LETTER FROM THE EDITOR |
 |
|
Dear Colleague, On Tuesday, September 23, 2025 at 9 am PT, the Yole Group will present the free webinar “Infrared Imaging: Technologies, Trends, Opportunities and Forecasts” in partnership with the Edge AI and Vision Alliance. Infrared (IR) cameras are increasingly finding adoption in a diversity of applications, as an adjunct or alternative to visible light cameras. Their ability to detect and measure energy below the visible light region of the electromagnetic spectrum makes them useful as thermal imagers, sensing objects (and their movements) whose temperatures differ from that of the ambient environment. Infrared light can also penetrate materials which visible light cannot, such as dust, fog, smoke, and thin walls. Combining visible light and infrared data results in a richer multispectrum understanding of a scene. The infrared spectrum subdivides into short-wave (SWIR), mid-wave (MWIR) and long-wave (LWIR) regions, each with unique associated sensor implementation technologies. Both uncooled and cooled sensor subsystems, each with unique tradeoffs, contend for market acceptance. This webinar, presented by Axel Clouet, Ph.D., senior market and technology analyst for imaging at the Yole Group, will discuss the history, current status and technology and market trends for the IR imaging ecosystem, including comparisons with alternative imaging approaches. Clouet will cover both high-end applications such as military and surveillance and high-volume opportunities like automotive and consumer electronics, and will share Yole’s latest market forecasts. A question-and-answer session will follow the presentation. For more information and to register, please see the event page. Brian Dipert |
OPTIMIZING EDGE AI DEVELOPMENT AND DEPLOYMENT |
.jpg?width=1120&upscale=true&name=unnamed%20(54).jpg) |
|
Why It’s Critical to Have an Integrated Development Methodology for Edge AI The deployment of neural networks near sensors brings well-known advantages such as lower latency, privacy and reduced overall system cost—but also brings significant challenges that complicate development. These challenges can be addressed effectively by choosing the right solution and design methodology. The low-power FPGAs from Lattice are well poised to enable efficient edge implementation of models, while Lattice’s proven development methodology helps to mitigate the challenges and risks associated with edge model deployment. In this 2025 Embedded Vision Summit presentation, Sreepada Hegade, Director of ML Systems and Software at Lattice Semiconductor, explains the importance of an integrated framework that tightly consolidates different aspects of edge AI development, including training, quantization of networks for edge deployment, integration with sensors and inferencing. He also illustrates how Lattice’s simplified tool flow helps achieve the best trade-off between power, performance and efficiency using low-power FPGAs for edge deployment of various AI workloads. |
.jpg?width=1120&upscale=true&name=unnamed%20(55).jpg) |
|
How to Right-size and Future-proof a Container-first Edge AI Infrastructure In this 2025 Embedded Vision Summit presentation, Carl Moberg, CTO of Avassa, and Zoie Rittling, Business Development Manager at OnLogic, provide practical guidance on overcoming key challenges in deploying AI at the edge, including remotely managing containerized models in resource-constrained environments using scalable, purpose-built infrastructure. They also cover selecting and integrating the right hardware and software for high-performance edge vision systems that bridge the gap between edge inference and cloud management to enable seamless AI operations. In the demanding landscape of embedded vision and edge AI, maximizing performance and efficiency is paramount. Moberg and Rittling explore how to leverage rugged industrial PCs with scalable container orchestration to avoid an over-resourced edge stack that overwhelms with manual operations and lacks operational overview. You’ll learn how to build an edge AI stack that unlocks real-time, efficient and reliable vision-based solutions and discover how a best-of-breed approach enables rapid iteration and future-proofs your edge AI solutions. |
EFFICIENT VOICE AND IMAGE UNDERSTANDING |
.jpg?width=1120&upscale=true&name=unnamed%20(56).jpg) |
|
