extracting optimized minutia for feature extraction in fingerprint classification

Today, human’s behavioral and physical characteristics are used to identify them. Themost important biometric feature used to validate or identify individuals is fingerprint. On theone hand, the increase in the number of individuals, and the development of fingerprintdatabases, make it impossible to identify fingerprints traditionally. So, automatic fingerprintidentification is needed. In automatic fingerprint identification systems, an input fingerprintimage should be compared to all the images in the database, which is a very costly and timeconsuming task. As a result, the first step in the fingerprint automatic identification process isto classify the database.Classifying a large collection of fingerprint images into several sub-collectionsdramatically reduces the time to search and identify an unknown fingerprint image in thatcollection. A content-based retrieval system usually involves two aspects of extractingfeatures and classifications. The first step in the proposed method is to perform preprocessingon fingerprint images. Afterwards, various properties of the images are extracted.Also, in the proposed method, minutia (split and end points) are extracted from the imagesand stored as a set of two-dimensional points as templates. Then, based on the distancecomparison, similar minutia are found and returned as the result of the matching operation.

کلیدواژه ها:

Classification ، fingerprint images ، minutia ، feature vector