One Investor's View on Sony CIS Business

Forbes contributor Stephen McBride publishes his view on Sony CIS business:

"The growth in sensor sales over the past three years has been nothing short of remarkable. This year Sony will generate more profits from imaging than any of its other business lines. Sensors are on track to generate $1.92 billion in profits… 10% more than Sony’s long-established gaming arm.

Sony absolutely dominates the image sensor industry, accounting for over 50% of global sales.

The battle among smartphone firms to make better cameras has been a boon for Sony. It controls over 70% of the smartphone sensor market. And it’s been the exclusive maker of image sensors for every iPhone since 2010.

And when it comes to quality, Sony is in a league of its own. Its image sensors are so far ahead, it charges 2X as much as its closest competitor.

Right now, Sony’s world-class imaging business is flying under the radar. But as the computer vision boom takes off, I expect the stock to attract a lot of hype. And with image sensors becoming a larger part of Sony’s business, it could easily soar 300%+ in the coming years."

3D Camera and Analytics Startup Raises $51M

VentureBeat, Forbes: Density, a San Francisoco-based startup building AI-powered 3D cameras, closes a $51M investment round, following $23M in previous funding.

Density uses 3D depth cameras and an AI cloud-based software to enable social distancing and occupancy analytics. Pepsi, Delta, Verizon, Uber, Marriot, and ExxonMobil are among its clients that use its service to figure out which parts of their offices get the most use and which the least and to deliver people-counting metrics to hundreds and thousands of their employees.

The Density 3D camera consists of over 800 components sourced from 137 supply chains. The camera attaches above a doorway and tracks movement with two Class 1 infrared lasers. The data is transferred via Wi-Fi to Density’s cloud-hosted backend, where it’s processed and analyzed. A web dashboard, SMS messages, signage, and mobile apps provide insights like the real-time capacity of a room and historical crowd sizes, while an API allows third-party apps, services, and websites to make use of the data in many other ways.

Density’s infrared tracking method offers a major advantage over other approaches: privacy. Unlike a security camera, its sensors can’t determine the gender or ethnicity of the people it tracks, nor perform invasive facial recognition. “It’s far easier to do a camera,” says Density CEO Andrew Farah. “But we believe the data pendulum has swung too far in one direction. It’s good to see people ask about data being collected … We knew that the right market was corporate clients with office space because our sensor can do occupancy detection inside of a room where a camera can’t go.”

Tower Updates on its Q2 2020 Imaging Business

SeekingAlpha: Tower Q2 2020 earnings call has few updates on the foundry's imaging business:

"Moving to our Sensor Business unit, we forecast single-digit year-over-year growth in spite of several segments being strongly adversely impacted by COVID-19. For example, dental x-ray sensors where we have major market share. The impact is mainly seen in the form of customer requests to push out orders, dental clinics were closed for a long period and are still not back to full speed hence customers are very inventory cautious. Industrial sensors namely manufacturing line inspections is also down as a function of reduced new manufacturing line build out. Due to our very strong CIS platform and compatibility, our fingerprint sensors namely the under OLED 101 sensor and the under LCD lens type modules, we were developed and qualified in very short times. We target our first volume production revenue in the fourth quarter wrapping throughout 2021.

Our time-of-flight program is moving along very well. We are prototyping the first sensor to our lead customer this quarter and plans for after production in the first half of next year. This would be our first product moving to mass production utilizing our stacked wafer BSI pixel level bonding platform.

And as stated, for us, one of the beauties of that market is that we were able to very, very quickly develop and qualify flows because of a very, very strong and diverse CIS core that we have throughout the company. So CMOS image sensor throughout the company, and take advantage of a market need. As far as the time-of- flight, there's a variety of applications are going out for time of flight. It's predominantly at 300 millimeter. And it's focused with backside illumination and stack wafers. And that is also more than one customer with blue chip customers that are signed up on long-term agreements. Of course, we have to perform, but that is a very, very substantial market for us.

But we have seen a very, very big reduction in the forecast of the image sensors for dental X-ray."

Yole Forecasts 2020-25 Camera Module Market

Yole Developpement report "CMOS Camera Module Industry for Consumer & Automotive 2020" forecasts the revenue of the global camera module market to expand from $31.3B in 2019 to $57.0B in 2025, at 12.8% CAGR.

