Research On Adaptive Image Enhancement And Segmentation For Microarray Image

Gene chip is a kind of biological chips,which also known as DNA chip.The technology fixed mass probe molecules on solid support and then hybrid with the sample molecules which was marked,thus it can obtain Gene expression information.Gene chip technology is widely used in drug research and screening,disease diagnosis,environmental protection,modem agriculture and so on.The gene chip image processing is a critical step of gene chip technology,which include image preprocessing,spot gridding,spot segmentation and information extraction.The paper is focus on the image preprocessing and spot segmentation research,details are as follows:Firstly,image preprocessing is been done.In this section,current stat of art denoise methods which has good performance is been realize,and then analyze advantages and disadvantages of the methods.Furthermore,the noise level function is using to assessment and judge roughly noise categories,the corresponding denoise method is using to get rid of the noise after noise category is known,and then leads to adaptive denoising.Secondly,based on the characteristic of gene chip image,a new adaptive contrast enhancement method based on histogram equalization is introduced,which can estimate the intensity of the background and enhance the spot contrast only.According to the Visual analysis of human eyes and gridding results,it is obvious that the proposed method perform better than traditional contrast enhancement methods.Finally,this paper summarize and analyze the traditional spot segmentation methods which are sensitive to image quality such as noises,low contrast and so on.a new clustering method based on multi features,including intensity features,spatial features is used to spot segmentation.After that a adaptive adjust is adopted to reduce noise and eliminate missing spots.Experimemets based on simulated and real datasets results show the proposed algorithm is better than traditional clustering methods.