Omnivision Proposes DTI with Charge Transfer Assist

Omnivision patent application US20150048427 "Image sensor pixel cell with switched deep trench isolation structure" by Sing-chung Hu, Rongsheng Yang, Gang Chen, Howard E. Rhodes, Sohei Manabe, and Hsin-chih Tai proposes a BSI pixel with DTI, filled by a lightly doped poly. A negative pulse applied to the DTI poly allows increasing PD doping and full well while maintaining a low image lag:

ISSCC 2015 Report: Shizuoka University Compressive Imager

Albert Theuwissen continues to post his reports from ISSCC. Shizuoka University presents a 200 million fps compressive imager. The imager consists of 15 pixel sub-arrays, each of them exposed at the differing times. In the end, a 200Mfps image sequence is recovered.

Cista and SMIC Start Mass Production of 0.13um-Process BSI Sensors

PR Newswire: Remember Cista, a mysterious image sensor startup founded by ex-Omnivision employees? SMIC and Cista System Corp. jointly announce a mass production start of two CIS-BSI products, of 1.3MP resolution with 1.75um pixel and 8MP resolution with 1.4um pixel, respectively. Both sensors are based on SMIC's 0.13um BSI process.

SMIC's 0.13um CIS-BSI technology is said to be independently developed and offers competitive performance. Based on a low leakage process, it only uses three aluminum metal layers for reduced cost and supports pixel sizes down to 1.4um for 8MP resolution CIS. SMIC also provides full in-house turn-key service which includes CIS wafer fabrication, color filter & micro-lens processing, TSV-CSP and testing to help customers shorten the supply chain with fast cycle time and low cost.

"Through working with our partner, Cista System Corp., we are very pleased with the achievement of the production phase for BSI technology," said Dr. Shiuh-Wuu Lee, EVP of Technology Development of SMIC. "Tests on the two sensors have shown great performance which demonstrates our readiness in 0.13-micron BSI technology platform. SMIC is also developing 1.1-micron pixel BSI for 13MP resolution and above, and 3D stacked BSI for high-end applications. With these new sets of products, we hope to provide high-quality CMOS Image Sensors to our customers at a competitive price."

"We are excited to partner with SMIC on launching the CIS-BSI sensors," said Wilson Du, CEO and President of Cista System Corp. "This partnership draws us one step closer to our goal of becoming more integrated with domestic industry resources in developing the image sensor sector. As we move forward, we hope to see more of our designs used in wider applications such as consumer electronics, telecommunications, medical equipment, automotive industry, automation and other applications."

So far, Cista web site only lists 1MP and 2MP products, with others, up to 13MP, are said to come soon. The 2MP C2580 and C2590 SoC specs look similar, and both sensors appear to be frontside-illuminated:

The 1MP C1680 has a similar block diagram and spec, but uses BSI technology:

CEVA Announces 4th Gen Vision Platform

PR Newswire: CEVA unveils the CEVA-XM4, its fourth-generation imaging and vision processor IP. CEVA-XM4 achieves up to 8x performance improvement with 35% greater energy efficiency, compared to the previous generation CEVA-MM3101.

The new IP's capabilities include real-time 3D depth map generation and point cloud processing for 3D scanning. In addition, it can analyze scene information using the most processing-intensive object detection and recognition algorithms, ranging from ORB, Haar, and LBP, all the way to deep learning algorithms that use neural network technologies such as convolutional neural networks (CNN). The architecture also features a number of unique mechanisms, such as parallel random memory access and a patented two-dimension data processing scheme. These enable 4096-bit processing -- in a single cycle -- while keeping the memory bandwidth under 512bits for optimum energy efficiency.

In comparison to today's most advanced GPU cluster (points to NVIDIA, I guess), a single CEVA-XM4 core will complete a typical 'object detection and tracking' use-case scenario while consuming approximately 10% of the power and requiring approximately 5% of the die area.

Taking computer vision one step closer to human vision, the CEVA-XM4 also supports a wide range of computational photography algorithms that enhance the video or image, including refocus, background replacement, zoom, super-resolution, image stabilization, noise reduction and improved low-light capabilities.

Omnivision Reports Quarterly Results

PR Newswire: OmniVision reports results for its fiscal quarter ended on January 31, 2015. Revenues for the quarter were $292.3M, as compared to $394.0M in the previous quarter, and $352.0M a year ago. GAAP net income in the quarter was $14.0M, as compared to net income of $28.0M in the previous quarter, and $30.6M a year ago. GAAP gross margin for the quarter was 22.1%, as compared to 22.0% for the previous quarter and 19.6% a year ago. The Company ended the period with cash, cash equivalents and short-term investments totaling $512.8M, a sequential decrease of $12.4M that resulted primarily from the use of cash in operating activities.

