Image Sensors World        Go to the original article...

Canon News site refers to Canon Japan web page showing that the company expands its CMOS sensor lineup by 5 new products for security applications:


With these new sensors addition, Canon portfolio looks somewhat more full with products:


12MP 1-inch CMOS 4K Sensor

• Filter Type - RGB
• Sensitivity - 22,000 e/lx/sec @Analog gain x1 (TBD)
• Number of Effective Pixels - 4000 x 3000
• Pixel Size - 3.2 um x 3.2 um
• Shutter - Rolling shutter 12 bit, 24 fps (All Pixels)
• Maximum Frame Rate - 10 bit, 60 fps (4K2K)
• Saturation 22,000 e (TBD) @ Analog gain x 1
• Dark Random Noise - 2.8 e rms @Analog gain x16 (TBD)
• Dark Current - 17 e/sec (TBD) @package reverse side 60℃
• Power Consumption - 540 mW (Typ.) @All pixels readout 24fps (12bit)

12MP 1/1.7-inch CMOS Sensor

• Filter Type - RGB
• Sensitivity - 8,900 e/lx/sec @Analog gain x1 (TBD)
• Number of Effective Pixels 4000x 3000
• Pixel Size - 1.86 um x 1.86 um
• Shutter - Rolling shutter 12 bit, 15 fps (All Pixels)
• Maximum Frame Rate - 11 bit, 30 fps (4K2K)
• Saturation - 12,000 e (TBD) @Analog gain x1
• Dark Random Noise - 1.3 e rms @Analog gain x16 (TBD)
• Dark Current - 7 e/sec (TBD) @package reverse side 60℃
• Power Consumption - 520 mW (Typ.) @All pixels readout 15 fps (12bit)

2.8MP 1/2.32-inch HDR CMOS Sensor - 3U3MRXSAAC

• Filter Type - RGB
• Sensitivity (e/lx/sec) 25,000 (Green) @Analog gain x1 (TBD)
• Number of Effective Pixels 1936 x 1456
• Pixel Size - 3.2 um x 3.2 um
• Shutter - Rolling shutter 120 dB, HDR
• Dynamic Range - 75dB, Normal
• Maximum Frame Rate (All Pixels) - 60 fps, Normal, 30 fps, HDR
• Operating Temperature - -40℃ ~ 105℃（-40°F ~ 221°F）
• Saturation - 23,000 e @Analog gain x1 (TBD)
• Dark Random Noise - 2.7 e rms @Analog gain x4 (TBD)
• Dark Current - 13 e/sec @Analog gain x 1, 60℃ (TBD), TBD @room temperature
• Output Format - @in all-pixel operating mode 12bit, 60fps
• Power Consumption - 300mW (Typ.) @using all pixels 60 fps (TBD)

Super 35mm 4K 9.34MP CMOS Sensor with 60 fps all-pixel readout

• Filter Type - RGB
• Sensitivity - 72,000 e/lx/sec @Analog gain x1 (TBD)
• Number of Effective Pixels 4112 x 2248
• Pixel Size - 6.4 um x 6.4 um
• Shutter - Rolling shutter
• Maximum Frame Rate 12bit, 60 fps (All Pixels)
• Saturation - 39,000 e(TBD) @Analog gain x1
• Dark Random Noise - 2.7 e rms @Analog gain x 8 (TBD)
• Dark Current - 54 e/sec(TBD) @package reverse side 60℃
• Power Consumption - 2 W (Typ.) @All pixels readout 60 fps

Full Frame 50MP CMOS Sensor

• Filter Type - RGB
• Sensitivity - 32,000 e/lx/sec @Analog gain x16 (TBD)
• Number of Effective Pixels 8688x 5792 (Horizontal x Vertical)
• Pixel Size - 4.14 um x 4.14 um
• Shutter - Rolling shutter
• Maximum Frame Rate - 6.8 fps (All Pixels)
• Saturation - 38,000 e (TBD)
• Dark Random Noise - 2.5 e rms @Analog gain x16 (TBD)
• Dark Current - 9.6 e/sec (TBD) @package reverse side 60℃
• Output Format - 16 ch analog outputs
• Power Consumption - 1.5 W (Typ.) @All pixels readout 6.8 fps

A couple of new videos showing the resolution power of Canon 120MP sensors:





Thanks to TG for the link!

