Saturday, August 19, 2017

Sony Proposes BCMD with Reduced Dark Current

Sony patent application US20170229493 "Pixel circuit and imaging apparatus" by Kouichi Harada and Toshiyuki Nishihara proposes dark current reduction in bulk charge modulated device (BCMD).

"...the above image sensor needs to read 1000 times in one frame for example, and has problems of increase in read voltage and increase in read time. Also, as the number of reads per frame increases, dark current of the FD increases proportionally. As a result, the dark current of the FD becomes the main component of the dark current of the pixel. The dark current of the FD is unable to be reduced easily, and thus even if the conversion efficiency can be set to 600 μV/e−, the accuracy of detecting one photon is reduced. If there is no FD, the accuracy of detecting one photon is improved.

...According to the present technology, it is possible to obtain an excellent effect that the dark current of the FD in the image sensor is eliminated, and that the conversion efficiency of converting the electric charge to a voltage can be improved.

Thursday, August 17, 2017

TrinamiX 3D Sensor Paper

BASF spinoff Trimamix publishes an paper "Focus-Induced Photoresponse: a novel optoelectronic distance measurement technique" by Oili Pekkola, Christoph Lungenschmied, Peter Fejes, Anke Handreck, Wilfried Hermes, Stephan Irle, Christian Lennartz, Christian Schildknecht, Peter Schillen, Patrick Schindler, Robert Send, Sebastian Valouch, Erwin Thiel, and Ingmar Bruder.

"Here we introduce Focus-Induced Photoresponse (FIP), a novel method to measure distances. In a FIP-based system, distance is determined by using the analog photoresponse of a single pixel sensor. This means that the advantages of high-density pixelation and high-speed response are not necessary or even relevant for the FIP technique. High resolution can be achieved without the limitations of pixel size, and detectors selected for a FIP system can be orders of magnitude slower than those required by ToF based ones. A system based on FIP does not require advanced sensor manufacturing processes to function, making adoption of unusual sensors more economically feasible.

In the FIP technique, a light source is imaged onto the photodetector by a lens. The size of its image depends on the position of the detector with respect to the focused image plane. FIP exploits the nonlinearly irradiance-dependent photoresponse of semiconductor devices. This means that the signal of a photodetector not only depends on the incident radiant power, but also on its density on the sensor area, the irradiance. This phenomenon will cause the output of the detector to change when the same amount of light is focused or defocused on it. This is what we call the FIP effect.

Qualcomm on VR Motion Tracking Setup

Qualcomm presentation "On-device motion tracking for immersive mobile VR" discusses 4-camera setup in the company's VR headset:

First Report on CIS Reproducibility, Variability and Reliability

Albert Theuwissen releases the first report on "Reproducibility, Variability and Reliability of CIS" in a 5-year series. The 175-page report contains 118 figures and 98 tables with data on QE, FPN, DSNU, Qsat, DR, SNR, DC, and more.

Wednesday, August 16, 2017

Coolpad CEO Quantifies Dual Camera Advantages

Coolpad, one of the large China-based smartphone makers that uses 13MP monochrome + 13MP RGB dual camera in its devices, says about the benefits of that configuration:

"The two lenses may look the same, but they have very different functions. One shoots in RGB to produce a color image, while the other takes care of the monochrome images. The monochrome lens brings out the detail, and engage more light than the RGB lens when in low-light condition, which takes care of the colors. The Dual camera 2.0 technology in Cool Dual actually enhanced the overall clarity of the image by 20%, help reduce image noise by 8% and improved brightness by 20%. “With these, we believe the real dual 13MP cameras brings us smart framing and the 6P lens gives customers the best quality of pictures”, said Jeff Liu, Coolpad Group CEO."

Digitimes on Automotive LiDAR Adoption

Digitimes Research comes up with its analysis of LiDAR adoption in the car industry, forecasting first LiDAR-equipped production cars appearing this year:

"...Audi will take the initiative to launch car models equipped with LiDAR sensors in 2017 and Mercedes Benz, Cadillac, Ford and Volvo are expected to follow suit in 2018-2019...

...only one million LiDAR sensors worth US$200 million will be shipped globally in 2019. However, global shipment value for LiDAR sensors will fast grow to US$500 million in 2024 along with decreasing cost and increasing adoption.

