Information Extraction Of DNA Microarray Image

Microarray technology provides a powerful approach for genomics research, since itcan assays large-scale gene sequences once a time and assists gene expression in analy-sis.This unique technology allows for molecular biologists and bioinformaticians to identifysimultaneously thousands of genes and predict their functionality within a larger system.Now according with experimental technology development, the processing precisionand efciency of the experimental data is becoming more and more demanding. Many ofthe existing methods in order to achieve more suitable and accurate rate with the method ofartificial participation, But which not only reduces the work efciency but also may be in-evitable to the additional error. so how to efciently and accuratly extract image informationhas been a hot spot research to many reseachers. par After microarray experiment all theinformation have been stored in DNA microarray image, the pixel value of each gene pointreflects expression level to the corresponding gene, so the focus of processing microarrayimage is to find genes point position and extract its pixel values. In this paper,it introducesin detail the automatic measuring method of the level to gene expression based on waveletanalysis put forward by Ghislain Bidaut etc. Ihe major ways are to make projection in hor-izontal and vertical direction and using Fourier transform to find the best Correction Angleto image and achieve the interval s between genes; Applying s can do some morphologi-cal close operation to a image, segment subarray; Dealing with wavelet decomposition ofimage in two directions, which respectively process pattern recognition to find genes pointcenter by a graph H in horizontal direction and a graph V in vertical dirction, accordingto the center position of attachment drawing grid for the first time, at the same time gettingthe corresponding gene point diameter; But for some image griding only a time may not beable to find the corresponding gene point diameter all at, So using the average distance from s to correct grid lines, along with increasing grid lines, it can be thinked that gene pointsexist in grid intersection point; Finally, extracting image information by the center of genepoint and diameter of, whicn is worth and need to be corrected by using the backgroundvalue. par One of the main reasons that it is now difcult to recognise DNA microarrayimage nition difcult is due to all kinds of noise, so the appropriate denoising method hasvery big efect to the final results. par In this paper, having a attempt to the wavelet thresh-old denoising, KSVD denoising method and the combination of these two denoising method,combining Ghislain Bidaut information extraction method based on wavelet, a lot of contrastexperiment have been done, which shows that the first wavelet denoising to KSVD denoisingefect is th best one.