problem of high precision. First, the segmentation of ulcer images with the complex background noise must be solved. We used wavelet transform to extract the image edge in multi - scale to get different scales＇extracting verge map. Through different scales of edge effect and ulcer border connectivity, different scales were obtained by comparing the edge again and extraction. The reducing of noise impacts and the maintaining of the continuity of border enabled better extracting of the edge map. On this basis, the gradient vector flow model was used to converge the outline to he border of ulcer and the size of ulcers was calculated within the area . The results show that byusing wavelet transform and gradient vector model, the calculated ulcer area, , is able to meet the requirements of clinical medicine.

关 键 词：动态轮廓模型 小波变换 图像分割 梯度矢量流 多尺度 胃溃疡 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]