Voice Interfaces on a Budget: Building Real-time Speech Recognition on Low-cost Hardware In this 2025 Embedded Vision Summit talk, Pete Warden, CEO of Useful Sensors, presents Moonshine, a speech-to-text model that outperforms OpenAI’s Whisper by a factor of five in terms of speed. Leveraging this efficiency, he shows how to build a voice interface on a low-cost, resource-constrained Cortex-A SoC using open-source tools. He also covers how to use voice activity detection as a first step before running speech-to-text to avoid false positives on noise that isn’t speech. In addition, he demonstrates how to use Python to control speech recognition and take actions based on recognized words. The Moonshine model’s compact size (as small as 26 MB) and high accuracy (<5% word error rate) make it ideal for embedded applications. All code and documentation are available online, allowing you to replicate the project. This presentation showcases the potential for voice-enabled interfaces on affordable hardware, enabling a wide range of innovative applications. |
.jpg?width=1120&upscale=true&name=unnamed%20(57).jpg) |
|
Image Tokenization for Distributed Neural Cascades Multimodal LLMs promise to bring exciting new abilities to devices! As we see foundational models become more capable, we see compute requirements grow as well. It is not uncommon to see LLMs grow to tens of billions of parameters, at a rate faster than what embedded processors can provide. In this 2025 Embedded Vision Summit talk, Derek Chow, Software Engineer at Google, and Shang-Hung Lin, Vice President of NPU Technology at VeriSilicon, introduce the concept of a “neural cascade,” a scheme that allows division of computation across devices. They present a recipe for constructing a neural cascade from a pre-existing LLM and they show how this system harmonizes edge and cloud devices to enable new experiences. |
UPCOMING INDUSTRY EVENTS |
|
Infrared Imaging: Technologies, Trends, Opportunities and Forecasts – Yole Group Webinar: September 23, 2025, 9:00 am PT |
FEATURED NEWS |
|
RealSense Completes Spinout from Intel, Raises $50 Million to Accelerate AI-powered Vision for Robotics and Biometrics Basler’s Stereo ace Supplies Precise 3D Images Even with Challenging Surfaces VeriSilicon’s AI-ISP Custom Chip Solution Enables Customers’ Mass Production AMD Acquires Brium to Strengthen Its Open AI Software Ecosystem NVIDIA Unveils NVLink Fusion for Industry Partners to Build Semi-custom AI Infrastructure |
EDGE AI AND VISION PRODUCT OF THE YEAR WINNER SHOWCASE |
.jpg?width=1120&upscale=true&name=unnamed%20(58).jpg) |
|
Airy3D DepthIQ (Best Camera or Sensor) Airy3D’s DepthIQ is the 2025 Edge AI and Vision Product of the Year Award Winner in the Camera or Sensor category. DepthIQ from Airy3D offers a simple and versatile 3D imaging solution that generates near-field depth data using just a single camera. This technology can be applied to a broader range of applications compared to current 3D imaging solutions, all at a fraction of the cost, resource requirements, and power consumption. DepthIQ is built upon a transmissive diffraction mask (TDM) that is applied over a CMOS image sensor, utilizing standard semiconductor technology to facilitate streamlined mass production. TDMs leverage diffraction to measure depth effectively. Because DepthIQ employs a single sensor, it avoids stereoscopic occlusions, making it an excellent choice for short-range applications. Airy3D’s DepthIQ technology integrates a TDM with a proprietary lightweight inline processing method. It produces near-field 3D data that is precisely aligned with the traditional 2D data captured by the sensor, all from a single device. The TDM acts as an optical filter applied on top of any CMOS sensor during the final stage of imager post-production, following the application of color filters and micro lenses. This TDM creates a unique raw data set that inherently combines 2D images with depth information into one device, enhancing reliability and stability over conventional 3D capture methods. Depth information is extracted from these unique raw data sets using proprietary imaging algorithms that demand very low computational power and do not require frame buffering. The TDM stack is placed directly on top of the micro lens layer and is fabricated using conventional semiconductor materials and processes. Since the stack is only a few microns thick, the resulting 3D sensor can easily integrate into the downstream manufacturing process, including module and end-product assembly. This patented solution transforms any CMOS imaging device into a 3D sensor for use in various applications, including Advanced Driver Assistance Systems (ADAS), security, robotics, augmented reality/virtual reality (AR/VR), and the Internet of Things (IoT). Please see here for more information on Airy3D’s DepthIQ. 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. |
