Visual coding (i.e. image/video compression) is an indispensable technology in audiovisual communications and can be found in countless environments including cameras, smartphones, browsers, platforms such as YouTube and Netflix, and even in the videoconferencing software which will broadcast the masterclass.
Recently, deep learning techniques used in computer vision have begun to be applied to visual compression, leading to methods which improve the results of traditional formats (e.g. H.264, HEVC).
In this masterclass we will review the basic concepts of image and video compression, comparing traditional approaches and the latest ones using deep learning and looking in particular at the latter.