Go to the original article...

Image Sensors World        Go to the original article...

ST publishes a couple of videos promoting its VG5761 sensor for in-cabin monitoring. Interestingly, the sensor is marked as obsolete at ST site:

Go to the original article...

Image Sensors World        Go to the original article...

Omnivision announces on Twitter: "Our OV6948 imager was just named by @GWR as “The Smallest #ImageSensor Commercially Available” at just 0.575 x 0.575mm. The sensor brings #FutureInSight to medical applications."

Go to the original article...

Image Sensors World        Go to the original article...

Fudan University, Shanghai, PISD "successfully combines all functionalities in one transistor. This can largely simplify the pixel design and improve its efficiency. Considering the huge market of image sensor, the development of PISD is very helpful for China to break the monopoly in the global market."

Fudan University's IEEE EDL paper "A Novel One-Transistor Active Pixel Sensor With In-Situ Photoelectron Sensing in 22 nm FD-SOI Technology" by Yong-Feng Cao, M. Arsalan, J. Liu, Yu-Long Jiang, and J. Wan "is commented as “revolutionary” by its reviewer." From the abstract:

"For the first time, a novel active pixel sensor (APS) with 22 nm fully depleted silicon-on-insulator (FD-SOI) technology is experimentally demonstrated. The APS in-situ integrates photo sensing, charge integration, buffer amplification, and random access in one transistor without charge transfer and assistance of additional transistors. The deep depletion effect in the substrate of the SOI MOSFET is used to collect photoelectrons which are then sensed by the top Si channel. A sensor array is proposed and high sensitivity is experimentally verified in a simplified circuit."


IEEE JEDS publishes Peking University, Beijing, open-access paper "UTBB-Based Single Transistor Image Sensor of Submicron Pixel Using Back Gate Modulation" by Liqiao Liu, Xiaoyan Liu, and Gang Du:

"Image sensor has developed for decades. Now, submicron photo sensor device with high performance is required. In this work, a UTBB (ultra-thin body and box) based single transistor image sensor has been investigated. The light collection and signal readout are accomplished by a single transistor, so the pixel of the UTBB image sensor can shrink down to the submicron. The main parameters impacting the performance of the UTBB image sensor such as back voltage, the thickness of the BOX, well doping concentration and well depth are investigated. Besides, the UTBB image sensor can achieve multi-resolution to adapt to different requirements. The performance of the UTBB image sensor is evaluated by TCAD simulations."

Go to the original article...

Image Sensors World        Go to the original article...

Fudan University, Shanghai, PISD "successfully combines all functionalities in one transistor. This can largely simplify the pixel design and improve its efficiency. Considering the huge market of image sensor, the development of PISD is very helpful for China to break the monopoly in the global market."

Fudan University's IEEE EDL paper "A Novel One-Transistor Active Pixel Sensor With In-Situ Photoelectron Sensing in 22 nm FD-SOI Technology" by Yong-Feng Cao, M. Arsalan, J. Liu, Yu-Long Jiang, and J. Wan "is commented as “revolutionary” by its reviewer." From the abstract:

"For the first time, a novel active pixel sensor (APS) with 22 nm fully depleted silicon-on-insulator (FD-SOI) technology is experimentally demonstrated. The APS in-situ integrates photo sensing, charge integration, buffer amplification, and random access in one transistor without charge transfer and assistance of additional transistors. The deep depletion effect in the substrate of the SOI MOSFET is used to collect photoelectrons which are then sensed by the top Si channel. A sensor array is proposed and high sensitivity is experimentally verified in a simplified circuit."


IEEE JEDS publishes Peking University, Beijing, open-access paper "UTBB-Based Single Transistor Image Sensor of Submicron Pixel Using Back Gate Modulation" by Liqiao Liu, Xiaoyan Liu, and Gang Du:

"Image sensor has developed for decades. Now, submicron photo sensor device with high performance is required. In this work, a UTBB (ultra-thin body and box) based single transistor image sensor has been investigated. The light collection and signal readout are accomplished by a single transistor, so the pixel of the UTBB image sensor can shrink down to the submicron. The main parameters impacting the performance of the UTBB image sensor such as back voltage, the thickness of the BOX, well doping concentration and well depth are investigated. Besides, the UTBB image sensor can achieve multi-resolution to adapt to different requirements. The performance of the UTBB image sensor is evaluated by TCAD simulations."

