基于HSI颜色模型的多层彩色细胞提取技术  

Technology of Multi-storey Color Cell Extraction Based on HSI Color Model

作　　者：侯发忠[1,2] 邹北骥[1] 周天亮[2] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083 [2]怀化医学高等专科学校,湖南怀化418000

出　　处：《计算机仿真》2009年第6期267-271,共5页Computer Simulation

摘　　要：为了克服传统图像分割方法在识别复杂背景下的细胞时存在的捕捉不全、噪声敏感和计数不准确的缺憾,构造了一种基于HSI空间下的细胞识别新法,通过在复杂背景图片中实选出几部分目标点的RGB模型信息,转换成优化的HIS模型,再用多图像平均法求出其H分量的直方图范围,并依之提取目标特征,实现细胞识别及计数。仿真实验结果表明:在处理染色不均、光照过强、背景复杂、细胞粘滞等特殊细胞图片,尤其是目标细胞与背景的色彩特征相似度高且多层分布的细胞图片时,方法输出的特征参数与实数吻合度高、计数准确,解决了临床应用上多层细胞难识别的问题,弥补了传统图像分割方法对此类图片处理的不足。In order to overcome the defects of the traditional image segmentation method in recognizing complex cells such as incomplete capturing, noise - sensitive and inaccurate count, the article constructes a new algorithm based on HSI space. Through selecting RGB model information of some target points, and transforming it into the optimization HIS model, obtaining its H -component of the histogram,the cell recognition and count are realized. The simulation results show that this method can solve the multi -cell recognition problem better.

关 键 词：颜色模型 分量 分割 特征 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]