Implementación de un sistema de identificación facial utilizando procesamiento digital de imágenes para el personal del bioterio de la Facultad de Ciencias de la ESPOCH.

Abstract

The objective of this thesis was to implement a facial identification system through the digital image processing for the vivarium staff at the Faculty of Science of the Higher Polytechnic School of Chimborazo. The Eigenfaces method was the procedure used for digital image processing, built on techniques of Principal Component Analysis (PCA), and together with the main variables correlated as: luminosity, user-camera distance, facial position, projection thresholds and Euclidean distance, for data storage, a local database and the Matlab software were used to develop the identification system because it allows to use the necessary facial recognition tools for identification. A total of 180 tests were carried out in three different days with a total number of 20 users, registering 3 access attempts for each one. The final outcome was an average percentage of 86% of system validity, registering a rating scale of at most 96% and at least 80%. Additionally, the combination was made with an electrical system for the controlled opening of an electric lock, which was implemented on the Vivarium door. It is concluded that the facial identification system implemented at the faculty of sciences Vivarium has an efficiency of 86% with the different characteristics established according to the correlation variables. It is recommended to analyze the respective scenario to determine the digital image processing method in order to implement a facial identification system.