VicarVision develops state of the art solutions and products for computer vision. Plugging a camera into a computer is an easy first step, but how can you make a machine see things? How can algorithms extract meaningful interpretations from a stream of pixels? The technologies we develop are aimed to create vicarious perception. We want our machines to perceive and name the objects and events presented to them through cameras, in the same way that we would; so they can keep an eye on things for us.

Field of interest

VicarVision has a special interest in perceiving people’s behavior. FaceReader™ is the premier product for automatically evaluating emotional expressions in the human face. It is used in behavioral research institutes all over the world in studies assessing the usability and user experience of software systems, food preference and video conferencing interactions, to name a few. PIRES™, first introduced in 1996, was the first ‘mugshot’, or forensic portrait analysis, system with the ability to fill in a full police suspect description form; in fact, it often performed better than most police officers. BodyReader™ gives real time estimates of human pose and other articulations of the human body. GazeTracker™ follows the direction of attention of a person watching a screen, using a conventional webcam. AIVOS™ is a generic platform for intelligent video surveillance systems that analyses actions and interactions of multiple persons.


Over the years, apart from people detection, we have developed vision solutions to detect objects such as cars, horses, shoe traces for forensic analysis, and evaluate the context of those images. SewerViewer™ is our visual analysis system for a sewer crawling robot. It is able to detect, classify and interpret various abnormalities, cracks and obstructions in a sewer system. We are also currently developing a foot sole analysis system that will be able to perform a screening of footsole images and detect early stage foot ulcerations.