Análisis de Métodos de Reconocimiento de Rostros Utilizando Descriptores SIFT y HOG Aplicado en la Escuela Nicanor Larrea.

Abstract

The research focuses on the analysis of face recognition methods using descriptors Scale Invariant Feature Transform and Histograms of Oriented Gradients, in order to know what is the most effective, applied at Dr. Nicanor Larrea School. The scientific method to perform a comparative analysis of the descriptors, for which two prototypes based on the theoretical foundations investigated and focused on the area of control attendance record of the teachers was built was used as a guide. Number of lines of code, runtime, percentage of use of processor and percentage of memory usage, the indicators were assessed with the help of tools: Line Tally, Process Explorer, Resource Monitor, and Matlab. Histograms of Oriented Gradients descriptor obtained 93.75% effective against 68.75% of Scale Invariant Feature Transform descriptor. Histograms of Oriented Gradients descriptor was selected for developing the system for recording attendance of teachers, the methodology was used in cascade, tools Matlab as a programming language and database engine MySQL. In conclusion Histograms of Oriented Gradients has 25% more effective than Scale Invariant Feature Transform descriptor. Building a face recognition device it is recommended for Electronic students using the Histograms of Oriented Gradients descriptor.