For Level 2 self-driving... LiDAR sensors are required to detect objects as far as 100-150 meters ahead. For Level 3, the required range is 200-300 meters

...Velodyne LiDAR and Quanergy Systems, and Germany-based Ibeo Automotive Systems are the main vendors globally of LiDAR sensors, with the former two focusing on solid-state models to reduce LiDAR sensor sizes.

Qualcomm Unveils Structured Light Depth-Sensing Platform

Qualcomm announces an expansion to the Qualcomm Spectra Module Program to incorporate a biometric authentication and high-resolution depth sensing for a broad range of mobile devices and head mounted displays (HMD). This module program is built on the 2nd generation Spectra embedded ISP family.

Now, the camera module program is being expanded to include new camera modules capable of utilizing active sensing for biometric authentication, and structured light for a variety of computer vision applications that require real-time, dense depth map generation and segmentation.

The low-power, high-performance motion tracking capabilities of the Qualcomm Spectra ISP, in addition to optimized simultaneous localization and mapping (SLAM) algorithms, are designed to support new extended reality (XR) use cases for VR and AR applications that require SLAM.

It also features multi-frame noise reduction for superior photographic quality, along with hardware-accelerated motion compensated temporal filtering (MCTF), and inline electronic image stabilization (EIS) for superior camcorder-like video quality.

The Spectra family of ISPs and new Spectra camera modules are expected to be part of the next flagship Snapdragon Mobile Platform.

Qualcomm Emerging Vision Technologies presentation gives some use cases for its 3D depth sensing module.

Qualcomm also publishes a Youtube demo if its structured light depth sensing module:

Tuesday, August 15, 2017

Pinnacle Introduces HDR ISP Core

Pinnacle Imaging Systems launches Denali-MC HDR ISP IP core said to preserve a scene’s color fidelity and full contrast range throughout the tone mapping process, all without producing halos, color shifts, and undesired motion artifacts.

Denali-MC provides a 16-bit data path capable of producing 100 dB or 16-EV steps of dynamic range. Denali-MC HDR IP completely eliminates halo artifacts and color shifts, and mitigates the ghost artifacts and transition noise often seen when merging multiple exposures. This allows Denali-MC to capture up to four exposure frames from 1080p video at 120 fps, while merging and tone mapping at 30 fps in real time. For applications requiring faster output frame rates, Denali-MC also supports a two frame merge mode exporting at 60 fps. Furthermore, Denali-MC can support up to 29 different CMOS sensors, including 9 Aptina/ON Semi, 6 Omnivision and 11 Sony sensors, and 12 different pixel-level gain and frame-set HDR methods, and is said to be easily ported to the most widely-used logic platforms.

HDR-Specific Features:
  • Advanced motion compensation algorithms virtually eliminate HDR merge artifacts and transition noise
  • Proprietary Locally Adaptive Tone Mapping technology preserves color fidelity through the entire tonal range without creating halo artifacts or color shifts
  • Automatic EV bracketing
  • Automatic or manual contrast adaptation for global or local video correction
  • React concurrent still frame and video capture feature, non-destructively extracts four source LDR Bayer images, merged Bayer HDR, tone mapped Bayer or HDMI RGB still frames without interrupting video
  • Ability to capture separate HDR and tone mapped output video streams concurrently (ideal for ADAS applications)
  • Two or four frame multiple exposure merge (with Sony IMX290 implementation)
  • HDR + Low illumination capabilities with Sony IMX290 sensor enable 24/7 round the clock video capture capabilities for any contrast and lighting condition

At Fairchild Imaging, we have been very impressed with the Denali-MC ISP state of the art locally adaptive tone mapping (LATM) functionality,” said Vern Klein, Director of Sales and Marketing Fairchild Imaging. “I’ve seen first-hand how they can make a great sensor perform even better in its native WDR mode. Camera manufacturers will benefit from this technology, which provides high quality HDR functionality without requiring companion chips or additional hardware cost to support the algorithms.

Pinnacle’s new Denali-MC HDR ISP is a significant achievement addressing HDR video requirements in surveillance, monocular camera automotive markets and machine learning with its customization, artifact compensation, color accuracy and quantifiable high dynamic range of 100 dB,” said Paul Gallagher, image sensor industry veteran and futurist. “Camera system developers in these markets would benefit from utilizing these attributes of the Denali-MC ISP as a standalone ISP or integrate Pinnacle’s HDR IP blocks within their existing ISP.

