Qualcomm said it plans to begin testing its new Cloud AI 100 chip with partners such as Microsoft Corp later this year, with mass production likely to begin in 2020.
Qualcomm’s new chip is designed for what artificial intelligence researchers call “inference” – the process of using an AI algorithm that has been “trained” with massive amounts of data in order to, for example, translate audio into text-based requests.
Analysts believe chips for speeding up inference will be the largest part of the AI chip market.
Qualcomm announced a working project to apply quantum field theory to deep learning-vision AIs using Qualcomm computing technology.
This is existing Qualcomm computing technology. Quantum field theory now provides a unique way of working on and solving complex problems using massive parallel processing resources.
Qualcomm Technology AI Projects
Qualcomm Technologies, a subsidiary of Qualcomm announced at the International Security Conference & Exposition, broad ecosystem support for the Qualcomm® Vision Intelligence Platforms. Leading OEMs, ODMs and ecosystem collaborators have integrated the platforms to deliver powerful computing for edge and on-device camera processing and machine learning, with exceptional thermal efficiency, across a wide range of IoT applications.
For AI and machine learning uses such as tracking, identification and classification of people, objects, and sounds the Qualcomm Vision Intelligence 400 & 300 Platforms use the latest purpose-built QCS605 and QCS603 SoCs that include:
Integrated advanced image signal processor (ISP) and the Qualcomm® Artificial Intelligence (AI)
Engine Heterogeneous compute architecture, including a cutting-edge ARM-based multicore Qualcomm® Kryo™ CPU, Qualcomm® Hexagon™ DSP with a vector accelerator and Qualcomm® Adreno™ GPU. These chipsets are ideal for running machine learning and other edge compute use cases that require video processing and analytics support, with real-time stitching for simultaneous multi-video streams while running concurrent applications.
The platform series also integrates proven connectivity, superior audio and hardware-based security to support the development of IoT devices with advanced security features and includes:
Android™ and Linux-based software platforms and advanced camera software development kits (SDKs) are optimized to support exciting new possibilities for enterprise, industrial, and consumer applications
“We are really just scratching the surface, and the variety of applications and possibilities are enormous, including smart security cameras, sports cameras, wearable and body cameras, robotics, inventory management, smart displays, dash cams and many more,” continued Torrance.
Qualcomm Vision Intelligence Platforms Momentum
“We have a long working relationship with Qualcomm Technologies, Inc. on the Qualcomm Vision Intelligence Platform. Altek has developed a series of Vision edge AI products including the AI edge camera and AI edge box using the Qualcomm Vision Intelligence Platform to enable AI inference on the edge. These products allow new experiences in different verticals like smart retail, smart cities and smart factories with quick detection response and superior data privacy.” said Alex Hsia, Founder & CEO, Altek.
“The work between AnyVision and Qualcomm Technologies, Inc. has catalyzed a shift in the industry, where a vision will now become a reality, and the smart edge will open the world to a truly intelligent future,” said Max Constant, Chief Commercial Officer at AnyVision. “Our new solution is designed to continue making AI more accessible to the world.”
“The Cisco Meraki MV camera portfolio is based on the idea of the camera becoming a sensor for the enterprise, not just a tool for security. The advanced capabilities of the Qualcomm Vision Intelligence platform allows us to achieve this. By enabling machine learning computer vision capabilities entirely in the camera, Cisco Meraki can drastically reduce the cost and complexity of deploying this technology. In doing so, we see a paradigm shift toward democratizing access to AI, where even the smallest of organizations can benefit,” said George Bentinck, product manager, camera systems, Cisco Meraki.
“We are excited to be working with Qualcomm Technologies, Inc. on powering intelligence at the edge,” said Parag Mehta, Chief Business Development Officer at eInfochips (an Arrow Company). “We have developed reference designs based on the Qualcomm Vision Intelligence Platforms and our Cloud collaborators to accelerate our customer’s production time, utilizing the Qualcomm® Neural Processing SDK for intelligent edge connection with the cloud leveraging machine learning services. The intelligent edge cameras provide real-time on-device inferencing that are designed to help companies develop innovative vision edge analytics for use cases like defect detection, object recognition, consumer behaviour and many more.”
