This blog post was originally published at Ambarella’s website. It is reprinted here with the permission of Ambarella.

Video recording applications continue to play a key role in automotive.

Our history as a company began with recording. It’s an inextricable part of our organizational DNA. Nearly fifteen years ago, we enabled the first-ever 1080p solid-state pocket camcorders, bringing high-definition video recording to the masses for the first time. And we continued the trend with the original GoPro Hero line of sports cameras, helping give rise to a revolution in the way video is recorded and shared. Any time, any place—this concept quickly became a reality. And we were there from the beginning.

Before long, high-quality video sharing became ubiquitous, powered by an array of affordable, high-performance devices made possible by some of our earliest generations of encoding SoCs. It wasn’t too surprising, therefore, when these recording devices eventually made their way into vehicles.

Often called “dash cameras,” these interior windshield-mounted devices—used for event capture and drive logging—skyrocketed in popularity worldwide, particularly in countries where video clips of roadway incidents could be used as evidence for insurance companies or in courtrooms. While the initial wave of dashcams were sold as standalone products in the retail aftermarket, it wasn’t long before automotive OEMs began to offer digital video recorders (DVRs) as optional accessories to be installed by the dealership at the time of purchase (known as “dealer-fit” installations). And soon after that, OEMs began integrating DVRs into vehicles during the production process itself, using the infotainment system as a built-in video display. Cars were suddenly rolling off assembly lines with recording capabilities natively equipped.

With so many active recording devices on the road, DVR clip sharing became an internet phenomenon. Even non-automotive events were occasionally captured via automotive devices, such as the 2013 Chelyabinsk meteor in central Russia.

Ambarella is an established leader in automotive video recording worldwide. Our solutions are used to capture millions of hours of video on any given day using aftermarket, dealer-fit, and OEM-installed recording devices, in a variety of different applications.

Use cases for our recording solutions include:

• Passenger vehicles
  • Drive logging for entertainment
  • Accident event recording
  • Vehicle security
  • Combination ADAS and recording
    • For example: traffic light detection (TLD), traffic sign detection (TSD), front collision warning (FCW), lane departure warning (LDW), object detection

• Commercial fleet vehicles
  • Accident event recording
  • Combination ADAS and recording
  • Integration with driver monitoring systems (DMS) and/or occupant monitoring systems (OMS)
    • Encouraging drivers to limit distractions (e.g., handheld devices)
    • Driver fatigue alerts
    • Cabin monitoring for driver safety
  • Documentation of traffic conditions

What are our advantages in recording?

Our solutions typically outperform the competition on the following measures:

• Low power consumption
• Low-light performance (noise filtering)
• AVC/HEVC encoding quality (e.g., the elimination of compression artifacts)
• Image dewarping
• Multi-exposure high dynamic range (HDR) processing
• Multi-sensor and multi-channel processing
• Full product line consisting of dedicated silicon with a range of capabilities, including computer vision
• ADAS features (optional)
• CVflow^® computer vision performance
• Reliability
  • Fail-safe journal file system for recording
  • Automotive-qualified parts (AECQ-100)
  • Functional safety-certified options

For additional information, please contact us.

Udit Budhia
Director of Marketing, Ambarella

The post Setting Records in Recording appeared first on Edge AI and Vision Alliance.

Sabina Pokhrel, Customer Success AI Engineer at Xailient, presents the “Introduction to DNN Model Compression Techniques” tutorial at the May 2021 Embedded Vision Summit. Embedding real-time large-scale deep learning vision applications at the edge is challenging due to their huge computational, memory, and bandwidth requirements. System architects can mitigate these…

“Introduction to DNN Model Compression Techniques,” a Presentation from Xailient

Register or sign in to access this content.

Registration is free and takes less than one minute. Click here to register and get full access to the Edge AI and Vision Alliance's valuable content.

The post “Introduction to DNN Model Compression Techniques,” a Presentation from Xailient appeared first on Edge AI and Vision Alliance.

This market research report was originally published at Yole Développement’s website. It is reprinted here with the permission of Yole Développement.

The automotive industry faces new industrial and technological challenges while undergoing dramatic changes in its value chain.

OUTLINE

• Market forecasts:

  In 2035, C.A.S.E. will be a US$318 billion market.

  Semiconductor’s value (chip level) in car will reach US$78.5 billion in 2026, with a 14.75% CAGR 2020-2026.

• Technology trends:

  Connectivity: Future V2X communication platforms for 5G implementation are being designed today, and solutions are expected for 2024. In the meantime, initial solutions are starting to appear with dual-4G and forward-compatible 5G capabilities.

  ADAS: Radars and cameras are the main sensors used by OEMs as they are quite performant and relatively cheap. For a few years, LiDAR sensors have been slowly entering the automotive industry to provide more automated driving functions.

  Sharing: New habits related to cars are emerging and Gen Y now desire connectivity, convenience and the possibility to choose from different range of transportation from A to B. They are shaping an industry in which in-demand car service providers are growing.

  Electrification: OEMs have announced their investment plans for the next 5 years which exceed US$250 billion worldwide. The timeline for cars electrification is very aggressive, as in 15 years OEMs will have to develop an entire car portfolio that will be fully electric.

• Supply chain:

  Supply chain management will change due to chips shortage, an increasing semiconductor content and electrification.

  OEMs have to negotiate directly with chip manufacturers, learn from the consumer industry and keep “buffer stock”.

  Traditional OEMs (Audi, Hyundai, etc…) are facing disruptive OEMs (Tesla, Apple…).

  China is intensifying the competition and invests massively in semiconductors and in the automotive industry.

“The value of semiconductors, at the chip level, in cars will grow from US$34.4 billion in 2020 to US$78.5 billion in 2026, a 14.75% CAGR.” asserts Eric Mounier, PhD, Director of Market Research at Yole Développement (Yole). He adds: “The largest growth will be in EVs due to the major shift to electrification.”

A car has today, on average, US$450 worth of semiconductors. In 2026, it will be US$700. Automotive developments are driven by technological developments for C.A.S.E. Yole’s analysts have estimated the C.A.S.E. related electronics modules market evolution to be:

• Connectivity: from almost US$33 billion in 2020 to almost US$55 billion in 2026, with 14.55% CAGR 2020-2026
• ADAS will reach more than US$60 billion in 2026 with 6.50% CAGR 2020-2026
• Sharing will reach about US$3 billion in 2026 with 10.39% CAGR 2020-2026
• Electrification will reach US$28,804 million in 2026, with 53.45% CAGR 2020-2026
• In 2035, C.A.S.E. will be a US$318 billion market.

In this context, the market research and strategy consulting company investigates disruptive semiconductor technologies for the automotive sector in depth. Its aim is to point out the latest innovations and underline the business opportunities.

These investigations have been developed with Yole’s partner, System Plus Consulting. According to Wilfried Théron, Electronic System, Department Director & Quality Manager at System Plus Consulting: “System Plus Consulting’s Teardowns uncover innovative design features and new semiconductor components to guide enterprises toward more streamlined solutions in future designs”. They provide clients unmatched intelligence into 4 automotive tracks (ADAS, Electrification, Infotainment and Telematics).

In addition, Yole releases today, the Automotive Semiconductor Trends 2021 report to provide an in-depth understanding of the changing automotive industry ecosystem and supply chain players. Including market trends and forecasts, take away and outlook, this study proposes key technical insights and analyses about future technology trends and challenges.

What are the economic and technological challenges of the automotive industry? What are the factors affecting the durability of the car industry? What are the key market drivers? Who are the companies to watch, and what innovative technologies are they working on? What are the recent investments and collaborations?

Today, Yole’s analysts present their vision of the automotive semiconductor trends.

As analyzed by Yole’s team in the new Automotive Semiconductor Trends 2021 report, wafer shipments will grow from 20 million to more than 45 million, with 8’’ being the most used wafer size. The 20nm node and below will be driven by ADAS and infotainment applications.

For Eric Mounier: “Today, most of the wafer production for automotive is for 130/180 nm and more, and leading-edge technology is very scarce. But 40nm and 28nm are used for the Mobileye EyeQ3 and EyeQ4 for ADAS and autonomy. Memory for infotainment and ADAS use 10-14 nm. In the future, 7nm could be used for ADAS. The current chip shortage mainly affects nodes in the 40-180nm range”.

The development of electric vehicles and self-driving technologies naturally attracts OEMs and Tier-1 component suppliers. Therefore, new OEMs like Nio, Xpeng, and Lucid Motors, among others, have recently entered the industry. Other players coming from the semiconductor or the consumer industries will enter the field as well. In this race to full autonomy, large OEMs with many resources – like Volkswagen – will develop the necessary software by themselves or partner with or acquire robotic vehicle companies. Generalist OEMs with few resources are expected to rely on Tier-1s to develop basic automated driving features. These Tier-1s will have to master camera, radar, LiDAR sensors, and the computing.

According to Pierrick Boulay, Technology & Market Analyst, Solid-state Lighting at Yole: “Companies from the semiconductor side, like Qualcomm, Nvidia, and Intel-Mobileye, are positioning themselves, sometimes through acquisitions, at the center of automated driving systems. For example, Qualcomm is in talk to acquire Veoneer to reinforce its position in the automotive industry”.

Companies coming from the consumer industry, like Apple, Huawei, or Xiaomi, are also entering the market. Depending on their strategy, they could develop only the self-driving part or the entire electric car, like Huawei is doing. Foxconn, which is known for its assembly role in the Apple supply chain, is partnering with several companies such as Apple and Stellantis and is increasing its automotive-related business. This new role of subcontractor, like Foxconn, is growing, and a recent partnership between Fisker and Magna showed that Magna will assemble the car. In the future, it could be possible to see new automotive OEMs being fabless and relying on the experience of subcontractors.

