A hybrid technique for overlapped chromosome segmentation of G-band mataspread images automatic

Abstract

Currently, the cytogenetic disorder such as the Down syndrome and the Prada william syndrome are a serious public health problem. The chromosome G-band images are applied for the automatic Karyotype analysis because it's easy to created and low cost. In contrast, the common problems of the chromosome images are disentangle such as the overlapping chromosome images. In this paper proposed a hybrid method for chromosome segmentation of G-band images. First, the Wayalun method is used to pre-processing the chromosome images. Then the Unsharp Mask and Zhang-Suen method are used searching one node in the overlapping chromosome skeleton images, and the chromosome edge detection is used the chromosome images edge detection algorithm (CEDA). Finally, this method searched the four cut points, in which used to separate the chromosome images overlapping, and the parts of the chromosome segmentation between point to point are created by the linear least squares fit method. The experimental results give the best performance for the automatic chromosome segmentation. A success rate of proposed method achieved at 97%.