The Cognitive Vision Lab (CoViL) focusses on the understanding of the human visual and cognitive system and the development of artificial systems with similar structure and strength. So far, only biological systems show strong cognitive capabilities such as object categorisation, advanced reasoning, planning and communication. Based on an elaborated stage of 'early cognitive vision', CoViL aims at the modelling of such cognitive functions in embodied systems. This modeling is done by means of the C++ Library CoViS (Cognitive Vision Software)
The early cognitive vision system we have developped computes semantically rich, reliable and task independent visual information from stereo image sequences. We have been applying this system in a number of contexts ranging from cognitive research to industrial applications. In particular, CoViL has been partner in the two EU projects PACO+ and Drivsco. PACO+ (2006-2010) was focusing on the modelling of cognitive processes in the area of humanoid robotics while Drivsco (2006-2009) was concerned with learning in the driver assistance domain. CoViL has been also involved in a number of Danish projects in the area of industrial robotics (Handyman,MoveBot) and image analysis (NISA). The currently running project IRFO is concerned with intelligent robots grasping flexible objects (2009-2012) while the project 'BinPicker' Bin-Picking (2009-2012). In 2011 three new EU projects in the area of cognitive robotics will be starting, namely Xperience (2011-2015), IntellAct (2011-2014) and LearnBiP (2011-2012).
As a by-product of our research, we have developed a new form of digital art (called 'Symbolic Pointillism') that makes functional processes in the brain explicit.
Currently we organize a workshop on 'Bootstrapping structural knowledge from sensory-motor experience ' in combination with ICRA 2013 .