DPReview article on LOFIC

https://www.dpreview.com/articles/cmos-inventor-eric-fossum-helps-us-explain-lofic/

CMOS inventor Eric Fossum helps us explain LOFIC

 We worked with Prof. Eric Fossum to help explain the latest step forward in sensor design

It’s increasingly likely that you’ll have heard the term LOFIC, and some of the excitement around the capability it brings. We spoke to Professor Eric Fossum, inventor of the modern CMOS image sensor, who turns out to have some involvement in the early development of the concept, to explain how it works and what it could mean for cameras in future.

LOFIC stands for Lateral OverFlow Integration Capacitor and is a different way of designing pixel circuitry that allows them to capture more light. Conceptually, it stems from a similar idea to the dual conversion gain sensors we’ve seen over the past 12 years, but the benefits it promises appear even greater.

Fossum, now a Professor at Dartmouth College, uses the metaphor that capturing light is like collecting rainwater in a series of containers or buckets.

We’re going to slightly complicate the metaphor to encompass the concept of “conversion gain,” which is the relationship between the number of photoelectrons coming from the light-sensitive photodiode in each pixel and the size of the electrical voltage signal that’s then read out (measured in Volts per electron). 

In CMOS sensors, photoelectrons captured in a small bucket (smaller capacitance) give a proportionally bigger signal when read out than the same number would, if captured in a larger bucket (larger capacitance).

So a small bucket is well suited to low light (fewer photoelectrons) but you need a much larger one for capturing a lot of light, so that you don’t overfill the bucket.

In a LOFIC sensor, instead of one, we have both a small and a large bucket. The charge from the photodiode starts to collect in the small bucket, but if it exceeds its capacity, it overflows into the second, larger bucket. (Fossum called this “cascaded integration” in an early patent)

When it comes to readout, we measure the small bucket first. Because the charge was captured in a small bucket (with high conversion gain) we get a large signal that overwhelms any other noise that might get added later in the process. This gives us nice clean shadows and low-light images.

In a LOFIC sensor, we can then connect the two buckets together, combining their capacity (so, lower conversion gain) and adding in any overflowed charge. We then read out this combined signal. If we didn’t capture very much light, we can just use the first readout result. But if the small bucket was full, we can then use the second, combined readout, instead, giving us much more capacity for highlights.

Unlike the dual conversion gain system, Fossum points to the key benefit: “you don’t have to choose in advance.” And, as he points out, there’s no reason this system has to be limited to two buckets.

In practice, this means sensors that can tolerate a lot more light. And its implementation in relatively small, high-resolution sensors, suggests that space constraints aren’t a problem, at least not once you can use stacked sensor fabrication.

In the Osmo 4P we’ve seen DJI claim 17 stops of DR, which is an extraordinary number for a Type 1 sensor. This puts its DR on a par with much larger sensors (though doesn’t mean its image quality will match them). In its implementation, you can see how the additional dynamic range can be used.

The 4P’s camera that uses LOFIC technology is the only one that needs to use the new D-Log2 profile: a profile designed to accommodate more dynamic range. Interestingly, the base ISO of D-Log2 (the setting at which it captures maximum DR) appears to be 1600, suggesting that it’s being exposed to capture two additional stops of highlights compared with the ISO 400 base sensitivity of D-Log profile used for the conventional sensors.

However, the D-Log2 mode also lets you expose this same sensor state all the way down to ISO 100: clipping those additional highlights but using more exposure to capture the remaining tones, and thus boosting tonal quality. 

This gives us an idea of what LOFIC could bring to dedicated cameras: a choice of greater highlight capture or higher quality captures with the same highlight cutoff point as existing sensors.

After nearly a decade of sensors getting faster but not really offering improved image quality, LOFIC technology might be the step forward we’ve been waiting for.

Interestingly, Fossum – who is responsible for inventing the active pixel CMOS image sensor that virtually all modern cameras are based on – invented an early version of this concept, but the successors to the company he’d worked for instead concentrated their efforts on the simpler dual conversion gain system. A team at Tohoku University in Sendai, Japan, independently developed LOFIC as we’re now seeing it, citing Fossum’s Micron patent in their LOFIC patent, but proposing a more modern implementation.
 

ISSCC 2027 Late Breaking Industry News

ISSCC 2027 is introducing a new Late Breaking News session to provide an opportunity for companies to showcase significant product announcements that missed the regular ISSCC submission deadline. This new session will feature four industry papers describing production-grade chips or IP blocks incorporated into products launching in 2027.

Please register your intention to submit by uploading a paper title and a short abstract by October 7, 2026 using the submission site below. The final manuscript is due by December 1, 2026, after which papers will undergo an expedited review process, with acceptance notifications issued on December 9, 2026. This track is limited to four accepted papers and is reserved exclusively for production-grade industry submissions covering products launching in 2027.

The submission site and key dates are summarized below.

Submission Site

https://submissions.mirasmart.com/ISSCC2027


Key Dates

    October 7, 2026 — Intent-to-submit deadline
    December 1, 2026 — Paper submission deadline (camera-ready format)
    December 9, 2026 — Industry committee review and selection; notifications sent for the four accepted papers
    December 14, 2026 — Final accepted papers approved by the ISSCC Editorial Team
    December 14, 2026 — Session summary due for the Advanced Program
     

Image Sensors Asia Program Available Oct 28-29

Link: https://www.image-sensors.com/image-sensors-asia/agenda?EventId=21998 

Day 1 Oct 28, 2026

9:00am
Session 1: Market Trends & Next-Generation CMOS Image Sensors
CMOS Image Sensors: Linking Market Evolution with Teardown and Reverse Engineering Analysis
CMOS image sensors (CIS) are at the core of modern imaging systems, serving a wide range of applications from consumer electronics to automotive, security, and industrial markets. This presentation will begin with an overview of the CIS market, including its current value, key drivers, and a five-year forecast, highlighting the main trends shaping demand across applications. Beyond market analysis, Yole Group combines these insights with teardown and reverse engineering activities to closely track technology evolution and industry changes. By dismantling commercial devices from system level down to the semiconductor level, it is possible to identify key design choices, integration approaches, and process technologies used by leading players. The presentation will illustrate how products are analyzed from full device teardown to cross-sectioning and physical analysis to extract detailed information on sensor architecture, stacking, and integration. This approach provides a practical and up-to-date view of the CIS ecosystem, linking market trends with real product implementation.
Bullets: • CIS market overview and 5-year forecast • Key application and demand trends • Teardown and reverse engineering methodology • From system dismantling to cross-section and physical analysis • Linking market insights with real device analysis
 
Anas Chalak | Market & Technology Analyst – Imaging, Yole Group
9:30am
Asian Suppliers Reshaping the Image Sensor Market: Driving the Next Phase of Growth and Competition
The global image sensor market is undergoing a significant shift, with Asian suppliers playing a central role in this transformation. This presentation examines key market dynamics, technology evolution, and demand across mobile, automotive, and security applications, highlighting how Asian suppliers are strengthening their competitive positioning and driving the next phase of market growth.
 
Jeffrey Mathews | Senior Analyst, TechInsights
10:00am
Networking Coffee Break
10:45am
From Eye to Intelligence — Where Next-Gen Image Sensors Must Go
Image sensors have always been measured by how well they serve the human eye — resolution, color accuracy, low-light performance, and dynamic range that translates to a beautiful photograph or video. That mandate isn't going away. Users still expect stunning capture from every camera they carry. But a new, parallel demand has emerged: the same sensor must now also serve as the front-end for machine intelligence — powering real-time scene understanding, contextual AI, and spatial computing in form factors that are always on and severely constrained. The next generation of image sensors must therefore serve two audiences simultaneously: the human viewer who expects premium image quality, and the AI pipeline that requires temporal consistency, power-efficient readout, and perceptual richness beyond what any display can render. In this talk, we explore how sensors are deployed today across both photography and AI workloads, where the tensions and synergies lie between these dual objectives, and what architectural choices — from pixel design to readout strategy to on-sensor processing — will define sensors that excel at both. The challenge for our industry is no longer choosing between the eye and the algorithm. It's designing silicon that refuses to compromise on either.
 
Harish Venkataraman | Director of Camera and Sensing Architecture and Systems, Meta Inc.
11:15am
Session 2: Breaking the Performance Limits of CMOS Image Sensors
Advancement of LOFIC CMOS Image Sensor Technology
After twenty years since its introduction, LOFIC CMOS image sensor technology has become a foundational architecture for achieving high performance HDR imaging in both scientific and consumer applications. This presentation first reviews the fundamental structure, operating principles, and development history of LOFIC CMOS image sensors. It then highlights recent advancements, including improvements in low light and high illumination sensitivity, SNR enhancement, and integration with wide waveband detection, analog/digital global shutter functionality, and other emerging techniques. Finally, recent application examples employing LOFIC CMOS image sensors will be presented.
 
