Segmentation Study Of CDNA Microarray Image Based On EED Filtering And Snake Model

With the rapid development of bio-science and the human genome project has beenfinished, studying on the biological chip has become the common topic of life scienceworkers around the world. Analysis of the gene chip information has been put into the coreposition belonging to the technical field of biochip. Gene chip technology which principleis same to the classical nucleic acid hybridization method, a number of molecular nucleicacid was fixed on the carrier following the predetermined position to consist of themicroarray lattice. To take the advantage of molecular hybridization and parallelprocessing, it able to analyze and detect a large amount of information at the same time. Itis a significant innovation and leap compare to the traditional biological DNA detectiontechnique and make a great push in following field such as gene expression and analysis,virus detection, new drug discovery and equipment development. So the image processingas an effective analysis means is important to the gene chip research.After nearly thirty years of research, it has developed some mature analysisalgorithms of traditional gene chip. Mainly processing chain is made up of imagepreprocessing, gridding, Target segmentation and signal extraction, which the targetsegmentation is the key point of the processing and affect the signal extraction directly,in addition, it is also the difficult point especially with the gene chip in randomarrangement. Since they could not gridding, the traditional segmentation method will notwork any more. To solve this problem, this paper proposed a segmentation method basedon Snake model.This paper describes the whole process of gene chip image comprehensively beginfrom image gripping. Firstly,converting the image to gray scale, enhancing the imageby contrast stretch and Top-Hat transform, and According to the gene chip image has thecharacteristics of multi boundary, this paper proposed the edge enhancing diffusion(EED)filter to remove the noise,smooth and preserve the contours simultaneously. And thefiltering effect is better than other commonly used filtering method. Then the active contour segmentation based on Snake model has been used to segment the image, whichhas an excellent performance on intensity inhomogeneity and weak boundary. Finallyextracting the signal intensity and doing visual expression based on the segmentationresults.As the adhesion phenomenon of the target often found in the image, this paper mainlyuses Snake model and also joins the Watershed method, which generates the dividing lineby effectively solves this problem.