Based on current trends, the Company forecasts revenues for the next quarter in the range of $265M to $295M.

"As we continue to expand in Asia and find new opportunities in emerging economies, our results may continue to be volatile, at least in the near-term," said Shaw Hong, CEO of OmniVision. "However, it is important for us to participate in these markets as they are integral to our strategy. We remain confident about our long-term growth prospects."

SeekingAlpha publishes Omnivision's earnings call transcript. Shaw Hong says on the acquisition prgress: "I would like to comment on the proposed acquisition from Hua Capital Management Ltd., or HCM, a Beijing-based investment management company. A few months ago, we received a preliminary non-binding proposal letter from HCM, pursuant to which a group of investors led by HCM proposed us to acquire all of the outstanding share of common stock of the company in cash at US$29 per share. While the company's board of directors is reviewing and evaluating HCM's proposal, no decision has been made with respect to the proposed transaction. There is no assurance that this or any other transaction will be consummated."

Few other quotes from Aurelio Cisneros - SVP, Worldwide Sales & Sales Operations:

"In a special case, an innovative dual sensor main camera design featuring our 8-megapixel product line was launched by a major Tier 1 China customer. This represents a major potential technology shift where dual camera designs are built for a multitude of reasons such as improved sensitivity, slim compact design requirements and innovative 3D data capture. If this trend spreads, the total available opportunity in the already enormous handset market could surge to higher volumes because dual camera designs could be used for the main camera as well as for the front-facing camera slot. The benefits of a dual camera design are applicable to either case."

"In our third fiscal quarter, we shipped 198 million units as compared to 246 million units in our prior quarter. The average selling price in our fiscal third quarter was $1.46 as compared to $1.60 in our prior quarter. The drop in ASP was attributed to the product mix and aggressive price erosion in some product categories, primarily products shipped to the China market.

Unit sales of 8-megapixel and above represented approximately 10% of total shipments in the fiscal third quarter, compared to approximately 12% of total shipments in the prior quarter. Unit sales of 3-megapixel to 5-megapixel category represented approximately 41% of total shipments, as compared to 42% in the prior quarter for the same category. Unit sales of 2-megapixel and below represented approximately 49% of total shipments, as compared to 46% in the prior quarter."

TI Structured Light Kit Demo

Embedded Vision Alliance posted a TI Youtube demo of its structured light reference design platform, first announced half a year ago.

Framos Survey on Machine Vision Market

Framos 2014 survey on machine vision market paints an interesting picture. "The study is based on answers from 54 respondents based in 13 countries, 10 manufacturers and 44 users, with the main focus on German-speaking countries and Europe. Respondents were ranked in terms of relevance, based on the production or purchase volumes on which they were surveyed. 40% of the manufacturers produce up to 500 cameras a year, 20% over 10,000 cameras. Principal sales areas are North America and Asia. 80% of users purchase fewer than 100 cameras a year, only 5% require more than 1000 cameras. The main area of use is Europe, at 75%, due to the focus of this survey."

Here is the respondents view on CCD vs CMOS market evolution:

"In respect of sensor brands and their future use in imaging systems, camera manufacturers and users are agreed on their decline or rise. Sensors from Sony may have less usage in percentage, but continue to be the most trusted brand, with a predicted market share of 35% in 2016. In particular, considerable growth is expected for the brands Aptina and Truesense Imaging (both now under one roof, belonging to OnSemi), to a market share of 13% and 19% respectively."

AlliedSens to Represent Caeleste in Asia and America

AlliedSens becomes an official representative of Caeleste for custom design CMOS image sensor in Asia and America.

Patrick Henckes, CEO of Caeleste says “With the growing demand of custom design sensors, it is critical for Caeleste to rely on the best possible partners to expand our activities worldwide.“

Eiji Watanabe, CEO of AlliedSens says “We are happy to have Caeleste’s cutting edge custom image sensor into our product line in order to fulfill the rapid growing demand of high performance image sensors for industrial, medical and automotive applications in Asia and Americas.”

AlliedSens also represents CMOSIS, PixelPlus and Awaiba.

ISSCC 2015 Report - Forza & NHK Sensor, Samsung

Albert Theuwissen continues his report from ISSCC imaging session. The second part presents Forza and NHK 133MP 60fps sensor in 35mm FF format. 484 12b SAR ADCs support that high pixel rate. The pixel size is 2.45 um, 2×1 shared, 2.5T/pixel.

