Machine vision can assistApril 07, 2005Machine vision provides a unique capability for understanding human actions. Based on a passive sensing technology, it allows for the coverage of large areas with a single sensor, without need for physical contact. Professor Matti Pietik'¤inen is in charge of the machine vision research team whose aim it is to equip the environment with several camera-based modules that alone or in combination are capable of analysing scenes and events. The team are working under the Academy-funded Research Programme on Proactive Computing (PROACT). A key requirement for proactive computing solutions is the ability to understand human actions. Human-machine dialogue in proactive systems can take place via smart interfaces which provide the system with control data in the shape of gestures or facial expressions, for example. The researchers on Pietik'¤inen's team have been working on a solution where the machine monitors people's facial expressions and head movements and on this basis determines the direction of their vision. The machine can then use this information to move a cursor on the screen in the same as a mouse is used in ordinary computer interfaces. Smart system recognises its operator and responds accordingly Machine vision has been used among other things in developing a smart classroom. "We are working on an automatic camera-based remote teaching system where the active camera is always chosen according to the teacher's actions. In practice this means that the document camera will be switched on when the teacher wants to show a document," Professor Pietik'¤inen explains. It is often important for a proactive system to be able to identify people and to specify its activities on this basis according to the current user, or only to allow access to a machine or to certain facilities to certain people. The research team under Professor Pietik'¤inen are developing a system based on automatic face recognition which uses the Local Binary Pattern (LBP) texture method that the team have developed earlier and that has attracted much international interest. The team are also working to develop techniques with which a machine can follow the actions of a moving person and understand what this person is doing. "Our work has paved the way to a new, LBP-based method where moving objects can reliably be separated from the background for further interpretation," Professor Pietik'¤inen continues. In the future the method will have practical application in video surveillance systems for purposes of crime prevention, for example. Suomen Akatemia (Academy of Finland) |
|||||||||||||||||||||
Science Research Departments
Earth Science Alternative Energy | Anthropology and Archaeology | Earthquakes and Volcanoes | Environment and Nature News | Global Warming | High-Energy and Particle Physics | Ozone Hole | Scientists Slow Light | Tsunami Space Science Astronomy and Space News | Black Holes | Chandra X-Ray Observatory | Extrasolar Planets | Hubble Telescope | International Space Station | Jupiter Galileo Mission | Jupiter Cassini Mission Flyby | Mars Exploration | Mars Odyssey 2001 | Mars Global Surveyor | Mars Polar Lander | Mars Climate Orbiter | Mars Pathfinder | Meteors and Asteroids | Mir Space Station | NEAR Asteroid Probe Mission | Pluto Planet Debate | Search for Extraterrestrial Life | Space Shuttle Program | Space Shuttle Mission: STS-102 | Space Weather Life Science Animal News | Biotechnology and Genetics | Brain Research | Human Cloning | Dinosaur and Fossil Discoveries | Endangered Species | Gene Therapy | Genetically Modified Food | Stem Cell Research | Whales and Whaling |
|||||||||||||||||||||
|
|||||||||||||||||||||
|
|||||||||||||||||||||