The COVID-19 crisis has emphasized the increasing importance of semiconductors in cars. Companies coming from the semiconductor and software sides have strong financial power and could acquire some Tier-1s or Tier-2s’ companies. This could reshape the automotive landscape in the coming years.

All year long, Yole Développement and System Plus Consulting publish an impressive collection of reports and monitors.

Experts realize various key presentations, organize key conferences and interview leading industrial companies.

In this regard, do not miss SEMI Smart Mobility Forum 2021. Eric Mounier PhD, Director of Market Research at Yole Développement will present “Automotive Semiconductor Trends” on September 24, 2021. Register here!

Analysts aim is to deliver key results and technology and market trends and explain the major changes. Do not hesitate to read System Plus Consulting’s interviews by EETIMES:

• EETIMES – Under the Hood: the Innovation-Rich Golf 8
• EETIMES – Under the Hood: the ID.3 and VW’s electrification platform

Make sure to be aware of the latest news coming from the industry and get an overview of our activities, including interviews with leading companies and more on i-Micronews. Stay tuned!

Extracted from:

• Automotive Semiconductor Trends 2021 report, Yole Développement, 2021
• Automotive Teardown Track, System Plus Consulting

Acronyms:

• C.A.S.E.: Connectivity, ADAS, Sharing, Electrification
• CAGR: Compound Annual Growth Rate
• OEM: Original Equipment Manufacturer
• ADAS: Advanced Driver Assistance Systems
• LiDAR: Light Detection and Ranging

The post Automotive Semiconductor Industry: The Winds of Change appeared first on Edge AI and Vision Alliance.

Chris Yates, Director at Vision Ventures, presents the “Benchmarking Value in the Age of AI: Rational Decisions About Black Box Technology” tutorial at the May 2021 Embedded Vision Summit. The combination of AI and embedded vision promises compelling solutions to an almost infinite range of practical problems, often vastly outperforming…

“Benchmarking Value in the Age of AI: Rational Decisions About Black Box Technology,” a Presentation from Vision Ventures

Register or sign in to access this content.

Registration is free and takes less than one minute. Click here to register and get full access to the Edge AI and Vision Alliance's valuable content.

The post “Benchmarking Value in the Age of AI: Rational Decisions About Black Box Technology,” a Presentation from Vision Ventures appeared first on Edge AI and Vision Alliance.

On September 29, 2020 at 2:00 pm CET (Central Europe Time), Alliance Member Company e-con systems will deliver the webinar “How to Choose the Right Camera for Your Product”. From the event page:

Choosing the right camera for your end product from evaluation to prototyping to productization can eliminate costly redesigns and save valuable time in your product development journey. In addition to this, having the right camera can enhance your product’s performance and help you gain an edge over your competition.

To help you out, we’re organizing this joint webinar in collaboration with Silicon Highway, where we take you through the full cycle of integrating a camera into your product. Learn from the top 1% in the embedded vision industry and leverage a combined experience and expertise of more than 3 decades!

Key Takeaways

• Factors to consider while choosing the image sensor, optics and interface for a vision based application
• How to accelerate product development with readily available off-the-shelf camera components
• How to leverage customization on top of off the shelf solutions to build a world class vision based product
• How does choosing the right vision solution, solve common imaging problems for specific use cases
• How e-con helps customers smoothly integrate cameras into their products

An archive recording will be available for viewing after the webinar conclusion, for those who are unable to attend the live event. For more information and to register, please see the event page.

The post Upcoming Webinar Discusses Choosing the Right Camera for Your Product appeared first on Edge AI and Vision Alliance.

New site to support development of embedded vision systems for connected smart devices

Toronto, ON (September 14, 2021) – iENSO (www.ienso.com), a leading provider of embedded vision data systems, has chosen Kyiv, Ukraine as its boots-on-the-ground gateway to Europe.

“A growing number of customers in Europe are working with iENSO to enhance their products with embedded vision, Edge AI, cloud services and reliable cybersecurity features,” said Sebastien Dignard, President of iENSO. “This trend toward creating a new generation of smart devices and appliances makes it the ideal time for iENSO to establish a local presence.”

iENSO’s Kyiv engineering team will help to expand the company’s capabilities as a turnkey solution provider, with vision system-on-a-chip (SOC) development and design and service support for customers. This builds on the company’s existing operations in North America (the HQ in Toronto, Canada) and in Asia (Shenzhen, China).

Ukraine is an important trade and travel corridor between Europe, Asia and the Middle East. From Kyiv, iENSO’s team can operate from a time zone closer to that of its European customers.

Ukraine Invest – the country’s investment promotion office for attracting foreign direct investment – is working to position Ukraine as a new alternative for services and engineering support for Europe’s electronics industry. Many multinational electronics companies have established Ukrainian operations in recent years.

About iENSO

Established in 2003, iENSO (www.ienso.com) is the embedded vision collaboration partner for major international brands, including several Fortune 500 companies. The company designs, builds and integrates custom vision and multi-sensor systems for companies that need to embed cameras, sensing capability, image processing, Edge AI, and cloud services capabilities in their products. With iENSO, brand owners can embed intelligent data extraction and analysis, protected by robust cybersecurity technology, to support real-time decision making at the edge. This extends from appliances and devices around the home, to a range of industrial and commercial applications, including robotics, autonomous vehicles, agriculture, and security.

The post iENSO Opens First Office in Europe appeared first on Edge AI and Vision Alliance.

In 2019, OpenCV and Luxonis launched one of the most successful Kickstarter campaigns for any camera, raising $1.35M with the help of 6500+ backers for OpenCV AI Kit with Depth (OAK-D).

OAK-D is the world’s first Spatial AI camera. It can run multiple neural networks simultaneously for visual perception tasks like object detection, image classification, segmentation, pose estimation, text recognition, and more while performing depth estimation in real-time.

OAK-D won the Best Camera or Sensor of the Year award at this year’s Embedded Vision Summit conference. The biggest spatial AI competition in the world, the OpenCV AI Competition 2021 with $400k in prizes sponsored by Microsoft Azure and Intel, was also centered around OAK-D.

OAK-D was ambitious. What’s coming next is insane!

September 15th, 2021, will start a new era in spatial AI with the launch of OAK-D-LITE. It is OAK-D’s little sister. Powerful, compact, delightfully lightweight, and yet ridiculously inexpensive.

Its easy-to-use Python API, comprehensive documentation, free tutorials and courses, and numerous free models available through Modelplace.AI make it accessible to anyone with only a working knowledge of the Python programming language.

OAK-D-LITE can fuel a spatial AI revolution, but where is the spark?

Its insanely low price of $74 is a bolt of lightning that will shake up the industry.

Let the revolution begin!

The post Spatial AI in Half the Size of an Altoids Tin appeared first on Edge AI and Vision Alliance.

News Highlights:

• Arm is delivering critical resources to accelerate the automotive industry’s software-defined future with support from leading industry players including AWS, Continental, CARIAD and more

• SOAFEE, a new software architecture and open-source reference implementation, brings the real-time and safety needs of automotive together with the advantages of a cloud-native approach

• Two new reference hardware platforms will enable automotive workload exploration and testing on high-performance Arm-based silicon ahead of commercialization

Cambridge, UK, September 15, 2021 – Arm, in collaboration with leaders across the automotive supply chain, today announced it is delivering a new software architecture and reference implementation, Scalable Open Architecture for Embedded Edge (SOAFEE), and two new reference hardware platforms to accelerate the software-defined future of automotive.

As vehicle architectures and capabilities evolve, automotive developers today are challenged by the increasing code complexity needed to deliver Advanced Driver Assistance Systems (ADAS), In-vehicle Infotainment systems (IVI), electrified powertrains, and autonomy. To meet these evolving consumer demands, computing must become more centralized, and software is increasingly critical to allowing this. The resulting changes to how software is being developed, deployed, and managed means that cloud-native development, best known for driving reductions in cost, time and complexity across the cloud infrastructure industry, is more applicable to automotive development than ever before.

However, to address the software-defined needs of cars today quickly and seamlessly, it is imperative to deliver a standardized framework that enhances proven cloud-native technologies that work at scale with the real-time and safety features required in automotive applications. This same framework can also benefit other real-time and safety-critical use cases such as robotics and industrial automation.

“The automotive sector is at a critical inflection point and the supply chain – from IP design to carmakers – is being re-examined and redefined,” said Chet Babla, vice president of automotive, Automotive and IoT Line of Business at Arm. “The industry has asked Arm and its ecosystem to accelerate the vision for a software-defined future. By virtue of Arm’s unique position in the supply chain, we’re leading a collaborative effort that is delivering the standards, software, developer resources and specialized processing platforms designed for the safety and real-time needs of automotive applications.”

Delivering a roadmap for success

SOAFEE is the result of automakers, system integrators, semiconductor, software, and cloud technology leaders coming together to define a new open-standards-based architecture for the software-defined vehicle. In addition, the SOAFEE reference implementation, an implementation of the architecture defined by a Special Interest Group (SIG) of these leaders, will be free open-source software aimed at allowing broad prototyping, workload exploration and early development. Arm is working with leading commercial solutions providers to maximize compatibility and provide a faster route to functionally-safe designs.

Based on the success of Project Cassini and SystemReady from Arm, which enable a standards-based cloud-native experience at the edge, SOAFEE now builds on these initiatives to enable cloud concepts like container orchestration with automotive functional safety and in real-time – a first in the industry. An initial release of the SOAFEE collateral and reference implementation is available to download now here.

Unlocking critical development time today

The development timeline for new automotive applications needs to be as fast and seamless as possible. The immediate availability of SOAFEE will empower cloud-based developers to apply their expertise and contribute to the future of mobility, built on Arm technology.