Samsung presents an always on low power sensor having an ultra-low power, low resolution, low quality mode, but waking up as soon as there is any movement in the scene and switching to a normal higher resolution, higher quality mode.

ON Semi Announces PDAF Technology

Business Wire: ON Semiconductor (Aptina) successfully demonstrated its second generation Phase Detect Auto Focus (PDAF) technology featuring a unique pixel micro-lens technology that enables fast focus at 25 Lux light levels. This PDAF has been implemented on a 13MP test chip with 1.1µm pixels and will be utilized in a new product introduction for mobile end-market customers later this year.

The PDAF technology uses two pixels to measure phase information from the target image. This phase information is then used to calculate the direction and the amount the lens needs to move to focus in less than 0.3 seconds, depending on focus actuator speed. Furthermore, ON Semiconductor has implemented a unique pixel micro-lens structure which maintains sensitivity in the PDAF pixels and ensures enough light is captured for PDAF to perform in low-light conditions as low as 25 Lux (similar to a dimly lit room). Competing technologies on the market employ methods that compromise pixel sensitivity and low light auto focus performance.

“ON Semiconductor’s PDAF technology is the result of innovations in both pixel and optical stack design,” said Shung Chieh, VP Mobile and Consumer Division for ON Semiconductor. “The outcome is a reduction of auto focus time by nearly a factor of four and delivers a great experience to the mobile device user. The leading low-light capability of our technology enables smartphone manufacturers to deliver a fast AF experience across all light levels to their customers.”

Toshiba Starts Mass Production of 8.1MP 1.12um Pixel Sensor

Business Wire: Toshiba started mass production of the previously announced 8MP, 1.12um pixel T4KA3 BSI sensor.

The sensor has a new low power circuit design that cuts power consumption to 46% that of Toshiba’s previous generation 8MP T4K35, an improvement over the earlier announcement that only stated 15% improvement. The sensor fits into 6.5mm square camera module, said to be the smallest in the class of 8MP sensors. The 8MP output at 30fps is possible using 2 Lane MIPI interface. The sensor incorporates "Bright Mode" that boosts image brightness up to four times, realizing HD video capture at 240 fps equivalent.

Allied Vision Confirms Sony CCD Discontinuation

Allied Vision, vendor of machine vision cameras, confirms Sony plans to stop CCD production. Michael Cyros, Chief Commercial Officer of the company, says in an interview: "Many in the industry had expected this to happen sooner or later. As you know, CMOS sensors have become more and more popular in the last couple of years. With greatly improved image quality, sensitivity, dynamic range and speeds, they have outpaced CCD technology of late. Sony’s decision is a logical step in that context to provide further capacity for CMOS growth."

Thanks to PJ for quoting the link in comments to the earlier post.

ISSCC 2015 Report

Albert Theuwissen starts publishing his overview of ISSCC 2015 imaging session. The first part talks about the new Sony 20MP/1.43um stacked sensor with two ADCs per column. The dual ADC can be used:

• for increased readout speed (120fps in 16MP/10b mode)
• for 3db noise improvement
• to achieve 1.3e- noise level by multiple reads
• simultaneous stills and video capture

More on Sharp Colorized IR Camera

Optics.org publishes an article on AIT and Sharp development of a colorized IR-illuminated camera: "The technology performs the color processing based on the weak correlation between the reflectance properties in the visible light region of objects and those in the infrared region, and obtains color images of the objects in the same or similar colors as the colors of the objects in visible light." The AIT colorizing algorithms require a modification of image sensor in the camera, but article gives no details on what kind of modification is needed.

Left- nightime image with no active IR illumination
Right - IR illumination is on and colorizing algorithms enabled

OmniVision and eyeSight Partner on Gesture Recognition

PR Newswire: OmniVision and eyeSight announced their co-development of next-generation gesture recognition technology in the industry's smallest form factor. eyeSight's gesture recognition technology, together with OmniVision's OVM6211 global shutter camera solution brings gesture recognition to a wide range of applications. The OVM6211 CameraCubeChip is a complete camera solution with a lens, sensor and packaging that uses a 3um OmniPixel3-GS global shutter pixel to capture 400 x 400 pixel video at 120fps.

"One of the frustrations with gesture recognition solutions in the market today is that they tend to work poorly in low light conditions," said Gideon Shmuel, CEO at eyeSight. "OmniVision's easily integrated OVM6211 global shutter sensor, working in tandem with our gesture recognition technology, can overcome the inherent challenges of recognizing gestures in difficult lighting conditions. This is critical for enhancing user experience and further broadening the scope of potential applications."

