Phase-aggregated Dual-branch Network for Efficient Fingerprint Dense Registration

arXiv:2404.17159v1 Announce Type: new
Abstract: Fingerprint dense registration aims to finely align fingerprint pairs at the pixel level, thereby reducing intra-class differences caused by distortion. Unfortunately, traditional methods exhibited subpar performance when dealing with low-quality fingerprints while suffering from slow inference speed. Although deep learning based approaches shows significant improvement in these aspects, their registration accuracy is still unsatisfactory. In this paper, we propose a Phase-aggregated Dual-branch Registration Network (PDRNet) to aggregate the advantages of both types of methods. A …

abstract arxiv class cs.cv deep learning differences fingerprints improvement inference low network performance pixel quality registration shows speed type while