nViso is a leading provider of emotion recognition software that interprets human facial micro-expressions and eye movements captured through video. Its solutions use proprietary 3D Facial Imaging technology with artificial intelligence to track hundreds of different facial points in order to recognize human emotions. The company uniquely combines the latest advancements in computer science, engineering and behavioural sciences to make automatic emotion recognition a viable reality using any image based device. Its proprietary analytical techniques are based on theoretical work by Dr. Paul Ekman, in which he demonstrates that emotions can be precisely recognized by minor changes in micro-expressions in the face. nViso is based at the Swiss Federal Institute of Technology in Lausanne, Switzerland (EPFL).
Facial expressions are innate to all humans and have been proven to be one of the strongest ways in which we express our emotions. Resulting from the work of Dr. Paul Ekman and others, a core set of emotions have been shown to be universally conveyed by facial expressions around the world: happiness, surprise, fear, anger, disgust, and sadness. Based on Ekman’s Facial Action Coding System, nViso’s 3D Facial Imaging technology uses sophisticated artificial intelligence algorithms to capture and measure the response of the main facial muscles involved in the expression of these emotions in real-time, by tracking 170 facial points. Advanced machine and deep learning systems precisely decode facial movements into the underlying expressed emotions providing a wealth of information about the emotional state of the driver.
Most current face recognition technologies are based on the appearance of the face. One of the typical problems, where the appearance-based recognition approaches fail, is 3D pose variation. As the human face is a 3D object, its visual appearance under different view angles varies significantly and in an automotive environment this is a major challenge, as the sensing module and driver monitoring camera can be placed inside the vehicle in many locations that cause major pose variation. nViso has developed highly accurate, identity preserving, pose normalization technology based on accurately tracking 170 facial features in points on the face used in emotion recognition. Pose normalization by nViso can improve recognition accuracy by up to 40%.
Emotion recognition based on facial analysis is a powerful way to capture passively the emotional state of a driver, however there are situations where the face may not be visible and as such analysis may not be available. Voice analysis can allow classification of driver speech at specific moments of interaction within the car where the face may not be visible. nViso voice analysis technology will allow richer driver analytics to be performed.