Bookmark and Share

Auviz Systems Announces Availability of AuvizDNN, Optimized Library for Convolutional Neural Networks on FPGAs

Nagesh Gupta, CEO, Auviz Systems to present at the Embedded Vision Alliance Summit

Campbell, Calif. – May 8, 2015 – Auviz Systems, a leader in Middleware FPGA IP for Data Centers and Internet of Things today announced the availability of AuvizDNN, a library of OpenCL functions to create Convolutional Neural Networks (CNNs) on FPGAs. Other Middleware IP libraries from Auviz Systems include AuvizCV, an optimized version of OpenCV, and AuvizLA, an optimized BLAS for FPGAs.

CNNs, a class of Machine Learning algorithms, are widely used for tasks such as Image Classification, Natural Language Processing, and Speech Recognition. Nagesh Gupta, Founder & CEO of Auviz Systems will present how FPGAs can be used to accelerate image classification at the Embedded Vision Summit at the Santa Clara Convention Center on May 12, 2015. The presentation titled “Trade-offs in implementing Convolutional Neural Networks on FPGAs” will describe an implementation of AlexNet using Xilinx FPGA’s and AuvizDNN IP.

“Machine Learning has emerged as a major compute workload in the next generation cloud,” stated Hemant Dhulla, Vice President of Wired Communications and Data Center at Xilinx. “Only FPGAs deliver real time performance on machine learning applications like image recognition while fitting the power budget of data centers. Auviz’s libraries open the power and performance of Xilinx FPGAs to programmers with APIs that are easily integrated into existing applications.”

About Auviz Systems

Auviz Systems is the technology leader in Middleware Intellectual IP for FPGAs. For more information, visit www.auvizsystems.com

About the Embedded Vision Alliance

Founded in 2011, the Embedded Vision Alliance is an industry partnership that brings together technology providers who are enabling innovative and practical applications for computer vision. Its 50 members include suppliers of processors, sensors, algorithms and software across a range of growing market segments and applications, including automotive, consumer electronics, gaming, imaging, and more.  Membership is open to any company that supplies hardware, software, or design services for computer vision systems and applications. For more information on the Alliance, visit https://www.embedded-vision.com.