Feature Reduction of Local Binary Patterns Applied to Face Recognition

In recent years, Local Binary Patterns have proved to be a powerful local descriptor for microstructures of images, having been introduced in many facial recognition systems and intelligent environments. In this work, we present the implementation of a face recognition method based on the use of Local Binary Patterns. We used data mining tools to get a smaller feature vector and thus improve the computational cost of the system. The implementation was tested with the Color FERET database, obtaining a recognition rate of 94% and reducing 75% the original feature vector dimension.