THE IMAGE PROCESSING ALGORITHMS FOR BIOMETRIC IDENTIFICATION BY FINGERPRINTS

The article discusses image processing algorithms for biometric fingerprint identification. The identification features of the structure of papillary patterns on the fingers have been studied taking into account the fact that different pressure, speed, direction, ambient temperature and humidity level lead to different images. Due to various digital image processing and analysis algorithms such as: the SIFT descriptor, as well as the closest competitor, the SURF descriptor, it is possible to quickly obtain unique characteristics for each image. This study used a database of photographs obtained from open sources - the Fingerprint Verification Competition 2004 (FVC2004). As a result of the work, the graphic image of the matching key points, as well as the number of matched key points by fingerprints, have been investigated. Search key points is performed using the Hesse matrix. The determinant of the Hesse matrix (Hessian) reaches the extremum at the points of maximum variation of the brightness gradient. Fingerprints were obtained using the optical sensor "Cross Match V300". The experimental study showed that the developed software system has invariance to image rotations.