The Research Of FISH Image Analysis System

The paper on the basis of researching the existing medical image processing platform researches the fluorescence microscopy image of the gastric tumor cells labeled by FISH probe, and preprocesses&segments the cells in the FISH image by use of this medical image processing system, and extracts the signals of HER-2DNA specific probe and chromosome17Chromosome centromeric which is as its internal control, and carries on quantitative analysis, so as to achieve the purpose of guiding clinical outcome.As Fluorescence images is under the influence of the intake devices, surrounding environments and auto fluorescence of some organizations in the in-taking process that makes the in-taken image contains noise, the primary task is to reduce the noise of the extracted fluorescent region in the image. This paper firstly introduces the characteristics of the FISH image, and studies on the binary morphology and gray-scale mathematical morphology in summary. And proposes a new color mathematical morphology algorithm based on the HSL color space and strengthens the visual effects of image with the idea of contrast enhancement combined top hat transform and low hat transform. Fluorescent regions must be segmented from the image to analyze and process. By analyzing the existing image segmentation algorithm and combining with specificity of fluorescence images the paper proposes fluorescence image adaptive segmentation algorithm basing on the threshold method. Fluorescence region does not one-to-one correspond with cells and the quantitative calculation of the signal bases on cells, so there are a variety of feature-based extracting methods for the extraction of the cells. The ratio of the red and green signals is calculated by tracking the color in the cells, so as to achieve the diagnosis and prognosis of gastric carcinoma cell line gene amplification. There is practical significance and broad for medical research&teaching and clinical diagnosis.Figure32table2reference54...