No abstract

Developing algorithms for Ambient Intelligence (AmI) environments involves testing algorithms using both simulated environments and user data. This has been traditionally done with simulators. While there are several datasets for Ambient Intelligence most of them have do not have an open format and a way to use them easily afterwards. We implement a software framework on top of an existing hardware platform for writing AmI applications that can also export a data dump consisting of heterogeneous sensor data matching several criteria from an AmI lab. Researchers using this framework can easily write AmI applications on top of shared datasets and easily test several algorithms against each other using real-world-data, a previously impossible endeavour.

This electronic document describes one of the first full open-source frameworks for posture and face detection and classification using Kinect sensors, based entirely on an open-sourced software platform. It has been designed with the possibility of automatic monitoring of either an elderly person who is situated in a closed space (a room with furniture, appliances, and other rigid objects), or a patient who suffers from a physiological disorder. It is also based on an original hardware platform (the AmI-Platform), which includes more Kinect sensors, data acquisition stations, and crunches (data servers). The articles tackle two of the most important issues for ambient intelligence: accurate determination of the person and his or her facial landmarks (to understand emotions) and fast and reliable posture determination (to take appropriate measures in case of an accident).