Preprocessing on Fingerprint Image and Minutiae Extraction
|School||Dalian University of Technology|
|Course||Control Theory and Control Engineering|
|Keywords||Pretreatment Image Enhancement After-treatment Refinement|
Fingerprint unique and stability , \Fingerprint recognition has been a long period of time , has been officially accepted , becoming an effective means of identification in the legal profession . Fingerprint recognition has the advantages of convenient , efficient , safe , reliable , and have been widely used in the financial security , digital encryption , e-commerce and other fields , and will play an increasingly important role in our future life . Fingerprint recognition generally include the following steps : pre-processing , refining , post-processing , feature extraction , feature matching . In accordance with the above steps , first introduced the development status of the fingerprint identification and related technologies , and discusses in detail the composition of the Automated Fingerprint Identification System , as well as part of the algorithm , fingerprint identification works and pretreatment . Design a set of pre-processing algorithms , which focus on image preprocessing enhancement algorithms binarization algorithm based on directional information of the fingerprint image pattern is obtained , and then use this pattern design directional filter on the original image . filtering noise, this method is a good solution to the most prone to fingerprint image breakpoints and the fork even the phenomenon . After using the dynamic threshold binarization method greyscale image binarization . Last of eigenstructure refinement fingerprint image pretended , in accordance with the characteristics they have , using different algorithms targeted remove them one by one that is carried out after the treatment , resulting in a clear and accurate fingerprint dotted line in Fig . Fingerprint minutiae extraction is the 8 neighborhood coding to achieve through the target point , the specific process as the through 8 o domain coding to represent the type of the target point and its neighborhood states of the pixels , according to the gradation values ??of the surrounding points different determine the bifurcation point and endpoints of these two feature points , and they are subsequently extracted . Fingerprint identification system with VC programming on the Windows platform , test the good results of the algorithm processing fingerprint images , which can effectively enhance the fingerprint image , to remove most of the pseudo - feature points .