“Major technology leaps are transforming vehicle electronics and software-architecture requirements, and driving the industry to a software-centric future. At AWS, we’re committed to innovation with key industry players like Arm to help solve complex challenges for delivering Software-Defined Vehicles with a service-oriented architecture,” said Bill Foy, Director, Worldwide Automotive Business Development at AWS. “In collaboration with Arm, and AWS’s Arm-based AWS Graviton2 instances, we can bring significant cost savings, and make it possible to run applications on the same architecture in the cloud and at the automotive edge to simplify developer workflow.”

In partnership with ADLink, Arm is delivering unconstrained performance on a new SystemReady-compatible development platform to help accelerate time-to-market. The platform, powered by Arm Neoverse-based Ampere Altra cores, will allow workload exploration and development on Arm-based silicon using the SOAFEE reference software stack for applications such as cockpit, ADAS, powertrain and autonomous driving. The development platform is comprised of a developer workstation plus a rugged in-vehicle product, and both are available for pre-order here, with general availability expected in Q4 2021.

• The AVA Developer Platform is a high-performance, 32-core scalable compute system built for lab-based development and is capable of running autonomous workloads. It allows developers to leverage accelerator hardware to complement high-performance central processing units (CPUs).
• For in-vehicle prototyping and testing, the high-performance AVA-AP1 has 80 cores for increased CPU performance, extra input / output (IO) capabilities and includes a safety processor to enable in-vehicle execution using real sensors.

A unique position to lead the software-defined revolution, supported by industry leaders

Riclef Schmidt-Clausen, Senior Vice President, Head of Intelligent Cockpit and Body at CARIAD, a Volkswagen Group Company said, “The software-defined car is coming much sooner than anticipated and the infrastructure needed for this is being developed right now, however the industry faces some macro challenges. These include enabling software portability across a wide range of hardware platforms and building the cloud-native software infrastructure to ensure seamless deployment of applications developed in the cloud to a heterogeneous edge platform. As the pioneers of this new evolution, CARIAD and Arm are working closely together to solve key technical challenges and lay a strong foundation for the software-defined future.”

Software-defined functionality is a growing trend in many sectors, and the combined solution of SOAFEE and the new development platform will have applicability in other areas including robotics in medicine, manufacturing, and logistics. For automotive, software-defined functionality will deliver safe, new in-vehicle experiences and features that meet consumers demands and expectations, and crucially will unlock new revenue streams and customer engagement opportunities for automakers, tier ones, software vendors and cloud service providers.

In addition to support from AWS, ADLink, Ampere and CARIAD, the initiatives have received broad support from leaders across the supply chain including Apex.AI, Capgemini Engineering, Continental, Green Hills Software, Linaro, Marvell, MIH Consortium, Red Hat, SUSE, Woven Planet, Zing Robotics and others. The industry turned to Arm due to its unique position at the intersection of hardware and software. Arm and its ecosystem have answered by delivering the resources to bring safety and real-time aware cloud-native paradigms to automotive development, making the software-defined car a reality.

Additional Resources

• Arm Viewpoints Podcast: Accelerating the Transition to the Software-Defined Vehicle
• The Software-Defined Vehicle Needs Hardware That Goes the Distance

Supporting Partner Quotes

ADLINK

“The acceleration towards a software-defined future for automakers requires development platforms that enable exploration of tomorrow’s automotive compute platforms. Our collaboration with Arm on their reference hardware platform for SOAFEE, built on the requisite high-performance Arm-based silicon delivers that with criticality and security. We’re excited about the potential for innovation in automotive on these unconstrained 32 and 80 core hardware platforms.” – Joe Speed, Field CTO, Global Partners, ADLINK Technology

Ampere

“The future of software-defined vehicles relies on cloud-native development paradigms. Ampere® Altra® and Ampere® Altra Max power high-performance cloud-native compute platforms that are opening new doors for automotive developers to seamlessly move workloads between the Arm-powered cloud and edge. These end-to-end Ampere Platforms combined with open-source frameworks like SOAFEE accelerate technology development cycles dramatically. This will enable developers to address the future needs of the cloud-native autonomous vehicle market faster than ever before.” – Jeff Wittich, Chief Product Officer, Ampere Computing

Apex.AI

“At Apex.AI, we make the Vehicle OS for the automotive industry to develop and deploy safety-critical software. SOAFEE will enable cloud-native software development to accelerate software-defined functionality while addressing the safety and real-time requirements of the car, and we are proud to be working with Arm on this initiative.” – Sanjay Krishnan, VP of Product, Apex.AI

Capgemini Engineering

“To successfully position themselves in an evolving landscape, automotive OEMs, suppliers and Tier 1s are taking automotive software to the next level. Arm’s SOAFEE initiative sets a clear direction for developers with a cloud-native reference architecture to code entertainment and driving features, improve AI safety or add security services. As the auto industry changes because of the growing importance of software engineering, Capgemini intends to work with Arm and its hardware and software ecosystem to drive value in product development.” – Walid Negm, EVP & Chief Research & Innovation Officer, Capgemini Engineering

Continental

“Continental is working closely with Arm on their SOAFEE and reference hardware ecosystem initiatives to adopt and populate cloud-native software engineering and technologies within the automotive domain. Together with the Continental Automotive Edge (CAEdge) platform the journey towards an IoT mobility ecosystem and software-defined vehicles is accelerated significantly.” – Michael Huelsewies, SVP and head of architecture and software, Continental

Green Hills Software

“As the leader in safe and secure automotive software, with hundreds of millions of vehicles using our technologies, Green Hills Software values an open standards-based framework for reducing the complexity and porting effort in foundational software for mixed-criticality automotive systems. Through the improved software-portability this delivers, we can help our OEM and Tier 1 customers to more quickly and efficiently bring new software-defined vehicles to market.” – Dan Mender, Vice President of Business Development, Green Hills Software

Linaro

“Linaro has driven collaborative software innovation in the Arm ecosystem for over ten years and we believe the advancement of the software-defined vehicle is a huge opportunity for both users and automotive industry developers alike. We are delighted to work with Arm and our members on delivering the foundational technologies that enable the first SOAFEE reference implementations.” – Andrea Gallo, Vice President of Business Development, Linaro

Marvell

“It is paramount that the cloud, carrier networks and vehicles all work seamlessly together to enable smart, software-defined vehicles. With SOAFEE, Arm and our ecosystem offer an open-standards common framework that will enable industry innovators to develop scalable, flexible, and best-in-class car to cloud services and solutions.” – Will Chu, Senior Vice President and General Manager, Automotive BU, Marvell

MIH Consortium

“At MIH, we are creating an open EV ecosystem that promotes collaboration in the mobility industry. MIH is powered by the members, like Arm, in our Open EV Alliance. With collective wisdom, we develop reference designs and standards resulting in a lower barrier to entry that drives the growth of the overall EV market. The SOAFEE initiative led by Arm shares a similar vision and goal as they are also taking an open approach to foster collaboration. The cloud-native mechanisms in SOAFEE plays a critical role in the EVKit framework that we are developing in MIH that would help reduce development effort. We’re also involved the reference hardware platforms to accelerate our development.” – William Wei, CTO, MIH Consortium

Red Hat

“Red Hat believes that open hybrid cloud technologies will enable the future of the automotive industry, driven and defined by software. To accomplish this, we have engaged with many partners, including Arm, and are pleased to further this collaboration as part of the SOAFEE initiative. We look forward to helping SOAFEE apply Linux-based open source and cloud-native approaches to automotive applications as they are developed and tested on standardized prototyping solutions using Arm reference hardware.” – Francis Chow, Vice President, In-vehicle Operating System, Red Hat

SUSE

“Cloud-native technologies offer huge potential for the automotive industry. These technologies will not only provide new experiences and new services, but also access to a rapidly expanding ecosystem from innovative partners. We’re excited to join Arm in the SOAFEE initiative, to bring the expertise of SUSE in operating systems and container management to the reinvention of the software-defined car.”– Alan Clark, Office of the CTO, SUSE

Woven Planet

“Woven Planet is developing its Arene platform to enable software-first vehicle development, providing open vehicle programming to everyone without compromising safety. As the leading hardware designer, Arm is uniquely positioned to bring together vehicle manufacturers who will in turn shape a common future for software-first development which targets Arm’s hardware ecosystem. Woven Planet is thrilled to join the SOAFEE program because it will enable tight integration between vehicle development and the cloud, lowering the barrier to entry for innovation, and allowing coders to invent in new and exciting ways.” – JF Bastien, Chief Architect, Woven Planet

Zing Robotics

“Zing Robotics’ electric, autonomous vehicle platform has a range of applications, from aviation to industrial to delivery. The transformative power of the platform comes from its software-defined capabilities, which is why we’re excited to be working with Arm on their SOAFEE software and reference hardware initiatives. The platform’s capabilities require enormous processing power and fail-safe design, so cloud-native technologies like Arm’s solutions are a great fit for us. Arm’s development hardware will be at the core of Zing’s autonomous workforce platform. “ – Brent Shedd, CEO, Zing Robotics

About Arm

Arm technology is at the heart of a computing and data revolution that is transforming the way people live and businesses operate. Our energy-efficient processor designs and software platforms have enabled advanced computing in more than 190 billion chips and our technologies securely power products from the sensor to the smartphone and the supercomputer. Together with 1,000+ technology partners we are at the forefront of designing, securing and managing all areas of AI-enhanced connected compute from the chip to the cloud.

The post New Arm Technologies to Transform the Software-defined Future For the Automotive Industry appeared first on Edge AI and Vision Alliance.

Stealth company incubated by blockchain unicorn Bitfury Group and global nanoelectronics R&D center imec

Eindhoven, Netherlands – September 15, 2021 – AI semiconductor startup Axelera AI announced today that it has successfully closed a seed investment round of $12 million.

The round was led by emerging technologies leader Bitfury and joined by global nanoelectronics R&D center imec and deep-tech venture capital funds Innovation Industries and imec.xpand. The funding will support the company’s development of an industry-defining product powering AI applications at the edge.

