一个用于肺癌早期诊断的彩色图像处理系统

介绍了一个用于肺癌早期诊断的彩色图像处理系统，该系统能够识别痰液涂片彩色显微图像中的肺癌细胞。为了提高识别癌细胞的稳定性与有效性，系统采用了一种细胞分割与分类的分层处理结构，首先，在某一特定的归一化彩色空间中利用自适应阈值的方法对细胞核进行分割，利用细胞核的形态学特征检出可疑癌细胞，最后用细胞核区的色度学特征和纹理特征对可疑癌细胞进行分类与识别，从而确定是否存在癌细胞。

A Color Image Processing System for Diagnosis of Early Stage Lung Cancers

In this peper, a color image processing system for diasnosis of early stage lung cancers is introduced, which can recognize lung cancer cells in color microscopic image of sputum slides. To improve the stability and efficiency of the recognition of cancer cell, a hierarchical processing architecture is adopted for the segmentation and recognition of cancer cells. First, a new adaptive thresholding method, based on the Bspline technique, is used to segment the nuclei in a special color space. Then, the suspicious cancer cells are selected using their nuclear morphological features. As the last step, all the suspicious cancer cells are recognized using the chromatic and textual features of the nuclei to determine the real cancer cells. Experiment results are given to show the feasibility of the approach proposed here.