The Segmentation Of Cell Images Based On The Improved Watershed Algorithm

Different cells have different morphological structure, a biological cell image includes a variety of biological information. In order to accurately describe and analyze the information, it is necessary for quantitative biological cell image analysis. Different cell images show the different shape feature, therefore, when analyze the cells image, should extracted different target of interest according to the needs. Which laid the foundation for cell image segmentation technology. Using image segmentation techniques can extract the cell structure of the various structures to the next stage of analysis. Therefore, the image segmentation techniques in image processing and analysis will focus on the analysis of the image preparation phase.With the medical image processing and analysis technology, medical image segmentation techniques have been developed rapidly. Among existed and proposed variety of segmentation research, not an algorithm is common. The reason is that in practical applications in medical image, its complex with the diversity and complexity. Medical images for different people you want to extract the details of the targets vary in different ways by the nature of medical images. So far more than the segmentation method is the integrated use of various segmentation methods. For different image characteristics, segmented the image of specific purpose.Morphological watershed segmentation algorithm is a commonly used methods. As the watershed algorithm is the use of the image gray gradient, so sensitive to noise, that directly using watershed algorithm to separate images easily lead to over-segmentation. Therefore, in order to solve the problem over-segmentation, segmentation is often combined with other methods of image preprocessing or segmentation of regional integration.In this paper, for the characteristics of cell images, a ner algorithm is proposed based on the morphological watershed algorithm. The main method is combination of morphological operations, using alternating sequential filtering of image filtering and then using multi-scale morphological gradient instead of morphological gradient. Using the open reconstruction to reduce the minimum markers, thereby reducing the over-segmentation resulting from the regional. Analysis with specific examples, that marked-based watershed segmentation is a practical and effective cell image segmentation.