Video Feature Extraction Will Enable Machines to Deliver the New Services and Benefits to Users and Industries, right as the Internet of Things is Maturing.
MILPITAS, Calif. -- June 18, 2019 — 82% of all Global IP traffic will be video by 2022, according to a white paper updated by Cisco on February 27, 2019. This video traffic will come from growth of camera integration in consumer, automotive and surveillance applications and ubiquity of wireless enabling distribution of streamed data. As IoT devices continue rapid growth, this video is increasingly going to be used by machines than humans.
“Machines will be capturing, compressing, managing and using much of this video, leveraging features extracted from video to automate applications and services, enabling rapid discovery and accelerating access to original media,” said Marc Naddell, Vice President of Marketing for Gyrfalcon Technology, Inc. (GTI). “Google re-shaped internet search by indexing text used on websites, and with new internet data being mostly video, imagine how companies recognizing this opportunity will create a new landscape for apps and services across all industries.”
MPEG has been introducing new capabilities for media with the new CDVA standard (Compact Descriptors for Video Analysis, ISO/IEC 15938-15). It describes how video features can be extracted and stored as compact metadata for efficient matching and scalable search. Solution developers would be able to leverage this standard to integrate new capabilities into devices with camera sensors and data centers that extract data and use it to provide enhanced services.
“It is exciting to see the adoption of CDVA in GTI’s line of products, which makes the technology readily available to a range of applications across industries,” said Werner Bailer, Key Researcher for Smart Media Solutions at JOANNEUM RESEARCH, who co-chairs the development of CDVA and contributed technology to this standard, which is the first making use of the recent progress in deep neural networks for multimedia applications.
Tomorrow’s Technology Today.
GTI’s Lightspeeur® line of chips can already extract features from video and embed them using the CDVA standard, whether used in devices to execute feature searches or in equipment to process video to embed features in each frame. The Lightspeeur® chips first became available in 2017 and use a 2-dimensional Matrix Processing Engine (MPE™) to process in memory with high performance ranging from 9.3 to 24 TOPS per Watt. This makes the chip a feasible and currently available element for any company seeing their opportunity to lead in visual search in their industry. The Lightspeeur® chips are available in 16 chip configurations on PCIe cards for data center applications for processing existing video files, or as single chips that can be integrated into end-user devices for providing feature extraction and search using visual data.
GTI will be showcasing visual search and feature extraction capabilities using Lightspeeur® chips at the CVPR 2019 expo in Long Beach, California from June 18-21 in their Booth (#412).
About Gyrfalcon Technology Inc.
Gyrfalcon Technology Inc. (GTI) is the world’s leading developer of high-performance AI Accelerators that use low power, packaged in low-cost and small-sized chips. Founded by veteran Silicon Valley entrepreneurs and Artificial Intelligence scientists, GTI drives adoption of AI by bringing the power of cloud Artificial Intelligence to local devices, and improves Cloud AI performance with greater performance and efficiency, providing the utmost in AI customization for new equipment and a path to AI upgrade to customers. For more information on GTI, visit https://www.gyrfalcontech.ai/.