High performance videos surveillance system dedicated to videogames

Tasks with high computational requirements, such as tomographic or satellite image reconstruction, combinational problems or computer image recognition would benefit from the high computational capabilities at low price of the hardware devices used in videogame stations or the graphics processing units of personal computers.

This idea culminated in 2003, when Mark Harris, a PhD student at North Carolina University at Chapel Hill (www.unc.edu), now working for the multinational company NVidia, used the computational power offered by a graphics processing unit (GPU) of a personal computer to simulate and render real time images of the physical processes involved in the formation of clouds. Since then, the term GPGPU (General Purpose computing on Graphics Processing Units www.gpgpu.org ) was acquired for the use of graphic processing hardware in more general applications. Currently, this alternative use of the videogame stations and GPUs is being applied in collaborative projects such as Folding@Home (http://folding.stanford.edu).

The present focus of the researchers of the High Performance Computation and Optimization Team (www.gavab.es/capo) part of the GAVAB group (www.gavab.es) at the Rey Juan Carlos University (www.urjc.es) the is project V-ATRAP. The objective of this project is to prove the benefits that this type of technology has to offer as a computing unit for new tasks such as visual recognition and tracing, both very useful applied to real-time video surveillance and offering of up to 14 times faster performances than the existing solution based on the traditional computers.

At present, the technology used in the construction of graphical chipsets for videogame stations is comparable to the most up-to-date technology used by computer manufacturers. In fact, a processing unit of a PC can hold 400 million transistors per core, while an average graphics processing unit can hold as much as double that number. This offers a higher computational power at a lower cost for a wide range of applications, such as high definition video processing, medical image processing or assisted multiple camera video surveillance.

Author Antonio Sanz Montemayor