Axelera AI was incubated by Bitfury Group in 2019, formerly operating as Bitfury AI until the company came out as an independent entity this year. Building on its extensive research and development focused on integrating world-class hardware and software technologies for edge and cloud-to-edge enterprise solutions, Axelera AI joined forces with imec in early 2020 to develop groundbreaking computing architecture for high-performance AI. Headquartered in the AI Innovation Center of the High Tech Campus in Eindhoven, Axelera AI also has R&D offices in Leuven (BE) and Zurich (CH). The company has already recruited an elite team of more than 20 senior engineers and developers from world-leading AI companies and research centers, including Intel, Qualcomm, IBM, and imec.

“With our launch as a new independent company and the closing of this investment, Axelera AI is now claiming its place as a significant player in the global AI sector,” said Fabrizio del Maffeo, CEO and co-founder of Axelera AI. “Our extraordinary team merges complementary expertise in software development, image processing, dataflow architecture, in-memory computing, algorithms and quantization with a proven track record of business success. We look forward to building on our extensive R&D and introducing new solutions across the globe over the next few years.”

“Bitfury has a proven track record of incubating and bringing game-changing technologies to market,” said Valery Vavilov, CEO and co-founder of Bitfury Group. “We can look no further than LiquidStack, the emerging leader in two-phase liquid immersion cooling for the world’s most intensive computing applications, and Crystal Blockchain, the most advanced blockchain analytics platform for digital asset compliance. We could not be more excited to support the launch of Axelera AI, which is bound to accelerate the development and delivery of critical AI solutions across global markets.”

“Imec has spent years researching custom circuits and novel architectures for efficient AI using in-memory compute, algorithms and quantization”, said Diederik Verkest, Program Director for machine learning at imec. “We are proud that after working together with the Bitfury Group, we can leverage and contribute this expertise to ensure Axelera AI develops and brings the best possible product to the edge AI market.”

“Axelera AI is a unique Dutch start-up on a journey to deliver a disruptive AI chip that will accelerate the application of AI at the ‘edge of IOT’,” said Sander Verbrugge, partner at Innovation Industries. “Affordable AI compute power operating at low power is essential to unlock the full potential of AI in our everyday lives. We are proud to be part of this journey and to support this world-class team in achieving its ambitions.”

“Axelera AI has brought together a stellar team and unites several distinct in-memory compute efforts from across Europe under one roof. We very much look forward to working with Fabrizio and his team and helping them achieve their ambitious goals by leveraging imec.xpand’s cross-border, value-add platform and network,” said Cyril Vančura, partner at imec.xpand.

Axelera AI is designing the world’s most efficient and advanced solutions for AI at the edge. Its game-changing hardware and software product will concentrate the AI computational power of an entire server into a single chip at a fraction of the power consumption and price of AI hardware today. The company’s groundbreaking solution will empower thousands of applications of AI at the edge, making the use of AI more efficient and accessible than ever before. Axelera AI’s products will be fully integrated with the leading open-sourced AI frameworks when it launches to select customers and partners in early 2022.

More Information:

Axelera AI Leadership

Axelera AI is led by CEO and co-founder Fabrizio del Maffeo. Prior to founding Axelera AI, he was Head of Artificial Intelligence at Bitfury and Vice President/Managing Director of Asus Group company AAEON Europe. To its non-executive board, Axelera AI has welcomed Bitfury Executive Vice President George Kikvadze and Innovation Industries partner Sander Verbrugge. Valery Vavilov, CEO and co-founder of Bitfury, will also serve as a board member.

About Artificial Intelligence at The Edge

More than 125 billion “things” are expected to be connected to the internet by 2030 – including smartphones, cameras, smart city sensors, and more. Each of these “internet of things (IoT)” devices, even the most miniature sensor, creates an exponential amount of data for analysis. Currently, only 2% of that data has been analyzed due to a lack of expertise and AI tools. Artificial intelligence allows us to analyze and understand this data thoroughly. However, the sheer amount of information and increasingly complex algorithms require a new generation of more powerful, efficient and accessible computation hardware. Due to growing privacy, security and bandwidth concerns, data is increasingly processed close to its origin, or “the edge.”

The AI hardware available today has been designed primarily for cloud computing operations, a sector with cost, power, and scalability constraints. Computational hardware for edge applications needs an entirely new design that considers specific power needs, computational needs and economics. Axelera AI will capture this opportunity by developing and delivering a game-changing and scalable technology with superior performance and efficiency to accelerate AI applications at the edge.

About The Bitfury Group

The Bitfury Group is the world’s leading emerging technologies company. Bitfury is building solutions for the future, with the most significant technologies of the millennium. Our mission is to make the world more transparent and trusted by innovating at every level of technology – hardware, security, and software.

Founded in 2011, Bitfury is the leading security and infrastructure provider for the Bitcoin Blockchain. In addition to securing the Bitcoin Blockchain, Bitfury also designs and produces innovative hardware that keeps cryptocurrencies and blockchains secure, including custom semiconductor chips and mobile data centers. Bitfury is also a software provider for some of the world’s most cutting-edge applications through its Exonum private blockchain framework and its Crystal Blockchain, the most advanced blockchain analytics platform for digital asset compliance.  To learn more, visit www.bitfury.com.

About IMEC

Imec is a world-leading research and innovation centre in nanoelectronics and digital technologies. Imec leverages its state-of-the-art R&D infrastructure and its team of more than 5.000 employees and top researchers, for R&D in advanced semiconductor and system scaling, silicon photonics, artificial intelligence, beyond 5G communications and sensing technologies, and in application domains such as health and life sciences, mobility, industry 4.0, agrofood, smart cities, sustainable energy, and education. Imec unites world-industry leaders across the semiconductor value chain, Flanders-based and international tech, pharma, medical and ICT companies, start-ups, and academia and knowledge centers. Imec is headquartered in Leuven (Belgium), and has research sites across Belgium, in the Netherlands, Taiwan and the USA, and offices in China, India and Japan. In 2020, imec’s revenue (P&L) totaled 680 million euro. Further information on imec can be found at www.imec-int.com.

About Innovation Industries

Innovation Industries is a Deeptech venture capital investor that invests up to €30M in Hightech, Agri&Foodtech and Medtech businesses. Today, Innovation Industries manages over €275M across two funds and has a team of 15 investment professionals. Its largest investors are the Dutch pension funds PME and PMT. Innovation Industries invests from seed to late stage and has partnerships with all Dutch technical universities and research institute TNO. For more information on Innovation Industries see: www.innovationindustries.com

About Imec.xpand

Imec.xpand is an early stage and growth fund with a game-changing approach to turn hardware driven nanotechnology innovation into successful global companies. The fund is independently managed by a team of entrepreneurs, seasoned venture capitalists and business experts and is targeting ambitious tech start-ups were imec knowledge, expertise and infrastructure can play a determining role in the success. For more information about imec.xpand please visit us on www.imecxpand.com

The post Dutch AI Semiconductor Startup Axelera AI Launches With $12 Million Seed Round appeared first on Edge AI and Vision Alliance.

Rakshit Agrawal, Director of Research and Development at Camio, presents the “How Containerization Unblocks Barriers to Fast, Easy Deployment of AI-Driven Visual Recognition Applications” tutorial at the May 2021 Embedded Vision Summit. Recent progress in machine learning algorithms has been meteoric. Yet their implementation in real-world embedded applications crawls at…

“How Containerization Unblocks Barriers to Fast, Easy Deployment of AI-Driven Visual Recognition Applications,” a Presentation from Camio

Register or sign in to access this content.

Registration is free and takes less than one minute. Click here to register and get full access to the Edge AI and Vision Alliance's valuable content.

The post “How Containerization Unblocks Barriers to Fast, Easy Deployment of AI-Driven Visual Recognition Applications,” a Presentation from Camio appeared first on Edge AI and Vision Alliance.

September 14, 2021–PARIS–(BUSINESS WIRE)–GrAI Matter Labs (GML), a pioneer of brain-inspired computing solutions, announced today that it has entered a partnership with ERM Automatismes (ERM), a French market leader for didactic solutions in technological and professional training for 30 years, to create a ROS2-based robotics education kit.

The Life-Ready AI Kit for Education combines robotics with brain-inspired AI as the key ingredient for Industry 4.0. The kit is comprised of a robot arm with a 2D/3D camera for visual perception, and a GML GrAI VIP platform integrated with an ADLINK I-Pi SMARC module. The ERM curriculum includes modules for grasping, regression, classification and more. It was developed with teachers from the French academy, universities, and a community of AI specialists to offer a very practical and easy to use model for teachers and students.

GML and ERM will showcase the education kit at the SIDO Fair in Lyon starting on 22^nd of September. This unique kit, using GML’s “Life-Ready AI” chip, GrAI VIP will demonstrate a real-life application of classifying agricultural products on a conveyor belt.

“We are delighted to be collaborating with ERM to teach our new generation of engineers cutting edge AI technologies. The students will learn to apply energy efficient machine-learning algorithms in an industrial context and help to kickstart a Life-Ready AI ecosystem which helps to save time, money and vital natural resources for Industry 4.0,” said Christian Verbrugge, GML Europe – Senior Director Business Development.

Cyril Liotard, CEO of ERM, is excited about the partnership, saying “with GrAI Matter Labs, we are expanding our existing portfolio with Life-Ready AI technology. GML offers a one-of-a-kind brain-inspired computing architecture that has never been seen before in education; that is why we are thrilled to collaborate with GML, and academia to achieve this goal together.”

About GrAI Matter Labs

At GrAI Matter Labs we are in the business of Life-Ready AI. Artificial Intelligence as close to natural as it gets. AI that feels alive. We deliver brain-inspired chips that behave like humans. AI that makes machines act and react in real-time. AI that optimizes energy and maximizes efficiency, saving time, money, and vital natural resources. GML is led by a team of visionary and seasoned engineers and is backed by leading investors, including iBionext, 360 Capital Partners, 3T Finance, and Celeste Management.