Rihito Kuroda | Professor, Tohoku University
11:45am
Networking Lunch Break
1:00pm
Colloidal Quantum-dot Based SWIR Image Sensing Technologies
Abstract to be updated
Jiwon Lee | Assistant Professor, POSTECH
1:30pm
Breaking the Boundaries of CMOS Sensors Performances
As imaging applications evolve, the demand for enhanced CMOS Image Sensor (CIS) sensitivity has become a key challenge. From low-light and high-speed capture to depth sensing, today’s passive imaging technologies rely heavily on the quality and capabilities of CMOS image sensors. Maximizing photon collection in both visible and near-infrared wavelengths is now essential to unlock new capabilities and deliver superior image quality — even in the darkest environments.
3L-OPTRONICS – standing for Low-Light Level Optronics – wants to democratize dual day / night vision through the development of a low-cost CMOS technology in both large public applications – smartphones, automotive – and specific applications – space, scientific instrumentation, industrial vision and defence. To address low-light imaging needs, our efforts are directed at incorporating high-performance, near perfect anti-reflective solution: black silicon. Black silicon is a wideband spectral and large angular acceptance anti-reflective technology resulting from random surface texturing. 3L-OPTRONICS integrates this technology on CIS by a patented process highly scalable industrially – from 8” wafers for in-house processing to 12” wafers for licensed foundries – without the need to modify the surrounded electronics and the pixel architecture, for both large and small pixels pitch.
Our latest experimental results on a 10µm pitch CIS optimized for visible imaging show a sensitivity improvement across the entire silicon absorption spectrum, notably by doubling quantum efficiency, with a minimum dark current degradation. One visual method for revealing the added value of black silicon in image sensor is to capture images at low light levels – in accordance with French night-time lighting standards. Through this study, we demonstrate that black silicon pushes the limits of CIS night vision capabilities from night level 3 – 10 to 2 mlux – to level 4 – 2 to 0.7 mlux. Based on these results, our simulations show that black silicon could also represent a breakthrough in efforts to reduce sensor power consumption and increase readout speed.
To push furthermore the night vision capabilities sensor’s limits, hardware processing is combined with in-house contrast enhancement and denoising image processing algorithms powered by AI. On this part, experiments have been made in real night-time conditions – estimated as night level 4 between 2 and 0.7 mlux – and give us the opportunity to compare the performance of software processing on a reference CIS and black silicon CIS. This study shows not only that black silicon pushes the limits of low-light imaging capabilities but also enhances AI performance on the resulting images.
Our research shows that black silicon is a real game changer of CIS performance limits thanks to its impact on the entire acquisition chain. The next step is to integrate it into CIS with smaller pixels and higher resolution to demonstrate its technological compatibility in the smartphone sector.
 
Ludovic Escoubas | CEO, 3L-OPTRONICS
2:00pm
Optics Innovations in Deep Sub-Micron Pixel Scaling and the Paradigm Shift to Meta-Optics
It covers optic tech development history of mobile image sensors into deep sub-micron, and the new introduction of meta-surface technology to overcome the limit of conventional sensor optic technology, and their recent published results, then conclude with a few published result of extreme pixel scaling possibility with meta optics technology as an outlook.
 
Dr. In-Sung Joe | Principal Engineer, Samsung Electronics
2:30pm
Networking Coffee Break
3:15pm
Session 3: Robotics Vision & Industrial Inspection - Building Eyes for Intelligent Machines
Semiconductor Inspection: The Image Sensor Advantage
As demand for greater precision and yield across the semiconductor value chain continues to rise, image sensors have become a critical enabling technology for advanced inspection and metrology systems. In this session, Xinyang Wang, Founder, Chairman and CEO of Gpixel Changchun Microelectronics Inc., will discuss the role of image sensors in semiconductor manufacturing, packaging, testing and PCB inspection, and examine the market opportunity they represent within the broader inspection equipment industry. He will also explore how evolving technologies, including advanced packaging and high-performance computing applications, are driving demand for higher-resolution and next-generation sensing solutions. The presentation will conclude with an assessment of the competitive landscape and Gpixel's positioning and advantages in this fast-growing market.
 
Dr. Xinyang Wang | Founder, Chairman and CEO, Gpixel Group
3:45pm
Topic to be confirmed
Speaker TBC
4:15pm
The Eyes of Physical AI: Why the Next AI Revolution Begins with Infrared Sensors
For decades, image sensors have helped machines see the world. The next era will require them to understand it.
As AI moves from the digital world into the physical world, machines will need more than visible images. They will need sensors that can reveal material properties, chemical signatures, moisture, temperature, hidden structures, and information that the human eye cannot see. Infrared imaging, particularly SWIR sensing, is emerging as a critical interface between AI and the real physical world.
In this presentation, James Lee, CEO of STRATIO, will discuss how infrared sensors and AI can together enable a new generation of Physical AI systems — from robotics, automotive, and industrial inspection to agriculture, recycling, defense, and consumer devices. The presentation will explore why the future of image sensors is not only about higher resolution or lower power, but about giving machines a deeper sensory understanding of the world.
 
Dr. JaeHyung Lee | CEO and Co-Founder , STRATIO, INC.
4:45pm
Panel Discussion: Vision for Intelligent Robotics
Robots are rapidly moving beyond structured industrial environments into warehouses, hospitals, retail stores, homes, and public spaces. To operate safely and autonomously, they require increasingly sophisticated vision systems capable of perceiving, understanding, and interacting with the physical world.
This panel will explore how advances in image sensors, 3D sensing, edge AI, and perception algorithms are enabling the next generation of intelligent robots. Industry leaders will discuss the sensor technologies shaping robotics, the challenges of machine perception in real-world environments, and the future roadmap for robot vision systems.
Key Audience Takeaways
 Technology roadmap for robot vision systems
 Future image sensor requirements for robotics
 Impacts of Robotics vision on image sensors, machine vision, industrial automation, and defense
 Emerging opportunities in industrial, service, and humanoid robotics
 The convergence of sensing, perception, and AI 

Day 2 Oct 29, 2026

9:00am
Session 4: SPAD, ToF and 3D Sensing Technologies
Hybrid Time-of-Flight Image Sensors for High-Image-Resolution LiDARs
Various types of time-of-flight (ToF) imagers, employing either direct ToF (dToF) or indirect ToF (iToF), have recently been developed to perform effectively outdoors under strong ambient light. dToF imagers are well placed for long-range applications but still have difficulty achieving high depth image resolution. In contrast, iToF imagers using short-pulse (SP) illumination are more effective outdoors under strong background light, owing to reduced in-pixel ambient-light charge acquisition shot noise, while providing high range-image resolution. This work presents a hybrid ToF (hToF) CMOS image sensor employing a compact 6-tap and one drain, lateral-drift-photodiode pixel designed with triple n-type doping and that attains a high image resolution of 720x488 pixels while retaining long-range capability, with 100 m outdoor range imaging demonstrated under 100 klux direct sunlight at 10 fps using range shifted multiple subframes. A linearization method for the pixel's linear-logarithmic response expands the dynamic range, enabling accurate depth imaging of highly reflective and retroreflective targets such as road signs. Scanning-illumination techniques greatly improve the signal-to-background noise ratio and enhance the depth resolution or reduce the light power intensity as well as mitigate multi-path problem. These approaches offer a practical path toward high-image-resolution LiDAR systems for several outdoor applications as well as marine-observation and underwater use.
 
Dr. Kamel Mars | Associate Professor, Shizuoka Institute of Science and Technology & Shizuoka University
9:30am
From Perception to Action: Closing the Gap for Physical AI
Physical AI is advancing rapidly, but a fundamental challenge remains: closing the gap between perception and action. Robots can increasingly sense and understand their surroundings, yet the time and uncertainty between seeing, interpreting, deciding, and acting still limit their ability to interact naturally and reliably with the physical world.

Closing this gap requires more than improvements in sensors, compute, or AI models independently. It requires a tighter, continuous feedback loop between sensing and action, where perception informs movement, movement influences what and how the system senses, and new observations immediately refine the next action.

This presentation explores how advances in image sensing, 3D vision, multimodal perception, Vision-Language-Action models, and world models are enabling this transition. It will also examine the critical role of simulation and synthetic data. Real-world human and robotic training data alone can never represent every environment, interaction, and edge case that physical AI will encounter. But simulation is only as useful as its representation of reality. Incorporating vision models grounded in real-world sensor data is critical to reproducing the visual and spatial conditions robots will actually encounter, including sensor characteristics, environmental variability, and challenging edge cases. This connection between real-world sensing and simulation enables more representative training and more meaningful validation before deployment.

By bringing real-world sensing, simulation, intelligence, and action closer together, the industry can move beyond robots that simply understand their environments toward machines capable of interacting with the physical world in real time.
 
Michael Nielsen | Chief Marketing Officer, RealSense
10:00am
Bringing Reliable 3D Perception to Robotics, AR/VR, and Beyond
A compact, fully integrated solid-state LiDAR sensor is increasingly important for applications in robotics, augmented and virtual reality, and agri-tech, where robust depth perception is essential for real-world understanding and interaction. We present a compact time-of-flight 3D LiDAR solution providing up to 2.3k depth zones and strong ranging performance even under high ambient illumination. The sensor is based on a 3D-stacked backside-illuminated SPAD architecture with a 10µm pitch and a 220 × 198 detector array. The pixels are grouped into 2 × 2 macropixels, yielding a native readout resolution of 110 × 96, with an additional test row resulting in a 110 × 98 physical array. The device is wafer-to-wafer stacked with integrated front-end circuitry, a flash time-to-digital converter, and integrating counters, enabling a compact and highly integrated depth-sensing platform. The module also integrates meta-surface optical elements for lensing, VCSEL IR light sources and a VCSEL driver for a truly all-in-one solution.
 
