The Segmentation And Statistic For Human Chromosome Image

Human chromosome identification is an important research topic for medical genetics, which has widely been applied in clinical medicine, teaching aids and scientific research. It is the basic accordance to judge the human genetic disease either. Traditional chromosome identification is carried out artificially. The process of identification not only is trivial but also make mistake easyly. In order to relieve the medical staff and teaching staff and improve the efficiency of the recognition, the technology of chromosome image identified automatically has been applied. Up to now, this technology is still affected by the images quality and image processing methods. The reason, whith the identification result is not satisfactory, is resulted by the segmentation mainly . It is necessary to explore the chromosome image segmenting methods.In this paper, the chromosome image made by the main function include: Image preconditioning, which is transformation of chromosomes gray-scale and the removal of impurities of impurities of image. Image Segmentation, which is firstly adopting the largest-entropy method, regional growth method to select the image threshold . The regional growth method is based on the selected seed point which expands outside to the edge according to the guidelines; then adopting the boundary detection means, the Canny grads operators to check the chromosome images boundary. And then, according to the result of the analysis and comparation, the regional growth based on the boundary detection are adopted and the experiment result is satisfactory. In the end, binary chromosome image is eroded by the morphology image processing method. Then marks the connectivity part of the image, the correct number of chromosome is carried out.According to experimentation, the algorithm of the region growing based on the edge detection for chromosomes image segmentation are effective.