The Research of Multilevel Fingerprint Classification Based on Large-scale Database
|Course||Computer Software and Theory|
|Keywords||Fingerprint identification Fingerprint quality evaluation Singular point location Largecapacity fingerprint database Multilevel classification retrieval|
With the development of information society,Security of Personal Information faced tremendous challenge.Fingerprint recognition technology as the representative of biometrics technology for its high accuracy and reliability features is becoming one of the ways to solve the information security. When the application of fingerprint technology more is widely used, the research of large fingerprint database becomes a hotspot.If we don’t build a indexing mechanism for the databases input fingerprint image will be contrasted one by one with a large database of fingerprint and he system will be in a high-load operation state.To reduce the search time and complexity of the algorithm, the database must use the keyword to multi-levelly classify.Therefore.with the main research line of automated fingerprint classification and retrieval technology based on large-capacity fingerprint database, the paper designed a three-tier system of fingerprint classification.The efficiency of fingerprint retrieval is greatly improved.The paper gets a achievement in Ithe fingerprint image preprocessing, fingerprint feature extraction and fingerprint classification and retrieval.The main achievement are:1.In the pre-processing stage of fingerprints,the paper achieved the key process steps of Fingerprint preprocessing in the basis of analyzing and learning fingerprint preprocessing algorithm from domestic and foreign scholars.Including the fingerprint orientation computation, image segmentation, image enhancement, binarization and thinning to two and getting accurate results of image data.2.A new method based on the combination of global and local features is studied to evaluate the quality of fingerprints.The quality of fingerprint images is evaluated hierarchically through different evaluation indices. If the fingerprint image is not qualified, the process of evaluation is ended and the user is reminded to re-enter another one.3.Fingerprint feature extraction algorithm is studied, putting forward a method for extracting hierarchical singularity.In order to initially identify the singular region,the curvature of fingerprint image is estimated, then using improved algorithm of the Poincare index accurate positioning of the singular point value in the singular region.So that the modified version of poincare Index can locate the singularities quickly and exactly, and it can improve the speed and accuracy of Fingerprint Classification.4.The fingerprint classification algorithms is studied.In order to match fast matching recognition of large capacity fingerprint database,designing a three-level Fingerprint Classification System, which using fingerprint pattern,Ridge Count and the average ridge frequency as general characteristics of the fingerprint to classify fingerprint. Through the experiment data comparison and analysis algorithm, the retrieval algorithm of high efficiency, strong robustness, provides an efficient indexing mechanism for a large amount of data matching fingerprint database.