Experiences accelerating features selection in Viola-Jones algorithm

Abstract

Faces and facial expressions recognition is an interesting topic for researchers in machine vision. Viola-Jones algorithm is the most spread algorithm for this task. Building a classification model for face recognition can take many years if the implementation of its training phase is not appropriately optimized. In this study, several settings for implementing the training phase are analyzed. The aim was to share our experiences when we try to accelerate the training phase using one computer with a graphical processing unit (GPU). For each setting, the execution times were analyzed and compared with previous studies. Although we don't contribute to break new ground in topic or methodology, we decide to share our experience in order to show an antecedent working with a cheap GPU with the aim that this can be useful to another for to make comparisons.