Sunday, July 31, 2016

Samsung Works on BSI Global Shutter Pixel Improvements

Samsung patent application US20160211306 "Image sensors" by Hyuk Soon Choi, Jung Bin Yun, Jungchak Ahn presents a number of different ideas on light shielding of a charge storage node SD in BSI global shutter pixels:

Saturday, July 30, 2016

e2v Videos

e2v publishes a number of videos of its imaging division track record:



Friday, July 29, 2016

Sony, Sharp Quarterly Results

Sony quarterly report and forecast reflect Kumamoto earthquake results:


The company PR separates Semiconductors from the battery business (to be sold to Murata) and updates on its status:

"Sales decreased 22.9% year-on-year (a 14% decrease on a constant currency basis) to 144.4 billion yen (1,402 million U.S. dollars). This decrease was primarily due to a significant decrease in sales of image sensors, reflecting the impact of a decrease in image sensor production due to the 2016 Kumamoto Earthquakes, lower demand for image sensors for mobile products, and the impact of foreign exchange rates. Sales to external customers decreased 21.4% year-on-year.

Operating loss of 43.5 billion yen (423 million U.S. dollars) was recorded, compared to operating income of 32.7 billion yen recorded in the same quarter of the previous fiscal year. This significant deterioration was due to the impact of the above-mentioned decrease in sales, the 20.3 billion yen (197 million U.S. dollars) impairment charge against long-lived assets for camera modules, net charges of 13.6 billion yen (132 million U.S. dollars) for expenses resulting from the 2016 Kumamoto Earthquakes and the negative impact of foreign exchange rates. During the current quarter, there was a 8.2 billion yen negative impact from foreign exchange rate fluctuations.
"

The Semicondustor business updated forecast for Sony fiscal year ending on March 31, 2017: "Sales are expected to be lower than the May forecast primarily due to lower-than-expected image sensor sales reflecting the impact of foreign exchange rates, partially offset by a faster-than-expected recovery from the 2016 Kumamoto Earthquakes. Operating income is expected to be lower than the May forecast mainly due to the negative impact of foreign exchange rates, partially offset by the favorable impact of the above-mentioned recovery."

Sharp too reports its quarterly earnings showing camera module business shrink, although it retains its position as the company's second largest product group:

Thursday, July 28, 2016

Rumor: Ingenic Semiconductor "Acquires" Omnivision

Reportedly, Ingenic Semiconductor, a small public company in Beijing, is in the process of acquiring Omnivision, and, possibly, Superpix. The deal is assumed to be a reverse merger where Omnivision gets access to Shenzhen stock market under Ingenic stock code 300223.

The Omnivision deal has been reported as Ingenic Semi's official announcement at the Chinese-language EastMoney site.

Thanks to XW for the link!

Update: The parameters of the deal can be found here:

Samsung Reports CIS Sales Growth

SeekingAlpha's Samsung Q2 2016 earnings call transcript has a little update on image sensor sales: "Second quarter earnings improved quarter on quarter due to ...the increased sales of high megapixel [inaudible] megapixel and above CMOS Image Sensor in China. ...In the third quarter, we expect ...the sales of CMOS Image Sensor in China market to be stay solid."

e2v Sapphire Videos

e2v publishes Youtube videos of its 1.3MP, 2MP and WVGA sensors from Sapphire family:


Xiaomi Redmi Pro Features Samsung Depth Camera

Xiaomi Redmi Pro smartphone features a dual rear camera: a 13MP Sony IMX258 is combined with an unnamed 5MP Samsung sensor "to aid depth of field recording, real-time background blur:"

Wednesday, July 27, 2016

ST Imaging Business Update

SeekingAlpha: ST reports its Q2 2016 earnings. The CEO Carlo Bozotti and other executives talk about the imaging business results:

"Let me speak briefly of our Imaging Product Division which is reported in the line others. Here, we had strong sequential increase in sales, thanks to our Time-of-Flight specialized image sensors entering multiple smartphone models, 10 new flagship smartphones in Q2 alone. Among these, we can mention that Huawei P9 Plus and Honor V8. And we continue to see a very good momentum.

