Image Sensors World        Go to the original article...

Venturebeat, ElectronicsWeekly, EETimes report more details on AIStorm AI-on-Sensor startup:

• AIStorm has been founded in 2011 and has been in stealth mode till the recent announcement of $13.2M round A financing.
• AIStorm’s patented chip design is capable of 2.5 Tera Ops and 10 Tera Ops per watt, which said to be 5x to 10x lower than the average GPU-based system’s power.
• The company uses a technique called switched charge processing, which allows the chip to control the movement of electrons between storage elements
• "The TowerJazz pixel is part of our input layer, so the charge comes from sensors, they produce electrons, and we multiply and move them" - says AIStorm
• AIStorm tested its first chip this month and plans to ship production orders next year
• The company’s first products are to be made in 65nm or 180nm process
• AIStorm is planning Series B round for follow up products in 28-nm and possibly finer nodes
• The production chips are aimed to be compatible with popular AI frameworks such as TensorFlow

Go to the original article...

Image Sensors World        Go to the original article...

Digitimes: Shipments of Android phones with 3D cameras are set to boom in 2019, propelled by the increasing adoption of rear ToF cameras, according to Digitimes Research.

Oppo led all other Android phone makers with its rear ToF camera introduced in November 2018. Oppo competitors, including Huawei, Xiaomi, and Vivo are likely to follow with their ToF-enabled models in 2019.

Shipments of ToF camera Android smartphones are expected to reach 20M units in 2019, Digitimes Research estimates.

Go to the original article...

Image Sensors World        Go to the original article...

De Rijkste Belgen newspaper reports that LiDAR startup Xenomatix raises 5M euros. 2M euros comes through the conversion of a bond loan, while 3M is a fresh investment by Carl Van Hool and AGC Automotive Europe, a part of Asahi Japan. XenomatiX and AGC have partnered to develop a windshield-mounted LiDAR.

For a LiDAR commercialization, Xenomatix is said to need 10M euros. After the current round of financing, the company has 6.8M euros. In the fiscal year that ended in June 2018, Xenomatix was profitable with income of 0.6M euros. Xenomatix has 22 employees.

Go to the original article...

Image Sensors World        Go to the original article...

BusinessWire: San Jose, CA-based startup AIStorm raises $13.2M in Series A round from Egis Technology, TowerJazz, Meyer Corpo, and Linear Dimensions Semiconductor Inc.

“This investment will help us accelerate our engineering & go-to-market efforts to bring a new type of machine learning to the edge. AIStorm’s revolutionary approach allows implementation of edge solutions in lower-cost analog technologies. The result is a cost savings of five to ten times compared to GPUs — without any compromise in performance,” said David Schie, CEO of AIStorm.

Using sensor data directly—without digitization is said to enable real-time processing at the edge. AI systems require information be available in digital form before they can process data, but sensor data is analog. Processing this digital information requires advanced and costly GPUs that are not suitable for mobile devices: they require continuous digitization of input data, which consumes significant power and introduces unavoidable digitization delay (latency). AIStorm aims to solve these problems by processing sensor data directly in its native analog form, in real time.

“It makes sense to combine the AI processing with the imager and skip the costly digitization process. For our customers, this will open up new possibilities in smart, event-driven operation and high-speed processing at the edge,” said Avi Strum, SVP/GM of the sensors business unit of TowerJazz.

“The reaction time saved by AIStorm’s approach can mean the difference between an advanced driver-assistance system detecting an object and safely stopping versus a lethal collision,” said Russell Ellwanger, CEO of TowerJazz.

“Edge applications must process huge amounts of data generated by sensors. Digitizing that data takes time, which means that these applications don’t have time to intelligently select data from the sensor data stream, and instead have to collect volumes of data and process it later. For the first time, AIStorm’s approach allows us to intelligently prune data from the sensor stream in real time and keep up with the massive sensor input tasks,” said Todd Lin, COO of Egis Technology Inc.

AIStorm’s management includes CEO David Schie, a former senior executive at Maxim, Micrel and Semtech; CFO Robert Barker, formerly with Micrel and WSI; Andreas Sibrai, formerly with Maxim and Toshiba; and Cesar Matias, founder of ARM’s Budapest design center. AIStorm is based in San Jose, CA with offices in Austria, Taiwan, Phoenix and soon Dresden and Israel.

Go to the original article...

Image Sensors World        Go to the original article...

Adobe presents AI-powered demosaic algorithm for Bayer and Fujifilm X-Trans CFAs:

"...we’re introducing an all-new Sensei-powered feature, Enhance Details. Harnessing the power of machine learning and computational photography, Enhance Details... takes a brand new approach to demosaicing raw photos.

The new Enhance Details algorithm enables you to increase the resolution of both Bayer and X-Trans based photos by up to 30%. Applying Enhance Details to your photos can greatly improve fine detail rendering, improve the reproduction of fine colors, and resolve issues that some customers reported with their Fujifilm X-Trans based cameras."


Via Imaging Resource.

Go to the original article...

Image Sensors World        Go to the original article...

Electronic Design publishes an interview with Leo Bai, SmartSens’ AI BU General Manager. Few quotes:

"...single-frame HDR Global Shutter technology is better for image-recognition-based AI applications than conventional CMOS image sensors that use Multiple-Exposure HDR technology. Combined with a DVP/MIPI/LVDS interface, single-frame HDR Global Shutter technology can be adapted to various types of SoC platforms.

...adoption of global-shutter technology is growing rapidly, in comparison to rolling-shutter technology. One of the main reasons is that a global-shutter CMOS image sensor is able to achieve excellent real-time performance without the jelly effect, especially in AI and machine-vision applications. With advanced manufacturing process technology and reduced cost, it’s expected to see increasing market demand for global-shutter CMOS image sensors."

Go to the original article...

Image Sensors World        Go to the original article...

Autosens Detroit conference to be held on May 14-16, 2019 announces its agenda with a number of interesting imaging presentations:

• Infrared Camera Sensing for ADAS and Driverless Vehicles: Applications, Challenges and Design Considerations. Workshop by Rajeev Thakur, OSRAM Opto
• Material interactions for autonomous sensor applications. Workshop by Jim Howard and Jonah Shaver, 3M
• Keeping eyes on the passengers - developing an in-cabin omni-sensor, Guy Raz, Guardian Optical Technologies
• The FIR Revolution: How FIR Technology Will Bring Fully Autonomous Vehicles to the Mass Market, Yakov Shaharabani, Adasky
• The Future of Driving: Enhancing Safety on the Road with Thermal Sensors, Tim Lebeau, Seek Thermal
• RGB-IR Sensors for in Cabin Automotive Applications, Boyd Fowler, Omnivision
• The Next Generation of SPAD Arrays for Automotive LiDAR, Wade Appleman, ON Semi
• What’s in Your Stack? Why Lidar Modulation Should Matter to Every Engineer, Randy Reibel, Blackmore
• Addressing LED flicker, Brian Deegan, Valeo
• From Camera to LiDAR systems alignment and testing in mass production of ADAS Sensors, Dirk Seebaum, Trioptics
• The influence of colour filter pattern and its arrangement on resolution and colour reproducibility, Tsuyoshi Hara, Sony
• Highly Efficient Autonomous Driving with MIPI Camera Interfaces, Hezi Saar, Synopsys
• Tuning image processing pipes (ISP) for automotive use, Manjunath Somayaji, GEO Semiconductor
• Computational imaging through occlusions; seeing through fog, Guy Satat, MIT Media Lab
• ISP optimization for ML/CV automotive applications, Alexis Lluis Gomez, ARM

Go to the original article...

Image Sensors World        Go to the original article...

Finally, somebody found the use for black sun effect in a productive way. MDPI publishes Sungkyunkwan University, Korea paper "Accurate and Cost-Effective Micro Sun Sensor based on CMOS Black Sun Effect" by Rashid Saleem and Sukhan Lee.

"An accurate and cost-effective micro sun sensor based on the extraction of the sun vector using a phenomenon called the “black sun” is presented. Unlike conventional image-based sun sensors where there is difficulty in accurately detecting the sun center, the black sun effect allows the sun center to be accurately extracted even with the sun image appearing irregular and noisy due to glare. This allows the proposed micro sun sensor to achieve high accuracy even when a 1 mm × 1 mm CMOS image sensor with a resolution of 250 × 250 pixels is used."

Go to the original article...

Image Sensors World        Go to the original article...

University of Cambridge paper "Grayscale-to-color: Single-step fabrication of bespoke multispectral filter arrays" by Calum Williams, George Gordon, Sophia Gruber, Timothy Wilkinson, and Sarah Bohndiek proposes grayscal photolitography for multispectral color filter manufacturing:

"Conventional cameras, such as in smartphones, capture wideband red, green and blue (RGB) spectral components, replicating human vision. Multispectral imaging (MSI) captures spatial and spectral information beyond our vision but typically requires bulky optical components and is expensive. Snapshot multispectral image sensors have been proposed as a key enabler for a plethora of MSI applications, from diagnostic medical imaging to remote sensing. To achieve low-cost and compact designs, spatially variant multispectral filter arrays (MSFAs) based on thin-film optical components are deposited atop image sensors. Conventional MSFAs achieve spectral filtering through either multi-layer stacks or pigment, requiring: complex mixtures of materials; additional lithographic steps for each additional wavelength; and large thicknesses to achieve high transmission efficiency. By contrast, we show here for the first time a single-step grayscale lithographic process that enables fabrication of bespoke MSFAs based on the Fabry-Perot resonances of spatially variant metal-insulator-metal (MIM) cavities, where the exposure dose controls insulator (cavity) thickness. We demonstrate customizable MSFAs scalable up to N-wavelength bands spanning the visible and near-infrared with high transmission efficiency (~75%) and narrow linewidths (~50 nm). Using this technique, we achieve multispectral imaging of several spectrally distinct target using our bespoke MIM-MSFAs fitted to a monochrome CMOS image sensor. Our unique framework provides an attractive alternative to conventional MSFA manufacture, by reducing both fabrication complexity and cost of these intricate optical devices, while increasing customizability."

Go to the original article...

Image Sensors World        Go to the original article...

PRNewswire: After releasing a 4MP sensor with NIR-enhanced 3um pixels 2 months ago, Smartsens adds a similarly specked sensor with 2um pixels. Aimed to AIoT (AI+IoT) market, the new 1/3-inch SC4238 4MP BSI sensor supports a 2-exposure HDR mode with a DR up to 100dB, and has improved QE in 850nm to 940nm band.

Compared with other 1/3-inch 4MP HDR products on the market, SC4238 has an advantage of good performance in low light — SNR1 0.47 vs 1.22, which is about 2.6 times better performance than our competitor.

SC4238 is in mass production now.

Go to the original article...

Image Sensors World        Go to the original article...

Counterpoint Research forecasts the demand for image sensors for passenger cars globally to double by 2023. The shipments of these image sensors will grow at a 19% CAGR, crossing 230M units by 2023:

"With rear cameras for basic parking assistance in cars becoming almost a standard feature on newer models, we are now seeing the advent of a higher proportion of cars fitted with front-facing and side cameras to enable enhanced ADAS features. For example, surround view cameras, which are currently an option on higher-end models, will see significant adoption over the next five to six years. The key factors driving this trend are, firstly, government encouraging OEMs to integrate advanced safety features, and, secondly, growing awareness and preference for advanced safety features among customers.

With higher adoption of front ADAS camera solutions, by 2023, all new cars sold in the US are likely to have more than three cameras per car."

Go to the original article...

Image Sensors World        Go to the original article...

Cea-Leti spin-off Curve-One startup aims to commercialize curved sensor technology and products:

"Fruit of 6 years of research and development, the Curve-One Fish-Eye benefits from the most optimized process. Based on not less than 8 patents, mass-production oriented, Curve-One is the most technically advanced wide field camera.

In terms of technologies, this is the dawn of a new era for cameras, camera phones, autonomous cars, drones, military instruments and bio-medical instruments, with the access to wider fields and exquisite homogeneity of the optical properties across the images, and faster systems not possible with classical flat focal planes.

Also, fewer components are needed, and the remaining ones are less complex. This increases economical and technical performance for the optical systems optimizations.

Soon to be off-the-shelf components for civil applications (cameras, civil drone) these breakthrough components will blossom in the technical needs for autonomous cars, military drones and advanced bio-medical applications."

Go to the original article...

Image Sensors World        Go to the original article...

ElectronicDesign: Cepton CEO and co-founder Jung Pei presents 11 myths about LiDARs as a background to show the company advantages:

1. LiDAR is a very high-tech solution.
2. LiDAR is expensive.
3. Solid-state LiDAR is the best approach because it has no moving parts.
4. Flash LiDAR is the best LiDAR for imaging.
5. LiDAR must operate infrared wavelengths.
6. LiDAR isn’t safe for the human eye.
7. LiDAR can’t work in poor weather conditions.
8. LiDAR can only be used for automobiles.
9. LiDAR won’t be incorporated into vehicles for another decade.
10. LiDAR can be fully replaced by cameras, radar, or a combination of the two.
11. FMCW LiDAR is better than ToF LiDAR.

Go to the original article...

Image Sensors World        Go to the original article...

PRNewswire: LG, Infineon, and PMD have teamed up to introduce ToF camera in the LG G8 ThinQ smartphone. Infineon's REAL3 image sensor chip will be integrated into the front-facing camera of the upcoming LG G8 ThinQ.

"Infineon is poised to revolutionize the market," said Andreas Urschitz, division president of Infineon's Power Management & Multimarket division. "We have demonstrated service beyond the mere product level – specifically catering to phone OEMs, associated reference design houses and camera module manufacturers. Within five years, we expect 3D cameras to be found in most smartphones and Infineon is poised to contribute a significant share."

"Keeping in mind LG's goal to provide real value to its mobile customers, our newest flagship was designed with ToF technology from inception to give users a unique and secure verification system without sacrificing camera capabilities," said Chang Ma, SVP and head of product strategy at LG Mobile Communications Company. "With innovative technology like ToF, the LG G8 ThinQ will be the optimal choice for users in search of a premium smartphone that offers unmatched camera capabilities."

Go to the original article...

Image Sensors World        Go to the original article...

Canon presents its current and future image sensor offerings:





FLIR presents its new InGaAs cameras:



Photron demos its polarized imaging use cases:





PCO talks about sCMOS sensor evolution:

Go to the original article...

Image Sensors World        Go to the original article...

Australian University of Adelaide group of researches publishes an interesting arxive.org paper with further research of Sony IMX219PQ 1.12um pixel nonuniformities "Reverse Engineering the Raspberry Pi Camera V2: A study of Pixel Non-Uniformity using a Scanning Electron Microscope" by Richard Matthews, Matthew Sorell, and Nickolas Falkner. The paper undergoes a peer review now.

"In this paper we reverse engineer the Sony IMX219PQ image sensor, otherwise known as the Raspberry Pi Camera v2.0. We provide a visual reference for pixel non-uniformity by analysing variations in transistor length, microlens optic system and in the photodiode. We use these measurements to demonstrate irregularities at the microscopic level and link this to the signal variation measured as pixel non-uniformity used for unique identification of discrete image sensors."

Go to the original article...

Image Sensors World        Go to the original article...

Fairchild Imaging is introducing two new 4/3” 10MP sensors, LTN4323 and MST4323, the first to feature sCMOS 3.0 technology. The monochrome LTN4323 (for scientific and industrial cameras) has 0.7e read noise. A special BSI process enhancement delivers a broad spectrum NIR-QE with greater than 2x sensitivity. Dark current at 30C is less than 2e/s enabling compact camera designs without the need for TE cooling. The color MST4323 (for professional cinema cameras) provides 4k video at 120fps.

Go to the original article...

Image Sensors World        Go to the original article...

ON Semi publishes a short video on its automotive CMOS sensor operation in challenging lighting transitions on the road:

Go to the original article...

Image Sensors World        Go to the original article...

Melexis announces a new version of its FIR thermal sensor array with lower thermal noise compared to the current MLX90640, an increased refresh rate of 64 Hz and an elevated operating temperature up to 125 °C.

The new MLX90641 is a small 16 x 12 pixel IR array housed in a 4-lead TO39 package that is able to accurately measure temperatures in the range -40 °C to +300 °C. The factory-calibrated devices ensure an accuracy of 1 °C in typical measurement conditions. The high accuracy is further supported by a Noise Equivalent Temperature Difference (NETD) of 0.1 K RMS.

Two different FoV options are available, a standard 55° x 35° and a wide angle 110° x 75°. The device is simple to use, operating from a single 3.3 V supply and storing all results in internal RAM for access via an I2C compatible digital interface. A proprietary algorithm ensures high thermal stability, even in conditions where the temperature is changing rapidly.

Go to the original article...

Image Sensors World        Go to the original article...

PRNewswire: OmniVision announces an expansion of its OVMed medical ISP family, designed for quick integration into single-use and reusable endoscopes as well as catheters. Building on the success of OmniVision’s versatile OVMed ISPs introduced in 2018, these five new processors provide device designers with even more flexibility, creating additional options for resolution, frame rate and output, along with the best-in-class feature set.

“By expanding our OVMed family to nine different ISPs, we are providing the industry’s largest set of medical design options, catering to customers new to CMOS imaging as well as those wanting to speed development for their latest medical devices,” said Tehzeeb Gunja, principal marketing and business development manager at OmniVision. “One size and price do not fit all; hence we offer a range of ISP solutions with differentiated size, price, feature set and outputs. We are now adding stereoscopic 3D imaging and the best-in-class, professional ISPs to our growing OVMed family, enabling innovative, state-of-the-art product concepts.”

The OVMed product line includes three classes of ISPs:

• Mini for the smallest and most cost-effective option to integrate into an endoscope handle
• Advanced for higher image quality while fitting easily inside a camera control unit (CCU)
• Pro for the highest image quality and feature set that can also fit inside a CCU

The newly introduced Pro line provides resolution of up to 4K2K, while supporting streaming video at 60 fps, along with high-end ISP functions that enable superior noise reduction and color accuracy for the highest level of image quality. Products in the Pro line accept up to 4-lane digital MIPI/sub-LVDS input. Additionally, they can connect to image sensors via longer input cables of up to five meters. The Pro ISPs also feature dynamic shading control, distortion correction, advanced backlight compensation, an automatic exposure and gain control window, exposure weighting and compensation, advanced auto white balance, color enhancement, hue, halation, contrast adjustment/enhancement/curve, high dynamic range, crop/digital zoom, resize, frame-rate control, freeze, time lapse, dual image display, illumination control, RAW 8/10/12 image output, and RGB565/888 bypass mode.

Two new ISPs (Mini and Pro) in the OVMed family offer dual-channel processing that supports 3D stereoscopic imaging. This feature allows an ISP to take input from two image sensors and process it in parallel, providing surgeons with the ability to perceive depth during surgical procedures. The result is improved patient outcomes and reduced risk for endoscopic surgery, as well as better diagnostic accuracy.

The table below lists the nine currently available OVMed products:


PRNewswire: OmniVision partners with Lighthouse Imaging to create a reference platform that enables medical device manufacturers to integrate stereoscopic 3D imaging into endoscopes. The platform incorporates two CMOS sensors and an ISP from OmniVision, plus a light control board for LED illumination and camera control unit from Lighthouse:

Go to the original article...

Image Sensors World        Go to the original article...

Omnivision booth at SPIE BIOS EXPO 2019 shows a multitude of image sensor-based checks for a human body:

Go to the original article...

Image Sensors World        Go to the original article...

Arstechnica publishes an article "How 10 leading companies are trying to make powerful, low-cost lidar" by Timothy Lee:

"...over the last year, I've gotten a steady stream of pitches from lidar companies, and I've talked to as many of them as I could. Ars has now been in contact with senior executives from at least eight lidar companies as well as others involved in the industry as customers or analysts. These conversations have provided a lot of insight not only into trends in the lidar industry in general but also about the technology and business strategy of individual companies."


Thanks to TG for the pointer.

Go to the original article...

Image Sensors World        Go to the original article...

Facial recognition technology finds more and more applications. Ecns.cn reports that Baidu Brain engineer has incorporated AI into a nearly two-meter-tall shelter for abandoned cats. The multi-level cat shelter includes a camera at the entrance that only opens the door when it recognizes the approaching animal is a cat. The shelter's facial recognition software can identify the faces of 174 different cats, report on the cat's state of health and whether the feline has been spayed or neutered.

