Research On Technology Of Skull Identification

With the development of social modernization, rapid and accurate identificationtechnologies have gradually become one of the research hotspots. Skull as a new kind ofbiological characteristic are proposed attracted more and more attention of researchers. Theoutstanding advantage is that it can’t be interfered by facial expressions and light constraints.As one of the application fields of skull, skull identification technologies combine patternrecognition, image processing, data mining and biology. They are widely used in fields ofintelligent monitoring, machine intelligence, security and defense, and it is of profoundtheoretical value and broad application prospects.Methods based on the skull recognition technology and system framework are studied,the main research content includes: system design, skull image preprocessing techniques,contour extraction, feature extraction and clustering recognition technology. The main workof this paper is that the overall design of the system, methods improved on contour extractiontechnologies and clustering recognition techniques.First of all, in the system design, according to system requirements and the researchtarget, the overall framework of the system is designed, including the interface design,structure design, function module design and operation process design. Secondly, in cranialcontour feature extraction, an extraction method of X-ray skull lateral contour based on cornerdetection and Snake model is proposed. The improved Harris corner detection algorithm isused for extracting skull image stability of the corner, and then filters out contour corner andconnects each corner as the initial profile curve of GVF Snake model. GVF Snake modelalgorithm is used for face contour extraction. Then, in classification and identification, inorder to further accelerate the recognition speed, all samples are used for first clustering andthen classification. According to the initial clustering center sensitive problems of fuzzyC-means clustering algorithm, a new method of a sample density method for selecting theinitial clustering center is proposed. Finally, in system realization, a skull recognition systemis accomplished based on basis of the theory. The research techniques are verified and theexperimental results are analyzed and summarized.