基于遗传算法的熵算法在人类染色体图像的分割中的应用

染色体图像分割是染色体图像分析与识别的重要内容之一。文中提出了实现染色体的自动分割的熵分割算法,该算法充分考虑了每个视野中分割目标(染色体)的数量大以及人类显带染色体被明暗交迭的带纹分割为若干不连续的片段等特点。文中还对传统的分割方法和熵方法进行了对比。传统的方法容易受噪声的影响从而将染色体分割开来。1维熵方法没有考虑灰度空间在空间的分布。2维熵方法在图像分割时较为费时。为了改善分割的精度和速度,在遗传算法的基础上运用1维和2维的熵方法。运用遗传算法的方法一场景分割不到一分钟并且得到得很好的染色体形态。算法的优越性通过实验来体现。

Entropic Methods for Segmentation of Human Chromosome Images based on Genetic Algorithm

Segmentation of chromosome image is an important aspect in chromosome analysis and recognition. Entropic algorithms realizing automatic segmentation of chromosomes in one scene were proposed. These algorithms adequately consider properties of chromosome images, such as a large number of chromosomes in one scene, many broken pieces in one chromosome with bright and dark bands, etc. The traditional methods were contrasted with two entropy methods. The traditional methods are easily infected by noise and to fragment the arms. 1D entropy method doesn't take into account the spatial distribution of grey levels. 2D entropy method has to pay the price of time when applied to image segmentation. In order to improve the segmentation precision and speed, both entropy methods are based on genetic algorithm. With these algorithms, most chromosomes in one scene can be separated in less than one minute, and results of segmentation fit well for external morphology of most chromosomes. The superiority of these algorithms is experimentally demonstrated.