ON Semi Presents Imaging Solutions

Business Wire: ON Semiconductor presents 1/2.7-inch AR0230CS image sensor aimed to 1080p IP cameras with data rates up to 60fps. The sensor features 3um DR-Pix pixels, an integrated support for HDR video, improved motion compensation, advanced local tone mapping (ALTM), and digital lateral overflow (DLO) that minimizes motion artifacts and enhances HDR colors:

The company also presents AF controller LC898214XD and AF/OIS controller LC898123XC having quite a complex internal structure with 32b DSP and relatively large memories:

AF/OIS controller block diagram

Samsung Techwin Sells off Its Imaging Patents

PR Web: ICAP Patent Brokerage announces for sale of Samsung Techwin Co., Ltd. imaging patents. The portfolio is directed to methods and apparatuses for camera phone modules, including:

1. Nine patents directed to a lens system. These address reduction of an entire length, correction of distortion, increase in viewing angle, high resolution power, securing similar optical performance with fewer lenses, and reduction of variation of resolution power in automatic zooming, etc.
2. Six patents directed to the structure and driving mechanism of the lens module. These allow for the minimization of the size of an optical system, auto-focus or zooming module, cost-saving of materials, and facilitation of electrical connection, etc.
3. Four patents directed to the sliding structure. These address compact size with less thickness, reduction of friction, increase in sliding stroke, user convenience, smooth sliding operation, and better readability of a key pad, etc.
4. One software patent addresses improvement of user convenience in combining images for panorama image.

Kinect-2 ToF Sensor Demo

ISSCC posts the 2014 demo of Microsoft Kinect-2 ToF sensor on Vimeo:

Sony to Discontinue CCD Products?

This info originally appeared in a number of forums, such as this one, but it has no official confirmation. The info can be a hoax.

Said all that, it seems that Sony intends to discontinue its entire line of CCD products. The story started unfolding at the end of January with closing the production line for the complementary color filters and the Kagoshima 200mm wafer production line. Last week, Framos, one of the Sony CCD distributors, sent the following email to its customers:

"Dear Customer!

This message is to inform that SONY decided to discontinue all their CCD image sensor products! Attached please find the official letters from SONY!

The CCD End-of-Life plan of SONY is the following:
- customers need to order their last-time-buy quantities by end of Aug 2015
- SONY fab will shut-down March 2017
- deliveries can be made until March 2020

Please note: for new designs, especially with CMOS sensors we can offer in-depth technical support and are able to provide reference designs etc to you under certain conditions. Especially we are already working on reference designs of the upcoming IMX25x-series.

If there are any questions or any further need, please contact your FRAMOS account manager, or sales@framos.com as soon as possible!"

Before that, Sony reportedly sent to its customers two announcements dated by the end of January:

Thanks to NA for the info!

Update: According to Sony Semiconductor flyer dated by Feb. 4, 2015, CCD used to be manufactured at Kagoshima and Kumamoto fabs:

Sony Low-Cost Global Shutter Sensor

Sony IMX249 low-cost global shutter sensor has already appeared in Point Grey cameras. Now Sony publishes an official IMX249 flyer. There is a difference between Point Grey's announced speed of 41fps and Sony's official 30fps at full 2.35MP resolution. However, on the sample photograph Sony states 60fps for FHD resolution:

Omnivision Applies for Si Photomultiplier Patent

Omnivision patent application US20150041627 "Partitioned silicon photomultiplier with delay equalization" by Eric A. G. Webster proposes 3-wafer stacked photomultiplying imager. The top wafer has photomultiplier diodes with quenchers, the bottom wafer has the readout electronics, while the mid interposer-wafer ensures equal-length interconnects to ensure the equal delay between the PDs in each pixel, and so improve the timing resolution:

poLight TLens Datasheet

poLinght publishes its TLens brief datasheet.

The POL10-3215 TLens Silver features:

• Suitable of 1/4-inch sensors
• Supports 3MP to 8MP resolutions
• Easy to add on camera module
• ESD and stray light protection option
• Package size: 4.8mm x 4.8mm
• Minimize additional thickness at fixed focus camera module level
• Centering accuracy < +/-0,05mm
• Response time down to 1ms
• Power consumption as low as 1mW
• Small footprint 3.2mm x 3.2mm
• Focus distance from infinity down to 10cm
• No sensitivity to gravity
• Voltage range at Bottom electrode: 0V to 5V
• Voltage at Top electrode: 0V to 40V
• Maximum applied current: 1mA
• Operating temperature from -10°C to 60°C

Qualcomm Research Demos

Qualcomm Research publishes Youtube videos of its 8-axis video image stabilization:



and scene classification:

Videantis and Gestigon Partner

Videantis and Gestigon partner to team Videantis’s low-power vision processing IP platform with Gestigon’s tracking and gesture recognition. The main aim of the cooperation is to bring gesture and pose recognition technology to the automotive market, which is rapidly adopting a wide variety of vision-based technology. More and more image sensors are entering cars, including depth-sensing technologies. These sensors can be used for skeleton tracking inside and around the car. The driver and passenger’s poses, behavior, and signals can be used for “classical” gesture control but also for more advanced, UX-focused comfort and security features.

