TeraDeep Takes Rapid Video Analytics to the Cloud with Amazon Web Services

FPGA Architecture In Amazon EC2 F1 Speeds Classification of Objects

News provided by

Dec 06, 2016, 02:30 ET

CAMPBELL, Calif., Dec. 6, 2016 /PRNewswire/ -- TeraDeep, a developer of accelerated deep learning applications for the cloud and on-premise Artificial Intelligence (A.I.) appliances, announced today that it plans to use Amazon Elastic Compute Cloud (EC2) F1 instances to bring high-performance, low-power video analytics to AWS cloud customers.

The low-latency, power-efficient TeraDeep Amazon Machine Image (AMI) relies on Amazon EC2 F1's FPGA-optimized architecture.

"TeraDeep has developed FPGA-based accelerator technology to substantially reduce the time and cost for complex video processing," said Didier Lacroix, CEO of TeraDeep. "Massive amounts of video footage are captured on camera systems and streamed every second. Accurate, automated, and low-latency video processing is critical so users can be alerted and respond to events immediately."

"As a growing startup, we found that the capital expenditure and lead time for developing and deploying FPGA technology on-premises is time-intensive and is the major reason there aren't more FPGA startups today," added Lacroix. "Amazon FPGA Images (AFI) and Amazon EC2 F1 Instances have dramatically reduced the time and cost to deploy our technology, eliminating the large up-front expense and slashing time-to-market for custom FPGA cards from months or years to just days."

About TeraDeep

TeraDeep is a pioneer in the development of specialized neural networks and acceleration technology for deep learning. The company develops and markets high-performance, low-power deep learning solutions for the cloud as well as on-premise AI Appliances. Its complete solutions offer a creative combination of specialized hardware designed for purpose-trained networks, available through a robust set of APIs. TeraDeep AI Appliances are leased either directly or through its value-added partners. It licenses reference designs and pursues OEM models with qualified parties.

SOURCE TeraDeep