肺癌细胞形态量化和结构特征的抽取  

FORM QUANTIZATION AND STRUCTURE FEATURE EXTRACTIONOF LUNG CANCER CELLS

作　　者：王洪斌[1] 熊海涛[1] 苏万芳[1] 胡匡祜[1] 黄志英[2] 

机构地区：[1]中国科学院生物物理所 [2]北京结核病胸部肿瘤研究所

出　　处：《生物物理学报》1999年第1期79-83,共5页Acta Biophysica Sinica

基　　金：国家自然科学基金

摘　　要：研究了形态十分复杂的肺癌细胞的结构特征，建立了肺癌细胞图像的抽取方法和算法以及相应的ＣＡＮＣＥＲ．ＣＥＬＬ自动分析软件。采用数学形态学的多种逻辑运算和灰度梯度法，并建立高分辨率掩模数字图像扫描方法。该方法可有效地抑制图像中的噪声和干扰，并较好地保持图像的清晰度。应用结果表明所建立的方法提高了肺癌细胞的分析速度和精度，并获得了五种类型肺癌细胞形态的量化结果。In this paper, research on structure feature of lung cancer cells that have complicated form was reported. Image extraction method of lung cancer cells and it's automatic analysis software system CANCER. CELL have been developed in light of logical arithmetic of mathematical morphology, gray-gradient method and high resolution digital image mask scan method which effectively restrained noise and interference in the image and mantained the clearness. Applications demonstrated that the established method improved the speed and precision of analysis. And the quantitative result of five-type lung cells was obtained. The research provides basis of automatic recognition of lung cancer cell.

关 键 词：肺癌细胞 形态量化 特征抽取 

分 类 号：R734.202[医药卫生—肿瘤]