Thai Handwritten Verification System on Documents for the Investigation

Abstract

Forensic biology is the application of biology to law enforcement. Handwriting behavior is one of the biometrics that can be used to identify those who are ownership. This paper proposes an algorithm for Thai handwritten verification on documents. This work intended to prove the handwriting in the investigation. The main issue presented in this paper is divided into three processes including data preparation, classification and results. Consequential, two steps of neuron network algorithm are used to resolve in classification with high accuracy. Then, the models of handwriting are created. Finally, we define criterion to compare between unknown handwritten with our model. The result shows height accuracy 90.00%.