A Scene Fingerprint Classification and Point Pattern Matching Algorithm Research in Police Identification

Although the verification （1∶1 mode） model’ fingerprint products can be found in market, the identification （1∶N mode） model’ fingerprint system in police is just what to pursue for a long time, which is of much importance to fingerprint identification theory and application field. Based on the classification model put forward by K.?Karu, adding strict restrict conditions, concepts, and definitions, this paper demonstrates its correctness. And feature space structure of fingerprint is analyzed and the concepts of registration conditions, matching support, and maximum matching pair are presented, completing a fast and precise matching with invariability of rotation, translation, and scale. The scene fingerprint experiments with 5000 fingerprints have verified the validity and correctness of the above algorithm.