“Videantis is the perfect partner for gestigon”, gestigon CEO, Moritz von Grotthuss, says. “Their scalable, efficient processor IP already powers millions of cars on the road, and is a great target for low-power implementations of our algorithms that enable an innovative state-of-the-art user experience. The partnership is a perfect match -- both companies excel in their field of work and have already built a significant customer network among global automotive Tier 1s and OEMs.”

Videantis products

Broadcom Proposes to Involve Internet Data to In-Camera Processing

Broadcom patent application US20150042843 "Systems and methods for improving images" by Ike Aret Ikizyan and Noam Sorek proposes to use network connectivity to improve smartphone camera image processing:

"Systems and methods are provided for improving image accuracy and/or perceptual quality by using supplemental data from sources external to an image sensor and/or a light sensor of an imaging device without requiring users to manually enter this data. The supplemental data, for example, may include lighting condition information retrieved from databases that provide general weather information, color temperature models, images in a similar environment, control setting information of those images, statistical information of those images, and/or any other information that may be used to improve the accuracy and/or perceptual quality of an image that was or will be captured. In some aspects, if the image has not yet been captured, the supplemental data may be used by the imaging device to adjust appropriate control settings (e.g., a color balance setting, a flash setting, an aperture setting, a shutter speed setting, an exposure compensation setting, an ISO setting, a light frequency setting, a noise reduction setting, a sharpening setting, etc.) such that the image can be captured accurately and/or with higher perceptual quality. In some aspects, if the image has already been captured, the supplemental data may be used to adjust the image itself so that it accurately reflects the environment in the image and/or improves the perceptual quality.

...environment identification module 202 may determine the environment using a global positioning system (GPS) sensor, a Wi-Fi-based positioning system (WPS) sensor, a gyroscopic sensor, an accelerometer, a magnetometer, and/or other sensors. One or more of these sensors may be part of imaging device 102."

One could expect such kind of patents come from Apple, Broadcom is a surprising source for that.

Samsung Proposes Color Splitter

Samsung patent application US20150041940 "Image sensor and electronic device including the same" by Gae Hwang Lee, Kyu Sik Kim, and Yong Wan Jin proposes a micro color splitter instead of color filter in front of the pixels, somewhat similar to what Panasonic presented 2 years ago. Samsung's application shows nano-structures splitting the light:

Update: Incidentally, Nokia patent application US20150042850 "Apparatus and a method for imaging" by Radu Bilcu and Martin Schrader published almost simultaneously with the Samsung one. Nokia presents a hierarchical color splitter between different image sensors or parts of them:

Interview with David Stork, Rambus

Image Sensors Conference, to be held in a month in London, UK, publishes an interview with one of its presenters, David Stork, Fellow and Research Director of the Computational Sensing and Imaging Group at Rambus Labs. The interview is mostly devoted to the lensless image sensor with grating instead of lens, already shown earlier on other conferences. Few quotes:

"In short, the grating produces an image blob on the sensor array that is meant for computers, not humans and unlike in a traditional camera, in our lensless smart sensor the image is computed using special algorithms. I’m particularly intrigued and encouraged by our most recent work on designing gratings and subsequent digital processing to address particular image sensing tasks such as image change detection, visual flow estimation, face detection, object tracking, people counting, and so on."

"Several facts are clear, though: our resolution is lower than that typically discussed at Image Sensors—thousands or tens of thousands of effective pixels, not millions or billions. Moreover, because of our sensor small size, we do not collect much light compromising our low-light sensitivity. We are confident that there are many large application areas for our sensors, even given these constraints."

Sony to Spin-Off Image Sensor Business?

Among many comments on yesterday's Sony Corporate Strategy meeting, Financial Times stands out claiming that Sony will likely spin off its image sensor business in future. Here is the quote, believe it or not:

"Sony plans to make the department responsible for the Walkman brand a separate subsidiary in October. This would likely be followed by the spin-off of the device division including image sensors in the future, according to Kazuo Hirai, [Sony] chief executive."