Go to the original article...

Image Sensors World        Go to the original article...

AAA report on pedestrian detection in modern cars ADAS emphasizes a limited efficiency of the camera-based systems. The night time tests (with headlights) were especially disappointing:

Go to the original article...

Image Sensors World        Go to the original article...

Sony announces that the Sociable Cart SC-1, jointly developed with Yamaha Motor, will begin service on November 1, 2019. The first phase of this service launch will take place in Okinawa Prefecture's Kanucha Bay Resort in Nago City, and the Southeast Botanical Garden in Okinawa City.

Sociable Cart SC-1 features image sensors capable of vision beyond that of human capacity on all sides of the vehicle, allowing it to sense the surrounding environment using images that are in-focus in all directions. Moreover, the ultra-high sensitivity characteristics of the image sensors and the high-resolution display installed inside the vehicle allow the passengers to see their surroundings at night even without any headlights. The SC-1 also incorporates mixed reality (MR) technology developed by Sony that can superimpose computer graphics onto the surroundings being displayed on the monitor. This turns the area that used to be taken up by windows, where passengers could only see the scenery, into an entertainment area, thereby enabling a more enjoyable mobility experience.

The initial service, Moonlight Cruise, is a night-time showcase where passengers ride on SC-1 after dark. Rather than being controlled by the passengers, SC-1's ultra-high sensitivity image sensors and high-resolution display start by showing the night-time scenery and overlaying entertainment content generated by the MR technology, while passengers are free to sit back and enjoy the experience.

Go to the original article...

Image Sensors World        Go to the original article...

AnandTech attempts to figure out the technology behind Huawei Mate 30 Pro smartphone 7680fps slow motion:

One of the headline features of the Mate 30 Pro is its super slow motion capabilities. Most smartphones available on the market today, if they offer slow motion, do it at either 120 fps, 240 fps, or up to 960 fps. This typically comes in the form of a sensor and system that can take 120 frames per second video and do clever interpolation to appear as if it is a higher frame rate. We’ve seen some good and bad solutions here.

The way that traditional slow motion cameras work is on a sliding scale – the trade-off between frame rate is resolution. If you reduce the area of the sensor that needs to take the image by one quarter, then technically the fast memory storing the video data can store 4x as much – as long as you can poll data from the sensor at 4x the speed, it should be good to go.

It seems that Huawei is doing a mix of things here to get 7680 fps. The camera offers several settings, which comes with a slider from 2x to 256x. Any setting 32x or lower gives a 1080p video, while 64x and higher gives a 720p video:

• 1080p at 120 fps = 4x speedup
• 1080p at 240 fps = 8x speedup
• 1080p at 960 fps = 32x speedup
• 720p at 1920 fps = 64x speedup
• 720p at 7680 fps = 256x speedup

Go to the original article...

Image Sensors World        Go to the original article...

BusinessKorea: Samsung is trying to leverage its advantage in advanced processes to overtake Sony on image sensor market. "Samsung Electronics' System LSI Business Division is expected to boost its image sensor sales and market share as it secured LG Electronics as one of its corporate customers. Major smartphone makers including China's Xiaomi and Oppo also reportedly adopted Samsung's image sensors for their latest products. Therefore, Samsung's share in the image sensor market is expected to grow."

Go to the original article...

Image Sensors World        Go to the original article...

FLIR Firefly DL combines Sony image sensor with Intel-Movidius AI processor in a compact body for relatively low price of $299:




Charbax publishes a demo of similar compact AI camera based on Lattice SensAI FPGA framework:

Go to the original article...

Image Sensors World        Go to the original article...

IPVM publishes the results of its testing of lasers impact on surveillance cameras:

"Our testing showed it is difficult to permanently damage surveillance cameras because:

• Close range required: Permanent damage did not occur from distance of ~50 feet or further from the camera, regardless of the strength of the laser we tested.
• Aiming by hand difficult: Targeting a laser towards a camera is difficult from more than ~10' away. Attaching the laser to a stationary object for aiming makes sensor damage much more likely.

Striking or shooting a camera is far more likely to damage a camera than using a laser.

