NUMEDIART boasts internationally recognised expertise in deep learning and computational attention. These tools are now experiencing considerable growth and are the reason for the revival of artificial intelligence. They are at the heart of most of the work carried out within the Institute, which can be categorised around 5 main axes:
Motion capture (MOCAP) and analysis of the resulting data:
These analyses can be carried out for the purpose of recognition, evaluation and/or comparison of movements, for the animation of synthetic characters and for the gesture control of sound synthesis.
Expertise in the “tracking” and interaction of people in indoor spaces, in the analysis of visual attention based on non-invasive systems, and in the prediction of visual attention on images and videos using artificial intelligence algorithms.
Understanding and navigating multimedia data:
NUMEDIART has unique expertise and algorithms in the field of the analysis and automatic understanding of media content and, in this framework, offers customised and innovative human-machine interactions
Natural human-computer interaction and voice technologies :
NUMEDIART develops cutting-edge research into voice technologies, resulting in algorithms for voice analysis, transformation, recognition and synthesis. As well as voice analysis, non-verbal and expressive behaviours are also studied to facilitate interaction with the user or to influence their mood. This is the domain of affective computing.
Virtual reality and augmented reality:
Projection of multimedia/3D content onto complex or moving surfaces, in interaction with users; use of headsets for augmented and virtual reality with optimised positioning of virtual objects; automatic understanding of real 3D spaces for the rapid creation of virtual worlds.