Report on the EUVIP 2021 Panel Session : Visual Information Processing and Machine Learning at the Crossroads

Abstract

Over the last few years, we have witnessed a significant interest in developing powerful tools able to answer the ever-growing demands from traditional consumers to advanced technological companies, spanning applications from multimedia to medicine, and from geosciences to space exploration, to mention a few. In all these applications, the human perception of the environment has been a key factor. Within this panel session hosted at EUVIP 2021, we have discussed the synergies between recent machine learning advances and visual perception required to help the development of powerful visual information processing techniques. This session was designed around some key questions that were addressed by three panelists of different and complementary scientific and technical culture and experience. The main objective was to share the experience of these experts, coming from different backgrounds, with the audience on topical issues that concern the evolution of the themes of the EUVIP workshop directly related to the analysis, processing and coding of visual information and applications and how to associate and integrate new approaches and technologies coming from machine learning. At the end of this session, the aim was also to highlight a number of converging points of view, which would give a more or less realistic idea of the promising avenues for future research.