Finally, at close range (5-10'), even small, less than 5 mW laser pointers using AAA batteries (e.g., the ones used to play with pets) were able to damage sensors, albeit only a few pixels at a time. High powered 30,000 mW lasers more quickly damaged sensors and in larger "chunks" of pixels, but were also able to effectively "cook" the camera, creating smoke within a few seconds of steady aiming."

Go to the original article...

Image Sensors World        Go to the original article...

CBS reports that Researchers at the University of Central Florida and Imec propose to use hyperspectral camera to help Florida’s fight against the invasive Burmese python in the Everglades. The python is typically very difficult to detect because of its camouflage markings. Thermal imaging doesn’t work because pythons are cold-blooded and adopt the temperature of their surroundings.

The new hyperspectral camera works in conjunction with an LED light source that together detect a python’s reflection, rather than its heat signature:

Go to the original article...

Image Sensors World        Go to the original article...

Imec publishes a list of its spin-offs in chronological order. Quite many of them are imaging-related:

• FillFactory, high speed imagers (acquired by Cypress, then sold to ON Semi)
• XENICS - IR ROICs and imagers
• Camargus is a start-up centered around multicamera and video stitching technology for the broadcasting market
• Magcam - imaging of magnetic fields
• Morrow - tunable lenses optics for everyday people’s glasses
• Midiagnostics - lens-free-imaging based disposable blood tests
• Spectricity - spectral sensing sensors

Recently, Imec has setup a local VC fund imec.XPAND to support the creation of more start-ups.

Go to the original article...

Image Sensors World        Go to the original article...

PRNewswire: Smartsens wins the “China’s Most Promising Company Award 2019,” jointly hosted by Ernst & Young and Fudan University School of Management. SmartSens was founded as a company focusing on security applications. While it continues to advance its lead in the Security & Surveillance industry, SmartSens is expanding its portfolio of products for artificial intelligence (AI), automotive and ADAS applications.

“We are honored to have been recognized by Ernst & Young and Fudan University, and to have received the ‘China’s Most Promising Company’ Award,” said SmartSens Co-Founder Li Yue. “In recent years, China has seen historic growth in research capabilities at unprecedented speeds, and it is changing the competitive landscape of a multitude of industries, including that of CMOS image sensors. SmartSens will continue leveraging its technological edge and collaboration with industry partners to meet the demands of new applications, driving CMOS image sensing technology into a new era.”


SmartSens also announces the launch of its own LED flicker suppression technology. "LED flicker suppression is a must-have feature for automotive CMOS image sensors," said Chris Yiu, CMO of Smartsens. "This time, through the independent research and development, it has been implemented on domestic CMOS image sensors. This function is believed to be a great success in the development of China's automotive electronics industry. In the future, CMOS image sensor chip products launched by Smartsens for the automotive electronics market will be equipped with this technology to help customers improve automotive electronic systems."

Go to the original article...

Image Sensors World        Go to the original article...

Korea Joongang Daily quotes IHS Markit data on CIS market shares:


IFNews quotes IHS Markit report on smartphone camera market: "IHS Market states that though global smartphone market declines, yet the camera market is growing in 2019. Considering the image sensor size is increasing 40%~50% while upgrade from 16MP to 48MP or even 64MP, and total wafer demand will increase about 50% in 2019."

Go to the original article...

Image Sensors World        Go to the original article...

Caeleste announces its presentations at Toulouse CMOS Image Sensors Workshop to be held on November 26-27, 2019. One of them is about ELFIS project - a rad-hard, GS, true HDR, BSI image sensor, already reported earlier:

Go to the original article...

Image Sensors World        Go to the original article...

Nikkei, Bloomberg, IPVM: China's largest security camera makers and face recognition AI companies have been added to the US trade sanctions list.

Hikvision is the world's largest security camera company. Hikvision has a close cooperation with Sony. The rumor is that Sony makes special image sensor versions optimized to Hikvision requirements that are exclusive to the company. From the company's web site:

"According to yearly independent research data from IHS Markit, Hikvision accounted for 19.5% of market share in global video surveillance industry in 2015, up from 4.6% in 2010, and has been ranked the No.1 market share leader globally for video surveillance equipment for five consecutive years. In 2015, Hikvision was ranked first in EMEA market with 12.2% market share, and was ranked second in Americas market with 7.3% market share.