For more information, please visit www.graimatterlabs.ai.

About ERM Automatismes

ERM Automatismes, founded in 1990, is a French company that provides technical systems and services in education, robotics, Fab&Test, and energy; and has developed didactic solutions for technological training as well as robotic and mechatronic solutions for the food industry. ERM provides a wide range of manufacturing and testing solutions to meet the growing needs of industrial customers, schools, colleges, universities, and engineering schools, as well as local authorities, municipalities, departments, and regions.

For more information, please visit https://www.erm-automatismes.com/.

The post GML and ERM Partner to Develop Life-Ready AI Kit for Education appeared first on Edge AI and Vision Alliance.

New Tensilica AI engine boosts performance and AI accelerators provide a turnkey solution for consumer, mobile, automotive and industrial AI SoC designs

Highlights:

• Domain-specific, extensible and configurable AI platform built upon mature, volume production-proven Tensilica architecture

• Industry-leading performance and energy efficiency for on-device AI applications

• Comprehensive, common AI software addresses all target markets

• Low-, mid- and high-end AI product families for the full spectrum of PPA and cost points

• Scalable from 8 GOPS to 32 TOPS currently and 100s of TOPS for future AI requirements

SAN JOSE, Calif., September 13, 2021—Cadence Design Systems, Inc. (Nasdaq: CDNS) today unveiled its Tensilica^® AI Platform for accelerating AI SoC development, including three supporting product families optimized for varying data and on-device AI requirements. Spanning the low, mid and high end, the comprehensive Cadence^® Tensilica AI Platform delivers scalable and energy-efficient on-device to edge AI processing, which is key to today’s increasingly ubiquitous AI SoCs. A new companion AI neural network engine (NNE) consumes 80% less energy per inference and delivers more than 4X TOPS/W compared to industry-leading standalone Tensilica DSPs, while neural network accelerators (NNAs) deliver flagship AI performance and energy efficiency in a turnkey solution.

Targeting intelligent sensor, internet of things (IoT) audio, mobile vision/voice AI, IoT vision and advanced driver assistance systems (ADAS) applications, the Tensilica AI Platform delivers optimal power, performance and area (PPA) and scalability with a common software platform.   Built upon the highly successful application-specific Tensilica DSPs already shipping in volume production in leading AI SoCs for the consumer, mobile, automotive and industrial markets, the Tensilica AI Platform product families include:

• AI Base: Includes the popular Tensilica HiFi DSPs for audio/voice, Vision DSPs, and ConnX DSPs for radar/lidar and communications, combined with AI instruction-set architecture (ISA) extensions.
• AI Boost: Adds a companion NNE, initially the Tensilica NNE 110 AI engine, which scales from 64 to 256 GOPS and provides concurrent signal processing and efficient inferencing.
• AI Max: Encompasses the Tensilica NNA 1xx AI accelerator family—currently including the Tensilica NNA 110 accelerator and the NNA 120, NNA 140 and NNA 180 multi-core accelerator options—which integrates the AI Base and AI Boost technology. The multi-core NNA accelerators can scale up to 32 TOPS, while future NNA products are targeted to scale to 100s of TOPS.

All of the NNE and NNA products include random sparse compute to improve performance, run-time tensor compression to decrease memory bandwidth, and pruning plus clustering to reduce model size.

Comprehensive common AI software addresses all target applications, streamlining product development and enabling easy migration as design requirements evolve. This software includes the Tensilica Neural Network Compiler, which supports these industry-standard frameworks: TensorFlow, ONNX, PyTorch, Caffe2, TensorFlowLite and MXNet for automated end-to-end code generation; Android Neural Network Compiler; TFLite Delegates for real-time execution; and TensorFlow Lite Micro for microcontroller-class devices.

“AI SoC developers are challenged to get to market faster with cost-effective, differentiated products offering longer battery life and scalable performance,” said Sanjive Agarwala, corporate vice president and general manager of the IP Group at Cadence. “With our mature, extensible and configurable platform based on our best-in-class Tensilica DSPs and featuring common AI software, Cadence allows AI SoC developers to minimize development costs and meet tight market windows. By enabling AI across all performance and price points, Cadence is driving the rapid deployment of AI-enabled systems everywhere.”

Supporting Quotes

“Scaling low power on-device AI capabilities requires extremely efficient multi-sensory compute. Cadence and the TensorFlow Lite for Microcontrollers (TFLM) team have been working together for many years to co-develop solutions that enable the most cutting-edge, low-footprint use cases in the AI space. The trend for real-time audio networks to use LSTM-based neural nets for the best performance and efficiency is a key example. Working closely with Cadence, we are integrating a highly optimized LSTM operator on Tensilica HiFi DSPs that enables the next level of performance improvements for key use cases like voice-call noise suppression. We are excited to continue this collaboration and provide industry-level innovation in the low-energy AI space.”
– Pete Warden, Technical Lead of TensorFlow Lite Micro at Google

“On-device AI deployment on our KL720—a 1.4 TOPS AI SoC targeted for vehicles, smart home, smart security, industrial control applications, healthcare and AI of things (AIoT)—is key to both our customers’ success and our mission to enable AI everywhere, for everyone. Cadence’s high-performance, low-power Tensilica Vision DSPs pack a lot of compute capacity with AI ISA extensions plus the necessary AI software to tackle the latest AI challenges.”
– Albert Liu, Founder and CEO of Kneron

“Integrating a Cadence Tensilica HiFi 4 DSP into the NXP i.MX RT600 crossover MCU not only provides high-performance DSP capabilities for a broad range of audio and voice processing applications, but also increases inference performance, enabling AI even in very low-power, battery-operated products. The HiFi neural network library allows NXP to take full advantage of the AI capabilities of the HiFi 4 DSP and integrate it into NXP’s eIQ Machine Learning Software Development Environment supporting the TensorFlow Lite Micro and Glow ML inference engines.”
– Cristiano Castello, Sr. Director of Microcontrollers Product Innovation at NXP Semiconductors

 “As AI applications are rapidly propagating from the cloud to the edge, integrating an AI accelerator on-device has become a must to meet low-latency requirements in ADAS, mobile, intelligent sensors and the IoT. AI SoCs require mature accelerator IP that addresses the varying needs of each market and includes a comprehensive software solution. With its Tensilica AI Base, AI Boost and AI Max technology that provides a clear migration path as performance and energy requirements evolve, Cadence is a compelling IP provider for comprehensive on-device AI IP solutions.”
– Mike Demler, Senior Analyst at The Linley Group

Availability

The NNE 110 AI engine and the NNA 1xx AI accelerator family support Cadence’s Intelligent System Design strategy, which enables pervasive intelligence for SoC design excellence, and are expected to be in general availability in the fourth quarter of 2021. For more information on the Tensilica AI Platform and new AI IP, please visit www.cadence.com/go/TensilicaAI.

About Cadence

Cadence is a pivotal leader in electronic design, building upon more than 30 years of computational software expertise. The company applies its underlying Intelligent System Design strategy to deliver software, hardware and IP that turn design concepts into reality. Cadence customers are the world’s most innovative companies, delivering extraordinary electronic products from chips to boards to systems for the most dynamic market applications, including consumer, hyperscale computing, 5G communications, automotive, mobile, aerospace, industrial and healthcare. For seven years in a row, Fortune magazine has named Cadence one of the 100 Best Companies to Work For. Learn more at cadence.com.

The post Cadence Accelerates Intelligent SoC Development with Comprehensive On-Device Tensilica AI Platform appeared first on Edge AI and Vision Alliance.

New Topaz sensors are ideal for mobile applications, offering an attractive price point with no-compromise performance.

Grenoble, FRANCE, September 14, 2021 – Teledyne e2v, a part of Teledyne Technologies [NYSE: TDY], has introduced its Topaz series of industrial CMOS sensors with new 2MP and 1.5MP resolution devices. These new 1920 x 1080 and 1920 x 800-pixel format sensors use state-of-the-art low noise, global-shutter pixel technology to offer powerful solutions and enable compact mobile designs for many applications.

Housed in a tiny 4.45 mm wide Chip Scale Package (CSP), the Topaz sensors have an optical array centre which precisely matches the mechanical centre of the package, allowing for a slim camera design. This makes them particularly suitable for miniature OEM barcode engine designs, mobile terminals and sleds, IoT, contactless authentication systems, wearable devices, drones and robotics. Their 1/3” optical format is made possible due to the small 2.5µm global shutter pixel that employs in-pixel CDS (correlated double sampling) and advanced dual light-guides to achieve good SNR at low-light, with low crosstalk for crisp images.

Gareth Powell, Marketing Manager for CMOS Sensors at Teledyne e2v said, “Modern logistics, retail, and manufacturing applications all require enhanced productivity and throughput, with longer working ranges. The new Topaz sensors were developed with this in mind, offering an excellent cost/performance ratio for high volume adoption. They also have a tiny footprint making them ideal to drive the world’s smallest barcode OEM engines and thinnest mobile platforms.”

Key Features

• Advanced 2.5µm x 2.5µm global shutter pixel
• Low readout noise of typically 3.3 electrons
• Excellent low light SNR performance with short exposure times
• Ultra-low dark signal for exceptionally good high temp performance
• Frame rate of >100 frames per second in 8 bit output mode
• 2 lane MIPI outputs (1.2Gpix/sec each) for seamless connection with CPU, ISP and other application processors
• Fast Wake Up mode – decode within 10 milliseconds after power up and other useful application features

Samples and evaluation kits are available now. Please contact us for more information.

Find out more about Topaz at VISION, 5-7 October 2021 in Stuttgart, Germany, Teledyne booth 8B10.

