Abstract: In the multi-view domain, it is challenging to correctly label multiple people across viewpoints because of occlusions, visual ambiguities, appearance variation, etc. Deep learning, although ...