Active Appearance Models for People Recognition (Master Thesis)

AAMs is a computer vision procedure for statistical matching of object shape and appearance between images. The present work analyzes how AAMs can be employed with RGB-Depth technology, so robots endowing a Kinect sensor can perceive and recognize humans more effectively in order to unequivocally identify them. This research is associated to the RoCKIn@Home challenge, an initiative from the European RoCKIn project focusing on domestic service robots.