Fingerprint Classification Based on the Henry System via ResNet

A Comprehensive Review of Unimodal and Multimodal Fingerprint Biometric Authentication Systems: Fusion, Attacks, and Template Protection

“…Using unidentified scenarios with LivDet 2011 to LivDet 2017, experimental results showes a decrease in overall classification error rates approximately fourfold, with an accuracy of 96.17%(139). CNN models such as Darknet, Alexnet, Resnet, Deep Belief Network, and VGG16 were utilized to develop the Henry Classification System as discussed by Souza et al When tests were conducted using grayscale and previously processed images as input, the best accuracy of 95.1% on NIST database 4 is achieved when the Gabor filter and morphological thinning operation were combined(140).Nahar et al developed a CNN-based finger impression affirmation approach, without image preprocessing. The frame combines the coordination and extraction steps.…”