Arguably, one of the most important problems in Computer Vision today
is the lack of vast amounts of data required to train our
algorithms. Many problems (e.g. face detection and recognition in the
wild) have been characterised as solved during the past years, however
the data used to train such systems are not publicly available due to
privacy limitations. To this end, I will present several databases
created to tackle these problems as well as the way we have employed
Neural Networks towards the problems of affect/age recognition and 3D
face reconstruction. Finally, I will also present future research how
our brain perceives music.