摘要
提出了一种基于完备色差模型的分割方法.利用优化的颜色空间模型转换公式,将宫颈癌细胞图像从RGB颜色模型转换为HSV颜色模型,对H、S、V通道分别进行最大类间方差直方图均衡化,构造最优色差模型实现图层重组,然后通过阈值分割结合特征参数筛选完成细胞从复杂背景中的提取.在VC++和OpenCV开发系统予以实现.实验结果表明,在处理背景与目标细胞色彩相似的宫颈癌细胞图片时,能够得到准确、完整的分离结果,弥补了传统分割方法的不足.
In order to overcome the defects of the traditional segmentation about color cervical carcinoma cell, a method of segmentation based on majorized color model was presented. To begin with, the color image is transformed from RGB model to HSV model, and the biistogram equalization of maximum classes variance is used to deal with every channel of the color image. Through making up an optimized model about color aberration and regrouping the processed single channel image, cervical cell can be segmented under complex background by the threshold combining with some characteristic parameters. The experimental results show that this method is accurate and effective.
出处
《武汉大学学报（理学版）》
CAS
CSCD
北大核心
2014年第3期275-278,共4页
Journal of Wuhan University:Natural Science Edition
基金
国家自然科学基金(81272443)资助项目
关键词
色差模型
直方图均衡化
图像分割
特征筛选
color aberration model； histogram equalization； image segmentation； feature selection