Pinnacle also publishes a video demo of its HDR capabilities (short version, long version):

Technavio Low-Light Imaging Market Report

BusinessWire: Technavio comes up with another unusual report "Global Low Light Level Imaging Sensors Market 2017-2021." Few quotes:

"In 2016, the night vision devices segment accounted for close to 55% of the total revenues due to high adoption of low light level imaging sensors by the defense sector. The increasing focus on reducing road accidents drives the demand for night vision systems. Night vision systems are offered as built-in systems by Audi, BMW, and Toyota. The market is expected to grow at a CAGR of close to 16% during the forecast period.

The global low light level imaging sensors market by cameras contributed 25% of the total revenues in 2016. These sensors are used in applications such as home security cameras, small business monitoring, and infrastructure security.

In 2016, the global low light level imaging sensors market by optic lights accounted for around 13% of the total revenue in 2016. These are widely used in lighting, decorations, and mechanical inspections of obscure things. Optic lights save space and provide superior lighting, and are therefore used in vehicles. Low light level imaging sensors are crucial components of optic lights as these sensors intensify the range of light.

Saturday, August 12, 2017

SMIC Reports Rising CIS Sales

SeekingAlpha publishes SMIC Q2 2017 earnings call saying that the foundry sees a significant growth in its CIS process volume:

"By device, most of our year-on-year revenue growth came from CMOS image sensors, NOR Flash, application processors and the power ICs.

And for the existing customer, for... the CMOS imager applications, we also see the recovery, yes.

Friday, August 11, 2017

Sony Kumamoto Earthquake Video

Imaging Resource publishes a video on Sony Kumamoto fab earthquake damage and a recovery effort that brought fab back to life in a so sort time:

LG Announces Smartphone Camera with F1.6 Lens

BusinessWire: The upcoming LG V30 smartphone is said to have camera module with F#1.6 lens, said to be the brightest in smartphones. I wonder what is the effective aperture of the pixels in the sensor and whether it makes use of the so bright lens.

Thursday, August 10, 2017

Sony introduces 2.9um Pixel 1080p60 HDR Sensors

Sony introduces IMX307 and IMX327 sensors featuring 2.9um pixels. The 1080p60 sensors support multiple-exposure and DOL-HDR and have 12b ADC.

Wednesday, August 09, 2017

ESPROS CEO Interview

Autosens publishes an interview with ESPROS CEO Beat De Coi. Few quotes:

Q: Why did you decide to buck the trend in semiconductors to have your own foundry?

A: Simply there was or is no silicon CMOS technology available which offered high Quantum Efficiency in NIR in combination with high performance CCD. Our backside illumination technology OHC15L offers what is needed for powerful LiDAR, TOF and in general ultrafast gated imagers.

Q: Your session on time-of-flight sensors is about next generation technology – what’s different about it from existing ToF?

A: - Quantum Efficiency is greater than 75% @905nm
- active ambient light suppression
- single pulse operation vs. continuous light modulation in previous generation systems

Q: What is the timescale of this technology coming to the market?

A: Early 2018

Tuesday, August 08, 2017

Stacking Technologies Overview

Phil Garrou IFTLE 346 overviews image sensor staking history since Sony introduced it in 2013:

Grenoble-Lyon Imaging Valley Guide

EETimes publishes a useful article "Engineer's Guide to Imaging Valley" by Junko Yoshida. Most of the France’s imaging companies are concentrated in the Grenoble/Lyon area:

The article gives a very nice overview of the history and contemporary state of the French Imaging Valley.

Thanks to JD for the link!

ON Semi Exits Mobile CIS Business, Licenses-Out Related Patents

SeekingAlpha Q2 2017 earnings call transcript gives some interesting info about ON Semi image sensor business:

"During the second quarter, we exited the mobile image sensor market as the margin profile for that business was not comparable with our target financial model.

Furthermore, we monetize the value of highly differentiated mobile imaging technology, through an intellectual property licensing agreement with a third-party. We have excluded the gain of approximately $24 million related to this transaction from our second quarter non-GAAP results.

Second quarter free cash flow and operating cash flow included approximately $24 million from a licensing arrangement related to the mobile image sensor business.

For the second quarter, we again posted strong growth in our CMOS image sensor business for viewing and ADAS applications. We continue to gain market share in automotive image sensors and our design win pipeline for our CMOS image sensors for automotive applications continues to grow at a rapid pace.

