Join or Sign in

Register for your free asmag.com membership or if you are already a member,
sign in using your preferred method below.

To check your latest product inquiries, manage newsletter preference, update personal / company profile, or download member-exclusive reports, log in to your account now!
Login asmag.comMember Registration
https://www.asmag.com/showpost/34358.aspx
INSIGHTS

NVIDIA lauds GPU as enabler of deep learning, intelligent video

NVIDIA lauds GPU as enabler of deep learning, intelligent video
At Secutech 2017, NVIDIA is demonstrating its various GPU solutions, including the Tesla GPU accelerator for servers and Jetson platform which is more for edge devices.
More and more, intelligent video analytics employs the use of the deep learning technology where systems learn what certain objects are and recognize those objects. The concept had been unfathomable not too long ago yet has increasingly become a reality thanks to more advanced and powerful hardware such as the graphics processing unit (GPU).
 
That’s the argument raised by NVIDIA, an exhibitor at the CompoSec Pavilion at Secutech 2017.
 
Deep learning involves two key elements: Training, whereby an overwhelming amount of data is fed into the system to teach it to recognize objects; and inference, where the system recognizes whether the objects are the things that it has learned. Either way the system needs exceptional computational power that’s increasingly difficult to be handled by the CPU, which uses a serial computing architecture where one task is executed by one core at a time. A more significant role can be played by the GPU, which uses a parallel computing architecture that breaks down tasks into individual components that are executed by multiple cores at a time, making it more ideal to run complex deep learning algorithms.
 
At the show this year, NVIDIA is demonstrating its Tesla GPU accelerator for servers and Jetson platform which is more for edge devices. Tesla-enabled servers can be used in on-site or cloud solutions; as an example for the latter, video taken by multiple cameras can be transmitted to the server in the cloud where a complex deep learning algorithm is hosted, allowing the system to detect various objects or situations based on what it’s trained to do.
 
Applications can be manifold. “In safe city, instead of sitting in front of the monitors all the time, you can let the system identify suspicious individuals or vehicles and track them accordingly. In retail you can be informed of most frequented areas and make related adjustments accordingly,” said Ian Chen, Senior Sales Manager at NVIDIA, adding prospects for the Asian markets are quite good.
 
“The camera penetration rate is quite high in Asia. A high camera count will naturally lead to higher demands for intelligent video analytics. We predict 60 to 70 percent of the IVA market will be in Asia,” he said.


Product Adopted:
Others
Subscribe to Newsletter
Stay updated with the latest trends and technologies in physical security

Share to: