彩色血液细胞图像的分割

Segmentation of Blood Cell Image

本文系统回顾了对彩色血液细胞图像进行分割的彩色空间和传统方法(包括阈值法、边缘检测法和流域分割法),也介绍了若干新的方法,如基于形变模型的边缘检测法、基于小波变换的流域算法和基于数学形态学的颗粒分析法。指出了要更加准确的对细胞图像进行分割,需要采用统计模式法和神经网络法技术等方法。

This paper systemtically reviews the color space and traditional approaches (including threshold,edge detection , watershed and merging), and introduces several new methods, such as edge detection based on deformable template , watershed based on wavelet transform and grain analysis based on mathematical morphology. It is suggested that statistical pattern recognition and neural net technique should be applied to achieve image segmentation more accurately.