Duncan Hall | Product Marketing Manager, STMicroelectronics
10:30am
Networking Coffee Break
11:15am
Session 5: Breaking the Low-Light Barriers
From Emerging to Mainstream: SPAD Technology Trends and X-FAB's Innovation Platform
Single-Photon Avalanche Diode (SPAD) technology is undergoing a significant transition from enabling specialized applications to becoming a mainstream sensing solution across multiple industries. Advances in CMOS integration, array scaling, and system-level architectures have accelerated the adoption of SPAD-based image sensors well beyond traditional scientific and defense applications. Today, consumer electronics, automotive sensing, industrial inspection, healthcare, and quantum-related technologies are all contributing to the growing demand for high-performance SPAD platforms. As application requirements become increasingly diverse, foundry technologies play a critical role in enabling scalable and manufacturable SPAD solutions. The presentation will introduce X-FAB's SPAD technology platform and highlight recent innovations in device architecture and process integration that enable customers to develop next-generation SPAD image sensors for a wide range of applications.
 
Heming Wei | Technical Marketing Manager, X-FAB
11:45am
Networking Lunch Break
1:00pm
A Novel Starting Material for CMOS Image Sensors with Built-in gain Stage for Low Light Application
In this paper we present a novel concept for a new starting material for CMOS image sensors with superior low light performance. We propose an avalanche gain silicon sensor for image sensors, incorporating a continuous buried gain junction as starting material. The structure separates photon absorption, charge multiplication, and signal collection, enabling internal gain that is largely independent of the pixel architecture. This innovative separation of function supports smaller pixels and thinner absorption regions while improving signal-to-noise ratio and low-light sensitivity.
 
Benoit Dupont | CEO, Etesian Semiconductor
1:30pm
Session 6: Automotive Vision Systems
From Requirements to Reality: Rethinking Performance and Cost in Automotive Image Sensors
Automotive vision systems are evolving rapidly as vehicles transition toward higher levels of intelligence and automation. While demand for improved perception continues to grow, maximizing every image sensor specification no longer guarantees the best system value. The industry is shifting from pursuing peak performance to delivering the right performance—meeting application-specific requirements while optimizing cost, scalability, supply chain resilience, and time-to-market. This keynote explores how the industry is redefining image sensor requirements through scalable platform strategies that balance high dynamic range (HDR), low-light sensitivity, LED flicker mitigation, high speed interface standardization, frame rate, cybersecurity, power consumption, and edge-AI computing against system cost and supply chain resilience. The discussion will examine how common sensor architectures can support multiple vehicle programs while remaining adaptable to future perception algorithms and regulatory requirements. Ultimately, the next generation of automotive image sensors will be judged not by peak specifications, but by their ability to deliver the right performance, at the right cost, for safer and more intelligent mobility.
 
Dr. Hieu Tran | Technical Product Marketing, Ecosystem Platforms and OEMs, onsemi
2:00pm
CMOS Image Sensors and SPAD Direct ToF Sensors for Automotive Applications

    Introduction - Sony Group Corporation’s sales, automotive market trend, and Sony Semiconductor Solutions Corporation’s market share target
    CMOS Image Sensors - Customer expectations, product lineup, and introducing the latest products
    Functional Safety and Cybersecurity - ISO26262 standard, patent examples and product line-up
    SPAD Direct ToF Sensors - Customer expectations, product lineup, and introducing the latest products


Naoki Kawazu | Senior Analog Design Manager of Automotive Development Department, Automotive Business Division, Sony Semiconductor Solutions Corporation
2:30pm
From Embodied AI to Autonomous Driving — How Global Shutter Vision Sensors Empower Every Scene
Large foundation models are transitioning from the digital realm to the physical world, positioning vision sensors as the essential gateway for robots and autonomous vehicles to perceive and understand their surroundings. SmartSens' SmartGS™ global shutter technology, based on a BSI backside‑illuminated pixel architecture, ensures that all pixels are exposed simultaneously, effectively eliminating rolling shutter distortion and motion blur even during rapid movement. This is complemented by Micro‑Pillar™ near‑infrared enhancement for superior low‑light sensitivity, along with Knee Point HDR and InSensor HDR® innovations that deliver exceptional dynamic range in challenging lighting conditions. These capabilities enable robust performance across a wide range of applications—from V‑SLAM navigation, obstacle avoidance, and object manipulation in embodied robots, to driver monitoring, occupant sensing, and traffic perception in autonomous vehicles. SmartSens offers SmartGS™ global shutter sensors covering resolutions from 0.3MP to 14MP, tailored to meet the varied needs of different scenarios. By providing reliable, high‑quality visual data, our solutions empower intelligent systems to advance from mere functional tools to true cognitive partners in every scene.
 
Jessie Zhao | Marketing Manager, SmartSens Technology (Shanghai) Co., Ltd.
3:00pm
Session 7: Mobile Application Updates
From Image Sensor Innovation to Product Value
Smartphone cameras continue to evolve through the integration of advanced image sensors, optics, sensing functions, and AI-based image processing. As a result, it is becoming increasingly difficult to predict the contribution of an image sensor to the final user experience based solely on device-level specifications. At the same time, even promising sensor technologies with the potential to create significant value do not necessarily make their way into products.

In this talk, drawing on my experience in both image sensor development and smartphone camera development, I propose a “deep technical collaboration” in which current challenges and use cases from the smartphone camera side are shared early, together with a “customer-driven pull approach” to overcome these challenges. Using examples such as AF, color reproduction, OIS, and HDR, I will discuss the importance of evaluating the value of sensor technologies at the system level, rather than based solely on sensor performance. Furthermore, I will discuss a practical collaboration framework in which the results of joint evaluation are developed into joint demonstrations and proposals, necessary and sufficient performance targets are jointly defined, and the smartphone maker provides a clear “pull” (explicit demand from the customer side). This approach can reduce uncertainty in suppliers’ development and investment decisions and help drive decisions to initiate development and move new technologies toward productization.
 
Masahiko Nakamizo | Chief engineer, Honor Technologies Japan
3:30pm
Session 8: Camera Systems, Characterization & Integration
Beyond the Datasheet: EMVA 1288 as a Hardware Integration Diagnostic for Application-Grade Camera Systems
In application-grade camera development, a sensor that meets all datasheet specifications on an evaluation kit can exhibit significant, reproducible IQ degradation on a custom PCB and pass every post-ISP functional test undetected. The instinct is to blame ISP tuning. The root cause frequently lies upstream — and standard image quality benchmarking is structurally incapable of finding it. This presentation reports findings from a structured EMVA 1288 Release 4.0 characterisation study conducted on an identical AR0235 sensor across two hardware configurations at matched gain, exposure, and illumination conditions. Despite the sensor being common to both platforms, EMVA analysis revealed a 3.7× difference in dark current variance — 5.95 e⁻/s versus 21.8 e⁻/s — alongside a sign reversal in mean dark current, indicating a platform-dependent black level offset that ISP calibration would silently inherit. Dark-frame covariance analysis further revealed structured, platform-specific row-correlated noise that collapses to near zero at half-saturation — the exposure regime in which standard SNR and noise benchmarks are typically acquired. This exposure dependence is the precise mechanism by which post-ISP metrics can report equivalent performance across two hardware configurations while concealing fundamentally different sensor-level noise behaviours. These findings illustrate a repeatable diagnostic pattern: integration failures that are invisible to post-ISP sharpness and noise evaluation become immediately apparent in raw-frame EMVA measurements, specifically dark current variance, dark-frame covariance structure, and DSNU spatial frequency analysis. We present a practical pre-tuning gate methodology based on these measurements, expressed as platform deltas rather than absolute values, and discuss the conditions under which post-ISP metrics will and will not surface sensor-level integration pathology — including the role of ISP noise reduction kernel scale, black level calibration timing, and exposure-level selection...
 
Tina Agnes Ruth K | Senior Project Engineer - Imaging, e-con systems
 

OpenAI buys Glass Imaging

Several news outlets are reporting that OpenAI has bought computational photography startup Glass Imaging for ~USD 300 million. Glass Imaging develops AI algorithms that improve image quality from small image sensors such as those found on smartphones and mobile devices.

WSJ: https://www.wsj.com/tech/openai-buys-startup-developing-smartphone-camera-63590370 

TechCrunch: https://techcrunch.com/2026/09/14/openai-buys-smartphone-camera-maker-glass-imaging-for-300-million-report-says/

Yahoo Finance: https://finance.yahoo.com/technology/ai/articles/openai-acquires-glass-imaging-camera-134907735.html 

Singular Photonics releases SPAD sensor with in-pixel processing

Singular Photonics launches world-first SPAD-based image sensor

Litavis offers programmable multi-modal capabilities on-chip, accelerating time-to-market for imaging applications.

Edinburgh, UK – September 10, 2026 – Singular Photonics, an innovator in single-photon imaging, today announced the launch of its most advanced image sensor to date. Litavis is the world's first single-photon avalanche diode (SPAD) sensor using in-pixel processing to unify imaging, timing, histogramming and photon statistics on a single chip.

Designed to bring unprecedented flexibility and intelligence to imaging applications, Litavis introduces a software-configurable sensing architecture capable of combining photon-counting imaging, programmable time gating, and advanced photon timing within a scalable SPAD array platform.

Litavis is also the first SPAD-based sensor to enable simultaneous intensity, timing and histogramming modes. The system also enables users to dynamically adapt sensor parameters in software without requiring new hardware designs.

By processing photon events directly on-chip, Litavis significantly reduces the amount of raw data that must be transferred and processed externally. This enables more scalable real-time imaging systems through lower latency, improved power efficiency and faster decision-making.