We continue to see strong growth in machine vision applications with our PYTHON line of CMOS image sensors. As I indicated earlier, we are engaging at the very early stage with key players in artificial intelligence for machine vision and robotics applications.

The company's spreadsheet shows the proportion of image sensor business over the years:

Sunday, August 06, 2017

UNC Chapel Hill Asynchronous Sensor Architecture

University North Caroline at Chapel Hill presents "A Frameless Imaging Sensor with Asynchronous Pixels: An Architectural Evaluation" by Montek Singh, Pintian Zhang, Andrew Vitkus, Ketan Mayer-Patel, and Leandra Vicci at 2017 IEEE Intl Symposium on Asynchronous Circuits and Systems.

"The goal of this work is to develop a novel CMOS camera sensor that provides frameless capture, and has significantly higher dynamic range, finer color sensitivity, and lower noise as compared to the current state-of-the-art sensors. The strength of the approach lies not in developing new types of photodetectors or amplifiers, but in the manner in which information is extracted from the pixel sensor, transported to the processing logic, and processed to yield intensity values. At the heart of the sensor is an asynchronous network to transport events from the pixel sensors to the off-grid processing circuitry. The asynchronous nature of pixel communication is the key to achieving frameless image capture."

The UNC Chapel Hill research group is looking for industrial partners who might be interested in the IP and in partnering with the group for further development.

Saturday, August 05, 2017

Chronocam in Novus Light

NovusLight publishes an article "Vision Inspired by Biology" based on the talk with Luca Verre, the CEO and co-founder of Chronocam. Few quotes:

"Based on the new technology concept, the company recently released a QVGA resolution (320 by 240 pixels) sensor with a pixel size of 30-microns on a side and quoted power efficiency of less than 10mW.

Because the sensor is able to detect the dynamics of a scene at the temporal resolution of few microseconds (approximately 10 usec) depending on the lighting conditions, the device can achieve the equivalent of 100,000 frames/sec.

In Chronocam’s vision sensor, the incident light intensity is not encoded in amounts of charge, voltage, or current but in the timing of pulses or pulse edges. This scheme allows each pixel to autonomously choose its own integration time. By shifting performance constraints from the voltage domain into the time domain, the dynamic range is no longer limited by the power supply rails.

Therefore, the maximum integration time is limited by the dark current (typically seconds) and the shortest integration time by the maximum achievable photocurrent and the sense node capacitance of the device (typically microseconds). Hence, a dynamic range of 120dB can be achieved with the Chronocam technology.

Due to the fact that the imager also reduces the redundancy in the video data transmitted, it also performs the equivalent of a 100x video compression on the image data on chip.

The company announced it had raised $15 million in funding from Intel Capital, along with iBionext, Robert Bosch Venture Capital GmbH, 360 Capital, CEAi and Renault Group.

From the company presentation at Event-based Vision Workshop 2017:

Friday, August 04, 2017

Smartphone Cameras Go from Double to Triple

GSM Arena reports that the upcoming Vivo X20 (Xplay 7) smartphone features 5 cameras - double on the front and triple on the rear sides:

IISW 2017 Papers On-Line

International Image Sensors Workshop 2017 held in Hiroshima, Japan on May 30-June 2, publishes all 110 presented papers on-line: 61 regular papers, 45 posters, and 4 invited papers. Some of the papers also have presentation slides published. It's a very good read over the weekend!

Thursday, August 03, 2017

Yole Uncooled Thermal Imaging Report

Yole Developpement releases "Uncooled Infrared Imagers market & technology trends" report: "Today the uncooled IR camera market is showing an 8% CAGR between 2016 and 2022 reaching almost US$ 4.4 billion at the end of the period. Only few players control this industry. In 2016, two leading companies, FLIR and ULIS, both with different market strategies and solutions, owned more than 75% of the total market (in volume).

Besides FLIR and ULIS, many others players are also benefiting from IR imaging market growth:

• SEEK Thermal has introduced its new, higherperformance RevealPRO, and CompactPRO products as the company moves from consumer products to more high-end products.

• Players such as BAE Systems or Leonardo DRS are benefiting from the defense market growth cycle that could still last for a few more years.

• Newcomers are introducing their products, for example, Teledyne Dalsa released its first Vox microbolometers in 2017.

• Many companies in China are developing their own microbolometers. They do not produce large volumes today but the domestic market has great potential.

