Traditionally, computer vision has been defined as a technology that duplicates human visual perception abilities. Many applications of computer vision were designed to replace or support the human operator in repetitive and time-consuming tasks, such as the computer-assisted diagnosis of medical images, long-term surveillance of extended premises etc. However, if we rethink computer vision as a technology meant to enhance human perceptual capabilities rather than to replicate them, we uncover new horizons for this field.
In this talk, I will provide an overview of my ongoing computer vision projects that serve to augment human capabilities in various ways. These projects include, but are not limited to the visual analysis of human motion performance, perceptual interfaces, and analysis of non-visual data with computer vision techniques.