Edge AI and Vision Insights: June 10, 2020 Edition

Last week, the Edge AI and Vision Alliance announced the winners of the 2020 Vision Product of the Year Awards. The Awards recognize the innovation and excellence of the industry’s leading technology companies that are enabling practical visual AI and computer vision.This year’s winners included NVIDIA’s Jetson Nano (Best AI Processor), iniVation’s Dynamic Vision Platform (Best Camera or Sensor), Morpho’s Semantic Filtering (Best AI Software or Algorithm), Intel’s DevCloud for the Edge (Best Developer Tool), and Horizon Robotics’ Journey 2 (Best Automotive Solution). Check out the archive recording of the awards presentation, and join us in congratulating the winning companies and products!

The 2020 Embedded Vision Summit, taking place online September 15-25, is not an academic research conference or a glittery marketing event. It’s the one place where you can find the latest practical techniques, technologies, applications and ideas in computer vision and visual AI—a place where you can hear about advanced techniques that actually work in the real world, such as cascaded DNNs, real-time data augmentation and guide-follower algorithms. Don’t just take my word for it; check out the already announced talks for yourself. (We’re adding new talks daily, and it won’t be long before we’ll have nearly 100 talks across four tracks.) And then be sure to register today with promo code SUPEREBNL20-V to receive your Super Early Bird Discount!

Brian Dipert
Editor-In-Chief, Edge AI and Vision Alliance

Data Annotation at Scale: Pitfalls and Solutions
In many real-world use cases, deep learning algorithms work well if you have enough high-quality data to train them. Obtaining that data is a critical limiting factor in the development of effective artificial intelligence. In this talk, Nikita Manovich, Senior Software Engineer at Intel, identifies common pitfalls encountered in obtaining and using public and private data for training and evaluating deep neural networks for visual AI—and presents techniques to overcome these pitfalls. He also presents the open source Computer Vision Annotation Tool (CVAT) (github.com/opencv/cvat), illustrating techniques his company has implemented to streamline annotation of visual data at scale. He discusses challenges faced in developing CVAT, how they were addressed, and plans for further improvements.

Snapdragon Hybrid Computer Vision/Deep Learning Architecture for Imaging Applications
Advances in imaging quality and features are accelerating, thanks to hybrid approaches that combine classical computer vision and deep learning algorithms and that take advantage of the powerful heterogeneous computing capability of Qualcomm Snapdragon mobile platforms. Qualcomm’s dedicated computer vision hardware accelerator, combined with the flexible deep learning capabilities of the Qualcomm AI Engine, help developers meet performance goals for the most complex on-device vision use cases with lower power consumption and processor load. In this presentation, Robert Lay, Computer Vision and Camera Product Manager at Qualcomm, reviews the Snapdragon heterogeneous architecture and how Qualcomm’s dedicated computer vision processor and tools accelerate the realization of high performance, efficient imaging and vision applications.

Fundamental Security Challenges of Embedded Vision
As facial recognition, surveillance and smart vehicles become an accepted part of our daily lives, product and chip designers are coming to grips with the business need to secure the data that passes through their systems. Training data, the resulting model data and how decisions are made and acted on can be proprietary information for the product, and important to keep out of competitors’ hands. Inputs from sensors and cameras can contain legally protected data, and the data may create ethical and privacy concerns as cameras and microphones in homes, cars and public settings explode in number. This presentation from Mike Borza, Principal Security Technologist at Synopsys, describes typical security concerns in vision systems today, including potential weaknesses in training-to-inferencing systems where data can be compromised, and discusses different approaches to security.

Potential Impacts of Privacy Regulation and Litigation on Vision Technology
In this talk, Robert Cattanach, Partner at Dorsey & Whitney LLP, provides insights into the fast-changing world of privacy regulation and litigation.

FRAMOS Launches Industrial 3D GigE Cameras Based on Intel’s RealSense Technology

Basler Introduces Embedded Vision Solutions for the NVIDIA Jetson Platform

IDS Imaging Development Systems’ uEye SE Cameras Contain 20 MPixel Sensors

Gyrfalcon’s Lightspeeur 5801 is Named as a Top 10 Processor for AI Acceleration at the Endpoint

FLIR Releases European Thermal Imaging Dataset for Automotive Driver Assistance Development

More News