Automatic Extraction Of The Data From Gene Microarray Images

With the development of human genome sequence plan and the breakthrough of molecular biology correlation technique, millions of biological information science data is waiting for extraction and analysis urgently. At the same time, computer and automation technology enhanced unceasingly, they are playing an irreplaceable role in data processing in many fields. The goal of my study is realizing the integral automatic extraction on the premise of the accurate gene localization and data of gene spots.The gene microarray chip technology has provided the highly integrated experimental tools. It can detect and analyze the massive gene samples. The chip preparation and scanning have realized the automation generally at present, but it always hard to realize the automatic extraction of the data from gene microarray image completely. The basic reasons are the microarray image's big data quantity, the high spot density, anomalous shape, the strong jamming noise and unobvious contrast ratio. Therefore, the difficulty and goal of my study is realizing the integral automatic extraction on the premise of the accurate gene localization and data of gene spots.In order to achieve this goal, firstly, this paper made the optimized improvement aiming at the existing non-automatic image processing flow. Drew the gene edges, located by grids, and then carried on flaw compensation. Secondly, the automatic image enhancement based on gray-scale morphology and automatic binarization methods is presented to automatic pre-processing, and made the edge detection using gene's own morphological feature. Thirdly, to realize the automatic image segmentation, carried on the grid localization by fast tilt method which based on the angle projection on two value charts, and solved the problems of flaw compensation and the adhesion division. Finally, according to the data contrast with the international authoritative software analysis result in the experiment, this paper used massive actual image data further confirmed the reliability, validity and integrity of the microarray image data extraction.