Research And Implementation Of Biochip Image Analysis

With the advance of Human Genome Project, bioinformatics has been promptly developed in recent 20 years. As a branch of a new class of bioinformatics which allows the monitoring of expression levels for thousands of genes and proteins simultaneously, biochip technologies are increasingly applied to broader fields.Image analysis is an important aspect of biochip technologies, which faces vast amount of data processing. Since the performance of image analysis impacts biochip technology deeply, how to construct the biochip image analysis methods with cost-effectivity, accuracy and rapidness has become the chief problem which should be solved eagerly.By analyzing the present developing situation of the field of biochip image analysis, this thesis put forward that to realize the automatic methods is the key issue of biochip image analysis. The thesis presents the detail research on biochip image analysis, including:Research and construct the working flow of biochip image analysis and data processing.Research on the gridding methods for biochip arrays, then the automatic gridding algorithm of biochip image based on projection is presented.Aiming at the common problems about the incline of biochip image, the automatic image rotation correction is realized by analyzing the power spectrum density of image projection.Based on analysis and comparison of the existed biochip image segmentation methods, the adaptive circle segmentation algorithm combined with the neighborhood searching algorithm is realized to biochip image segmentation.Data extraction and analysis, the main topics of biochip data processing, are researched, and used to solve the data processing problem of protein chips.Design the software structure of biochip image analysis workstation, then the biochip image analysis workstation based on PC and the software based on embedded system is developed.The result with biochip image analysis on clinical experiments proves that the methods in paper are cost-effective, accurate and rapid. With the pattern of automatic analysis, it is possible to achieve to process large-scale biochip data effectively.