Rozpoznání typu onemocnění kůže v otiscích prstů

Abstract

This thesis focuses on the analysis of fingerprints affected by dermatological conditions. The aim of the work was so design and verify an algorithm capable of automatically evaluating whether a fingerprint is suitable for further analysis and estimate the type of damage affecting it either localized or widespread. The algorithm was implemented in the Python programming language and employs standard image processing methods such as filtering, binarization, skeletonization and minutiae detection. A real-world dataset of fingerprint images from patients with various diagnoses was used for testing. The results show that the proposed approach can serve as a useful tool for preliminary fingerprint quality assessment in the context of skin disease, with the best performance achieved on fingerprints with localized damage.