And the other good opportunity is that in the imaging and particularly in our specialty image sensors with the Time-of-Flight technology, we see momentum and we are starting also new opportunities, and this is another important contributor of our sequential growth expectation.
"

And then comes a question on Apple cooperation rumors:

Veysel Taze - ODDO SEYDLER BANK AG:

"Basically, not much questions left. But one will be interesting around your image sensor business. There were recently rumors that one of the North American smartphone OEMs is ramping a new R&D in France, a fab for camera modules and image sensors. And I was wondering if that's an option for you? I know you discontinued the camera module business, but is that an opportunity, or yeah?"

Jean-Marc Chery - COO:

"...first of all, I confirm that ST decided to discontinue the camera module, the commodity camera module, so we have no intent to come back in this camera module. So we are focusing on specialized imaging sensor and especially our Time-of-Flight technology where we see a very good traction, especially with a nice product called Ewok (???) where we have mainly Asia Pac and especially in China smartphone opportunities. And this is where we are focusing and focus a lot on that."

Another interesting question:

Janardan N. Menon - Liberum Capital Ltd.

"Just a couple of quick follow-ups from me. One is on your imaging business, the Time-of-Flight sensors. If I remember right, you've made a loss of about $80 million last year on the imaging side. I was just wondering, given the very strong growth that you've seen so far in the first half of this year and what you expect to see through the second half based on your design wins, would you be able to sort of almost eliminate that loss by the end of this year or is that going to be work in progress into next year as well?"

Carlo Bozotti - President & CEO:

"To respond to the first question, the answer is yes, and this is certainly an area where we can turnaround because we have the result of the camera modules and the commodity that was also linked to certain needs in terms of fab loading. And this now is done. So, it was a massive R&D effort. It was an important R&D effort with very limited sales. So now sales are coming and, of course, there is a very significant improvement."

13MP Panasonic Sensor in Mass Market Smartphones

Panasonic image sensors have design wins in mainstream smartphones. Chinese UMi Vienna and Super smartphones use 13MP Panasonic MN34172 sensor:

Tuesday, July 26, 2016

Color Night Vision with Sony ISX017

KB ViTA publishes a PAL-format night scene video with new Sony ISX017 Starvis sensor. The video was acquired through the imagers's analog output at 1/50s exposure time, with 25mm F1.2 lens:



ISX017 has not been officially announced yet, but there is quite a lot of info about it on the Net:


Thanks to AB for the info!

Monday, July 25, 2016

e2v Capabilities

e2v publishes a Youtube video on its imaging division capabilities:

Nissan Presents Automatic Driving Minivan

Nikkei: Nissan ProPilot automatic driving technology will be used for the Serena minivan to be released in late August 2016. The technology supports automatic driving on a single lane of a motorway. The ProPilot uses a monocular camera attached to the windshield (inside the car). The camera recognizes vehicles on the road, white lines, measures the distance to a preceding vehicle and detect the position of the user's vehicle in the lane. The accelerator, brakes and steering are controlled based on those data.

Propilot display

Sunday, July 24, 2016

OptoPACK IPO on KOSDAQ

Korean-language BusinessPost and JAFCO-Asia report that image sensor chip-scale package maker OptoPACK has been listed on KOSDAQ. The company has filed a 31.7 billion won last year's sales (~$28M) and an operating loss of 30 million won (~$26K).

Magic Leap Aims to Replacing Smartphones

RoadToVR quotes Magic Leap CMO Brian Wallace saying that the company is working on an AR headset that’s portable, lightweight, and can be worn all day. He further said that the aim is to completely replace smartphones someday.

Talking about the AR devices, James Mackie publishes a Youtube video claiming that Microsoft Hololens is "the absolute gamechanger."