Update: EETimes too quotes FT article on the image sensor business spin-off.

Update #2: Nikkei quotes Sony CEO intending to spin-off most businesses and keep the HQ small:

"We will spin off businesses that are currently run by the main body one by one," Hirai said. "(By leaving only departments related to management, R&D, etc to the main body), we will make small headquarters to increase the speed of making management strategies and decisions."

TI ToF Solutions

TI has updated its ToF imaging web page. The new version includes a fairly detailed ToF camera system design guide, and a comparison of different 3D depth imaging approaches:

MIPI Officially Releases CSI-2 v1.3

Business Wire: MIPI Alliance introduces CSI-2 v1.3, an update to its Camera Serial Interface (CSI) spec. The new interface operates CSI-2 on either of two physical layer specs: D-PHY, which CSI-2 has used traditionally, as well as MIPI C-PHY, a new PHY that MIPI Alliance released as v1.0 in September 2014. Products may implement CSI-2 solutions using either or both PHYs in the same design.

MIPI C-PHY uses 3-phase symbol encoding of about 2.28 bits per symbol to transmit data symbols on 3-wire lanes, or “trios,” with embedded clocking, facilitating longer trace reach and maximizing camera port configurations on mobile platforms. MIPI CSI-2 v1.3 with C-PHY provides performance gains, increased bandwidth delivery of 22.7 Gbps over four lanes at 2.5 Gsps (Giga-symbols per second) for realizing higher resolution, better color depth, and higher frame rates on image sensors while providing pin compatibility with MIPI D-PHY. Popular imaging formats including 4K video at 30fps using 12 bits per pixel may be delivered using a single MIPI C-PHY lane.

CSI-2 with D-PHY v1.2 can achieve a peak transmission rate of 2.5 Gbps over a single lane or 10 Gbps over four lanes. A 12bpp, 30 FPS 4K video can be transmitted using two data lanes.

Sony Announces New Sensors

Sony announces 4 new image sensors.

IMX290LQR and IMX291LQR are 1/2.8-inch, 2.3MP 2.9um BSI pixel CMOS sensors featuring an improved IR sensitivity and WDR modes. The new 2.9 µm BSI pixel is said to provide two or more times the sensitivity in the visible-light region and three or more times the sensitivity in the NIR region than that of the existing Sony product (IMX236LQJ). In addition, two types of WDR are provided: IMX290LQR has multiple exposure and DOL (Digital Overlap) -WDR function, while IMX291LQR has only multiple exposure WDR.

The multiple exposure-type WDR outputs one set of two or four frames with different exposure times. The gain can also be set separately for each frame in addition to the exposure time.

The DOL-type WDR function outputs the data for up to three frames with different storage times line by line. By performing special signal processing with an ISP or other device at the image sensor rear-end, this is said to improve the picture quality under low illumination compared to the multiple exposure-type WDR function.

Sony publishes quarter-resolution low-light images comparing the new sensors with the older IMX236LQJ, both shot at 0.08 lux, F1.4 lens 1080p30 mode (click to expand):

Older IMX236LQJ, gain 48dB

New IMX290LQR, gain 63dB

2.35MP IMX302LQJ is yet another member of 5.86um pixel-based global shutter sensor (Pregius) aimed to industrial and machine vision market. The sensor's speed is 64.1fps with 12b ADC.

Update: Talking about Sony global shutter sensors, IMV Europe quotes Thad Smith, director of business for image sensor products at ON Semiconductor: "IMX174 can indeed be considered a disruptive technology. The technology demonstrates the possibilities of the new charge-based pixel approach and sets a new benchmark for image quality. However, while IMX174 is a single part in the portfolio, On Semiconductor’s image sensors offer a larger range of solutions for the industrial market. This is a key differentiator." In recent months, Sony appears to release quite a few IMX174 derivatives featuring the same 5.86um global shutter pixels. And there is a 2nd generation of its global shutter pixel coming.

The 6.84MP ICX816AQG 1/1.1-inch CCD targets traffic monitoring cameras and features 3.69um pixels with high sensitivity and low smear:

Update: For some reason, as of Feb. 21, 2015, Sony has withdrawn the ICX816AQG CCD announcement. The links to it are broken.

Omnivision Parnters with Almalence

PR Newswire: OmniVision is collaborating with Almalence to bring DSC-like performance to mobile applications. The coupling of 10MP OV10823 and 13MP PureCel OV13850 image sensors with Almalence's computational imaging algorithms allows for the capturing of high resolution images or videos in low-light conditions.

