Webface: design of an identity verification system in PHP

Abstract

[en] As web browsers improve their capabilities and customers get access to faster internet connections, the market is experiencing an increase on the services located in the web. This has opened the gate for the implementation of systems that, not long ago, used to seem unfeasible due to the technical limitations intrinsic to client-server applications. In this context, the project presents a design of an identity verification system integrated on a web application programmed in PHP. The implemented web-server user registration makes use of face detection and optical character recognition techniques to gather the information needed to verify the identity of the user. Such decision is made by using a deep neural network, and requires the user to provide an identity document. In order to simplify, the only document supported is the most recent version of the Spanish national identity document, referred as DNI 3.0. The scope of this study is use existing libraries and machine learning techniques (including trained classifiers) for the purpose of identity validation, and not not to modify and re-train the classifiers.