Xilinx’s reVISION expands into great deal of Vision-Guided Machine Studying under Edge to Cloud
At Embedded World, Xilinx, Inc. on March 14, 2019, announced expansion right wide variety of vision guided machine learning applications when using the Xilinx reVISION stack. This announcement complements the recent?Reconfigurable Acceleration Stack, significantly broadening the deployment of machine learning applications with Xilinx technology with the edge to your cloud.
The new reVISION stack enables a much wider range of software and systems engineers, with no hardware design expertise to formulate intelligent vision-guided systems easier and faster.? These engineers is now able to realize significant advantages when combining machine learning, computer vision, sensor fusion, and connectivity.
reVISION is enabling a rapid growing number of applications in markets where differentiation is critical, systems must be extremely responsive, and the latest algorithms and sensors has to be quickly deployed. This can include approximately two-thirds in the applications for vision focused semiconductors. Applications span quite a few markets just like high-end consumer, automotive, industrial, medical, and aerospace & defense. Next generation applications include collaborative robots or ‘cobots’, ‘sense and avoid’ drones, augmented reality, autonomous vehicles, automated surveillance and medical diagnostics.
reVISION enables the swiftest option to quite possibly the most responsive vision systems, with as many as 6x better images/second/watt in machine learning inference, 40x better frames/second/watt computer system vision processing, and 1/5th the latency over competing embedded GPUs and typical SoCs. Developers with limited hardware expertise will use a C/C++/OpenCL development flow with industry-standard frameworks and libraries like Caffe and OpenCV to develop embedded vision applications on a single Zynq SoC or MPSoC.
Leveraging the unique great things about reconfigurability and any-to-any connectivity, developers are able to use the stack to rapidly develop and deploy upgrades. Reconfigurability is important to ‘future proof’ intelligent vision-based systems as neural networks, algorithms, sensor technologies and interface standards continue to evolve for an accelerated pace.
The Xilinx reVISION stack contains a wide range of development helpful information on platform, algorithm and database integration. This consists of support for popular neural networks including AlexNet, GoogLeNet, SqueezeNet, SSD, and FCN. Additionally, the stack provides library elements including pre-defined and optimized implementations for CNN network layers, essential to build custom neural networks (DNN/CNN). The device learning elements are complemented by way of a broad number of acceleration-ready OpenCV functions for computer vision processing. For application level development, Xilinx supports industry-standard frameworks including Caffe for machine learning and OpenVX for computer vision.? The reVISION stack includes development platforms from Xilinx and other companies, including different sensors.
“Our eye tracking technologies have greatly benefited individuals struggling with ALS or some other different types of paralysis, run by the Zynq SoCs for high-resolution vision based analytics,” said Robert Chappell, CEO and Founder, Eyetech Digital Systems. “The new reVISION stack offers new opportunities for algorithm development by leveraging the power of machine learning. This may enable us to inflate our offering of human interaction hardware as well as improving our core eye tracking products.”
“The embedded industry is an evolving application space where adjustments in algorithms, neural networks and sensors require reconfigurability of your target platforms,” said Lakshmi Mandyam, senior director of segment marketing, ARM. “ARM-based Zynq technology from Xilinx will enable these applications to be deployed efficiently while accelerating the adoption of innovative machine learning applications within the edge to your cloud.”
“We are seeing tremendous fascination with machine listening to advice from exploding for the cloud, and believe our ongoing investment in development stacks will accelerate mainstream adoption.” said Steve Glaser, SVP of Corporate Strategy at Xilinx. “Today, a huge selection of embedded vision customers have realized higher than 10x performance and latency advantages with Xilinx technology. By building reVISION, those same advantages will now become designed to numerous customers.”
The reVISION stack might be accessible in the next quarter of 2019. For additional information and join the reVISION community visit at?www.xilinx.com/revision.