The post Teledyne e2v Introduces the Industry’s Smallest 2MP & 1.5MP CMOS Sensors, Featuring a Low-noise Global Shutter Pixel appeared first on Edge AI and Vision Alliance.

LETTER FROM THE EDITOR

Dear Colleague, This Thursday, September 16 at 9 am PT, Edge Impulse will deliver the free webinar “Introducing the EON Tuner: Edge Impulse’s New AutoML Tool for Embedded Machine Learning” in partnership with the Edge AI and Vision Alliance. Finding the best machine learning (ML) model for analyzing sensor data isn’t easy. What pre-processing steps yield the best results, for example? And what signal processing parameters should you pick? The selection process is even more challenging when the resulting model needs to run on a microcontroller with significant latency, memory and power constraints. AutoML tools can help, but typically only look at the neural network, disregarding the important roles that pre-processing and signal processing play with tinyML. This hands-on session will introduce the EON Tuner, a new AutoML tool available to all Edge Impulse developers. You’ll learn how to use EON Tuner to pick the optimum model within the constraints of your device, improving the accuracy of your audio, computer vision and other sensor data classification projects. The webinar will include demonstrations of the concepts discussed. For more information and to register, please see the event page. Brian Dipert Editor-In-Chief, Edge AI and Vision Alliance

DEEP LEARNING ADVANCEMENTS

New Methods for Implementation of 2-D Convolution for Convolutional Neural Networks The increasing usage of convolutional neural networks (CNNs) in various applications on mobile and embedded devices and in data centers has led researchers to explore application specific hardware accelerators for CNNs. CNNs typically consist of a number of convolution, activation and pooling layers, with convolution layers being the most computationally demanding. Though popular for accelerating CNN training and inference, GPUs are not ideal for embedded applications because they are not energy efficient. ASIC and FPGA accelerators have the potential to run CNNs in a highly efficient manner. In this talk, Tokunbo Ogunfunmi, Professor of Electrical Engineering and Director of the Signal Processing Research Laboratory at Santa Clara University, presents two new methods for 2-D convolution that offer significant reduction in power consumption and computational complexity. The first method computes convolution results using row-wise inputs, as opposed to traditional tile-based processing, yielding considerably reduced latency. The second method, single partial product 2-D (SPP2D) convolution, avoids recalculation of partial weights and reduces input reuse. Hardware implementation results are presented. Imaging Systems for Applied Reinforcement Learning Control Reinforcement learning has generated human-level decision-making strategies in highly complex game scenarios. But most industries, such as manufacturing, have not seen impressive results from the application of these algorithms, belying the utility hoped for by their creators. The limitations of reinforcement learning in real use cases intuitively manifest from the number of exploration examples needed to train the underlying models, but also from incomplete state representations for an artificial agent to act on. In an effort to improve automated inspection for factory control through reinforcement learning, Nanotronics’ research is focused on improving the state representation of a manufacturing process using optical inspection as a basis for agent optimization. In this presentation, Damas Limoge, Senior R&D Engineer at Nanotronics, focuses on the imaging system: its design, implementation and utilization, in the context of a reinforcement agent.

EDGE DEVICE OPTIMIZATIONS

Deep Learning on Mobile Devices Over the last few years, convolutional neural networks (CNNs) have grown enormously in popularity, especially for computer vision. Many applications running on smartphones and wearable devices could benefit from the capabilities of CNNs. However, CNNs are by nature computation- and memory-intensive, making them challenging to deploy on a mobile device. In this presentation, independent AI architect Siddha Ganju shares easy-to-use, practical techniques that can improve CNN performance on mobile devices. DevCloud for the Edge In this video, Jeff Bier, Founder of the Edge AI and Vision Alliance, interviews Monique Jones, Senior Software Engineer Team Lead at Intel, about her company’s DevCloud for the Edge developer toolset. Bier and Jones compare and contrast a DevCloud-based development flow versus the traditional approach that leverages a development board located at the engineer’s desk and directly connected to a computer, highlighting the advantages of the cloud-based approach and showcasing some of the toolset’s key features and capabilities. DevCloud for the Edge allows you to virtually prototype and experiment with AI workloads for computer vision on the latest Intel edge inferencing hardware, with no hardware setup required since the code executes directly within the web browser. You can test the performance of your models using the Intel Distribution of OpenVINO Toolkit and combinations of CPUs, GPUs, VPUs and FPGAs. The site also contains a series of tutorials and examples preloaded with everything needed to quickly get started, including trained models, sample data and executable code from the Intel Distribution of OpenVINO Toolkit as well as other deep learning tools.

UPCOMING INDUSTRY EVENTS

Introducing the EON Tuner: Edge Impulse’s New AutoML Tool for Embedded Machine Learning – Edge Impulse Webinar: September 16, 2021, 9:00 am PT Securing Smart Devices: Protecting AI Models at the Edge – Sequitur Labs Webinar: September 28, 2021, 9:00 am PT How Battery-powered Intelligent Vision is Bringing AI to the IoT – Eta Compute Webinar: October 5, 2021, 9:00 am PT More Events

FEATURED NEWS

STMicroelectronics’ New 8×8 Multi-Zone Ranging Time-of-Flight Sensor Brings Distance Discernment to a Spectrum of Consumer and Industrial Products MediaTek’s Latest AI-enhanced Application Processors Include the Kompanio 900T for Tablets and Notebooks and Kompanio 1300T for Premium Tablets Efinix Has Announced AEC-Q100 Qualification and an Automotive Product Line Initiative for Its Programmable Logic Devices Ambarella’s Partnerships with KeepTruckin and Yandex Target Vehicle Fleet Applications for Computer Vision Unikie and Ericsson’s Combined Development Leverages Visual AI and a 5G Network for Real-time Remote Vehicle Steering More News

EDGE AI AND VISION PRODUCT OF THE YEAR WINNER SHOWCASE

Simbe Robotics Tally 3.0 (Best Enterprise Edge AI End Product) Simbe Robotics’ Tally 3.0 is the 2021 Edge AI and Vision Product of the Year Award Winner in the Enterprise Edge AI End Products category. Tally is the only robot on the market that combines computer vision, machine learning, and RFID technologies to audit store shelves across a range of retail environments. Tally detects misplaced, mispriced, and out-of-stock items, arming retailers with stronger insights into shelf availability, and ensuring that items are more quickly restocked and corrected, improving the customer experience. Tally 3.0 combines both edge and cloud computing, enabling it to transfer some of its AI and machine learning workloads to the edge. This hybrid system better optimizes data collection and processing, getting insights to store teams more quickly. By operating both on the edge and in the cloud, Tally can more quickly use deep learning to help with tasks like autofocus and barcode decoding, ensuring stores have the most up-to-date data. Please see here for more information on Simbe Robotics’ Tally 3.0. The Edge AI and 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 a company’s leadership in edge AI and computer vision as evaluated by independent industry experts.

The post Edge AI and Vision Insights: September 14, 2021 Edition appeared first on Edge AI and Vision Alliance.

Dinesh Balasubramaniam, Senior Product Marketing Manager at Ambarella, presents the “When 2D Is Not Enough: An Overview of Optical Depth Sensing Technologies” tutorial at the May 2021 Embedded Vision Summit. Camera systems used for computer vision at the edge are smarter than ever, but when they perceive the world in…

“When 2D Is Not Enough: An Overview of Optical Depth Sensing Technologies,” a Presentation from Ambarella

Register or sign in to access this content.

Registration is free and takes less than one minute. Click here to register and get full access to the Edge AI and Vision Alliance's valuable content.

The post “When 2D Is Not Enough: An Overview of Optical Depth Sensing Technologies,” a Presentation from Ambarella appeared first on Edge AI and Vision Alliance.

This market research report was originally published at Yole Développement’s website. It is reprinted here with the permission of Yole Développement.

Yole Développement (Yole) identifies new applications in a dynamic market bringing growth and opportunities for lots of companies…

OUTLINE:

• Market forecasts by applications:

  Growth in datacom & telecom applications will drive the InP market in the next 5 years. CAGR between 2020 and 2026 are respectively 22% and 7%.

  Emerging InP applications such as LiDAR for automotive, 3D sensing and wearables are showing impressive CAGR: 210%, 467% and 112% respectively.

• Technology trends:

  The main technology used today in optical communication applications is DFB with more than 63% of the market share.

  InP EEL process yields are still low, hampering economy-of-scale.

  The demand for transceivers with high transmission rates exceeding 400GBps will drive technology advancement in single lasers and integrated devices.

• Supply chain:

  II-VI and Lumentum are leading the InP bare-die market.

  LandMark continues to dominate the InP epiwafer market, with more than 60% of the open epiwafer market share.

  Sumitomo and AXT continue to lead the wafer market with more than 75% of combined market share.

  China has been investing heavily in optical infrastructure with a massive deployment of 5G transceivers.

  Due to US-China trade war and COVID 19 pandemic, many laser & photonics companies lost big customers…

“At Yole, we expect an initial slight market penetration of InP in 2022 in wearables, followed by a significant increase to US$255 million in 2026 with a CAGR2022-2026 of 112%.” asserts Ahmed Ben Slimane, PhD, Technology & Market Analyst, Compound Semiconductors and Emerging Substrates at Yole Développement (Yole). He adds:“For LiDAR applications, InP could be promising, enabling eye safety at higher wavelengths. Leading companies such as Volvo, ZF, Continental, Daimler, etc. are interested in adopting InP-based LiDAR .”

For smartphones, OLED displays are transparent at wavelengths in the range of 13xx to 15xxnm, explains Yole’s analyst. OEMs interested in removing the camera notch on mobile phone screens and integrating the 3D-sensing modules under OLED displays are considering moving to InP EELs, replacing the current GaAs VCSELs. Even though this trend is currently in an early R&D phase, Yole sees a strong interest from several players, such as ams, Infineon Technologies, STMicroelectronics and several laser manufacturers and sensor players.