Thursday, July 21, 2016

1st Executive Infrared Imaging Forum

Yole Développement and its partner CIOE announce the 1st Executive Infrared Imaging Forum: From Niche to Large Volume Applications. Organized by Yole and hosted by CIOE, this event will take place on September 8, 2016 in Shenzhen, alongside the 18th China International Optoelectronic Expo 2016. The Forum agenda:
  • Thermal Imager with Microbolometer for Smartphone: Evolution & Comparison on the last Technologies Trends
    Sylvain Hallereau, Project Manager, System Plus Consulting
  • Molded, Wafer Level Optics for High Volume Applications
    John Franks, Technical Director, Umicore Electro Optic Materials
  • Fabless Development of CMOS Micro-bolometer FPA for High Performance Consumer Application
    Kwyro Lee, CTO of Sirius Inc. & Prof. in School of EECS, KAIST, Daejon, Korea
  • NETD, uncooled FPA, Electro-Optic characterization, TWS, SmartIRTM
    Frédéric Mathieu, Chief Operating Officer & Project Manager, Device-ALab
  • Thermodiode Infrared Technology for Mass Market Applications
    Andreas Krauss, Product Management Sensor Components (AE/PRM-S) , Robert Bosch GmbH
  • Keynote - TBD
    Stuart Klapper, Autoliv
  • Uncooled IR Imaging Market Perspectives
    Eric Mounier, Senior Analyst, Yole Développement
  • Tracking Illuminator: Novel Instrument Combining Thermal IR Detection with VIS Illumination
    Hubert Jerominek, INO
  • ULIS inside "Smart Buildings"
    Cyrille Trouilleau, Product Manager, ULIS
  • Wuhan GST's very Low Cost Uncooled Infared Detector Development
    Gao Jianfei, Detector Center Director, GUIDE
  • Thermal and Hyperspectral Imaging solutions for Integrators
    Eric Guyot, Director French Office and IR Cameras Product Line Manager, Telops

LFoundry on Image Sensor Production Challenges

FBK publishes a LFoundry's Antonio Venezia short talk on CIS production challenges in EU Supertwin project on entangled photon imaging (part 1, part 2):



MIT on Importance of High Full Well in Interferometric Microscopy

Adimec interviews Poorya Hosseini of MIT to discuss the use of extreme high full well capacity camera in interferometric microscopy:

"There has been some debate on what typically limits the sensitivity of the phase measurements in interferometric microscopy and holographic microscopy that has been a few nm in terms of optical path length. We hypothesized that photon shot noise sets the current limit rather than other noise sources such as mechanical vibrations or power fluctuations of the illumination source. After some theoretical calculations and showing shot noise is indeed the limiting factor, we needed a tool to see how far we can push shot noise down before we hit the mechanical limits. This is only possible when you can collect a huge number of photons in a short period of time using this camera, and of course lots of light."

Wednesday, July 20, 2016

Tractica on 3D Imaging Market

According to a report from Tractica, the worldwide market for 3D imaging hardware and software will grow from $3.2b in 2014 to $24.9b in 2024 at a compound annual growth rate (CAGR) of 23%. The consumer and mobile and the automotive markets were the two largest in 2014, accounting for $1.2 billion and $1.0 billion, respectively. The robotics and industrial market will sport the highest growth rate of 26.9% among all the application markets to become worth $3.9 billion by 2024. Overall, the consumer and mobile market will continue to be the largest market for 3D imaging technology, accounting for $10.1 billion in 2024 and growing at a CAGR of 23.3%:

Leap Motion on 3D Imaging for VR

Leap Motion CTO and co-founder David Holz shares his 1.5-hour long AltspaceVR presentation on imaging in VR applications:

Sony vs Samsung in Galaxy S7

EETimes publishes TechInsights reverse engineering findings comparing Sony and Samsung 12MP camera modules extracted from Galaxy S7 smartphones:

Sony IMX260
Samsung S5K2L1SX

Few quotes:

"Sony had used TSVs in their earlier CMOS image sensors and we had expected the same for the IMX260. But we don’t see them, as they have been replaced by a direct wafer bonding process that we will discuss later.

The Samsung CMOS image sensor has arrays of TSVs along its perimeter and these are used to make the electrical connections to the underlying ASIC.

Simply joining the two wafers together will allow Van der Waal forces (hydrogen bonding) to hold the wafers together, but this bond strength is insufficient for the task. High temperature anneals can convert these bonds to covalent bonding, but high temperature anneals can cause problems with thermal mismatches between the materials making up the two wafers. This is where Ziptronix’s 7,109,092 (‘092) patent comes into play. This patent describes a method for subjecting the wafer’s oxide surface with a fluorinating treatment to promote the covalent bonding of the two wafers in a room temperature process.

This wafer bonding process is likely done prior to the backside thinning of the image sensor wafer that might be about 150 µm thick, as is the control ASIC. The two wafers would brought into alignment using IR microscopes to see through the wafers to their respective alignment marks. This process has yielded a misalignment of the image sensor’s metal 6 pad and the control ASICs metal 8 pad of less than 0.25 µm."

Samsung uses arrays of through silicon vias to connect its S5k2L1SX image sensor to its underlying ASIC. Figure 7 shows a series of Through Silicon Vias (TSVs) arranged on an approximately 5 µm x 8 µm grid pattern. The TSVs are the dark oval shapes seen at the center of each grid point.

