生物分子水平动态图像信息处理软件系统的建立与应用

The research is a kind of fundamental applied researches, which fuses many scientific disciplines, such as computer vision, neural networks, mathematical morphology, wavelet analysis, and biology. The methods, algorithms, databases and software for automatic analysis of biological images on the level of nanometer are also built in the research. Based on previous neural network models and algorithms, a modified network model and algorithms are established to process biological images rapidly, which are suitable to enhance dynamic biological images and recognize their textures. They also overcome the deficiencies of slow speed and low precision for the serial analysis of traditional image segmentation, and are applied to solve the problems for processing dynamic biological image on the level of molecule, e.g. the ultrastructure of nucleolus and the location of nucleolar DNA in situ, automatic analysis of image of surface structure of cell wall-deficient EVC, automatic analysis of high-resolution G band of chromosomes with neural networks, automatic quantitative analysis of bipolysaccharide liposome by image processing method. Among 15 scientific papers have made during the research, 7 papers have been published, among which there are 3 papers are published in the international journals that are embodied in SCI. The research will benefit the furthermore applied research in the biological fields and have great prospects.

神经网络、数学形态学和小波理论等学科与生命科学相结合，研究光钳和近场光学中生物活体、高分辨的动态结构图像信息处理、识别的方法、算法,建立自动分析软件系统。用于生镅е醒窍赴头肿铀降脑硕刺碌耐枷裥畔⒌募庸ず透骼喽问奶崛Ｌ岣咄枷裎评淼那逦群投员榷?保证图像定量分析精度，为生命科学基础理论研究提供先进手段。