Litavis supports a broad suite of configurable operating modes, including:

•  Multiple TCSPC and high throughput TDC configurations using internal or external clocks
•  High-dynamic-range photon counting modes with selectable resolution
•  Windowed and coincidence-based detection for advanced temporal analysis
•  Combined timestamping and imaging modes

Developers can reconfigure the same hardware platform for different applications with minimal redesign effort, accelerating development cycles and reducing system complexity. The sensor is suitable for a broad range of applications including machine vision, robotics, physical AI, depth sensing, spectroscopy, scientific imaging, medical imaging, quantum technologies and industrial automation.

“Litavis combines the sensitivity advantages of SPAD technology with the adaptability and scalability increasingly required by modern multimodal imaging,” said Shahida Imani, CEO of Singular Photonics. “It condenses these into one intelligent sensor, enabling a new generation of imaging systems built on a single, configurable platform. We designed Litavis to serve multiple applications and markets, and we've already received significant pre-order volumes from companies and institutions around the world.”
As with Singular’s other products, Litavis is named after an ancient Celtic deity, an earth goddess, whose name means “the broad one” – reflecting the extensive capabilities of the new sensor.

Litavis will be demonstrated at SPIE Sensors + Imaging in Edinburgh next week, and at VISION, the world's largest machine vision tradeshow in Stuttgart in October.

Technical details
Litavis is a CMOS SPAD-based imaging sensor platform built on an architecture that combines single-photon sensitivity with integrated digital photon processing directly on-chip and in-pixel. This enables precise capture of spatial and temporal information, while reducing reliance on external processing hardware, allowing the sensor to extract additional scene information including depth, timing and event characteristics.

Litavis delivers continuous 256 × 256-pixel photon-counting imaging under low-light conditions. In parallel, time-stamped photon events are generated at a lower rate on a 64 × 64 macropixel grid, providing picosecond-resolution photon timing alongside full-resolution photon counting.

The sensor also supports multi-event timing, allowing multiple detected photons per excitation cycle to be individually time-binned, and in-pixel histogramming. These capabilities provide access to temporal statistics beyond conventional intensity measurements. This supports applications such as fluorescence lifetime imaging, dynamic light scattering, and quantum sensing.

Litavis integrates configurable acquisition and timing capabilities directly within the pixel array. This allows developers and system integrators to rapidly prototype, evaluate and deploy advanced imaging solutions using a single adaptable platform.

More coverage of Sony TSMC collab

Link: https://www.techtimes.com/articles/323736/20260810/sony-splits-sensor-manufacturing-tsmc-63b-deal-capture-physical-ai-market.htm

Some excerpts below. 

Sony Splits Sensor Manufacturing With TSMC in $6.3B Deal to Capture Physical AI Market
Joint Venture Keeps Pixel IP at Sony While TSMC Supplies Logic Process for Cars and Robots 

Sony Group and Taiwan Semiconductor Manufacturing Co. announced Monday that they plan to invest approximately ¥1 trillion (approximately $6.3 billion USD) in a joint venture to manufacture next-generation image sensors at Sony's Koshi City fab in Kumamoto Prefecture, with commercial production targeted for 2029, according to the Japan Times. The investment is not an incremental expansion of Sony's existing manufacturing footprint — it is an acknowledgment that the sensors required to power robots, autonomous vehicles, and AI-driven industrial systems operate on fundamentally different technical principles from the smartphone sensors that built Sony's 43% market lead, and that meeting those requirements demands a manufacturing partnership Sony cannot replicate alone.

Why Sony Could No Longer Do This Alone

[...] the sensors required for physical AI applications — cameras inside self-driving cars, robotic manufacturing lines, and industrial inspection systems — are a categorically different product from the sensors inside a smartphone, and building the manufacturing infrastructure to produce them at scale requires both the capital and the process technology that Sony's in-house fabs cannot provide on their own.

[The difference is threefold.]

The first is the global shutter. [...]

The second is dynamic range. [...]

The third is temperature tolerance and functional safety certification. [...]

What the Fab-Light Model Actually Splits

[...] Sony will not share [the pixel layer] manufacturing process with TSMC under the joint venture[...]. Sony retains full control of the pixel wafer.

The second wafer is the logic layer [which will benefit] enormously from advanced logic process technology [from] TSMC's logic process expertise, specifically its ability to manufacture the logic die at nodes Sony cannot match in-house [...].

The two wafers are then bonded using copper-to-copper hybrid bonding — a technique Sony pioneered for commercial image sensor production around 2016 [...].


The Sony-TSMC sensor joint venture supplies the perception layer that physical AI systems require. Whatever a robot or autonomous vehicle is doing with its AI model, it is perceiving the world through an image sensor first. 

Sony's IP Bet and the Competitive Clock

The joint venture is also, implicitly, a response to a competitive clock [...] Apple confirmed in August 2025 that it was working with Samsung at the latter's Austin, Texas facility [...].

Samsung + KAIST nanostructures improve image sensors performance

Open access paper link: https://www.cell.com/iscience/fulltext/S2589-0042(26)02211-X

Yoon et al., "Angle- and polarization-adaptive aperiodic-anisotropic metasurfaces for broadband reflectance suppression" iScience (2026)

Optical reflectance at the air-silicon interface degrades CMOS image sensor (CIS) performance, causing signal loss and image artifacts. Conventional single-layer anti-reflective coatings (1-ARCs) offer CMOS-compatible simplicity, but their isotropic nature and limited spatial tunability make them poorly suited to handle angle- and polarization-dependent characteristics of incident light, particularly where the chief-ray angle varies across the sensor surface. Here, we present an aperiodic-anisotropic metasurface (AAM) composed of subwavelength TiO2 nanodisks with spatially varying geometric anisotropy, enabling broadband, angle- and polarization-resolved impedance matching tailored to local incidence conditions. Its performance was verified through unit-cell-level optimization and further validated by full-area simulations on a 20 × 20 μm2 Si substrate under Gaussian beam illumination, where the AAM achieved ∼1.40% average reflectance across 400–700 nm for both polarizations, outperforming conventional 1-ARC and double-layer anti-reflective coating (2-ARC). This offers a practical solution for CIS and other systems such as LiDAR (light detection and ranging) receivers under spatially varying illumination. 

 

Sony & Mitsubishi Electric collab

https://www.sony-semicon.com/en/news/2026/2026072201.html

Mitsubishi Electric and Sony Semiconductor Solutions Agree to Establish a Joint Venture to Build AI Vision Sensor Solutions for Manufacturing Applications
To Enhance Recognition, Decision-Making and Control in Manufacturing and Contribute to Labor Savings and Unmanned Operations 

Mitsubishi Electric Corporation (“Mitsubishi Electric”) and Sony Semiconductor Solutions Corporation (“Sony”) have entered into a definitive agreement on a strategic partnership aimed at accelerating the automation and advancement of manufacturing equipment and manual operations in the manufacturing industry. Under the partnership, the two companies will establish a newly formed joint venture (“JV”), Advanced Vision Solutions Co., Ltd., which is scheduled to begin operations in October 2026, subject to the receipt of necessary approvals and clearances from relevant authorities.

Going forward, through the JV, the companies will develop and offer vision-sensor*1  solutions leveraging cutting-edge technologies that enable AI-based analysis of visual data directly on image sensors. The solutions will precisely capture conditions and changes across a broad range of manufacturing sites that have traditionally been difficult to visualize, and connect them to decision-making and control. They will support automation and autonomous operation of a wide range of manufacturing equipment, contributing to labor savings and unmanned operations.

This partnership will enable the two companies to integrate Sony’s industry-leading image sensors and edge AI technologies with Mitsubishi Electric’s extensive expertise in FA control systems cultivated over many years. The solutions will precisely capture conditions at manufacturing sites, product quality, equipment operations, and maintenance-related information that were previously difficult to fully understand and utilize. Furthermore, it will efficiently extract the necessary information from the acquired data and connect it to appropriate decision-making and subsequent control, enabling timely and autonomous improvements tailored to equipment and operational conditions with minimal manual intervention. 
In addition to vision data, the solutions will integrate with a variety of data from manufacturing sites, contributing to the development of multimodal production systems capable of detecting changes, early signs and other conditions that cannot be fully identified from a single type of data alone. Through this initiative, the companies aim to drive the deployment of physical AI in the FA domain and provide solutions that can be used without specialized expertise in optics or AI, across broad industries seeking labor savings, autonomous operations, quality improvement, and more advanced maintenance in production environments.

*1 Vision sensor: A high-precision sensor that captures images of products and parts on manufacturing lines using a camera and detects the condition and positional information of target objects in real time. By working in conjunction with factory automation (FA) equipment, it enables automation at manufacturing sites.

“Mitsubishi Electric has long provided a wide range of FA equipment to customers, primarily in the manufacturing sector,” said Takayuki Tsuzuki, Executive Officer (Associate); Group President, Factory Automation Systems at Mitsubishi Electric Corporation. “Today, the need for labor-saving, unmanned systems and autonomous optimization at production sites is greater than ever. Through this partnership with Sony Semiconductor Solutions, we will develop AI vision sensors as a new core component and build a new FA digital solutions business by integrating vision data with our digital platform, ‘Serendie™*2.’ Furthermore, by realizing advanced autonomous control with physical AI and other applications, we aim to provide new value to customers and address social challenges such as labor shortages.”

“Sony has contributed to solving challenges and creating new value across a wide range of industries by accurately capturing real-world information through edge‑AI sensing technologies centered on image sensors,” said Shinji Sashida, President and CEO of Sony Semiconductor Solutions Corporation. “Image sensors, which serve as the ‘eyes’ that generate vision data, play an important role in enabling AI to understand on-site conditions and connect them to decision-making and control. Through this partnership with Mitsubishi Electric, which has deep expertise and a strong customer base in the FA domain, we will bring solutions that leverage our sensing technologies to the market, create value across an even broader range of industries, and further accelerate the evolution of our sensing technologies.”