In this context, the market research and strategy consulting company investigates disruptive technologies and related markets in depth. Its aim is to point out the latest innovations and underline the business opportunities.

Released today, the InP Wafer, Epiwafer and Device Market 2021: Photonics and RF Application report delivers a comprehensive and detailed understanding of the InP industry, covering markets from wafers and epiwafers to bare dies. Including market trends and forecasts, supply chain, technology trends, technical insights and challenges analysis, take away and outlook, this study proposes an in-depth description of the ecosystem and main players’ strategies.

What are the economic and technological challenges of this industry? What are the key market drivers? Who are the companies to watch, and what innovative technologies are they working on? What are the recent investments, mergers, and acquisitions?… Today, Yole’s compound semiconductors and emerging substrates analysts present their vision of the InP industry.

As analyzed by Yole’s team in the new InP Wafer, Epiwafer and Device Market 2021: Photonics and RF Application report, as an indispensable building block for high-speed and long-range optical transceivers, InP laser diodes remain the best choice for telecom & datacom photonic applications. However, following the COVID-19 outbreak and the US-China trade tensions, telecom infrastructure deployment was disrupted, resulting in a minor InP market slowdown in 2020. The requirement for more data transfer at higher speed in datacom is increasing, with technology migrating to single InP lasers targeting state-of-the-art 100 Gbps output, making them preferable in 400Gbps and 800Gbps transceivers.

Driven by high volume adoption of high data rate lasers, the datacom bare die market reached around US$963 million in 2020. It is expected to be worth US$3.1 billion in 2026 at a 22% CAGR during this period.

Meanwhile, the cyclic InP telecom market will continue its growth thanks to 5G deployment. Yole sees a comfortable increase from US$1.1 billion in 2020 to US$1.7 billion in 2026 at a CAGR2020-2026 of 7%.

The InP industry is a dynamic market with lots of opportunities for legacy players and new entrants. In its InP technology & market analysis, Yole reports a fragmented market, with numerous companies, especially at the device level.

According to Ahmed Ben Slimane from Yole: “Two American players are leading the InP market: II-VI and Lumentum. Both have increased their market share and strengthened their position thanks to strategic mergers and acquisitions. Indeed, II-VI acquired Finisar in 2019, and Lumentum acquired Oclaro in 2018.”

II-VI and Lumentum have both vertically integrated business models: they generate revenues at bare die, device, and module-level. II-VI also offers epiwafer products. Their combined bare die market share is around 30%.

Facing II-VI and Lumentum, Yole identifies major Chinese players like Hisense and Accelink. Both companies are in the global top five and are increasing step by step their share of the market. The two Chinese companies take advantage of the US-China trade tensions and the massive 5G transceiver deployment in Asia, explain Yole’s analysts in the InP report.

Sensing applications targeting the mass consumer and automotive markets are attracting new players. Yole has identified several players interested in entering this market:

• Vertically integrated InP players with the know-how and an already established structure, as they can easily switch to sensing applications as soon as the market becomes bigger. II-VI and Lumentum are part of this segment.
• GaAs players with foundry capabilities could leverage the existing GaAs tools to switch to similar InP processes. Yole sees for example, ams and Trumpf.
• Emerging foundries or companies already working on InP-based solutions.

According to Poshun Chiu, Technology & Market Analyst, Compound Semiconductors & Emerging Materials at Yole: “In the last category, we witnessed an increase in private investments and SPACs in the last year. These include, in Q4-2020, Luminar raised US$590 million and went public, then acquired OptoGration; in Q2-2021, Aeva went public with an initial valuation of US$1.7 billion; and in Q2-2021, Rockley Photonics announced its intention to go public at an initial valuation of US$1.2 billion with an Apple supported project for smartwatches”.

All year long, Yole Développement publishes an impressive collection of compound semiconductors & emerging substrates reports and monitors. InP technology & market report is part of these activities.

Experts realize various key presentations, organize key conferences and interview leading industrial companies all year long. Their aim is to deliver key results and technology and market trends and explain the major changes.

Make sure to be aware of the latest news coming from the industry and get an overview of our activities, including interviews with leading companies and more on i-Micronews. Stay tuned!

Extracted from:

• InP Wafer, Epiwafer and Device Market 2021: Photonics and RF Applications report, Yole Développement, 2021

ACRONYMS:

• InP: Indium Phosphide
• CAGR: Compound Annual Growth Rate
• DFB: Distributed Feedback Lasers
• EEL: Edge Emitting Laser
• LiDAR: Light Detection and Ranging
• OEM: Original Equipment Manufacturer
• GaAs: Gallium Arsenide

The post With More Than US$5 Billion in 2026, the InP Bare Die Market is Offering Opportunities and Profits appeared first on Edge AI and Vision Alliance.

Now available for on-demand viewing is the archive recording of the webinar “Building Scalable Edge ML Solutions with NXP i.MX Applications Processors and AWS Cloud Services,” co-presented by Ali Osman Örs, Director of AI ML Strategy and Technologies for Edge Processing at NXP Semiconductors, and Jack Ogawa, responsible for Strategic Semiconductor Partnerships for IoT at Amazon Web Services. From the event page:

Machine Learning (ML) technology is fast becoming ubiquitous as the heart of differentiated IoT devices, often defining the smart capabilities delivered. ML model monitoring, maintenance and updatability are essential to ensure that these smart capabilities continue to deliver the value promised throughout the lifecycle of the device, and require an MLOps strategy tailored for IoT scale.

In this session, NXP and AWS explore how to build and deploy ML solutions to many edge devices at scale and securely support MLOps for maintaining models through their lifecycle. The audience will learn how to address common MLOps challenges in the context of an IoT application running across multiple embedded devices.

What You Will Learn:

• Building ML solutions with AWS cloud services such as Amazon SageMaker, AWS IoT Greengrass
• Integrating cloud services with edge inference solutions from NXP
• Deploying to an edge i.MX 8M Plus device based board
• Managing a multi device fleet with Amazon SageMaker Edge Manager

For more information and to register, please see the event page.

The post Webinar Explores Building Scalable Edge ML Solutions Using NXP Processors and Amazon Services appeared first on Edge AI and Vision Alliance.

This blog post was originally published at Deeplite’s website. It is reprinted here with the permission of Deeplite.

Words can have multiple meanings. As a result, people involved with adjacent areas of the technology industry often misunderstand each other, even when using the same terms. Each group has its own frame of reference, history and technical domain expertise.

In particular, areas of technology cross-overs and convergence are often fraught with category errors, flawed assumptions – or just poor communications. There is a significant risk that this is occurring in the area of Edge AI. At least four different groups interpret that term in very distinct ways.

For example, many professionals in the cloud and network world have no idea about what can be achieved with optimized Edge AI on devices, either currently or what is likely soon – and what that implies for their own visions of the future. The telecoms industry, in particular, appears to be at risk of missing an important “disruption from adjacency”.

This article is aimed at helping these people talk to each other, better understand each others’ needs and expectations – and also help them avoid poor decisions through a lack of awareness of broader tech trends.

If you asked representatives of the following industries to play “word association” with the phrase “Edge AI”, they might suggest very different explanations:

Deep Neural Network (DNN) specialist

Someone involved in image detection or speech analysis might mention the trends towards “model compression” or “AI optimization”, with heavy, resource-consuming or slow cloud-based inferencing shrunk down to work more efficiently on a CPU or microcontroller on a device – for instance a camera or smartphone. This is “AI at the edge” for them. It may solve multiple problems, from lower latency to reduced energy consumption (and better economics) for AI.

IoT System developer

Someone involved with building a connected vehicle, or the quality-control for a smart factory, might think about a local compute platform capable of combining feeds from multiple cameras and other sensors, perhaps linked to autonomous driving or closed-loop automation control. Their “edge AI” resides on an onboard server, or perhaps an IoT gateway unit of some sort.

Mobile network operator (MNO)

A telecom service provider building a 5G network may think of Edge AI both for internal use (to run the radio gear more efficiently, for instance) and as an external customer-facing platform exploiting low-latency connections. The “mobile edge” might be targeted at a connected road junction, video-rendering for an AR game, or a smart city’s security camera grid. Here, “Edge AI” is entwined with the network itself – its core functions, “network slicing” capabilities, and maybe physically located at a cell-site or aggregation office. It is seen as a service rather than an in-built capability of the system.

Datacentre & cloud providers

For companies hosting large-scale compute facilities, AI is seen as a huge source of current and future cloud demand. However, the infrastructure providers often won’t grasp the differences between training and inferencing, or indeed the finer details of their customers’ application and compute needs. “Edge” may just mean a datacentre site in a tier-3 city, or perhaps a “mini datacentre” serving users in a 10-100km radius.

These separate visions and definitions of “Edge AI” may span as much as 10 orders of magnitude in terms of scale and power – from milliwatts to megawatts. So, unsurprisingly, the conversations would be very different – and each group would probably fail to recognize each other’s “edge” as relevant to their goals.

These are not the only categories. Others include chip and module vendors, server suppliers, automation and integration specialists, cloud/edge platforms and federation enablers and so forth. Added to these are a broad array of additional “edge stakeholders” – from investors to government policymakers.

Why does this matter? Because AI applications ultimately fit into broader ecosystems, transformation projects, consumer and business products or even government policy and regulatory regimes. In most cases, all of these groups will need to organize themselves into a value chain, or at least depend on each other.

The developer perspective

Often, edge-AI market participants focus – understandably – on what they perceive as their unique capabilities, whether that is their preferred models, their physical premises, network/system speeds and their existing customer relationships. And internally, they are looking for new revenue opportunities and use-cases to help justify their investments, as well as gain more “customer ownership”.