The TSVs connecting the image sensor die to the metal 7 traces of the control ASIC was formed after the two dies are affixed to each other, and after the image sensor die had undergone its backside thinning, using a TSV last process. The TSV metallization appears to be tungsten with probably a titanium nitride barrier and possibly titanium adhesion layers.

Using TSVs to electrically connect the image sensor die to the ASIC certainly gets the job done, but at the cost of the added real estate needed for the TSV arrays.
"

Sony IMX260 Wafer Bonding Surfaces
Top Surface of Samsung S5K2L1SX with TSV Array
Samsung S5K2L1SX TSVs

Tuesday, July 19, 2016

Varioptic Demos its Lens Vibration Tolerance

Varioptic publishes a Youtube video showing that its liquid lens performance is quite insensitive to a vibration:



An older video shows how its electrowetting liquid lens operation:

Analog Devices Occupancy Sensor

Analog Devices presents its occupancy vision module reference design based on a smart image sensor that reduces the volume of data transferred to a Blackfin DSP. It sounds like the image sensor is based on the acquired Swiss Snap Sensor startup's technology:

Meizu Phone Features Sony IMX386 Sensor

PRNewswire: Meizu MX6 phone features Sony new 12MP 1/2.9-inch IMX386 rear camera sensor with 1.25um pixel. Meizu says its quality is similar to the one in Galaxy S7 camera. "The high-density phase detection pixel embedded in the sensor brings enhanced focus accuracy and speed," says Meizu. The camera has 6-element F2.0 lens and achieves the autofocus speed of 0.2s. An upgraded ISP is said to achieve the "camera speed of 319ms."

Monday, July 18, 2016

Pixart Works on VR Imager for Next Gen Nintendo NX

Digitimes quotes Chinese-language Commercial Times saying "Pixart is expected to complete tape-out of chips supporting VR (virtual reality) technology by year-end 2016, said the paper, adding that the chips will support next-generation Nintendo NX gaming machines."

Making 250,000 Machine Vision Cameras a Year

PointGrey publishes a video on what it means to manufacture 250,000 machine vision cameras a year:

Sunday, July 17, 2016

Friday, July 15, 2016

Vision Apps with Xilinx Zynq UltraScale+

Xilinx publishes a Youtube video demoing embedded vision capabilities of its Zynq UltraScale+ MPSoC:

Thursday, July 14, 2016

Rumor: Apple to Open Imaging Lab in Grenoble, France

Macrumors reports that Apple is to open a research lab in Grenoble, France focused on developing image sensors, according to French newspaper Dauphiné Libéré, with details posted in iPhon.fr.

Apple is said to have a team of dozen researchers and engineers working on imaging R&D for more than a year at Minatec, in collaboration with STMicro. Apple plans to have approximately 30 engineers in the new research lab, which will occupy about 800 sq. m.


Thanks to RP for the link!

ST VL53L0 Reverse Engineering

Chipworks publishes a reverse engineering report of the latest ST ToF sensor:

1. SPAD TOF sensing circuits
2. PMU to the SPAD circuit
3. IR Emitter Driver

Wednesday, July 13, 2016

ToF Cameras Shootout

Aalto University, Espoo, Finland, publishes MsC Thesis "Performance Evaluation of Time-of-Flight Depth Cameras" by Matti Laukkanen. Three ToF cameras are compared: MESA SR4000 (released in 2010), Panasonic DIMager EKL-3106 (2012) and Microsoft Kinect v2 (2014). Few comparisons:


"Overall, Kinect v2 can be considered as the winner of the trio. Compared to the other two devices, it offers approximately ten times more pixels and a significantly larger FOV at a fraction of the price. It is also the most accurate, as it can achieve up to millimetre precision. SR4000 is at its best almost as accurate, but on the other hand it is highly sensitive to background light. D-IMager differs significantly from the other two, as its depth accuracy is nearly tenfold lower."

Basler on CCD to CMOS Conversion

René von Fintel, Head of Product Market Management at Basler, says about transition from CCD to CMOS cameras after last year's Sony's announcement that production of CCD sensors would be ending:

"The switchover is moving forward surprisingly quickly in terms of the adoption of CMOS sensors for new design-ins within facilities and systems. Most customers have taken the imminent discontinuation of Sony's CCD sensors as an occasion to reconsider their current CCD cameras and replace them at the next possible opportunity with CMOS-based cameras. This makes sense for two reasons: New CMOS cameras can deliver faster speeds and better image quality on the one hand, yet through their low prices also offer a better price/performance ratio on the other."

