Category: Seminars and Conferences
State: Archived
27 May 2019 at 2:30 pm

Computer vision for perception, representation and learning

Sala Teatro, Links foundation

Computer vision and machine learning have largely contributed to understanding data and scenes and have achieved impressive results in detection and recognition of scenes and objects, mostly mature for industrial transfer. The research field is however still limited when it comes to modelling and learning from videos, to reasoning on objects, parts and scenes, and to developing universal representations and memory which fit multiple tasks. I would present my research and interests in the field, with particular reference to most recent own work on people trajectory forecasting and re-identification.