Hikvision has established partnerships with world technology leaders including Intel, Texas Instruments, Ambarella, Sony, Hisilicon, Western Digital and Seagate."

From Dahua web site: "Dahua Technology has been ranked 2nd in the global CCTV & video surveillance equipment market according to an IHS report since 2014, and was ranked 2nd in “2018 a&s Security 50”."


According to the recent IDC report, "China's video surveillance market is relatively concentrated, with the top three vendors (Hikvision, Dahua and Uniview) having a combined market share of 56.4%. The concentrated market enhances top vendors' economies of scale effects."


Alibaba-backed Megvii filed for an IPO this summer of at least $500m in Hong Kong, while SenseTime raised $620m in a second round of funding in just two months last year and is one of the world's most valuable unicorns in artificial intelligence.

Go to the original article...

Image Sensors World        Go to the original article...

Veoneer presents some of the challenges in design and use of thermal camera in automotive applications:

Go to the original article...

Image Sensors World        Go to the original article...

New Imaging Technologies presentation "SWIR imaging and space applications" by Simon Ferré shows the company's solutions for space:

Go to the original article...

Image Sensors World        Go to the original article...

Counterpoint Research publishes its analysis of smartphone image sensor market. Its key points are:

• An annual revenue growth rate of up to 10%
• The total revenue from smartphones is likely to grow above $12b
• Smartphones is the single largest end-market accounted for nearly 70% of the overall CIS revenue in 2018
• In 2018, each smartphone featured 2.4 image sensors on average
• By 2020, this number is expected to be over three
• In 1H 2019, over US$5b worth of CIS were sold in smartphones.
• Sony Image Sensor Solutions accounted for over 20% of Sony’s total operating income during the quarter ending June 2019
• In Q2 2019, Sony earned 195 billion yen or over US$1.8b from image sensor sales
• Sony is planning to invest in CIS ~700 billion yen (~$6.5b) during the period of 2018~2020
• Samsung LSI is to invest KRW 133 trillion (roughly $10.8b) for non-memory semiconductors
• Samsung CIS business stands at ~$2.46b in terms of revenue
• Samsung’s image sensors are used in its own phones as well as in Chinese brands like Xiaomi, Vivo, and OPPO
• During H1 2019, Sony and Samsung together occupied approximately 85% of sales revenue
• Smaller players like OmniVision and SK Hynix are primarily used in front cameras or multi-camera setups as supplementary sensors
• SK Hynix plans to transform part of its DRAM fabs to production of CIS

Go to the original article...

Image Sensors World        Go to the original article...

Fujitsu publishes a video on how its 3D camera and AI help with scores in gymnastics and more:

Go to the original article...

Image Sensors World        Go to the original article...

Autosens Brussels publishes interviews with Aeye, Fraunhofer, Xenomatix, and Hamamatsu.

Go to the original article...

Image Sensors World        Go to the original article...

IEDM 2019 program has briefly come up online and put down shortly after that yesterday with a promise to be published again in mid-October. Many interesting image sensor papers have been exposed for this short time:

• A 1/2inch 48M All PDAF CMOS Image Sensor Using 0.8µm Quad Bayer Coding 2×2OCL with 1.0lux Minimum AF Illuminance Level
  Tatsuya Okawa, Susumu Ooki, Hiroaki Yamajo, Masakazu Kawada, Masayuki Tachi, Kazuhiro Goi, Takatsugu Yamasaki, Hiroki Iwashita, Sony Semiconductor manufacturing Corporation, Masahiko Nakamizo, Takayuki Ogasahara, Yoshiaki Kitano, Keiji Tatani, Sony Semiconductor Solutions Corporation, Sony Semiconductor Manufacturing Corporation
  We created the world's first all PDAF CMOS image sensor using 2x2 on-chip lens architecture. That had 1/2 inch 48M pixels with 0.8µm Quad Bayer coding for high resolution and HDR function, and all PDAF pixels achieved a minimum AF illuminance level of 1 lux.
• Three-layer Stacked Color Image Sensor With 2.0-μm Pixel Size Using Organic Photoconductive Film
  Togashi Hideaki, Sony Semiconductor Solutions Corporation
  A three-layer stacked color image sensor was formed using an organic film. The sensor decreases the false color problem as it dose not require demosaicing. Furthermore, with the 2.0-μm pixel image sensor, improved spectral characteristics owing to green adsorption by the organic film above the red/blue photodiode, were successfully demonstrated.
• High-definition Visible-SWIR InGaAs Image Sensor using Cu-Cu Bonding of IIIV to Silicon Wafer
  Shuji Manda, Sony Semiconductor Solutions Corporation
  We developed a back-illuminated InGaAs image sensor with 1280 x 1040 pixels at 5-um pitch by using Cu-Cu hybridization connecting different materials, a III-V InGaAs/InP of photodiode array, and a silicon readout integrated circuit (ROIC). A prototype device showed high sensitivity at visible to SWIR wavelengths and low dark current.
• Nanophotonics contributions to state-of-the-art CMOS Image Sensors (Invited)
  Sozo Yokogawa, Sony Semiconductor Solutions Corporation
  Recent progress of Back-illuminated CMOS image sensors (BI-CISs), focusing on their pixel improvements with the design of optical properties using subwavelength sizescale strcutures and photonics technologies, are reviewed. These technologies contribute not only improving BI-CIS basic performance but also adding new functions for versatile sensing applications.
• 14nm FinFET Process Technology Platform for Over 100M Pixel Density and Ultra Low Power 3D Stack CMOS Image Sensor
  Donghee Yu, Choongjae Lee, Myoungkyu Park, Samsung, Samsung Electronics
  CMOS Image Sensor(CIS) products need higher voltage device and better analog characteristics than conventional SOC & Logic products. This work presents newly developed 14nm FinFET process with 2.xV high voltage FinFET device characteristics showing excellent analog and low power digital characteristics comparing to 28nm planar process.
• A 0.8 µm Smart Dual Conversion Gain Pixel for 64 Megapixels CMOS Image Sensor with 12k e- Full-Well Capacitance and Low Dark Noise
  Donghyuk Park, Seung-Wook Lee, Jinhwa Han, Dongyoung Jang, Heesang Kwon, Seungwon Cha, Mihye Kim, Haewon Lee, Sungho Suh, Woong Joo, Yunki Lee, Seungjoo Nah, Heegeun Jeong, Bumsuk Kim, Sangil Jung, Jesuk Lee, Yitae Kim, Chang-Rok Moon, Yongin Park, Samsung Electronics
  A 0.8 μm-pitch 64 megapixels CIS has been demonstrated for the first time. 6k e- full-well capacity (FWC) was achieved, and the advanced color filter isolation was introduced. Dual conversion gain enhanced the Tetracell FWC to 12k e-. Highly refined deep trench isolation and photodiode also improved dark noise characteristics.
• Low-Latency Interactive Sensing for Machine Vision (Invited)
  Paul K. J. Park, Jun-Seok Kim, Chang-Woo Shin, Hyunku Lee, Weiheng Liu, Qiang Wang, Yohan J. Roh, Jeonghan Kim, Yotam Ater, Hyunsurk Ryu, Samsung Electronics
  We introduce the low-latency interactive sensing and processing solution for machine vision applications. The event-based vision sensor can compress the information of moving objects in cost-effective way, which in turn, enables the energy-efficient and real-time processing in various applications such as person detection, motion recognition, and Simultaneous Localization and Mapping.
  