Monday, July 11, 2016

ON Semi Enhances NIR Performance of its CCDs

BusinessWire: The 8MP KAI-08052 CCD is the first device in ON Semiconductor’s CCD portfolio with enhanced NIR response, provides up to twice the sensitivity in NIR as the company’s standard Interline Transfer CCD pixel design.

The new 5.5um CCD pixel design used in the KAI-08052 extends the electron capture region deeper in the silicon to better capture electrons generated by long wavelength photons. This deeper pixel well improves detection of NIR wavelengths by up to a factor of two depending on the specific wavelength studied. And since the well structure also isolates the photodiodes from each other, this increase in NIR sensitivity comes without any reduction in image sharpness.

Camera manufacturers and end customers continue to confirm that products based on both CCD and CMOS technologies will be needed in these markets, making it important that we continue to develop and advance both of these technologies,” said Herb Erhardt, VP and GM, Industrial and Security Division, Image Sensor Group at ON Semiconductor. “The KAI-08052 provides a significant improvement in NIR sensitivity compared to our standard pixel design, and we look forward to utilizing this technology in additional products in the future.

As always, ON Semi is very open in releasing the actual performance info:

Sunday, July 10, 2016

Samsung Applies for Color Splitter Patent

Samsung patent application US20160054172 "Image sensor for producing vivid colors and method of manufacturing the same" by Sookyoung Roh, Sunghyun Nam, Seokho Yun, and Hongkyu Park proposes a color splitter to improve sensitivity of image sensor:

"Although organic dye color filters are widely used in color image sensors, such organic dye color filters absorb light in all but one certain wavelength band. Therefore, using such color image sensors may cause a relatively large amount of light energy loss and a low light efficiency."

So, a color separating beam splitter is proposed. The splitter is made of high refractive material diagonal stripe located inside a low refracting material layer:


Apparently, the simulation show quite high efficiency of the proposed color division scheme comparing with the traditional die based CFA:

Saturday, July 09, 2016

Omnivision Proposes RGB-IR CFA

Omnivision's patent application US20160104735 "Dual-Mode Image Sensor With A Signal-Separating Color Filter Array, And Method For Same" by Jin Li, Qian Yin, and Dyson Hsinchih Tai proposes RGB-IR CFA:

ON Semi's SuperPDAF and GS Demos

ON Semi's new Mobile Consumer Division demos PDAF performance of its recently announced 13MP 1.1um pixel AR1337 sensor:



Another video shows Python 25MP, 82fps global shutter operation:

Friday, July 08, 2016

Large Area Image Sensors Tutorial

Albert Theuwissen publishes an excellent series of posts "Butting Versus Stitching" comparing different ways to combine large format image sensor out of smaller parts, some of them quite exotic - part 1, part 2.

Thursday, July 07, 2016

Sony 4K 480fps GS Sensor

Imager No-Mania quotes Watch.Impress article on Sony presentation at 2016 VLSI Symposium. The 4K, 480fps sensor based on 5.86um global shutter pixel is used in a newly announced HDC-4800 video camera. The sensor features a dual gain ramp ADC stripes placed on top of the image sensor die:

Japan Allows Cameras to Replace Mirrors in Cars

Automotive News: Japan became one of the first countries allowing vehicles to use cameras instead of mirrors, beginning June 17, 2016. EU too is expected to revise its local regulations this year, while the US is seen adopting the mirrorless standard in 2018, and China is expected to do it too in the coming years.

The promise of mirrorless cars is sparking a rush of suppliers to the technology, including Japan's Ichikoh Industries and Germany's Robert Bosch GmbH. By 2023, about 29 percent of the Japanese market or about 2.3M vehicles, are expected to have video monitors as interior mirrors. At the same time, it's forecasted that about 12% of the market, or about 900,000 vehicles, will have cameras in place of exterior sideview mirrors.

