Construction of Rolled Digital for Newborns Using Image Registration

• Wesley Augusto Catuzzo de Bona UTFPR http://orcid.org/0009-0008-6743-6892
• Jefferson Tales Oliva UTFPR https://orcid.org/0000-0003-1574-1293
• Luiz Fernando Puttow Southier UTFPR https://orcid.org/0000-0003-2420-4094
• Fabio Favarim UTFPR https://orcid.org/0000-0001-7490-8167
• Marcelo Teixeira UTFPR https://orcid.org/0000-0002-1008-7838
• Dalcimar Casanova UTFPR https://orcid.org/0000-0002-1905-4602

Abstract

This work aimed to reconstruct rolled fingerprint images from multiple flat fingerprints using image registration and composition techniques. The methodology involved selecting high-quality frames through a quality network, the application of preprocessing with remapping, cropping the area of interest, and adjusting colors. In addition to image registration steps, like key point identification, feature matching, and image distortion. Results indicated that while classical image registration had limitations, the composition technique significantly increased the number of features extracted, enhancing fingerprint identification. It was demonstrated that the composition method effectively captures important fingerprint details, offering a potential improvement in newborn biometric identification.

Keywords: Fingerprint, Newborn, Image Registration, Composition