Company name: Advanced Vision Solutions Co., Ltd.
Business description: Development of AI-powered vision sensors and vision-sensor solutions for factory automation (FA)
Shareholding ratio: Mitsubishi Electric 60%, Sony Semiconductor Solutions 40% 

 

Elio raises $21m for AI sensing

Elio: https://www.eliolabs.com/technology 

News Link: https://www.businesswire.com/news/home/20260723118244/en/Elio-Raises-%2421M-to-Build-Sensing-for-the-AI-Era

SILICON VALLEY, Calif.--(BUSINESS WIRE)--Elio, a company building sensors designed for artificial intelligence rather than the human eye, today announced a $21 million funding round led by Innovation Endeavors and Xora, with participation from Kevin Weil and Scribble VC. Existing investors UpWest and Resolute Ventures, who led the company's previous round, are continuing their support.

Today's sensors capture one fixed view of the world, decided in advance, shaped by what the human eye can see rather than what a machine needs. Elio is changing that: its sensors let AI decide what to capture, in real time, based on what it is trying to figure out - delivering exactly the information needed, when it's needed, instead of a fixed picture to interpret after the fact.

The technology puts computation inside the optics itself. Before light reaches the sensor, dynamic optical layers of micromirrors process it directly - acting more like a neural network than a lens - pulling out signals a conventional lens would flatten and lose. AI learns how the optics behave and corrects them live, reading objects and materials by their physical signature, not just their pixels.

As a result, a single module can do the job of many, and it keeps gaining new abilities after it's already shipped. Sensing starts to behave like software: the same unit gets more capable over time, instead of becoming outdated.

Elio can be used in any context where sensing is required. In microscopy, Elio lets researchers watch living cells respond to a drug over time without staining or fixing the sample, work that today requires destroying the sample for a single datapoint. In semiconductors, Elio enables engineers to see through a chip's stacked layers to catch buried defects without cutting the wafer apart. In robotics, Elio gives machines multiple senses in one module that self-calibrates and switches as the scene changes. In defense, Elio can detect small, fast-moving threats like drones at long range across conditions such as darkness or haze that challenge even the best conventional sensors. Across every application, Elio is expanding the universe of what sensors are meant to do - building dynamic, real-time, label-free methods that capture the complexity of the real world and allow AI to find the patterns that matter.

Elio is led by Founder and CEO Nadav Grossinger and Chief Technology Officer Nitay Romano, who have worked at the intersection of optics and AI for more than 20 years. The two previously co-built Pebbles Interfaces, acquired by Meta, then spent seven years leading the physical-sensing stack behind Meta's AR and VR headsets. Grossinger's earlier ventures include ColoRight, acquired by L'Oréal. Romano, a leading figure in diffractive optics, previously served as Chief Optical Scientist at Holo/Or, the science underpinning Elio's core technology. 

NovoViz releases variable frame rate SPAD sensor

Link: https://novoviz.com/variable-frame-rate-spad-camera/

The NVx5VFR-HW variable frame rate camera was developed for applications requiring high sensitivity and/or frame rate but with reduced output bandwidth. The camera combines the benefits of a SPAD camera, namely the single-photon resolution and fast operating speeds, with those of an event camera—low output data rates. 

Features: 

• 480 x 320 or 256 x 192 SPAD pixels
• Event-based readout
• Adjustable event threshold
• 20ns global shutter
• On-chip pixel binning
• USB 3.0  

Event camera emulation
[...] because there is a single sensor and a single acquisition, the two modalities are inherently synchronized in time and space, without genlocking, cross-sensor calibration, or timestamp drift. The SPAD sensor thus functions as a flexible substrate that can be reprocessed into multiple imaging modalities entirely in software. 

HDR Imaging
[...] photon-counting approach lets each pixel build its own exposure curve, avoiding the clipping and noise floor issues that limit standard CMOS or CCD imaging. 

Instantaneous adaptation to lighting variation
[...] photon-counting approach lets the sensor adapt instantly to drastic lighting shifts, with no exposure recalibration, no motion blur, and no saturation, capturing both the bright and near-dark scenes with equal clarity in the same brief 2ms window. 

Singular Photonics raises US$2.15mn for next-gen SPADs

Singular Photonics Lands Funding to Launch Next-Generation Image Sensors as Commercial Traction Grows

Fabless semiconductor startup appoints former Arm CTO to Board as demand for SPAD-based image sensors accelerates.

Edinburgh, UK – August 19, 2026 – Singular Photonics, a developer of SPAD-based image sensors, today announced it has closed an oversubscribed $2.15 million funding round, led by ACF Investors, with participation from Wren Capital, Cambridge Angels, Scottish Enterprise, Quantum Exponential and Old College Capital. The new capital will accelerate development of a new generation of sensors, expanding the company's product roadmap in response to demand from its growing customer base.
Singular has already doubled its full-year 2025 sales in 2026 to date, driven by customer engagements spanning multiple sectors, including a recent collaboration with instrumentation leader Renishaw. The investment will be used to scale engineering capacity and speed time-to-market for new image sensors capable of advancing new applications in markets ranging from machine vision, industrial automation and physical AI to scientific discovery and medical imaging.
 
“This has been a phenomenal year for Singular Photonics,” said CEO Shahida Imani. “We've already doubled our 2025 sales, we're approaching break-even, and our customers are telling us exactly what they want from future products. This oversubscribed round means we can respond faster – expanding our sensor portfolio and bringing new features to market on our customers’ timelines, not just our own.”
 
“Singular Photonics is an exciting addition to our portfolio, combining world-leading sensor science with genuine commercial promise,” said Tim Mills, Managing Partner at ACF Investors. “The industry is moving from capturing images to generating actionable insights directly from light itself, and Singular has both the technology and the team to lead that shift. We're delighted to back them into their next phase of growth.”
 
“Scottish Enterprise has supported Singular Photonics ambition to develop SPAD-based image sensors over several years,” said Derek Shaw, Director of Company Funding & Investment at Scottish Enterprise. “The Photonics, Quantum, Sensing and Connectivity and Semi-conductors Sector is a key opportunity area for growth for Scotland. Companies such as Singular Photonics can play a vital role in transforming our economy by scaling-up, creating high value jobs and competing internationally.” 
 
“In a strong signal of confidence in Singular’s trajectory, existing shareholders followed the new investors into the round, meeting or exceeding their pre-emption rights,” said Singular Photonics Chairman Pete Hutton. “Combined with the appetite shown by new investors, the round was met with robust demand – a clear vote of confidence in the team, the technology and the commercial momentum the company has built.”

Dipesh Patel joins Board
Singular also announced that Dipesh Patel, the former CTO of global semiconductor giant Arm, has joined the company’s board of directors. During his 25 years at Arm, Dr. Patel held a number of general and technical management positions culminating in his role as CTO where he was responsible for the Research and Digital IT functions.
 
“Singular Photonics is transforming image sensors from passive imaging devices into intelligent sensing platforms,” said Dr. Patel. “SPADs have long been viewed as highly sensitive photon-counting devices, but Singular is showing the world that a far greater opportunity is now emerging, and I’m looking forward to helping the company capitalise on that opportunity.”
 
Powering the SPAD revolution
Every digital camera, industrial vision system and scientific imaging instrument depends on image sensors to capture light and convert it into useful information. Over the decades, these sensors have improved, but they are now approaching the limits of how much information can be captured. The arrival of sensors based on single-photon avalanche diodes (SPADs) is changing that.
Unlike conventional image sensors, which essentially measure the intensity of light, SPAD sensors can precisely detect and time individual photons (the smallest units of light). This opens the door to an entirely new generation of imaging systems that can capture information that conventional cameras simply cannot see.
 
Singular Photonics is pioneering the development of SPAD-based image sensors that also integrate advanced on-chip computation. This allows information to be analysed at the point where light is first detected, making the company’s sensors capable of extracting meaningful information in real time. As sensors become increasingly capable of understanding the information they capture, they can enable smaller, faster and more efficient systems that meet the growing demands of edge computing and AI-driven applications.

Sony TSMC collab

Link: https://www.sony-semicon.com/en/news/2026/2026081101.html

Sony Semiconductor Solutions and TSMC Agreed to Establish Joint Venture for Next-Generation Image Sensors

Atsugi, Japan and HSINCHU, Taiwan, R.O.C., August 11, 2026 - Sony Semiconductor Solutions Corporation (“Sony”) and Taiwan Semiconductor Manufacturing Company Limited (“TSMC”) (TWSE: 2330, NYSE: TSM) today announced the execution of a legally binding definitive agreement for the establishment of Advanced Vision Semiconductor Manufacturing Corporation as a joint venture (“JV”) in Koshi City, Kumamoto Prefecture, Japan. The JV will serve as a core hub for the development and manufacturing activities required for the volume production of image sensors for smartphones utilizing advanced manufacturing process technology. The JV is expected to commence volume production in 2029.

Sony will be the sole controlling shareholder and the JV is planned to operate as a consolidated subsidiary of Sony Group Corporation. The Representative Director of the JV is planned to be appointed by Sony.

Pursuant to the definitive agreement, Sony plans to contribute approximately 465 billion yen to the JV through a combination of cash contributions and the transfer of its assets* through a company split. TSMC plans to make cash contributions of approximately 282 billion yen. These capital contributions are expected to be made in phases based on market demand and other relevant business conditions. In addition to these capital contributions by Sony and TSMC, investments required to realize the production capacity planned by the JV are being considered on the premise that they would receive support from the Japanese government.

