Check out these upcoming webinars from the Edge AI and Vision Alliance and its Member companies:
On Tuesday, March 23, 2021 at 9 am PT, Nota will deliver the free webinar “Selecting and Combining Deep Learning Model Optimization Techniques for Enhanced On-device AI,” in partnership with the Alliance. Model optimization techniques such as pruning, quantization, filter decomposition, and NAS (neural architecture search) are becoming increasingly important in efficiently implementing deep learning on the edge. Determining the optimum technique (potentially involving the combination of multiple different techniques) and defining hyperparameters for peak performance, historically demanded deep technical expertise and significant resources in order to successfully achieve the optimization objective. In this presentation, Nota will describe its experiences with (and resultant perspectives on) various deep learning model optimization techniques, as well as demonstrating how multiple techniques can be combined to further improve performance. The company will also discuss NetsPresso, its automatic model optimization platform. For more information and to register, please see the event page.
On Wednesday, March 24, 2021 at 9 am PT, Jeff Bier, founder of the Alliance, will deliver the free webinar “Deep Learning for Embedded Computer Vision: An Introduction” in partnership with Vision Systems Design. Deep learning for embedded computer vision has made it possible to deploy high-performance, low-cost embedded systems capable of running powerful AI algorithms for tasks ranging from image classification to object detection and tracking, in applications from cars to kitchen appliances. The end result is products that are safer, easier to use, more efficient and more capable. But visual AI and computer vision are quite different from traditional embedded technologies, and for many product development groups, these new technologies bring unfamiliar challenges and unexpected risks. In this session, Bier will cover:
What a typical embedded vision system looks like
Training and test data — your new best friend, and maybe also your worst nightmare, and
Choosing a processor for embedded vision applications
For more information and to register, please see the event page.
On Thursday, March 25, 2021 at 9 am PT, Microchip Technology will deliver the free webinar “Enabling Small Form Factor, Anti-tamper, High-reliability, Fanless Artificial Intelligence and Machine Learning,” in partnership with the Alliance. Microchip’s FPGAs offer a power-efficient 4K video and imaging fabric with in-built anti-tamper protection, reliability against single-event upsets and the industry’s first integrated quad core RISC-V microprocessor subsystem in a FPGA. Along with the easy-to-deploy VectorBlox AI SDK, users can develop next-generation smart vision platforms that do not require a heat sink or a thermal fan. In this presentation, Microchip Techology’s Diptesh Nandi will describe how his company’s products help you overcome the challenges of designing edge devices that need security, reliability, have a small form factor, and are able to run in battery backup and be field-upgradable. For more information and to register, please see the event page.
And on Tuesday, March 30, 2021 at 9 am PT, Algolux will deliver the free webinar “Optimizing a Camera ISP to Automatically Improve Computer Vision Accuracy,” in partnership with the Alliance. Cameras are the most ubiquitous sensor used for both display and computer vision in ADAS and autonomous vehicle designs. Typical applications include surround-view, object detection for collision warning and automatic emergency braking, and traffic light and sign recognition. To achieve subjectively “good” image quality (IQ), the camera’s image signal processor (ISP) parameters must be tuned for each specific lens / sensor configuration by experienced imaging engineers, typically over many months. However, good visual IQ isn’t necessarily what’s needed for specific computer vision (CV) algorithms. In this session, Marc Courtemanche, Product Architect at Algolux, will describe and demonstrate how to use a breakthrough workflow to automatically optimize an ISP to maximize computer vision accuracy in only days. Easy to access and deploy, the workflow can improve CV results by up to 25 mAP points while reducing time and effort by more than 10x versus expert manual tuning approaches. For more information and to register, please see the event page.
