CanControls at a glance
CanControls is a young and innovative company, which was founded in February 2005 as a spin-off of the Chair of Technical Computer Science of the Aachen University of Technology in Germany, where it currently resides. The company specializes in cutting-edge man-machine interfaces, real-time image processing, computer vision and scene analysis. Being one of the most promising spin-offs from universities to date, CanControls has won several awards and is supported by the German government.
Through its close ties to the university CanControls combines the skills to remain on top of leading-edge research with the ability to create industrial-strength, marketable products. In the following we will give a short description of some of our core areas of expertise:
Facial Feature Analysis and Interpretation for Man-Machine-Interfaces
Advancements in the processing power of commodity hardware allow the development of new, innovative man-machine interfaces. In particular non-intrusive facial feature analysis offers a great potential for the application in many areas such as speech recognition, special controls or medical applications. CanControls has developed a system that can detect and track the face and facial features such as eyes, mouth, chin, etc. in real time while requiring only commodity processing and sensing hardware (e.g. webcams). The system is robust against changes in the appearance of the user (e.g. beard, glasses) as well as changes in the environment (e.g. lighting variations).
We have applied this knowledge to the field of medical rehabilitation by developing OPSIS, the first facial expression driven control system for a power wheelchair. In cooperation with a leading international manufacturer we are finalizing the product for the mass-market. Furthermore we are using our technology to develop a vision-based driver assistance system which can identify the driver and monitor her state (e.g. level of tiredness).
Person and Object tracking for Surveillance and Security Applications
CanControls has developed a vision-based system that can track multiple people in highly cluttered environments such as flight cabins or office rooms. These kinds of environment impose strong restrictions on the placement of the cameras in that the view angle on the scene is usually very low and the field of view is partially obstructed by furniture, seats or other obstacles. The resulting, frequently occurring person-person and person-scene occlusions are resolved by the system while maintaining the respective person ID. The system has been enhanced to track arbitrary objects such as cars.
Scene Understanding and Environment Control Systems
Scene Understanding describes the process of not only detecting certain objects or features in an image but to understand the meaning of the observed image content. CanControls has long-time experience in the field of vision-based sign language recognition used by mute people, including involvement in the EU project “WISDOM” within FP5. The semantic of this language can only be inferred by a joint classification of facial expression, hand shape and the trajectory of its movement. The system developed by CanControls can currently recognize and distinguish around 150 different signs.
In a related project we have developed a two-dimensional, articulated shape model of the human body, including head, trunk, arms, legs and their respective joints. The model can automatically be adapted to a person in a camera image and thus allows a precise understanding of the person’s orientation and pose.
|Phone:||+49 241 9632200|
|Fax:||+49 241 9632209|