"When using Almalence's algorithms, our high performance OV13850 PureCel image sensor can enable high quality video recording in full resolution at 30 frames per second for zero shutter lag, and also can support high frame rate 1080p HD video at 60 frames per second with electronic image stabilization," said Tehzeeb Gunja, senior partnership and business development manager at OmniVision.

1.4um OmniBSI-2 pixel OV10823 performance is augmented by Almalence's low-light enhancement technology, which can enable 10dB reduction in noise without washing out desirable image details. Cameras using super resolution can experience a 1.5 to 2x effective resolution increase. Almalence's super resolution and low-light enhancement technologies are pure software solutions.

"As a result of this technology partnership between Almalence and OmniVision, consumers can benefit from the advantages of marrying advanced hardware and software solutions," said Eugene Panich, CEO of Almalence. "Our companies have developed complementary technologies and our passions are very much aligned. We are delighted to see our computational imaging solution paired with image sensors that operate at this level of performance."

Sony FY2015-17 Strategy

Sony mid-term corporate strategy for FY2015-17 declares its image sensor business one of the main growth drivers. "In Devices, Sony aims to further bolster its competitive edge in the area of CMOS image sensors by investing to increase production capacity and enhance R&D."

From Sony strategy presentation:

Pixart Reports its Sales Down

Pixart Q4 2014 earnings report reveals that "the revenue in 2014 Q4 decreased by 7.7% QoQ to NT$1,097.0 million. Due to favorable product mix and the appreciation of US dollar, 2014 Q4 gross margin increased to 47.8% from 46.4% in 2014 Q3." The company's investor presentation shows the segment sales proportions:

Qualcomm Presents New Camera Features for its AP

Qualcomm Snapdragon blog presents few camera features implemented in its recent ISP pipeline. The presented features exploit modern sensors ability of taking few snapshots in a quick succession, and use it for flash and no-flash images combination, all in focus feature, and ghost-free HDR imaging. A new Optizoom feature is supposed to rival the optical zoom quality "by capturing 12 images in a burst capture, and then creating a sharper image by comparing the differences and similarities in the micro movements of the pixels you’ve captured:"

For some reason, Optizoom presented a year ago works in a different way:

Microspectrometer Use Cases

nanoLambda publishes a Youtube animation showing many possible uses for its Apollo microspectrometer:

ISO-less Camera Presentation

Aggregate.org publishes slides on ISO-less camera by Hank Dietz and Paul Eberhart from University of Kentucky presented at SPIE Electronic Imaging conference:

FT: Sony Aims at Automotive Cameras

Financial Times: Sony owns about 40% market share in CMOS sensors for smartphones, digital cameras and tablets, followed by roughly 16% share of OmniVision and 16% share of Samsung, according to 2014 TSR report. But Sony only ranks fifth with a 5% share in automotive image sensors.

Shigeo Ohba, GM of Sony’s image sensor business, said "We have to be No 1 in automotive sensors" by the time self-driving cars are expected to hit the roads in early to mid-2020s. Recently, Sony established a new business group targeting image sensors for cars and is actively hiring engineers. “We were latecomers in smartphone cameras but we still created this market. Now we can do the same with automotive cameras,” said Shoichi Kitayama, Sony’s GM of the automotive group.

Sony CEO Kazuo Hirai is expected to disclose his automotive image sensor plans on the strategy briefing on Wednesday.

Toshiba Presents its Camera Modules for Google Project Ara

Modular Phones Forum: Toshiba presents its work on camera module design for Google Project Ara modular phones. Shardul Kazi, SVP from Toshiba America, says:

"We’ve been engaged in Ara program for more than a year now. Engagement started because of Toshiba’s involvement in the standards committee of MIPI UniPro, and UniPro is a backbone of the Ara phone. So we know UniPro very well. To make developers life easier we developed bridges (two types of bridges and the switch chip), so module developers can use those chips and they don’t need to learn about complexity of UniPro protocol."

So, Toshiba has started from the camera modules as Phase 1 of its Ara designs:

Toshiba has created a special web site devoted to its Ara modules progress and featuring a Youtube video with the first 5MP module:

Backside Plasma Damage Reduction

ST and Commissariat A L'energie Atomique Et Aux Energies Alternatives patent application "Back side illumination image sensor with low dark current" by Jean-Pierre Carrere, Patrick Gros D'aillon, Stephane Allegret-Maret, and Jean-Pierre Oddou proposes an additional conductive layer on the BSI image sensor backside to screen the back of the imager from plasma processing-introduced defects:

"The degradation comes from plasma-based treatments used during the treatments of the back side of the sensor, particularly for forming the back side metallic contact pads, of aluminum for example, and for forming the layer protecting the colored filters, which aims to protect these filters during the slicing of the sensors along the slicing lines of the semiconductor wafer.