Hitron Systems Inc.
“We are thrilled to work with Qualcomm Technologies, Inc. on our next-generation edge AI solutions, based on the Qualcomm Vision Intelligence Platform featuring the QCS605 that will revolutionize the future of the security camera industry,” said Nick Choi, strategic planning department. Director, Hitron Systems Inc. “We continue to work together to deliver exciting new AI/machine learning based camera products with exciting new user experiences.”
“Qualcomm Technologies, Inc.’s products are helping us meet our goal to deliver innovative 5G on demand cameras to customers all over the globe. AI and deep learning technology for next gen cameras open a variety of new IoT applications in a variety of segments such as smart home, smart shop (smart retail), smart logistics and smart city. We look forward to continuing building our solutions using the Qualcomm Vision Intelligence 400 Platform based on the QCS605 and with the resulting products, hope to create safer cities,” said Myung Ho Yoo, president of Intellivix.
Intrinsyc Technologies Corporation
“We are excited to have utilized the capabilities of the Qualcomm Vision Intelligence Platforms to create production-ready computing solutions that enable OEMs to rapidly build smart cameras with superior image processing and enhance Artificial Intelligence (AI) capabilities,” stated Cliff Morton, vice president, client solutions, Intrinsyc Technologies Corporation. “In addition to two powerful platforms based on the Qualcomm Vision Intelligence Platforms, the Open-Q™ 605 SBC and Open-Q™ 626 SOM, Intrinsyc provides comprehensive design and development services, including expertise in camera development and tuning.”
“Linkflow is transforming the wearable camera industry by offering the FITT360, based on the Qualcomm Vision Intelligence Platform, a radically new design with three cameras to create seamless 360 scenes for consumer as well as enterprise customers,” said Yongkuk Kim, CEO, Linkflow. “The Qualcomm Vision Intelligence Platform offers superior multi sensor stitching of FULL HD camera streams along with robust EIS for smooth 360 capture. The long battery life has allowed us to differentiate our product while providing a solid commercial product with a high-value combination of advanced features and power efficiency.”
“LITE-ON Video Surveillance Strategic Business Unit proudly utilizes the hardware features, software development, and ecosystem relationships delivered by Qualcomm Technologies, Inc.” said by Allen Hsu, general manager. “As dedicated ODM and JDM supplier, our use of the Qualcomm Vision Intelligence Platform based on the APQ8009 and APQ8053 SoCs provide us the latest camera features and technologies to enable LITE-ON to deliver versatile products for the IoT. By offering solutions with superior RF performance and image quality tuning, LITE-ON can support our customers with the difficult challenges of hardware design and allow OEMs to focus on the development of software capabilities and user interface while being able to reduce development time to commercialization.
“Our Vision AI Developer Kit for Microsoft Azure IoT solutions featuring the Qualcomm Vision Intelligence 300 Platform helps customers and developers innovate by unlocking the benefits of AI at the edge,” said Rodney Clark, vice president, IoT Sales, Microsoft Corp. “The device uses the Qualcomm QCS603 SoC for hardware acceleration of the AI model to deliver superior inferencing performance that can be utilized to create new solutions for a variety of scenarios, from industrial safety, retail, home and enterprise security cameras, smart home devices and more.”
“Motorola Solutions uses the Qualcomm Vision Intelligence Platform in our Si500 body worn cameras,” says Scott Mottonen, corporate vice president of devices, Motorola Solutions. “It provides best-in-class video quality for the demanding needs of first responders.”
“The Qualcomm Vision Intelligence Platform 200 provides connectivity for Owlcam’s AI, so video security can go where it’s needed most; cars and trucks. While either driving or parked, the Qualcomm Vision Intelligence Platform’s power efficiency allows Owlcam to be always-on and to deliver real-time video and alerts for vehicle crashes and break-ins,” said Andy Hodge, Founder and CEO, Owlcam, Inc.
