Analýza vlivu dermatologických poškození u otisků prstů na jejich kvalitu

Abstract

This thesis deals with the analysis of a database containing fingerprints influenced by of skin diseases and evaluates their quality using the NFIQ 2.3 algorithm. Subsequently, one of the approaches for generating damage to synthetic fingerprints was selected and further improved to support grayscale images and extended to simulate low pressure, interruption of papillary lines, and influence of acrodermatitis continua disease. After implementing these methods, 16 000 synthetic fingerprints with corresponding damage were generated, and their quality was evaluated using the NFIQ 2.3 algorithm. The results showed that the newly designed simulations lead to a realistic decrease in fingerprint quality. Fingerprints affected by acrodermatitis continua the worst quality scores. A significant difference was also observed between binary and grayscale images, with grayscale images exhibiting lower quality, greater variance, and higher susceptibility to quality degradation when damaged.