基于细胞显微图像特征的ROI自动提取算法  

AN AUTOMATIC ROI EXTRACTION ALGORITHM BASED ON THE FEATURE OF MICROSCOPIC CELLS IMAGE

作　　者：黄良孟[1] 孙玲玲[1] 李训根[1] 周磊[1] 

机构地区：[1]杭州电子科技大学微电子CAD所,浙江杭州310018

出　　处：《计算机应用与软件》2009年第3期41-43,共3页Computer Applications and Software

基　　金：浙江省科技计划重点项目(2007c21015)

摘　　要：为了在细胞显微图片中快速有效地自动提取细胞聚集的区域,提出了一种基于细胞显微图像特征的感兴趣区域(ROI)自动提取算法。该算法以图像的自动阈值分割为基础,利用图像膨胀后边缘的融合和重叠消亡原理,以图像特征的相似度为判别手段,最后用八邻域轮廓跟踪法取得轮廓线。对大量细胞显微图片的实验表明,该算法成功实现了对细胞显微图像的感兴趣区域沿边缘全自动快速提取。In order to quickly extract the effective regions in a microscopic cells image automatically, an automatic ROI （ regions of interest） extraction algorithm based on the feature of microscopic cells image, inspired by the theory of amalgamation and overlap decease of the image edges when the image has dilated, is presented. The algorithm is based on image automatic threshold segmentation, uses the similarity of image features as the discrimination means to finally obtain contours with eight neighbourhood contour tracking algorithm. Experiments on a set of microscopic cells images demonstrate that this method can get the precise ROI automatically and fleetly.

关 键 词：感兴趣区域 提取 膨胀 图像特征 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] Q343[自动化与计算机技术—计算机科学与技术]