Biometric recognition model using deep convolutional neural networks and computer vision

Ovalle C.; Qquenta D.S.; Sucapuca J.V.; Qquenta R.S. · Proceedings of the LACCEI international Multi-conference for Engineering, Education and Technology (2023)

Resumen

Authentication of the person by means of unique features such as finger veins is used in various fields such as security. In this research, a verification method based on convolutional neural networks with the help of computer vision is proposed. Through experimentation, it was possible to create an artificial intelligence model that shows the measurements of loss and precision when verifying the images of a data set. Finally, it is concluded that the loss of biometric recognition, the lower the percentage, the better the model performs. For the modified VGG16 model, 40 epochs were carried out for training, while the Mobilenet was 50 epochs. Additionally, at the end of the execution, the proposed architecture finished in 13 minutes and 45 minutes.

Instituciones: Universidad Privada del Norte; Universidad Peruana Unión