More precisely, the plasma charges the silicon oxide-silicon nitride stack triggering a trapping of positive charges (holes) in the anti-reflective silicon nitride layer, holes which will then migrate to the interface with the substrate to create a dark current by recombination with electrons."

In order to reduce the dark current, the application proposes: "an additional layer situated above the anti-reflective layer and including at least a lower part containing hydrogenated amorphous silicon (a-Si:H) or hydrogenated amorphous silicon nitride (a-SiNx:H), in which the Si/N ratio of the number of silicon atoms per cubic centimeter to the number of nitrogen atoms per cubic centimeter is greater than 0.7, and preferably greater than or equal to 1.2.

The function of this additional layer is to absorb, or even to stop, ultraviolet radiation, and/or to evacuate the trapped charges in the anti-reflective layer during subsequent plasma treatments, which will therefore reduce the residual positive charges possibly trapped in the anti-reflective layer and consequently reduce the dark current.

The thickness of this lower part of the additional layer is preferably less than a few tens of nanometers, for example a hundred nanometers, or even less than ten nanometers."

Apple Proposes 3D-integrated Pixel with Global Shutter

Apple patent application US20150035028 "Image sensor with buried light shield and vertical gate" by Xiaofeng Fan, Philip H. Li, Chung Chun Wan, Anup K. Sharma, Xiangli Li proposes a 2-layer stacked global shutter pixel:

The pixel is split into 2 layers: "a buried light shield 710 separates the sensing layer 702 into a first substrate layer 738 and a second substrate layer 740. The sense region 716, the storage region 712, and the pixel transistors can be in the second substrate layer 740 while the photodetector 708 resides in the first substrate layer 738. The buried light shield 710 prevents unwanted charge accumulation in the storage region 712 and in the sense region 716." The vertical transfer gate 714 is made of poly in the trench and has a contact pad 718.

Himax Reports 2014 Revenues

Himax reports its 2014 business results: "Revenues from CMOS image sensors also experienced significant growth in 2014 and should continue to be a fast growing product area for Himax in 2015. Sales of the Company’s 2 and 5 megapixel products were particularly strong, mainly due to robust demand from several international customers as well as from Chinese white-box customers. Looking into 2015, Himax expects its sales of 8-megapixel sensors, now the mainstream design for smartphones, to accelerate while high end 13 megapixel sensors to start mass production later in the year. This puts Himax firmly on the map as one of a small handful of companies capable of offering a comprehensive product portfolio for smartphone cameras.

In addition to the a fore-mentioned consumer applications, Himax is also making significant progress in CMOS image sensors for non-consumer applications, which typically enjoy higher margins and have relatively little direct competition. Himax already started initial shipments to a Korean automotive end customer and is actively engaging more module houses to penetrate into Chinese before-market-installation automotive applications. As the Company grows sensor sales for automotive and other non-consumer applications and continues to turn the bulk of its CMOS image sensor sales from 2 and 5 megapixels to higher end, higher margin 8 and 13 megapixels, Himax expects its CMOS image sensor business to accelerate in 2015."

"CMOS Image sensor sales nearly doubled year over year."

"Furthermore, the Company continues to partner with numerous industry leading companies using its cutting edge and industry-dominant wafer level optics, or WLO, for the development of ...technologies of the future, namely array cameras. ...To meet the anticipated demand growth for its LCOS and WLO products, the Company is expanding its production capacity starting in Q1 2015."

From Himax Feb 2015 investor presentation:

Himax customers

Canon Shows 120MP Sensor Prototype

DPReview: Canon exhibits 120MP, 2.2um pixel image sensor prototype at CP+ photography expo in Yokohama, Japan. To boost the sensor speed, Canon utilizes "parallel processing, which reads signals at high speed from multiple pixels." It seems to be the same sensor that Canon announced 4.5 years ago (see the official announcement here).

CMOSIS Rings Bell at Euronext

Euronext TV: CMOSIS CEO, Luc Demey, rang the closing bell at Euronext Stock Exchange in Brussels, Belgium:

Yole on 3D Hybrid Pixels

EETimes publishes an interview with Pierre Cambou, one of the authors of Yole's recent report on CMOS sensor market. Yole predicts that the next trend will likely be 3-D Hybrid CMOS image sensors that by-pass the TSVs for high precision direct copper-to-copper wafer bonding, allowing per-pixel interconnects between the wafers:

Another interesting point is that Yole seems to be too conservative estimating the image sensor market growth in its reports: