AUTOMATIC ANALYSIS OF G BAND OF CULTIVATED WHEAT CHROMOSOME USING NEURAL NETWORK
小麦染色体高分辨率G带的神经网络自动分析的研究

This paper presents a Double Kohonen Neural Network (DKNN) lased on the Self-Organize Feature Mapping theory to analyge chomosome automatiolly. The DKNN consists of two layers, each of which is a Kohonen network. The first layer maps a 2-d plane to a 2-d feature space to extract the high-resolution bands of chromosome and to compute the parameters of bands. The second layer maps the arrays of high dimension feature parameters to a 2-d plane to pair and classify the chromosomes automatically. The result indicates that this method can extract the feature parameters of G band in the chromosomes of cultivated wheat and then pair them automatically, rapidly and accurately.