Computer vision is likely to bring a revolution to advertising, enabling a two-way interaction between surfaces and the crowds. Here we develop a solution to engage people in a game with the animated floating balloons. The advertisement has been made for the marketing agency UbachsWisbrun/JWT, and installed on one of Amsterdam’s main squares – Rembrandtplein.

Pictures taken with a smart-phone are overlaid with old pictures or paintings of the same scene. In this augmented reality setting, the user is able to fade the old picture/painting away to see the differences.

Drowsiness is one of the most common causes of car accidents. It is estimated that 30% to 40% of all accidents are related to drivers falling asleep. Our system is able to detect driver’s drowsiness based on input from a basic camera, which can even be the one embedded in your mobile phone or navigation device. If critical level is reached, the driver is alerted.

The issue with surveillance cameras is that they usually give fish-eye images as the output. In this demo, we show how such images are first flattened, and then further analyzed for surveillance purposes. People, as well as accompanying objects, are being observed and tracked; unattended objects are spotted.

We demonstrate a system which estimates direction of gaze for users in a controlled environment (e.g. sitting in front of a screen). Computational costs are reduced to a minimum by using the same technology both for eye corner locator and eye center locator. If high mapping precision is not a priority of the application, the system can achieve desirable accuracy without additional dedicated hardware.

The most expressive way humans display emotions is through facial expressions. Humans detect and interpret faces and facial expressions in a scene with little or no effort. Still, development of an automated system that accomplishes this task has proved rather difficult. We present a fully integrated system, which is able to accurately detect and recognize facial expressions in real time, representing a major step forward in achieving a human-like interaction between the man and the machine.

For more information, see Visual Recognition.

This system is analyzing driver’s awareness by tracking head pose information. It can allow studies of driver behavior and report dangerous outcomes such as longer distractions by events away from the windshield.

The ubiquitous application of eye tracking has so far not been possible without dedicated and expensive hardware, such as infrared high definition cameras. With this system, we perform very accurate eye center localization and tracking in existing images or with a simple webcam.

Head pose and eye location estimation are two closely related problems. We demonstrate the system designed to (1) enhance the accuracy of eye location estimations in low resolution videos, (2) extend the operating range of the eye locator and (3) improve the accuracy and re-initialization capabilities of the head pose tracker.