Under the strategic partnership, Sony intends to take a leading role in the development of core image sensor technologies, as well as product planning and product design. The JV plans to leverage TSMC’s advanced process technology and manufacturing expertise to drive development and manufacturing activities required for the volume production. Through this collaboration, the companies intend to accelerate the commercialization of image sensor products driven by customer needs while advancing technological innovation and strengthening manufacturing capacity.

The establishment of the JV and completion of the transaction are subject to the receipt of required regulatory approvals and the satisfaction of other customary closing conditions. 

All other aspects of the partnership remain unchanged from the announcement by Sony and TSMC titled “Sony Semiconductor Solutions and TSMC Enter Preliminary Agreement for Next-Generation Image Sensor Strategic Partnership” issued on May 8, 2026. 

IISS 2027 Awards Nominations Open

Call for 2027 IISS award nominations is now open. 

IISS AWARD NOMINATIONS:

    Walter Kosonocky Award: call for nomination and nomination form (Due Jan 15, 2027)
    IISS Awards: call for nomination and nomination form (Due Nov 30, 2026)

IEEE 4001 Standard announced

See also IEEE blogpost: https://standards.ieee.org/beyond-standards/how-good-is-the-camera-ieee-4001-gives-answers/ 

How good is your camera, actually? This question is addressed by the new IEEE 4001 standard, developed by a large group of people from the hyperspectral imaging community, recognizing that an improvement over current practice in this field is needed. This standard defines a set of performance characteristics that are directly comparable between cameras. It can be seen as an analog to the EMVA 1288 for image sensors, by defining a similarly comprehensive specification for complete cameras. To serve this purpose, some new characteristics have been developed for IEEE 4001. The standard also defines camera-related metadata that quantify the effects imprinted by the camera on an image. The formal scope of the standard is hyperspectral imaging, but it can be readily adapted to other imaging modalities where it also has potential to be useful. The standard can be downloaded from the IEEE XPLORE site (for free with library access, or through a $133 paywall). Here is an excerpt from the IEEE press release:


Comprehensive, yet minimal, camera characteristics

IEEE 4001 incorporates many conventional camera characteristics, but also some newly developed performance metrics, in order to arrive at a minimal (necessary and sufficient) set that includes parameters of a complete signal model, as well as measures of actual resolution and various nonidealities. Many aspects of camera performance are covered:

Spectral: Spectral range, Number of bands, Spectral peak width, Spectral sampling interval, Equivalent Band Count

Spatial: Pixel format, Field of view, Geometrical pixel field of view, Peak width, Equivalent Pixel Count, Effective spatial fill factor, Spatial distortion, Object distance

Radiometric: Effective throughput, Response nonuniformity, Read noise, Normalized read noise, Dark current, NESR at the noise floor, Saturation capacity, Linearity error, Signal to noise ratio, Dynamic range, Dark correction

Temporal: Maximum frame rate, Integration time range, Synchronization between bands, Dead time

Coregistration: Spatial coregistration error, Spectral coregistration error

Stray light: Stray light apparent reflectance

Data output: Output data level, Real time capability, Synchronization capability, Invalid data flagging

Calibration: Statement of radiometric, spectral, and geometrical accuracy

Other characteristics: Bad pixels, Polarization dependence, Detector type/temperature

Metadata enhancing image exploitation

IEEE 4001 also defines camera-related metadata describing the influence of camera properties on a given image. With IEEE 4001-compliant metadata, it is for example possible to estimate the signal-dependent physical noise amplitude in all pixels and all bands, and to estimate the local magnitude of artifacts due to coregistration error. This can enable more accurate extraction of information from imagery.

 

 

ST releases dToF sensor with 54x42 zones and 100fps

STMicroelectronics unveils new compact direct Time-of-Flight 3D LiDAR module bringing high-resolution spatial awareness to compact edge AI systems with industry-leading resolution and performance

VL53L9 is the first direct Time-of-Flight (dToF) 3D LiDAR all-in-one module in ST’s portfolio, offering a resolution of 2.3K zones, wide field of view, on-chip processing, 100 frames per second, and sensing range from 5 centimeters to 9 meters.

Designed to meet the evolving needs of customers and partners across diverse industries, including robotics, industrial automation, smart buildings, AR/VR, and healthcare.

STMicroelectronics (NYSE: STM), a global semiconductor leader serving customers across the spectrum of electronics applications, today [June 23, 2026] announced the launch of the VL53L9, a compact direct Time-of-Flight 3D LiDAR all-in-one module that sets a new benchmark in high-resolution sensing. The VL53L9 combines state-of-the-art features in a compact and cost-effective package, delivering AI-ready output data for low-compute edge AI systems on small microcontrollers (MCUs) and high-performance sensing across a wide range of applications across robotics, industrial automation, smart buildings, AR/VR, and healthcare.

“VL53L9 demonstrates how far Time-of-Flight sensing has evolved, combining high-resolution depth data, up to 100 frames per second, and a fully integrated architecture in a single compact module. By simplifying integration and reducing system complexity, we enable customers to accelerate the development of applications such as robotics, smart infrastructure, and healthcare monitoring,” said Alexandre Balmefrezol, Executive Vice President and General Manager of the Imaging Sub-Group at STMicroelectronics. “This launch reflects our strategy to move beyond standalone sensors and deliver integrated sensing systems that support real-world edge AI.”

“3D sensing demand accelerates across robotics, industrial automation, XR, and intelligent consumer devices. Time-of-Flight technology is expanding beyond smartphones into applications requiring compact, affordable, and precise depth perception, from navigation and people monitoring to gesture recognition and safety. Higher resolution multizone dToF modules are now emerging as key enablers for this next wave of 3D sensing adoption,” said Anas Chalak, Market & Technology Analyst at Yole Group.

ST FlightSense™ VL53L9 is designed for multiple industry use cases:

– Robotics: enhanced small object detection, SLAM (simultaneous localization and mapping) and obstacle avoidance for autonomous navigation.

– Industrial automation: accurate volume measurement in tanks and bins, improving operational efficiency and inventory management.

– Smart buildings and homes: reliable human presence detection and people counting while preserving user privacy.

– AR/VR and consumer electronics: advanced gesture recognition, body tracking and finger skeleton for immersive user experiences.

– Healthcare: fall detection and monitoring solutions for eldercare and patient safety.

Enhancing 3D sensing with precision and efficiency

The VL53L9 offers an unprecedented 2,268 resolution zones (54×42) with a wide 54°x42° field of view, enabling detailed 3D depth mapping and precise detection of small objects, contours, and edges. Leveraging ST’s proprietary stacked BSI SPAD sensor technology and innovative metasurface optical elements (MOE), the module delivers fast and accurate ranging from less than 5 cm up to 9 meters with up to 1% accuracy and a frame rate of 100 frames per second.

All-in-one sensing data for edge AI and easy integration

The VL53L9’s dual-scan flood illumination replaces traditional dot scanning, reducing motion artifacts, eliminating dead zones, improving small-object detection, and capturing complementary 2D infrared and 3D depth images. Compared to competition, this greatly simplifies post-processing and enables a broad range of edge AI use cases to run efficiently on small MCUs with low compute requirements. The all-in-one module further integrates on-chip dToF processing, a dedicated power management IC, and is fully calibration-free, simplifying integration and reducing system cost and complexity.

Compact form factor

Measuring just 12.8 mm x 6.1 mm x 4.6 mm, the VL53L9 is a reflowable, single-component module compatible with a wide range of cover glass materials. It supports dual power supply operation (1.2 V and 3.3 V) and outputs data via MIPI or I3C interfaces, ensuring compatibility with diverse CPU architectures. The module is certified as Class 1 laser safe, providing reliable and secure operation for end users.

Availability

ST FlightSense™ VL53L9 in mass production in early July 2026 with samples and volume shipments available to customers globally. For more information, visit the VL53L9 product page: https://www.st.com/en/imaging-and-photonics-solutions/vl53l9cx.html 

Single Photon Workshop 2026 materials available

Abstracts from the recently concluded SPW2026 in Naples, Italy are now available online:

https://spw2026.org/programme

https://spw2026.org/poster-session-1

https://spw2026.org/poster-session-2 

indie / ams OSRAM deal

Link: https://www.businesswire.com/news/home/20260511959181/en/indie-to-Acquire-CMOS-Image-Sensor-Product-Line-from-ams-OSRAM

indie to Acquire CMOS Image Sensor Product Line from ams OSRAM 

ALISO VIEJO, Calif.--(BUSINESS WIRE)--indie Semiconductor (Nasdaq: INDI), an automotive solutions innovator, has announced the signing of a definitive agreement to acquire the fabless CMOS image sensor group from ams OSRAM AG for a total consideration of 40 million euros.

With primary operations in Belgium and Portugal, this product line includes intelligent, high‑performance CMOS image sensors for a broad range of industrial, automation, and physical artificial intelligence (AI) applications. This portfolio of products, IP, and designs aligns with indie’s automotive ADAS sensing solutions and further strengthens the Company’s multimodal sensing capabilities across radar, vision, LiDAR, and ultrasonic.

