Local Binary Pattern with Hyperparameter Tuned Support Vector Machine for Fingerprint Classification
Abhijeet Chougule, Medha Y. Shah · 2019
In recent times fingerprint authentication has become essential part of many organisation. With increase in use the risk of spoofing the fingerprint is also increased.Previously many software approaches has been used to solve this problem. Hyperparameter tuning shows improvement by Increase in Accuracy Score(IAS%) upto 27.43%. The proposed model of Local Binary Pattern with Hyperparameter tuned Support Vector Machine achieves Average Accuracy Score(AAS) of around 98.69% on LivDet 2009, 96.39% on LivDet 2011, 97.57% on LivDet 2013, 94.84% on LivDet 2015 dataset.