Boosted projections and low cost transfer learning applied to smart surveillance

Abstract

Computer vision is an important area related to understanding the world through images. It can be used in biometrics, by verifying whether a given face is of a certain identity, used to look for crime perpetrators in an airport blacklist, used in human-machine interactions and other goals. Deep learning methods have become ubiquitous in computer vision achieving many breakthroughs, making possible for machines, for instance, to verify if two photos belong to the same person with human-level skill. This work tackles two computer vision problems applied to surveillance. First, we explore deep learning methods for computer vision in the task of face verification and second, we explore dimensionality reduction techniques for the task of detection.