“By integrating ams’ CMOS imagers with our sensor‑fusion hardware and perception software, we’re able to deliver unparalleled sensing systems for next‑generation autonomous machines, including emerging applications such as humanoid robots, cobots, and AMRs,” said Mark Tyndall, executive vice president of corporate development and investor relations at indie. “This unique carve-out extends our position in sensor‑fusion technology and significantly expands our portfolio of GaN SLED light‑source solutions. Together, these technologies broaden our offerings, opens new customer opportunities, and positions indie to capture a larger share of the rapidly emerging physical AI market.”

Image sensors are a key component of sensor-rich platforms within high-performance visual applications such as humanoids, cobots, and industrial automation. As noted by Research and Markets, the image sensor market is forecasted to grow to over $40 billion by 2030. This market growth is being driven by rising autonomy, safety regulations, and increased adoption of AI-based vision ADAS systems, as well as applications that require multimodal sensing, low-latency, and high-resolution.

The consideration includes a cash payment of 35 million euros paid at closing, and a 5 million euros vendor debt note provided by ams OSRAM. The transaction is subject to customary closing conditions, including regulatory approvals. It is expected to close in the third quarter of 2026 and be immediately accretive.  

Recent earthquake disrupts Sony's image sensor fab

Yahoo news July 28, 2026 

https://www.yahoo.com/news/world/articles/sony-fujifilm-plants-evacuate-7-160448320.html

Sony, Fujifilm plants evacuate as 7.1-magnitude earthquake hits Japan's image sensor hub 

Sony press release July 31, 2026
 
https://www.sony-semicon.com/en/info/2026/2026073101.html

Status of Sony Semiconductor Solutions Group Manufacturing Operations Affected by 2026 Kumamoto Earthquakes (Second Update) 

The current status of the impact on our operations from the earthquake that occurred in the Kumamoto region on July 28, 2026, is as follows:

Sony Semiconductor Manufacturing Corporation's Kumamoto Technology Center (located in Kikuyo-machi, Kumamoto Prefecture), which temporarily was halted after the earthquake, has completed safety inspections of its buildings and facilities, and has been carrying out recovery work. As a result, the site is scheduled to resume operations in stages beginning August 4, 2026. The Kumamoto Technology Center is expected to return to its pre-earthquake operating level by mid-August 2026.

Further updates will be provided promptly should any new information become available.

Digital Camera World July 31, 2026 

https://www.digitalcameraworld.com/cameras/an-earthquake-has-halted-production-at-the-worlds-largest-camera-sensor-hub-sony-confirms-could-camera-shipments-be-delayed

The largest image sensor plant from Sony – the world’s biggest manufacturer of image sensors – has halted production after an earthquake hit Japan’s Kumamoto Prefecture on July 28. As of July 31, production still has not resumed, Sony said during an earnings report. 

Sony and Samsung 200MP sensors compared

 

 

200MP camera sensors are everywhere on modern smartphones, but are all 200MP cameras the same?

In this video, we explain why smartphone manufacturers keep adopting 200MP sensors, then run blind camera tests featuring Samsung HP3, HP5, HP9 series, Sony LYTIA 901 and SmartSens SCC80XS. We compare official specifications with real-world daylight and low-light lab samples to reveal how sensor hardware, ISP tuning and optical modules affect image quality.

A larger sensor does not always produce better photos. Final imaging performance depends on the entire camera system instead of sensor specs alone. Let’s check the actual gaps between these mainstream 200MP sensors.

Video on CMOS image sensor basics

iDS published a YouTube video on CMOS image sensor basics:

 

This video provides a comprehensive introduction to modern CMOS sensor technology and explains how hardware innovations have transformed image quality, efficiency, and performance. You’ll learn why pixel sizes continue to shrink, how this impacts imaging results, and which techniques manufacturers use to compensate for these limitations.

The presentation covers key advancements such as microlenses, backside illumination, deep trench isolation, wafer stacking, and smart features like event‑based imaging and sensor‑level AI (Artifcial Intelligence). You’ll also gain insights into image pipelines, FPGA (Field  Programmable Gate Array) versus ISP (Internet Service Provider) (Image Signal Processor) processing, and why system‑level design decisions matter for real‑world machine vision applications. This is essential knowledge for anyone evaluating sensor performance or designing imaging systems that require reliability, speed, and high‑quality results.

Topics covered

    Trends driving pixel size reduction and sensor scaling
    Hardware innovations such as BSI, micro lenses, and deep trench isolation
    Improvements in quantum efficiency and dynamic range
    Image pipelines with FPGA, ISP, and host‑based processing
    Smart sensor features including event‑based imaging and AI at the edge
    Cost, yield, and design considerations for modern CMOS sensors

eyeo raises €40M series A

[News from May 2026]

https://eyeo-imaging.com/knowledge/news/eyeo-raises-e40-million-to-fix-the-flaw-that-has-kept-every-camera-70-blind/ 

eyeo raises €40 million to fix the flaw that has kept every camera 70% blind

 Eindhoven (Netherlands), May 11, 2026 – eyeo today announced it has raised €40 million in Series A funding, bringing eyeo’s total funding to €55 million. The round is led by Innovation Industries, with participation from existing investors imec.xpand, Invest-NL Deep Tech Fund, QBIC fund, High-Tech Gründerfonds (HTGF) and Brabant Development Agency (BOM).

eyeo has developed the world’s most advanced nanophotonic color-splitting technology, redefining imaging for consumer, industrial, XR, smart city and mobile applications with color-splitting photonics technology that triples light sensitivity and breaks sensor resolution limits, delivering unprecedented picture quality, color accuracy and resolution.

The fresh funding will accelerate eyeo’s sensor design capabilities, deepen its OEM partner ecosystem, and drive commercial deployment across a $30 billion global imaging market.

A new era for imaging

All seven billion image sensors sold each year, across smartphones, autonomous vehicles, XR devices, industrial machines, smart cities and more, share a fundamental flaw: color filters work by rejection, blocking red, green and blue light in turn and discarding 70% of the light in the process.

eyeo’s Nanophotonic Color Splitting (NCOS®) technology platform works differently. Instead of filtering light, it splits it: separating light into its colors and guiding every photon directly to the pixel where it belongs. Backed by 26 patents, the NCOS® platform triples light sensitivity and breaks resolution limits. This breakthrough approach unlocks picture quality, color accuracy, resolution, and cost efficiency, which was never-before possible.

Compatible with existing CMOS sensor platforms, eyeo’s technology unlocks sub-micron pixels with 3x the light for ultra-compact, high-performance imaging, removing the trade-off between image quality and sensor size that has constrained product design for decades.

From breakthrough science to production-ready sensors

This Series A round marks a decisive step forward in eyeo’s commercialization journey. Building on the foundation established and proven at imec over the last seven years, multiple tier one customer engagements, and successful process integration at a commercial foundry, eyeo is now investing in the full in-house engineering capability required to deliver sensors to customers at volume scale. The funding will be deployed across three strategic priorities:

•     Building a world-class IC and system architecture design team at the Antwerp sensor design center opened earlier this year
•     Developing and releasing next-generation color-splitting image sensors in 3D-stacked CMOS technology
•     Growing sensor sales and commercial partnerships with OEM customers across smart city, industrial, XR and mobile markets worldwide

 

Videos: Galaxycore, Emberion

 

 
Transforming Infrared Imaging with Emberion’s Cutting-Edge Sensors 

 

 

 Galaxycore: What Is LOFIC? The HDR Technology Behind Advanced Image Sensors  

 

Prophesee raises €20 Million

Link: https://www.prophesee.ai/2026/06/15/prophesee-launches-mantara-event-based-drone-detection/

Prophesee raises €20 million and launches Mantara, the first fully integrated drone detection system built on event-based vision and AI 

PARIS, June 15, 2026 – As Eurosatory and Vivatech open their doors this week in Paris, Prophesee – inventor and world leader in event-based perception – announces the launch of Mantara®: the first integrated system built natively event-based for drone detection and tracking. Mantara runs on Hearth®, a new software platform with built-in AI processing.

Field-validated in June 2026, Mantara is powered by Hearth, the platform succeeding OpenEB and the Metavision SDK. These announcements coincide with a €20 million capital raise led by French fund Critical Path Ventures, with participation from existing shareholders. Prophesee is now held by a substantial majority of French shareholders.

Mantara is a camera whose performance rests on integrated sensors that do not capture images but motion and events. Each pixel responds independently, in microseconds, the instant movement occurs. Fast, erratic targets that blur or vanish between frames on a conventional camera remain perfectly visible and fully characterizable by Mantara: in low light, in backlit conditions, against cluttered backgrounds. Low latency is the heart of its edge: in the time a conventional system takes to capture, transfer, and process a single frame, Mantara has already detected a drone, characterized it, and triggered a response or initiated tracking.

This capability is rooted in a bio-mimetic design inspired by the human visual system. The eye does not record the world in full. Instead, it captures sparse, essential signals, and the brain reconstructs and interprets what matters. If Mantara is the eye, Hearth is the brain.

Hearth is built with cybersecurity requirements, GDPR compliance, and AI governance standards baked in. It enables sensor fusion and will ensure backward compatibility across successive generations of Prophesee sensors. Applications built on Metavision and OpenEB will migrate to Hearth seamlessly.

Recent conflicts have put drones at the center of modern warfare, and the threat is moving fast beyond the battlefield: airports, power plants, critical infrastructure, public gatherings. Drones fly low, change direction sharply, and increasingly carry a near-zero signature: minimal electromagnetic emissions, little acoustic output, little thermal trace. They arrive faster, and in greater numbers, with coordinated swarms capable of overwhelming defenses built to track a single target.