• On the other hand companies like Bosch, long involved in the MEMS and infrared businesses, have changed their strategies.

"2016 was a good year for the microbolometer market. There were almost 900,000 uncooled IR camera shipments, worth $2.7B in revenues thanks to a dynamic commercial market and continued growth for military applications. Many commercial applications drove this growth, including thermography, surveillance, PVS and firefighting. In 2022, we estimate there will be 1.7M units shipped.

Thermography is still the leading commercial market by far, in both value and volume. We estimate that there will be 500,000 thermography units shipped annually by 2022. As camera prices continue to fall, with several new products below $1000, sales are growing.

Surveillance is another interesting market. Until recently, thermal cameras have primarily been used in high-end surveillance for critical and government infrastructure. New municipal and commercial applications with lower price points are now arising, including traffic, parking, power stations and photovoltaic planning. We estimate this market will grow at almost 17% over 2017-2022 to reach 300,000 units by 2022.

Night vision in cars, including autonomous vehicles, could boost the microbolometer market. China is already a large market for automotive night vision, absorbing 25% of the total number of systems produced. In coming years, China will continue to account for a high share of this market.

Himax Q3 2017 Update Focuses on 3D Imaging and NIR

Himax Q3 2017 report has an update on the company's imaging business:

The Company believes 3D sensing is among the most significant new features for the next generation smartphone. The Company’s SLiM product line, based on structured light technology, is a state of the art total solution for 3D sensing. Himax’s goal is to provide total solutions with performance, size, power consumption and costs all suitable for smartphones and tablets. Himax offers fully integrated structure light modules, with the vast majority of the key technologies inside the module also developed and supplied by the Company. These critical in-house technologies include advanced optics utilizing the Company’s world leading WLO technology, laser driver IC, high precision active alignment for the projector assembly, high performance near-infrared CMOS image sensor and, last but not least, an ASIC chip for 3D depth map generation. The fact that all of these critical building blocks are developed in-house puts the Company in a unique position. Himax is able to react quickly and tailor its solutions to customers’ specific needs.

It also represents a very high barrier of entry for any potential competition and a much higher ASP for the Company. While the Company prefers to offer a total solution, it can also provide the aforementioned individual technologies separately to select customers so as to best accommodate their specific needs.

Thanks to the Company’s absolute technology leadership, its progress made with the fully integrated structure light 3D sensing total solution module is very exciting. Himax is seeing strong demand for 3D sensing solutions from numerous tier 1 customers. The Company is in close collaboration with select leading smartphone makers and partners right now, aiming to bring its total solution to mass production as early as early 2018 to meet the customers' aggressive launch timetables. Moreover, given that the Company is offering highly integrated solutions with ASPs much higher than those of individual components, by the time the Company starts shipping its total solutions, they will be a major contributor to both Himax’s revenues and profit, and consequently create a more favorable product mix for the Company.

Himax continues to make great progress with its two machine vision sensor product lines, namely, near infrared (“NIR”) sensor and Always-on-Sensor (“AoS”). The Company’s NIR sensor is a critical part in the structured light 3D sensing total solution. The Company’s NIR sensors’ overall performance is far ahead of those of its peers in 3D sensing applications. Himax currently offers low noise HD, or 1 megapixel, and 5.5 megapixel NIR sensors and is planning to add more to further enrich its product portfolio. Himax’s NIR sensors deliver superior quantum efficiency in the NIR range, especially over 940nm band which is critical for outdoor applications.

The Company’s AoS solutions provide super low power computer vision, which enables new applications across a wide variety of industries. The ultra-low power, always-on vision sensor is a powerful solution capable of detecting, tracking and recognizing its environment in an extremely efficient manner using just a few milliwatts of power. The Company is pleased to report that it already has one major global brand leveraging its AoS in their new high end TV models, which have already hit the market.

For the traditional human vision segments, Himax sees strong demand in notebooks and increased shipments for multimedia applications such as car recorders, surveillance, drones, home appliances, and consumer electronics, among others

Sony on Automotive Image Sensing

AutoSens publishes an interview with Abhay Rai, Director Product Marketing: Automotive Imaging at Sony. Few quotes:

Q: What are sensor problems for autonomous driving?

A: Due to the emergence of safety critical application, basic performance requirements have changed. Sensors need to reliably work for long time. Reliable operating temperature range, functional safety, security, power, and heat are very important performance parameters that a safety critical system design need to address.