Wednesday, July 06, 2016

Light Raises $30M in C-Round

Light, a startup developing its L16 multi-aperture computational camera, announces it has closed a $30M series C round of funding, led by GV (formerly Google Ventures). On the way to its first product, the company releases a Vimeo video showing how the L16 prototypes are being manufactured:

Toshiba Reports Success of its Image Recognition Processor

Toshiba reports a success with its image recognition LSI for ADAS applications:

Bioinspired Sensor Can See in the Dark

Design News: Researchers at the University of Wisconsin in Madison (UW-Madison) developed the bioinspired photosensitivity enhancer (BPE), an "artificial eye" inspired by creatures that live deep under the ocean, where light is scarce. Their paper is published in Proceedings of the National Academy of Sciences (PNAS):

"Artificial eye for scotopic vision with bioinspired all-optical photosensitivity enhancer" by Hewei Liu, Yinggang Huang, and Hongrui Jiang

From the abstract: "to date, strategies toward improving low-light imaging primarily focus on developing electronic image sensors. Inspired by natural scotopic visual systems, we adopt an all-optical method to significantly improve the overall photosensitivity of imaging systems. Such optical approach is independent of, and can effectively circumvent the physical and material limitations of, the electronics imagers used. We demonstrate an artificial eye inspired by superposition compound eyes and the retinal structure of elephantnose fish. The bioinspired photosensitivity enhancer (BPE) that we have developed enhances the image intensity without consuming power, which is achieved by three-dimensional, omnidirectionally aligned microphotocollectors with parabolic reflective sidewalls. Our work opens up a previously unidentified direction toward achieving high photosensitivity in imaging systems."

The main idea is to condense the light into a smaller area by means of "thousands of microphotocollectors (μ-PCs). The miniaturized, low-cost, and zero-power-consumption device presented here can be implemented independently in imaging systems, or combined with other image enhancement technologies." A 2x-3x improvement in light sensitivity has been reported:

e2v Sensor Guided Juno to Jupiter, ON Semi Sensor Will Observe Jupiter Clouds

e2v’s image sensors have accurately guided NASA’s Juno spacecraft to reach its destination, Jupiter. e2v’s TH7890 512 x 512 17 micron pixel front illuminated area array image sensors for star-trackers were provided by e2v’s manufacturing centre in Grenoble. They have been used to determine the orientation of the spacecraft by measuring its position relative to stars. These optical devices were specifically designed to ensure very high pointing accuracy and stability over long periods and to survive the severe radiation surrounding the planet, as Juno flies closer to the gas giant than any other spacecraft ever before.

JunoCam will use a ON Semi's KAI-2020 CCD to provide the first color pictures of Jupiter’s cloud tops from orbit. The imaging capabilities of JunoCam will be impressive. The camera will capture data in four separate wavelength bands: blue (420–520 nm), green (500–600 nm), and red (600–800 nm) for reconstruction of a color image; plus a band from 880–900 nm to image a narrow absorption of methane centered at 889 nm. And the 3 km resolution available at closest approach (an order magnitude better than that available from the Voyager, Galileo, and Cassini probes) will enable viewing of small-scale structures in the atmosphere.

CEA-Leti and EPFL Pre-Announces 0.5e- Noise Pixel

CEA-Leti and EPFL pre-announces its paper on fast and low noise image sensor based on thin-oxide pmos SF, to be published in September issue of JSSC:

A Sub-0.5 Electron Read Noise VGA Image Sensor in a Standard CMOS Process

Boukhayma, Assim; Enz, Christian; Peizerat, Arnaud

Abstract:

A sub-0.5e−rms temporal read noise VGA (640H×480V) CMOS image sensor has been integrated in a standard 0.18μm 4PM CMOS process. The low noise performance is achieved exclusively through circuit optimization without any process refinements. The presented imager relies on a 4T pixel of 6.5μm pitch with a properly sized and biased thin oxide PMOS source follower. A full characterization of the proposed image sensor, at room temperature, is presented. With a pixel bias of 1.5μA the sensor chip features an input-referred noise histogram from 0.25 e−rms to a few e−rms peaking at 0.48 e−rms. The imager features a full well capacity of 6400 e− and its frame rate can go up to 80 fps. It also features a fixed pattern noise as low as 0.77%, a lag of 0.1% and a dark current of 5.6e-/s. It is also shown that the implementation of the in-pixel n-well does not impact the quantum efficiency of the pinned photo-diode.

Monday, July 04, 2016

CEVA Announces 2nd Generation Neural Network Framework

CEVA introduces CDNN2 (CEVA Deep Neural Network), its second generation neural network software framework for machine learning. Coupled with the CEVA-XM4 intelligent vision processor, CDNN2 offers significant time-to-market and power advantages for implementing machine learning in embedded systems for smartphones, ADAS, surveillance equipment, drones, robots and other camera-enabled smart devices.