Face Recognition: Learn About GPU Acceleration
Professor Brian Lovell of the University of Queensland, Australia, who's also Chief Technical Officer at Imagus Technology, is a well-known figure in the fields of fields of computer vision and pattern recognition. Lovell is also a long-time advisor to (and advocate of) the Embedded Vision Alliance. On Tuesday November 5 at 9AM PT, Lovell and Imagus co-worker Stephen Brain will co-present a free webinar associated with NVIDIA's GPU Technology Conference, entitled "Accelerating Face-in-the-Crowd Recognition with GPU Technology." From the event page:
Face-in-the-crowd recognition capability is at the heart of ambitious city-wide rollouts such as the Safe City Test Bed in Singapore. City-wide CCTV processing means the analysis of many thousands of video feeds requiring enormous computing power. A key component of the implementation and rollout strategy is NVIDIA GPU technology.
GPUs can speed up both face recognition and face detection. In one implementation, the Robot Operating System (ROS) from Willow Garage was used to distribute the processing tasks across potentially thousands of networked CPUs and GPUs to form a completely scalable architecture.
Join Brian Lovell and Stephen Brain from Imagus Technology as they introduce the recently released iFace Library and show how GPU technology can be used to address bottlenecks at each stage of processing. The presentation will also cover the speakers’ use of ROS, OpenCV, OpenMP, and Armadillo libraries to develop fast reliable distributed video processing code.
Imagus Technology has been researching and developing non-cooperative face-in-the-crowd recognition algorithms for many years and their system was awarded CCTV System of the Year at IFSEC2011. The technology is now available commercially via the release of the iFace Libraries.
To register for this free webinar, please visit the event page.