Mantara emits nothing. It cannot be detected or jammed. And it learns: Hearth delivers regular updates that allow every deployed unit to adapt to shifts in threat behavior. Ukraine has made this clear: tactics evolve far faster than traditional development cycles can follow.
 

Civil and Military Missions

Mantara serves both civil and defense use cases. Airports, stadiums, energy and industrial sites, prisons, and ports can integrate it into existing security infrastructure for continuous, automated airspace monitoring. Governments and defense organizations have access to the same system for force protection and counter-UAS operations.

For civil operators, Mantara also offers a compelling privacy advantage: because the sensors detect motion events rather than recording images, Mantara can track moving objects without ever producing footage that could identify individuals. This is a decisive asset for deployment over public spaces and for GDPR compliance.

Hearth: the Reference Platform for the Event-based Ecosystem

Mantara is part of a broader strategic shift encompassing Prophesee’s entire product line.

Hearth will give the event-based ecosystem a trusted reference environment for building scalable, production-grade modules and applications. It will manage sensor updates, backward compatibility, and unlock capabilities that were previously out of reach for developers and integrators working with Prophesee’s patented technology.

As part of this transition, Prophesee announces the end of life of OpenEB, its open-source framework, and of the standalone Metavision SDK. Hearth is the path forward for the developer community. Prophesee will support the migration.

Prophesee’s partners and customers have built a wide range of applications – from eye trackers and ADAS (Advanced Driver Assistance Systems) to ultra-fast cameras for smartphones, ping-pong playing robots, industrial inspection systems, and drone and satellite tracking and targeting solutions. Event-based vision is reaching an inflection point. The market is maturing rapidly and demanding solutions that are robust, secure, and production-ready. With around a hundred patents and multiple sensors to its name, Prophesee has industrialized event-based technology and now intends to structure the broader application ecosystem by making Hearth widely available.

Built with a Network of European Industrial Partners

Mantara benefits from Prophesee’s partnership with IDS Imaging Development Systems GmbH, a leading industrial camera manufacturer. Building on the commercial success of the IDS uEye EVS camera line – based on Prophesee sensors – the two companies expanded their partnership in March 2026 with an agreement signed at Embedded World in Nuremberg to co-develop next-generation industrial vision systems.

Prophesee and Exosens have entered an MoU, paving the way for a long-term partnership to extend bio-mimetic technology beyond the visible spectrum, combining Prophesee’s event-based vision with Exosens’ detection and imaging technologies.

A €20 Million Raise to Anchor Prophesee in France

Prophesee announces a €20 million capital raise led by Critical Path Ventures, a new investor. The team has been reinforced with talent from Google, Microsoft, BCG, Thales, and companies working in drone, robotics, and quantum technologies, as well as from the armed forces. The funds will support the commercial scale-up of Mantara, the development of embedded intelligence within sensors, and the launch of Hearth. 

UV SPAD review article

"Ultraviolet Photon Detection in CMOS Single-Photon Avalanche Diodes: A Review" by Wang et al. in IEEE Sensors Journal (vol 26 no 12 June 2026)

DOI: 10.1109/JSEN.2026.3690120 

Abstract
Ultraviolet (UV) imaging is enabling a growing range of applications in industrial inspection, biomedicine, and aerospace. Single-photon avalanche diodes (SPADs), known for their exceptional sensitivity and fast timing resolution, have achieved remarkable success in visible and near-infrared (NIR) applications, e.g., in light detection and ranging (LiDAR). However, there remains a significant gap
in the development and deployment of UV SPAD detectors. This review presents the latest progress, challenges, and application potential of SPAD technologies in deep-UV (DUV) and near-UV (NUV) ranges. It reviews the challenges in UV detection and the currently reported NUV SPADs, including wide bandgap semiconductors (WBSs), avalanche photodiodes (APDs), silicon-based SPADs with front-side illumination (FSI) and back-side illumination (BSI), and silicon photomultipliers (SiPMs). By linking UV
imaging with single-photon detection, this review aims to support future silicon-based complementary metal-oxide-semiconductor (CMOS) high-speed, large-format UV SPAD arrays, and their integration into next-generation imaging systems. 

 

 

 

HESAI color lidar

Press release (April 2026): https://www.hesaitech.com/hesai-unveils-picasso-6d-full-color-spad-soc-next-gen-etx-and-innovations-in-spatial-intelligence-and-physical-ai/

Hesai Unveils Picasso 6D Full-Color SPAD-SoC, Next-Gen ETX, and Innovations in Spatial Intelligence and Physical AI

Hesai Technology (Nasdaq: HSAI; HKEX: 2525), a global leader in intelligent technology and 3D perception, hosted its 2026 Technology Open Day on April 17, showcasing its latest generation of lidar technologies spanning full-color perception, ultra-high-resolution imaging, and spatial intelligence applications.

A key highlight was Picasso, the world’s first 6D full-color ultra-sensitive lidar SPAD-SoC, enabling native fusion of spatial and color perception at the ASIC level. Built on this platform, Hesai’s ETX lidar series has been upgraded to support up to 4320 channels and is expected to enter mass production in the second half of 2026. The company also previewed Kosmo, a new spatial intelligence AI hardware product, and outlined a new direction in robotic actuation modules for physical AI systems. 

At the core of today’s announcement is Picasso, Hesai’s fifth-generation 6D full-color ultra-sensitive lidar ASIC platform, designed to enable vehicles to see both geometry and color natively on a single chip.

Unlike conventional lidar systems that capture only geometric information such as XYZ, Picasso integrates RGB sensing and time-of-flight (ToF) ranging at the chip-level, enabling the direct generation of colorized point clouds without the need for complex sensor fusion. This 6D full-color capability delivers precise spatial and temporal alignment, resulting in richer environmental understanding and significantly enhancing the perception capabilities of next-generation autonomous systems. 

With photon detection efficiency (PDE) exceeding 40%, Picasso delivers industry-leading sensitivity, enabling stronger long-range performance, better detection of small objects, and improved performance in low-light environments within the same power envelope.

Rather than simply combining camera imagery with lidar point clouds, Picasso enables true ASIC-level fusion, with each point inherently carrying color information. This allows autonomous systems to move from spatial perception to real-time understanding of the driving environment, improving performance in different scenarios such as traffic signals, lane markings, and construction zones. 

The introduction of Picasso marks a major step in lidar development, moving beyond incremental improvements in resolution toward a new generation of full-color perception. This breakthrough is supported by Hesai’s long-term vertical integration strategy and in-house development across core lidar technologies. Following the launch of Fermi C500, a RISC-V-based high-performance lidar controller SoC that integrates MCU, FPGA functions, and ADC into a single chip, the company has established full-stack in-house ASIC development capabilities. This technology foundation has already translated into large-scale commercialization, with Hesai’s upgraded ATX lidar achieving full-scale mass production and securing over 6 million units in orders from leading OEMs.

 

ETX is expected to enter mass production in the second half of 2026, with planned deployment across multiple flagship vehicle models, supporting the large-scale adoption of L3 autonomous driving systems. 

Sony TSMC partnership

Link: https://www.sony-semicon.com/en/news/2026/2026050801.html

Sony Semiconductor Solutions and TSMC Enter Preliminary Agreement for Next-Generation Image Sensor Strategic Partnership 

Atsugi, Japan — Sony Semiconductor Solutions Corporation (“Sony”) and Taiwan Semiconductor Manufacturing Company Limited (“TSMC”) (TWSE: 2330, NYSE: TSM) today announced the signing of a non-binding memorandum of understanding (“MOU”) to form a strategic partnership for the development and manufacturing of next-generation image sensors.

Under the proposed partnership, Sony and TSMC intend to establish a joint venture (“JV”), with Sony being the majority and controlling shareholder,to set up development and production lines in Sony’s newly constructed fab in Koshi City, Kumamoto Prefecture. Through the JV, both companies expect to leverage Sony's expertise in sensor design alongside TSMC’s strengths in process technology and manufacturing excellence as part of a broader collaboration aimed at enhancing image sensor performance.

With the MOU signed, Sony and TSMC are discussing potential investments by the JV. These investments, along with new capital investment by Sony in its existing plant in Nagasaki, are being considered on the premise that they would be implemented in phases based on market demand, and that they would receive support from the Japanese government.

This partnership also seeks to explore and address emerging opportunities in physical AI applications, such as automotive and robotics, paving the way for future innovations and expanded technological advancements. The establishment of the JV remains subject to the execution of a definitive legally binding agreement regarding the partnership and the satisfaction of customary closing conditions. 

“Building on the trust cultivated through our long-standing collaboration with TSMC, I am delighted that we have reached an agreement to advance our partnership to a new stage,” said Shinji Sashida, President and CEO of Sony Semiconductor Solutions Corporation.“This JV is a significant initiative that brings together the strengths of both companies and aims to drive further advancement in technology and business within the next-generation image sensor field. Building on this JV, Sony intends to further strengthen its business operations with a focus on creating high added value. Rooted in the Sony Spirit that has guided us since our founding, we intend to continue to take on the challenge of creating new markets through unprecedented ideas and distinctive technologies.”

“Sony has been our long-time partner in the CMOS image sensor business. We are excited to elevate our collaboration to the next level, which represents a key step forward in driving future sensing technology in the AI era,” said TSMC Senior Vice President and Deputy Co-COO, Dr. Kevin Zhang. “This partnership underscores our shared commitment and mutual vision of leveraging cutting-edge technologies and innovative solutions to deliver leading sensing technology and products. We look forward to working closely together to achieve impactful results and create lasting value for all stakeholders.”