Um sistema de reconhecimento facial aplicado a um ambiente virtual de aprendizagem composto por agentes pedagógicos

Abstract

Currently, most of Learning Virtual Environments (VLEs) has no efficient mechanisms that can identify students access to both the system and during the development of their activities. In general, this type of system adopts the use of security mechanisms for authentication and login password. The use of this simple type of authentication increases vulnerability to fraud as to the identity of that which carries the course. Based on this context, this work presents a face recognition system based on techniques of Principal Component Analysis and Autofaces, and investigates their accuracy in identifying the student during the entire course in an VLE. The results show an accuracy exceeding 90% in the identification of student achievement throughout the course in VLE consists of pedagogical agents.

Index Terms - Learning Virtual Environments, Face,EigenFaces, PCA, Pedagogical Agents.