Challenge for an image sensor for autonomous driving is: Can a sensor still see well under all real world conditions (reliable working and always working).

Q: Why can’t we just use dashcam footage on YouTube for finding problems?

A: Dashcam footage can be a datapoint but autonomous driving system requires more real world data and real world simulation to get to the best, robust and the safest autonomous driving car.

NHK Open House 2017

NHK Open House held in Tokyo in May 2017 revealed the latest TV innovations:

3D-integrated image sensor with per-pixel interconnect (together with University of Tokyo): "We have managed to shrink the pixels from the previous size of approximately 80 × 80 µm² to approximately 50 × 50 µm²."

Fast 8K 240fps image sensor: "We developed a 33 megapixel image sensor capable of high-speed operation and constructed a prototype 8K high-speed camera supporting shooting at 240 fps, which is four times the frame rate in 8K test satellite broadcasting. This camera enables the shooting of fast-paced action such as that in sports in 8K ultrahigh-definition video."

And many other innovations including organic image sensor "with a charge multiplication photoelectric conversion fi lm in order to achieve highly sensitive 8K cameras" and "organic image sensors with three organic films that provide sensitivity for each of the primary colors."

Wednesday, August 02, 2017

Pixart Q2 2017 Results

Pixart Q2 2017 revenue increased by 17.4% QoQ to NT$1,295.3 million. The gross margin increased from 52.8% in previous quarter to 55.8% in 2017 Q2 due to decrease in expenses and product mix change.

Yole on Apple Activity in Grenoble Area

EETimes article on fabs mentions Apple imaging group activity: "Or, take the example of the imaging-tech landscape growing rapidly in the Grenoble/Lyon area. Pierre Cambou, activity leader for imaging and sensors at Yole Développement, explained that imaging technology innovation often demands advancements in new manufacturing techniques. In return, it creates a tech-driven environment.

Reportedly, more than a dozen Apple engineers are moving to Grenoble to open an R&D center. This is happening precisely because the region has the expertise in image sensors and production — led by ST. “You need to have factories” to make an ecosystem, said Cambou.

Technavio Divides ToF Market to Regions

BusinessWire: Techvanio market reports often feature an interesting mix of quite obvious and quite unusual statements.

Sunil Kumar Singh, a lead analyst at Technavio, reports: "The growing popularity of augmented reality and virtual reality devices, 3D scanners, and gesture recognition technologies and the high investment in driverless cars by automobile manufacturers such as Ford, Nissan, and Tesla are expected to drive the global ToF market.

The demand for camera-enabled phones has been on the rise in South America and will drive the market for ToF sensors in this region. The replacement of CCD sensors with ToF sensors in many applications will also have a major impact on the ToF sensors market. The US, followed by Canada and Brazil, is the leading revenue generating country in the region owing to the early adoption of the technology

Global time of flight sensor market is expected to grow at a CAGR of 3% from 2017-2021. The consumer electronics segment accounted for close to 52% of the ToF sensor market share in 2016.

Tuesday, August 01, 2017

Digitimes Research: Sony Market Share is 45%

Digitimes Research believes that Sony took a 45% share of the global CIS market in 2016, while Samsung grabbed a 15% share of the market. The global CIS shipment value will grow to $11.2b in 2017 from $10.4b in 2016. The market is forecast to grow to nearly $13.8b in 2020.

Sony Image Sensor Sales Up but Forecast Down

Sony announces its quarterly earning results. The company updates about its image sensor business:

"Sales increased 41.4% year-on-year (a 38% increase on a constant currency basis) to 204.3 billion yen. This increase was primarily due to a significant increase in unit sales of image sensors for mobile products, as well as the absence of the impact of a decrease in image sensor production due to the 2016 Kumamoto Earthquakes in the same quarter of the previous fiscal year, partially offset by a significant decrease in sales of camera modules, a business which was downsized."

The forecast for 2017 fiscal year ending in March 2018 has been updated too:

"Sales are expected to be lower than the April forecast primarily due to lower-than-expected image sensor unit sales for mobile products, partially offset by the impact of foreign exchange rates. Operating income is expected to be higher than the April forecast mainly due to lower-than-expected production costs as well as the positive impact of foreign exchange rates, partially offset by the impact of the above-mentioned decrease in sales."