"Beyond the sensor itself, innovations in all the different subcomponents of the camera module are in high demand. The introduction of periscope lenses was a major event that allowed 5x or even 10x optical magnification within the existing thickness of mobile phones. Optical image stabilization (OIS) is another critical technology for photography especially for telephoto, hence players are also looking for innovations in this area, using new materials, MEMS or liquid lens to replace the Voice Coil Motor (VCM) approach.

Technical upgrades of camera modules include the “Active Alignment” process to align multiple cameras well. There will be several innovations in camera module integration, like pop-up cameras or side-up cameras, and under-screen cameras in future.

The industry leader LG Innotek continues to maintain its position due to the large orders from Apple. Closing in are Ofilm and Sunny Optical, who have climbed to the second and third positions by relying on the strong domestic market in China, replacing Semco and Foxconn/Sharp. In the years to come the US-China trade war could play a big role in reorganizing the ranking of Compact Camera Module (CCM) players."

Memristor-Based Smart Image Sensor

National Science Review paper "Networking retinomorphic sensor with memristive crossbar for brain-inspired visual perception" by Shuang Wang, Chen-Yu Wang, Pengfei Wang, Cong Wang, Zhu-An Li, Chen Pan, Yitong Dai, Anyuan Gao, Chuan Liu, Jian Liu, Huafeng Yang, Xiaowei Liu, Bin Cheng, Kunji Chen, Zhenlin Wang, Kenji Watanabe, Takashi Taniguchi, Shi-Jun Liang, and Feng Miao from Nanjing University, China, and National Institute for Materials Science, Japan, proposes a pixel array that can recognaze objects:

"Comparing to human vision, conventional machine vision composed of image sensor and processor suffers from high latency and large power consumption due to physically separated image sensing and processing. Neuromorphic vision system with brain-inspired visual perception provides a promising solution to solve the challenge. Here we propose and demonstrate a prototype neuromorphic vision system by networking retinomorphic sensor with a memristive crossbar. We fabricate the retinomorphic sensor by using WSe2/h-BN/Al2O3 van der Waals heterostructures with gate-tunable photoresponses, to closely mimic the human retinal capabilities in simultaneously sensing and processing images. We then network such sensor with a large-scale Pt/Ta/HfO2/Ta one-transistor-one-memristor (1T1R) memristive crossbar, which serves as the role similar to the visual cortex in human brain. The realized neuromorphic vision system allows for fast letter recognition and object tracking, indicating the capabilities of image sensing, processing and recognition in the full analog regime. Our work suggests that such neuromorphic vision system may open up unprecedented opportunities in future visual perception applications."

Brookman Presents its ToF Products Catalog

Brookman has updated its ToF products page with a number of cameras based on BT008D sensor:

Brookman also congratulates Shizuoka University Prof. Shoji Kawahito, one of the founding members and a chairman of the company, with reception of the “Tamba Takayanagi Award/Achievement Award”. "This is the result of excellent achievements in the pioneering research on high performance and high functionality of image sensors, which the professor has been working on for a long time, and their practical application. Currently, CMOS image sensors continue to evolve from conventional 2D imaging to 3D sensing. We will continue to do our utmost to return the accumulated research results to society by continuing to challenge the possibilities of image sensing."

e2v Unveils ToF Image Sensor

Teledyne e2v introduces Hydra3D, an 832 x 600 pixel resolution ToF CMOS sensor. Hydra3D is based on a 10µm three-tap pixel.

Update: GlobeNewswire: Ha Lan Do Thu, Marketing Manager for 3D imaging at Teledyne e2v says, “We are very pleased to announce our newest Time-of-Flight sensor, the first multi-tap high resolution sensor in the market. Our partnership with Tower allows us to provide customers with the highest level of 3D performance, including uncompromised image quality in both 2D and 3D mode, in all operation conditions.”

Hydra3D comes with an evaluation kit (Hydra3D EK), enabling customers to evaluate the sensor in multiple application setups. The kit includes a compact 2/3-inch optical format calibrated module, which includes a light source for near infrared illumination and an optic. Two versions will be available targeted at performing the Time-of-Flight principle at short-range distances (up to 5 metres) or mid-range distances (up to 10 metres) and with a field-of-view of 60° x 45° or a field of view of 40° x 30°, while capturing real-time 3D information at a full resolution.

Rafael Romay, VP of Professional Imaging at Teledyne e2v, says: “The great technology innovation and partnership with Tower has been key in the development of this innovative new ToF image sensor, helping us to bring to market this best-in-class solution.”

Avi Strum, SVP and GM of Sensors and Display BU at Tower, added “We are very excited about the release of Hydra3D. Our strong partnership with Teledyne e2v goes back more than 15 years and many of their state-of-the-art products are manufactured by Tower. The Hydra3D ToF product is aligned well with Tower’s strategic investment in the ToF market. We look forward to many other Teledyne e2v products utilizing our world class CIS technology.”

Samples will be available in August 2020 and evaluation kits will be available in September 2020.

Sony Transfers its Fabs Management to NEC Facilities

Sony: NEC Facilities and Sony Semiconductor Manufacturing Co. has agreed to jointly establish SSN Facilities as a new company that manages facilities at semiconductor production bases. SSN Facilities will undertake facility management, repair work, maintenance work, etc. of Sony Semiconductor Manufacturing from September 1, this year. “SSN” stands for Sustainable & Smart Next generation.

NEC Facilities has a wealth of experience and specialized human resources, knowledge, and know-how regarding factory facility management at manufacturing bases, mainly semiconductors and electronic components. With the establishment of SSN Facilities, as an outsourcer of facility management in the manufacturing industry, we aim to further expand our business, centering on the semiconductor manufacturing field, where demand continues to grow as digitalization continues.

Sony Semiconductor Manufacturing outsources facility management operations at seven production sites in Japan to SSN Facilities to ensure stable operation and maintenance of clean rooms and facilities, as well as to enhance and streamline operations, as well as its own business.

QURV Startup to Develop SWIR Image Sensors

Barcelona, Spain-based The Institute of Photonic Sciences (ICFO) launches a spin-off company, Qurv Technologies. The new company develops wide-spectrum image sensor technologies and integrated solutions for computer vision applications, addressing the needs of an autonomous and intelligent new world.

Qurv’s graphene/quantum dot image sensors platform technology allow operation from the visible to the SWIR range and can be integrated with current CMOS low-cost, high-manufacturability processes. Qurv’s "plug and play" approach aims to bring advanced machine vision capabilities to markets that are not accessible by the current state of the art SWIR sensors.

Qurv incubated in the KTT Launchpad for more than 6 years and holds a portfolio of more than 10 patent families. The incubation has received support from the Government of Catalonia, the Ministry of Economy, Industry and Competitiveness of Spain, the European Research Council, the Barcelona City Council and the Castelldefels City Council, the European Regional Development Funds allocated to Generalitat de Catalunya for emerging technology clusters, and the European Union’s Horizon 2020 research and innovation program.

Antonios Oikonomou, Qurv’s CEO, comments, "Nature itself hides a vast amount of information beyond what is visible. By harnessing and efficiently processing this information, a new era in health, security and decision-making will emerge. However, no mass-deployable solution exists to provide these capabilities at scale and to everyone. With the immense support of the KTT unit at ICFO, we are now ready to achieve precisely this- to bring a technology once available only in the lab to the world.”

Stijn Goossen, the company’s CTO adds, “Our unprecedented expertise of the graphene/quantum dot stack puts us in an optimal position to leverage the benefits of integration with silicon CMOS technology in terms of functionality, performance and addressable markets. World-renowned experts in graphene, Prof. Frank Koppens and quantum dots, Prof. Gerasimos Konstantatos, have been key in the early technology development. We are delighted to announce that they will take up the role of scientific advisors to the company while further maturing the technology.”

Photron Presents FastCam with 4MP 1440fps APS-C Image Sensor

Photron unveils FASTCAM NOVA R2 with "unique CMOS image sensor technologies:"

• 4MP resolution
• 2048 x 2048 pixels at 1,440fps
• 1920 x 1080 pixels at 2,560fps
• ISO 8,000 monochrome
• ISO 2,500 color
• Global Electronic Shutter: 1ms to 2.7μs independent of frame rate
• 12b ADC

US Blacklists O-Film

Nikkei, Reuters: Camera module maker O-Film is one of 11 Chinese companies added to the U.S. Commerce Department's Entity List over alleged human rights abuses involving China's Uighur Muslim minority. The Shenzhen-listed O-Film supplies its camera module to many companies, including Apple, Microsoft, HP, Dell, General Motors, Amazon, Samsung, Huawei, Oppo, ZTE, and Sony.

"Since the establishment of its business entity of CCM business unit in 2012, OFILM has been focusing on the development and manufacturing of image modules. In just four years, it has become the largest manufacturer of image modules in the world, and the gap with other module manufacturers has expanded year by year. According to the market research firm TSR, the market share of OFILM’s CCM shipments exceed 20% in 2018, ranking first of the industry."

O-Film publishes an official statement on the blacklisting:

"As a global leader in technological innovation and advanced high-technology manufacturing, OFILM has always abided by the laws and regulations of the nations where we operate. We treat our employees equally and protect their rights and interests. Every year, multiple times, including in 2020, OFILM has passed independent third-party Corporate Social Responsibility workplace audits organized by our customers, including surprise inspections. We have received the RBA certification of the Responsible Business Alliance.

Ofilm employees are never coerced to work for us.

With great respect, we call on the United States to re-examine its recent decision. We look forward to communicating fully with the relevant US government departments. We also look forward to a full airing of the facts and to a just outcome."

LiDAR News: Leddartech, Xaos, IDTechEx, Velodyne, Uber

Leddartech's Frantz Saintellemy, President and COO, features in a podcast:



Korea-based Xaos Sensor presents its MEMS-based LiDARs priced at $200:

IDTechEx publishes a nice video with review of different LiDAR approaches on the market:

Forbes contributor Sabbir Rangwala publishes his analysis of Velodyne merger with GRAF and going public:

• Velodyne's valuation post-deal close has grown from ~$1.8B (Velodyne was estimated at a valuation of $1.6B before the merger announcement) to ~$3B as of July 21, 2020
• The ASP per unit drops from $7K in 2019 to $600 in 2024 – which is ~ a 10X reduction, and therefore a 10X increase in volumes shipped
• The growth in market share and unit volumes is based on a transition from the 360° FOV LiDAR products (Surround LiDAR which Velodyne has traditionally dominated) to the Vela series of products in which there is significant competition, and Velodyne is just starting to develop
• Finally, profitability and cash flow – they currently lose about $50M/year, break even by 2023 as the Vela products kick in
• The above analysis indicates that a part of their growth will need to come from acquisitions

Bloomberg reports that Uber conciders a guilty plea by its former LiDAR engineer Anthony Levandowski is proof that he’s a liar, and supports its decision to make Levandowski alone shoulder a $180M legal award Google won against him.

He agreed to plead guilty to Google-Waymo LiDAR trade secret theft and was driven into bankruptcy when Google won a contract-breach arbitration case against him. Levandowski was counting on Uber’s promise when it first hired him to provide legal cover, known as indemnification, from his former employer.

Uber now says it has no obligation to reimburse Levandowski for the $180M.

Past, Present, and Future of Face Recognition

A preprint paper "Past, Present, and Future of Face Recognition: A Review" by Insaf Adjabi, Abdeldjalil Ouahabi, Amir Benzaoui, and Abdelmalik Taleb-Ahmed from University of Bouira, Algeria, University of Tours, France, and University of Valenciennes, France, presents the challenges for the facial recognition algorithms:

"Face recognition is one of the most active research fields of computer vision and pattern recognition, with many practical and commercial applications including identification, access control, forensics, and human-computer interactions. Significant methods, algorithms, approaches, and databases have been proposed over recent years to study constrained and unconstrained face recognition. 2D approaches reached some degree of maturity and reported very high rates of recognition. This performance is achieved in controlled environments where the acquisition parameters are controlled, such as lighting, angle of view, and distance between the camera-subject. However, if the ambient conditions (e.g., lighting) or the facial appearance (e.g., pose or facial expression) change, this performance will degrade dramatically. 3D approaches were proposed as an alternative solution to the problems mentioned above. The advantage of 3D data lies in its invariance to pose and lighting conditions, which has enhanced recognition systems efficiency. 3D data, however, is somewhat sensitive to changes in facial expressions. This review presents the history of face recognition technology, the current state-of-the-art methodologies, and future directions."