基于细胞图像肺癌诊断系统的设计与实现

基于细胞图像的肺癌诊断系统主要是利用数字图像处理和模式识别的技术对肺癌细胞的图像处理，并根据提取出的细胞特征对肺癌进行早期的病理诊断．不同于以往的细胞诊断系统，文中提出了新的细胞分割和重叠细胞重构的方法．首先把彩色肺癌细胞图像转化为灰度图像，对其进行平滑、去噪，然后用一种新的基于强化学习的方法寻找合适的灰度阈值，把细胞区域分割出来，形成二值图像，并对图像进行基于形态学的二值滤波，再针对此时分割出的重叠细胞，利用一种改进的deBoor—Cox方法分离与重构，最后进行特征提取，根据提取出来的多种特征对细胞分类，诊断出肺癌细胞．

Design and Implementation of Lung Cancer Diagnosis System Based on Cell Images

In cell image based Lung Cancer Diagnosis System, Digital Image Processing and Pattern Recognition are used to process the lung cancer images. Therefore, thus early pathological diagnosis can be made based on the features extracted from these lung cancer images. The paper proposes a novel method to extract cell area and reconstruct the overlapped ones. Firstly, the colored images are transformed to gray images and filtered. Secondly, a reinforcement learning based method is used to find optimum threshold to segment the images into Bi-level images after filtering the Bi-level images. Besides, the paper uses a B spline based and improved deBoor-Cox method to divide and reconstruct the overlapped cells, then some morphological and color features are extracted from the cell images and at last these features are used to classify the lung cancer cells.