Researchers at the Fraunhofer Institute for Factory Operation and Automation IFF developed a scanner and analysis software in collaboration with the Saxony-Anhalt State Office for Heritage Management and Archaeology, which digitally capture the visual features of coins and recognize them in a matter of seconds. The scanning system can be used to identify and recognize coin finds. The novel scanner O.S.C.A.R., short for Optical System for Coin Analysis and Recognition, not only scans coins’ visual features but also the minutest signs of wear such as scratching, clipping, contours, edges, pitting and denting, which render an object unique.

Coin surface scan

CNA Insider publishes 3 more use cases for facial recognition:



Oleksii Kharkovyna publishes many more facial recognition use cases in his Medium post "Facial Recognition And AI: Latest Developments And Future Directions," such as:

• Usage for workflow. In Japan, AI-equipped cameras will continuously analyze facial expressions of workers to improve their productivity. If they are caught sleeping during the work, they could be blasted with freezing air.
• At the Dubai Airport, a virtual aquarium is fitted with 80 facial recognition cameras that examine people as they walk through the aquarium. The system will identify a person who has problems with the law, even without his/her knowing.

BiometricUpdate reports that UK and France have made it illegal to hide your face to avoid facial recognition.

Go to the original article...

Image Sensors World        Go to the original article...

Sony quarterly earnings report reveals 8% YoY lower image sensor sales siting lower mobile image sensor demand. The next fiscal year forecast is reduced by 6%:

• FY18 Q3 sales decreased 8% year-on-year to 230.3 billion yen, primarily due to a decrease in unit sales of image sensors for mobile devices.
• Operating income decreased 14.0 billion yen year-on-year to 46.5 billion yen.
• We have downwardly revised our FY18 sales forecast to 870 billion yen and operating income to 130 billion yen.
• This downward revision is primarily due to a reduction in the unit sales forecast for image sensors resulting from a deterioration in the smartphone market.
• Although we expect the difficult market environment for smartphones to continue going forward, we are seeing high-end sensors adopted in both highend and mid-range models as smartphone makers seek to differentiate their products through camera functionality.
• Consequently, there is no change to our view that demand for the high-end image sensors that Sony excels at making will continue to increase due to the trend toward multi-lens cameras and larger die-sizes. There is also no change to the plan I recently outlined to increase our production capacity to nearly the maximum that can fit into our existing facilities.

SeekingAlpha publishes the earnings call transcript with few more details on Sony expectations from the 2019 business:

"So the third quarter rate of operation of the semiconductors as well as what's our view on the first half of fiscal '19. Currently for the third quarter fiscal '18 the wafers on an average 99,000, 99K and this is along the line of October forecast. In the fourth quarter of fiscal '18, wafer injection number is on an average 89K. And for fiscal 2019, it is very difficult to cite a specific number, so I like to communicate to you basic way of thinking.

So the smartphone market units for fiscal '19, it is expected to foresee any growth, probably the unit sales would decline. That is the forecast held up by most of the market related people. But on the other end multi-lens cameras as well as a large size imagers that's the trend which will be accelerated further over our October forecast throughout fiscal 2019, so those are both positive and negative factors for us.

So how we forecast the outcome of that? In any case as far as fiscal '19 is concerned, the market condition will be very tough. So we would like to be conservative and carefully manage the business in a steadfast manner.

In case of Semiconductor basically, the demand immediately around us is weakened. As you know, our image sensors takes about five months from the introduction of wafer to shipment that lead time five months. Therefore the level of inventory at the end of December in 2018 was a little heavy even in view of the smartphone shipment in 2019. And we are taking measures to adjust the inventory to the appropriate level but towards the year-end, in view of the immediate situation, inventory level appears to be a little heavier."

Go to the original article...

Image Sensors World        Go to the original article...

Arxiv.org paper "Determining Image Sensor Temperature Using Dark Current" by Richard Matthews, Matthew Sorrel, Nickolas Falkner from University of Adelaide, Australia claims that the authors were able to observe pixel variations in SEM:

Go to the original article...

Image Sensors World        Go to the original article...

Almalence compares its SuperSensor with Google Pixel 3 super resolution zoom. Both are based on multiframe image acquisition and processing:

Go to the original article...