Editor-In-Chief, Edge AI and Vision Alliance
APPLICATION AND MARKET TRENDS
Key Trends and Challenges in Practical Visual AI and Computer Vision
With visual AI and computer vision technologies advancing faster than ever, it can be difficult to see the big picture. This talk from Jeff Bier, Founder of the Edge AI and Vision Alliance and President of BDTI, examines the most important areas of recent progress that are enabling developers of vision-based systems and applications–as well as the most significant challenges inhibiting more widespread and successful development of vision-based products. Bier explores key practical aspects of algorithms, the new wave of edge AI processors, development tools and processes and emerging classes of sensors. In each of these domains, he highlights recent developments that illustrate where the industry is heading, and also identifies obstacles that still need to be addressed. Throughout the talk, he calls out examples of leading-edge commercial building-block technologies and end-products that illustrate these trends and challenges.
Market Trends in Automotive Perception: From Insect-Like to Human-Like Intelligence
Today there are two paths towards autonomous vehicles. Mass-market automobiles continue to add more sensors and more compute power to enable increasingly sophisticated ADAS functionality. Separately, developers of robotic vehicles utilize high-end, industrial-grade sensors (lidar, cameras and radars) along with massive centralized computing. Either way, the push towards autonomy demands more and more computational power as increasingly demanding algorithms process increasing amounts of sensor data. In this presentation, Pierre Cambou, Principal Analyst at Yole Développement, shares his company’s analysis and forecast of the ADAS and autonomous vehicle perception market. When will cars with L2 to L5 level automation become mainstream? What sorts of processing power will they require? What alternative innovation scenarios might disrupt current trends?
Lessons From the Start-up Trenches
This session with Oliver Gunasekara, founder and CEO of NGCodec, explores lessons learned during the journey of NGCodec—a video codec developer—from its founding in 2012 to its successful acquisition in 2019 by Xilinx. Gunasekara begins with a brief presentation summarizing NGCodec’s path, sharing some of the key lessons learned during this journey. These include the reality that ideas are easy but execution is hard, which means that sharing your idea is important. Gunasekara also explains why he believes that perseverance and grit are critical leadership skills. He shares why he has come to believe that ultimately, it’s all about the team—and that, given the competitive market in Silicon Valley, leveraging a remote team can really pay off. Finally, he touches on structuring the company foundation and stock options for optimum tax efficiency, and then delves deeper into key lessons learned in an interview conducted by Shweta Srivastava.
Can You Patent Your AI-Based Invention?
Patenting inventions is one way to capitalize on the hard work and creativity that drives an innovative product, but recent cases have made things difficult for innovators whose inventions are expressed in software systems and processes. Software itself frequently cannot be patented. But this is not to say that AI-based vision and embedded vision inventions are unpatentable. This presentation from Thomas Lebens, Partner at Fitch, Even, Tabin & Flannery LLP, gives you an update on the latest developments in software patentability and helps you navigate this landscape to best protect your innovative efforts.
NVIDIA Jetson Nano (Best AI Processor)
NVIDIA’s Jetson Nano is the 2020 Vision Product of the Year Award Winner in the AI Processors category. Jetson Nano delivers the power of modern AI in the smallest supercomputer for embedded and IoT. Jetson Nano is a small form factor, power-efficient, low-cost and production-ready System on Module (SOM) and Developer Kit that opens up AI to the educators, makers and embedded developers previously without access to AI. Jetson Nano delivers up to 472 GFLOPS of accelerated computing, can run many modern neural networks in parallel, and delivers the performance to process data from multiple high-resolution sensors, including cameras, LIDAR, IMU, ToF and more, to sense, process and act in an AI system, consuming as little as 5 W.
Please see here for more information on NVIDIA and its Jetson Nano. The Edge AI and Vision Product of the Year Awards (an expansion of previous years’ Vision Product of the Year Awards) celebrate the innovation of the industry’s leading companies that are developing and enabling the next generation of edge AI and computer vision products. Winning a Product of the Year award recognizes your leadership in edge AI and computer vision as evaluated by independent industry experts. The Edge AI and Vision Alliance is now accepting applications for the 2021 Awards competition. The submission deadline is Friday, March 19; for more information and to enter, please see the program page.
Here you’ll find a wealth of practical technical insights and expert advice to help you bring AI and visual intelligence into your products without flying blind.