The Research And Application Of Fingerprint Core Automatic Seek

The technology of Automated Fingerprint Identification is used widely in several occasions which need automatic personal verification , such as finance , information technology , public security , medical treatment and welfare etc . How to find the center of a fingerprint has still not been resolved completely. It is found by person or person and computer together . In this way , the processing speed is restricted . So in this paper the emphasis of the research is to find out the core of the fingerprint automatically . On the other hand this paper dose some research on fingerprint image pretreatment , fingerprint classification , fingerprint feature extraction and fingerprint matching . And we give some new arithmetics in the realization process .1. In this paper , we adopt the block direction filtering method , the self-adapt two-valued disposing method , the morphological two-valued image filtering and the morphological two-valued image thinning to complete the fingerprint image pretreatment . The pretreatment processing can wipe off the breaks and forks of the fingerprint gray image.2. In this paper , we researched and realized several fingerprint core automatic seek arithmetics , such as the Poincare Index method , the point direction consistency measurement, the block direction tracing arithmetic , the direction average difference method and the maximal block direction number . Poincare Index method may obtain several fingerprint cores . The improved Poincare Index method can obtain the core of many fingerprint images , but can not obtain the core of some images which are classified as arch . The block direction tracing arithmetic can find out the center area of every kinds of fingerprint images . The synthesis method is established on these two methods . The experimentations show that this method can seek the center point of the fingerprint image accurately.3. Based on the place of fingerprint core and the variation of direction in the fingerprint center field , the fingerprint images are classified as six categories , which are Whorl, Arch , Tented Arch , Left Loop , Right Loop and Sundry .4. Using templet , the detail features are extracted and stored in the corresponding fingerprint file . Twice-matching method is used to identify the fingerprints . The experimentations show that in a general way the identification veracity is more than eighty percent.5. Twice-matching method is very robust and it can ignore the affects of image rotation and distortion.