Study Of Gridding Gene Chip Image Based On Genetic Algorithms And Deformable Template

The technology of gene chip is a new technology with the development of life sciences and information technology and has become an efficient method of finding the information of biology molecule. The gene chip images processing belongs to the part of data analysis in the application of gene chip technology. It's an important part of the gene chip images processing gridding gene chip images. Previously the gridding was semi-automatic or by hand in applications and was affected by subjective factor. So it's not suitable to analyze a batch of gene chip images data.The method of deformable template matching based on genetic algorithms to auto-gridding is adopted in the paper. The method finds the optimum transforming parameters for the optimum match of deformable templates and original images by genetic algorithms. Then transforming the standard grid by the optimum transforming parameter and gridding the original gene chip image by the standard grid, the algorithm finishes the auto-gridding without interactive operation.First of all, the paper introduces the pre-prepare work for the algorithm, that is the pre-processing of gene chip images .The part includes denoise and segmentation .The method of denoising is wavelet denoising with phase preserving and the segmentation is adaptive thresholding. Finally, the acquired denoised binary image is referred to input image for the algorithm. The paper discusses the basic theory of genetic algorithms including schemate theorem, building block hypothesis, implicit parallelism, the analysis of astringency and so on, as the theoretical base of application. Then introducing template matching and deformable template matching, translating template matching into optimizing parameter and inducting genetic algorithms into deformable template matching, it's convenient to resolve the optimization problem.The algorithm is realized in VC++, concentrating on how to select genetic parameters and analyzing the result of the experiment. The auto-gridding method proposed in the paper acquired is better effect and has the promising value in research and application.