Socionext and Partners Taking Vision Learning to the Next Level at the Annual ISC-West in Las Vegas
Featuring Leading-Edge Image Processors and Advanced AI Edge Solutions for Video Monitoring, Surveillance and Security Camera Applications
Langen/Germany, 25. March, 2019 --- Socionext Inc., a global leader in advanced image processing technology, will feature a lineup of high-quality, high-performance imaging technology and solutions featuring the company’s Milbeaut®series of image signal processors and edge server for security, surveillance and video monitoring systems at the annual ISC West - International Security Conference in Las Vegas April 10 – 12, booth 32107.
Socionext’s Edge Solution
In security applications today, there are growing demands for processing power at the edge node, not only for low latency but also for the concerns of sending sensitive information to the cloud. At this year’s ISC-West, Socionext will showcase variation of low-power edge solutions, including AI capability built inside camera devices, and edge AI server that controls multiple cameras.
Robust, Multi-Face Detection and Recognition by In-Camera Devices
Socionext, in partnership with industry experts, has developed a convolutional neural network (CNN) / deep learning solution enabling the adoption of AI at the edge. The company will demonstrate a solution for multi-face detection and recognition suitable for surveillance applications with visual search functionalities. The vision AI solution performs extremely well under poor lighting conditions, with the capability to recognize occluded, oblique faces, as well as locate and identify multiple objects and people in a single scene. The new feature allows for instant face and metadata capture by devices installed in the cameras and then sent to the cloud for deep learning inference accuracies to the edge, similar to the deep learning models in the data center.
Featured in this demonstration is Socionext’s Milbeaut SC2000, which is designed for challenging computer vision applications such as high- performance security cameras. The SC2000 offers state-of-the-art image quality that allows for improved recognition by both humans and machines.
Fig.1: Socionext’s Edge AI Solutions
Real-Time Object Detection and VMS Systems for Edge Surveillance Applications
Socionext will showcase a high-speed, high-performance and energy-efficient AI solution designed for surveillance and security applications. The solution features AI accelerator edge solution and
Socionext’s Arm-based “SynQuacer” SC2A11 processor.
Socionext and partners have developed the video management system (VMS) with AI accelerator for delivering superior parallel processing performance with video analytics capabilities such as image classification, face recognition, and object detection. The system is capable of delivering multiple channels of video images captured from IP cameras in real-time. All video analytics are done by the AI accelerator and SynQuacer processor at the edge, with the capability of tagging metadata to the streamed videos for added intelligence in video surveillance and proactive detection.
Industry-Leading Image Processor with 3DNR and WDR
The company will also feature its ultra-low power SC2002 image processor which incorporates the world’s best 3DNR (Three-Dimensional Noise Reduction) and the WDR (Wide Dynamic Range) functionalities for capturing high-quality images under obscure, poor light quality conditions. The SC2002 consumes at a mere 1.5W in typical operating conditions and is equipped with a high-performance DSP with built-in intelligence, which enables a wide range of video monitoring and security camera systems optimized for crime prevention.
For ISC West website and programs, visit http://www.iscwest.com/Show-Info/
SC2000 product information: http://www.socionext.com/en/products/assp/milbeaut/SC2000.html
SC2002 product information: http://www.socionext.com/en/products/assp/milbeaut/SC2002.html
SynQuacer product information: https://www.socionext.com/en/products/assp/SynQuacer/
PR2019132 - Socionext ISC-West - ENGLISH FINAL.pdf
Size: 0.241 Megabyte | Date: 25.03.2019 08:30
For Press Inquiries
Socionext Europe GmbH