But the questions which don’t get asked often enough are “What does the developer – and the final end-user – really value? What are their constraints? And how will that drive their decision choices, now or in the future?”

For instance, consider an application developer working on an AI-powered object recognition tool. At the moment, their product has a few problems to resolve. In particular, the response times are laggy, which reduces the effectiveness and market opportunity of the overall solution. Given the round-trip time of video images to and from the cloud, plus the significant processing load and inference time, they can only get one reliable response per second, and the implied cost means it’s only suitable for certain high-value tasks.

That’s fine for monitoring crowds and lost property in a railway station – or detecting a particular parasitic beetle on a crop-leaf – but isn’t useful for spotting defects on a fast production-line conveyor or to react to a deer jumping in front of an autonomous vehicle.

For the next version of their product, they have a range of different improvement and optimization paths they could pursue:

• Aim for use-cases with reliable lower-latency network connections to the cloud, for instance, by using 5G, fiber or Wi-Fi6, which are designed to minimize end-to-end delays.
• Locate some workloads in regional mini-datacentres (say 100km away) rather than distant hyperscale facilities (1000+ km distant), reducing the speed-of-light and switching/routing delay.
• Use optimized inference models to reduce processing time and potentially fit on the end-device or nearby gateway itself.

However, latency is not the only criterion to optimize for. In this example scenario, the developer’s cloud-compute costs are escalating and they are facing ever more questions from investors and customers about issues of privacy and CO2 footprint. These bring additional trade-offs to the decision process.

Indeed, at a high level, there are numerous technical and practical constraints involved, such as:

• • Reliability and accuracy – does the new product maintain comparable levels of performance to the previous (slow but reliable) version? Does it introduce any new dependencies that could reduce uptime and availability?
  • Power consumption (either for environmental reasons or to preserve battery life)
  • Connectivity – is the network connection predictable (e.g., fiber to a fixed machine in a factory or hospital) or intermittent (for instance, an IoT system on a train going through tunnels and remote areas)?
  • Privacy/data sovereignty – how many organizations (or countries’ governments) can potentially see/intercept the data? Are there GDPR issues involved?
  • Contractual complexity involved with third parties – for instance, dealing with multiple providers of network and compute resource, especially where SLAs and guarantees may not be comparable.
  • Memory and processing resource – either on the device or in edge/cloud nodes. This aspect takes on greater resonance in an era of semiconductor shortages, where obtaining new devices or servers may be difficult.
  • Skills and personnel – can they, or their customers, deploy, maintain and troubleshoot the new platform? Do they need particular permissions or certifications to access specific sites such as cloud datacentres?
  • Single or multiple devices involved – does the application work on standalone sensors, or is the task inherently a multi-device process? For instance, a single camera might monitor a doorway or support a heating engineer’s infra-red detector. Still, a town square or robot’s path through a factory might need dozens of inputs to be combined.

Looking through this list – and also considering all the other AI-related tasks, from audio/speech analysis to big-data trend analysis for digital twins – there is no singular “answer” to the best approach to Edge AI. Instead, it will be heavily use-case-dependent.

The implications of on-device AI and model compression

There are numerous approaches to optimizing AI models, both for server-side compute and on-device optimization. From the previous discussion, it can be seen that if localized inferencing becomes more feasible, then it will likely expand to many use-cases – especially those that can run independently on single, standalone devices. This has possible significant benefits for AI system developers – but also less-favorable implications for cloud and low-latency network providers.

Consider something intensely private, such as a bedside audio analyzer that detects sleep-apnoea, excessive snoring and other breathing disorders. The market for such a product could expand considerably if it came with a guarantee that personal data stayed on-device rather than being analyzed on the cloud. The model could be trained on the cloud, but inferencing could be performed at the Edge. If appropriate, it could communicate results with medical professionals and upload raw data if the user then permitted it later, but local processing would be a good selling point initially.

Disruptive Analysis regularly speaks to representatives of the datacentre and telecoms worlds, especially in connection with new network types such as 5G. There is very little awareness or understanding of the role of on-device compute or AI – or how rapidly it is evolving, with improvements in processor hardware or neural network optimization.

Even in more camera-centric telecoms sectors such as videoconferencing, there seems to be little awareness of a shift back from the cloud to edge (or exactly where that Edge is). There has been some recent awareness of the conflicts between end-to-end encryption and AI-driven tasks such as background blurring and live audio-captioning – but that is just one of the trade-offs that might be shifting.

The medium-term issues that seem to be underestimated are around energy budgets and privacy. If model compression and on-device Edge AI can prove not just “greener” in terms of implied CO2 footprint, but also reduce the invasiveness of mass data-collection in the cloud, then it may be embraced rapidly by many end-user groups. It may also catch the attention of policymakers and regulators, who currently have a very telecom/cloud-centric view of edge computing.

Despite this shift, it is important not to exaggerate the impact on the wider cloud and network market. This changes the calculus for some use-cases (especially real-time analysis of image, video and similar data flows) – but it does not invalidate many of the broader assumptions about future data traffic and value of high-performance networks, either wireless or wired.

When all participants in the market understand each other’s language and technology trajectories, we should hopefully see fewer poor assumptions and less unrealistic hype. There are huge advances occurring across the board – from semiconductors to DNN optimization to network performance. But each alone is not an all-purpose hammer – they are tools in a developers’ toolkit.

Interested in the topic and want to learn more? Watch the archive recording of our panel discussion combining insights from experts in Edge AI, Cloud AI, Telecommunications, Deep Learning, Semiconductor Design and more, to fuel the dialogue around AI sustainability.

Dean Bubley
Founder, Disruptive Analysis

The post Edge AI – Speaking the Same Language appeared first on Edge AI and Vision Alliance.

Toby McClean, Vice President of AIoT Technology and Innovation at ADLINK Technology, presents the “Five Rights of an Edge AI Computer Vision System: Right Data, Right Time, Right Place, Right Decision, Right Action” tutorial at the May 2021 Embedded Vision Summit. Solutions builders and business decision-makers designing edge AI computer…

“The Five Rights of an Edge AI Computer Vision System: Right Data, Right Time, Right Place, Right Decision, Right Action,” a Presentation from ADLINK Technology

Register or sign in to access this content.

Registration is free and takes less than one minute. Click here to register and get full access to the Edge AI and Vision Alliance's valuable content.

The post “The Five Rights of an Edge AI Computer Vision System: Right Data, Right Time, Right Place, Right Decision, Right Action,” a Presentation from ADLINK Technology appeared first on Edge AI and Vision Alliance.

This blog post was originally published at e-con Systems’ website. It is reprinted here with the permission of e-con Systems.

One of the major challenges in surveillance and monitoring systems is inadequate availability of light. In the past, images captured in low lighting conditions were not clearly visible, had detail loss, and contained noise. Today, continuous improvements in sensor technology have given birth to a new generation of cameras that perform exceptionally well in low lighting conditions. This article intends to dive deep into the four factors that determine the low light performance of a camera.

What is low light performance, and the factors affecting it?

Low light performance is the ability of a camera to provide intelligible images even in low light conditions by preserving details and minimizing noise in the output image. Low light camera gives high-quality images even in extremely low lighting conditions as shown in the image below.

The factors that contribute to the low light performance of a camera are:

• Pixel Size
• Signal to Noise Ratio (SNR)
• Lens aperture
• Sensitivity

Let us now look at each of the above factors in detail.

Pixel Size

Pixel size is the physical dimension of a single pixel in an image sensor. Pixels are square wells that collect photons, and pixel size refers to the dimension of this square well. Pixel size in sensors that come with low light performance is typically 3µm or larger. The bigger the well, higher the number of photons collected, and hence better would be the performance of the camera in a low light environment.. This makes pixel size one of the most critical factors considered while choosing a camera for low light applications. However, it is to be noted that with increased pixel size, only fewer pixels can be accommodated in a compact imaging sensor causing either a reduction in resolution or an increase in the sensor size.

Signal to Noise ratio (SNR)

SNR or Signal to Noise Ratio, in simple terms refers to the ratio of the quantity of signal (valid image data) to the noise (invalid image data) in an image. A higher SNR indicates lesser noise. In an imaging sensor, dark current and transistor noise are some of the major noise components. A low-light camera comes with the ability to minimize these noise components to achieve a higher SNR thereby producing good quality images.

Lens aperture

Lens aperture determines the quantity of light that falls on the sensor. This is a design consideration than a specification of an imaging system. To capture images in low light conditions, a larger lens aperture is recommended allowing more photons to reach the sensor. However, increasing the lens aperture reduces the depth-of-field of a camera. Depth of field is a crucial consideration for an application with varying working distances, and hence a careful design is required while building a low light imaging system for such applications.

Sensitivity

Sensitivity is an indicator of the ability of an image sensor to convert incident photon energy to electrical signals. Cameras with high sensitivity allow efficient conversion of low light images to quality outputs, making it an inevitable feature for low light cameras. Responsivity is synonymous with sensitivity, and also indicates the efficiency of an imager in converting photons from a low light scene into detailed images.

Low light camera modules from e-con Systems:

• e-CAM83_CUMI415 – 4K IMX415 Low light Camera Module
• e-CAM31_MI0330_MOD – 3.4 MP Liquid Lens Camera Module
• e-CAM30A_CUMI0330_MOD – 3.4 MP Low Light Camera Module
• e-CAM220_CUMI327_MOD – Full HD IMX327 Lowlight Camera Module
• e-CAM21_CUMI290_MOD – 2MP IMX290 ultra-lowlight Camera Module
• e-CAM27_CUMI290_MOD – 2MP IMX290 ultra-lowlight Camera Module
• e-CAM20_CU0230_MOD – 2MP HDR Camera Module

Bremi Maruthaiyan
Business Development Executive, e-con Systems

The post 4 Major Factors that Determine the Low Light Performance of a Camera appeared first on Edge AI and Vision Alliance.