“PathPartner is excited to be a part of the ecosystem by providing design services and independent software on the Qualcomm Vision Intelligence platforms”, said Ramkishor Korada, executive vice president of business development and marketing at PathPartner Technology. “With our design services encompassing audio/camera pipeline tuning, on-device artificial intelligence algorithm development and complete system integration coupled with out-of-box algorithms for image processing and computer vision, we help customers develop advanced AI-based camera solutions on Qualcomm Technologies, Inc. platforms for various camera segments such as security, dash cam, action, wearable, home monitoring, industrial, retail and smart doorbell ”
“We’re excited to combine Pilot AI’s deep learning-based solutions and Qualcomm Technologies, Inc. accelerated hardware. Pilot AI’s lightweight software paired with the powerful, yet power efficient, Hexagon DSP is designed to enable a new generation of real-time applications at the edge, including detection, tracking, and identification of people, vehicles, and more,” said Jonathan Su, CEO, Pilot AI.
“We have been collaborating with Qualcomm Technologies, Inc. on multiple fronts. Next-gen cameras and video conferencing systems need a high level of edge compute for computer vision and machine learning based applications as well as high-performance video and ISP technology,” said Casey King, chief technology officer of Poly (NYSE: PLT). “We’re delighted to have worked with Qualcomm Technologies, Inc. on our Polycom Studio product and embrace our collaboration as we continue to build out additional solutions using the Qualcomm Vision Intelligence Platform.”
Qisda Presice Optics
“With the advanced Qualcomm AI engine of QCS605 from Qualcomm Technologies, Inc., Qisda is able to deliver a new surveillance platform for our customers to take advantage of all the benefits of the next-generation video analytics,” said Tony Lin, head of Qisda Presice Optics business unit.
RICOH COMPANY, LTD.
“We are excited to be using the Qualcomm Vision Intelligence Platform 200 with the APQ8053 SoC on our latest 360 camera RICOH THETA Z1, said Shinobu Fujiki, general manager, THETA Division, RICOH COMPANY, LTD. “The advanced technologies have enabled us to offer superior 23MP still, and 4K video 360-degree capture, while allowing a slim form factor optimized to produce unprecedented high performance and high quality at low power consumption.”
Security and Safety Things – a Bosch Start Up
“The security surveillance industry is going through a remarkable transformation driven by artificial intelligence, near-ubiquitous connectivity and the increasing compute power available to devices,” said Nikolas Mangold-Takao, VP Product Management & Marketing, Security and Safety Things – a Bosch Start Up. “We are excited to work with Qualcomm Technologies to meet this trend and bring the latest advancements in camera, machine learning, and connectivity to our ecosystem of partners.”
“We saw a rapidly increasing demand for intelligent cameras, and the Qualcomm Vision Intelligence Platform is an ideal choice for industries such as smart retail, smart logistics, smart buildings and wearables. Through our on-device AI technology and Qualcomm Technologies, Inc. leading platform, Thundercomm is committed to providing world-class AI, IoT and Automotive solutions to our customers worldwide,” said Larry Geng, Chairman, Thundercomm.”
“VIVOTEK, as a globally recognized leading IP surveillance solution provider, aspires to become the Eye in IoT. We are dedicated to edge computing and to developing a new Android-based camera by adopting the Qualcomm Vision Intelligence Platform. This new camera will generate insights and create greater added value for our customers utilizing the advanced AI-driven analytical engine. The first prototype will be displayed in VIVOTEK’s booth (#12089) at ISC West 2019. Together with Qualcomm Technologies, Inc.’s solutions, we are turning a new page in the IP surveillance industry,” said Dr. Steve Ma, vice president of engineering, VIVOTEK Brand Business Group.
SOURCES- Qualcomm, Reuters, eWeek
Written By Brian Wang, Nextbigfuture.com