Performance of MyNET Model on Handwriting Biometrics

Abstract

This paper is in the continuation of Writer Identification using Neural Network[1]. Offline signature identification is the challenging task till date. Unlike to verification problem of one-to-onemapping, identification maps single with rest them to identify the signer. For this process we model a model using convolution neural network. This paper explains the performance of MyNET model on MyNET offline signature dataset which consists of 434 writers 20 signatures each.