• A 2.2µm Stacked Back Side Illuminated Voltage Domain Global Shutter CMOS Image Sensor
  Geunsook Park, Alan Hsiung, Keiji Mabuchi, Jingming Yao, Zhiqiang Lin, Vincent Venezia, Tongtong Yu, Yu-Shen Yang, Tiejun Dai, Lindsay Grant, OmniVision
  This paper introduces a 2.2µm stacked BSI voltage domain global shutter CMOS image sensor displaying over 100dB shutter efficiency, as well as high NIR-QE of 38% at 940nm, 60% MTF Ny/2 at 940nm with stacked pixel level connections, high density MIM capacitors, and Full back-side Deep Trench Isolations.
• A Highly Reliable Back Side Illuminated Pixel against Plasma Induced Damage
  Yolene Sacchettini, Jean-Pierre Carrère, Célestin Doyen, Stéphane Ricq, Romain Duru, Vincent Goiffon, Pierre Magnan, Kristell Courouble, STMicroelectronics / ISAE-Supaero, STMicroelectronics, Univ. of Toulouse
  Plasma process interaction with BSI image sensor is for the first time presented. The backside dielectrics properties modulate the damage, this was characterized by measuring the dielectrics charge and the interface state density. Metal oxides present a better hardiness to plasma damage due to their negative charge even after plasma.
• Flexible, Active-Matrix Flat-Panel Image Sensor for Low Dose X-ray Detection Enabled by Integration of Perovskite Photodiode and Oxide Thin Film Transistor
  Taoyu Zou, Changdong Chen, Ben Xiang, Ya Wang, Chuan Liu, Shengdong Zhang, Hang Zhou, Peking University Shenzhen Graduate School, Sun Yat-sen University
  An image sensor based on low-cost two-step deposited perovskite photodiode arrays and oxide (IGZO) TFTs is demonstrated for direct and indirect X-ray imaging applications. The system can be fabricated on flexible substrates, and the perovskite photodiode exhibits a significant direct X-ray response, reaching a sensitivity of ~887 μCGy-cm^–2
• Intelligent Vision Systems – Bringing Human-Machine Interface to AR/VR (Invited)
  Chiao Liu, Andrew Berkovich, Song Chen, Hans Reyserhove, Syed Shakib Sarwar, Tsung-Hsun Tsai, Facebook Reality Labs
  An all-day wearable AR/VR device in a glasses form factor needs new input modalities. The candidates include voice, eye gazing, hand/body/head gestures, and BCI. This paper describes computer vision based modalities and the sensor and system specifications, and propose solutions to the extremely stringent power, form factor and performance challenges.
• High-speed Image Processing Devices and Its Applications (Invited)
  Masatoshi Ishikawa, The University of Tokyo
  We have developed a high-speed and low-latency image processing devices and systems. In this talk, their architectures and applications such as robotics, factory automation, fuman interface, bio/medical applications, 3D achieving, and vehicles will be explained by using videos.

IEDM tipsheet explains the impact of Sony InGaAs integration paper:

"Demand for imaging in the short-wavelength infrared range (SWIR, or 1,000-2,000nm wavelengths) has been increasing for industrial, science, medical, agricultural and security purposes. InGaAs has been used to build SWIR sensors because it can absorb light in this range that silicon cannot. With conventional back-illuminated InGaAs sensors, each pixel of a photodiode array is connected to a readout circuit on a silicon wafer by means of a microbump. But it’s difficult to scale these bumps, and so creating finepitch pixel arrays for greater image definition is difficult. A Sony team will describe an architecture in which each pixel in an InGaAs/InP photodiode array is connected to the readout circuit not with microbumps, but by means of copper-to-copper bonding, resulting in a much tighter pitch. They used the technique to build a prototype 1280 x 1024-pixel array with a 5µm pitch. Also, thinning of the InP layer and process optimization yielded a sensor that demonstrated high sensitivity and low dark current, respectively. The researchers say this work paves the way for high-definition SWIR imaging."

Go to the original article...

Image Sensors World        Go to the original article...

Nature article "Why deep-learning AIs are so easy to fool" by Douglas Heaven says:

"A self-driving car approaches a stop sign, but instead of slowing down, it accelerates into the busy intersection. An accident report later reveals that four small rectangles had been stuck to the face of the sign. These fooled the car’s onboard artificial intelligence (AI) into misreading the word ‘stop’ as ‘speed limit 45’.

Such an event hasn’t actually happened, but the potential for sabotaging AI is very real. Researchers have already demonstrated how to fool an AI system into misreading a stop sign, by carefully positioning stickers on it. They have deceived facial-recognition systems by sticking a printed pattern on glasses or hats.

These are just some examples of how easy it is to break the leading pattern-recognition technology in AI, known as deep neural networks (DNNs).

“There are no fixes for the fundamental brittleness of deep neural networks,” argues François Chollet, an AI engineer at Google in Mountain View, California. To move beyond the flaws, he and others say, researchers need to augment pattern-matching DNNs with extra abilities: for instance, making AIs that can explore the world for themselves, write their own code and retain memories. These kinds of system will, some experts think, form the story of the coming decade in AI research."

Go to the original article...