Image Signal Texture And Fingerprint Recognition Technology

In each field of image processing, the texture as an important imagetype has been paid much more attention day by day. The feature descriptionand detection of texture are very important to discern image signal,compress high-efficiently of image and adaptive control technology ofspeed of data stream. Texture analysis including various content of imageprocessing, has extensive foundation. Meanwhile, according to thecharacteristic of the texture, various kinds of pointed analysis theoriesare proposed. With the deepening study, it has already become anindependent research field gradually. So, the analysis and research ofthe texture are theoretically meaningful and practically valuable.In this dissertation, the detection and recognition technology oftexture characteristic are studied emphatically. Various kinds oftexture analytical method are applied to fingerprint recognition system,and the results of computer simulation indicated that the technique isfeasible and system is excellent.In texture feature description, the coarseness and orientationfeature of texture are analyzed from space auto-correlation and powerspectral density distribution of texture, two methods of measurecoarseness: space auto-correlation function and fractal dimension arepresented in this dissertation. Then, a conception of vectorialorientation field model of the striped texture image is proposed, andthe method of building vectorial orientation field is researchedemphatically.After building vectorial orientation field of the striped textureimage, the texture segmentation and enhancement technique based onorientation feature are studied. In texture segmentation technique,a8-D feature vector of texture is constructed, and fuzzy C means clustering...