Friday, April 19, 2013

Image Sensors at VLSI Symposia 2013

VLSI Circuit Symposium to be held on June 11-14, 2013 in Kyoto, Japan has Session 2 entirely devoted to image sensors:

A 1-inch Optical Format, 14.2M-Pixel, 80fps CMOS Image Sensor with a Pipelined Pixel Reset and Readout Operation
H. Honda*, S. Osawa*, M. Shoda*, E. Pages*, T. Sato*, N. Karasawa*, B. Leichner**, J. Schoper**, E.S. Gattuso**, D. Pates**, J. Brooks**, S. Johnson**and I. Takayanagi*, *Aptina Japan and **Aptina Imaging, Japan
A 1-inch optical format, 14.2M-pixel, 80fps, digital-output CMOS image sensor that employs a row-shared dual conversion gain pixel is presented. To achieve the 80fps readout rate, a pipelined pixel reset/readout scheme named "nesting scan" has been introduced, where the charge sense node inside a pixel is reset during the previous row. Readout noise and maximum handling signal charge of the sensor are 1.7e-rms and 16ke-, respectively, and the ratio of column FPN to readout noise is lower than 0.03 in all gain conditions.

A 5.9μm-Pixel 2D/3D Image Sensor with Background Suppression over 100klx
J. Cho*, J. Choi*, S.-J. Kim**, J. Shin***, S. Park*, J.D.K. Kim*** and E. Yoon*, *University of Michigan, USA, **Institute of Microelectronics, Singapore and ***Samsung Advanced Institute of Technology, Korea
A 2D/3D image sensor with reconfigurable pixel array and column-level background suppression scheme is presented for high resolution outdoor imaging. The proposed pixel array employs pixel binning and superresolution techniques for adaptable resolution. The sensor achieved a 5.9μm pixel and was able to capture full resolution outdoor depth images under daylight over 100klx.

An Ultra-Low Noise Photoconductive Film Image Sensor with a High-Speed Column Feedback Amplifier Noise Canceller
M. Ishii, S. Kasuga, K. Yazawa, Y. Sakata, T. Okino, Y. Sato, J. Hirase, Y. Hirose, T. Tamaki, Y. Matsunaga and Y. Kato, Panasonic Corporation, Japan
We developed an ultra-low noise image sensor in which an organic photoconductive film (OPF) is laminated on the entire surface of the pixel circuits. In order to suppress the kTC noise in the pixel circuit of a three transistor configuration, a high- speed column feedback noise cancel circuit is newly developed. An ultra-low noise of 2.9 electrons during the horizontal blanking period of only 5 μs has been achieved.

A 100-fps Fluorescence Lifetime Imager in Standard 0.13-μm CMOS
R.M. Field and K.L. Shepard, Columbia University, USA
A wide-field fluorescence lifetime imager capable of up to 100 frames per second (fps) is presented. The imager consists of a 64-by-64 array of low-noise single photon avalanche diodes (SPADs) in a standard 0.13-micron CMOS process, 4096 time-to- digital converters, and an application specific data path to enable continuous image acquisition at a total output data rate of 42 Gbps. These features combine to enable new lifetime-based diagnostic imaging.

820-GHz Imaging Array Using Diode-Connected NMOS Transistors in 130-nm CMOS
D.Y. Kim*, S. Park*, R. Han** and K.K. O*, *University of Texas and **Cornell University, USA
An 820-GHz 8X8 imaging array using diode-connected NMOS transistor detectors is demonstrated in 130-nm CMOS process. Measured mean responsivity of 3.4 kilo volt per watt and mean NEP of 28 pico watt per root Hz at 1MHz modulation frequency are achieved. The NEP is 3.5X lower than that of NMOS and slightly lower than that of Schottky diode terahertz imaging arrays implemented in CMOS. The minimum NEP is 15.5 pico watt per root Hz, which is the lowest for THz detector arrays fabricated in CMOS. The imaging array occupies 2.0X1.7mm2 and the power consumption is 9.6 mW.

Session 14 too has papers on image sensors:

A 100,000 fps Vision Sensor with Embedded 535GOPS/W 256x256 SIMD Processor Array
S.J. Carey, A. Lopich, D.R.W. Barr, B. Wang and P. Dudek, The University of Manchester, United Kingdom
A vision chip operating with 1.9pJ/op efficiency has been fabricated in 0.18μm CMOS. Each of the 256x256 pixel-processors (dimensions 32x32μm) contains 14 binary and 7 analog registers coupled to a photodiode, an arithmetic logic unit, diffusion and asynchronous propagation networks. At the chip's periphery, facilities exist to allow pixel address extraction, analog or digital readout. The chip has been exploited to conduct real-time image processing operations at 100,000fps, locating a closed-shape object from amongst clutter.

A 240x180 10mW 12us Latency Sparse-Output Vision Sensor for Mobile Applications
R. Berner, C. Brandli, M. Yang, S.-C. Liu and T. Delbruck, University of Zurich and ETH Zurich, Switzerland
This paper proposes a 0.18μm CMOS vision sensor that combines event-driven asynchronous readout of temporal contrast with synchronous frame-based active pixel sensor (APS) readout of intensity. The sensor is suitable for mobile applications because it allows low latency at low data rate and therefore, low system-level power consumption. The image frames can be used for scene analysis and the temporal contrast events can be used to track fast moving objects, to adjust the frame rate or to guide a region of interest readout. Sharing the photodiode for both readout types allows a compact pixel design that is 60% smaller than a comparable technology. The 240x180 sensor has a power consumption of 10mW; the temporal contrast pathway has a minimum latency of 12us, a dynamic range of 120dB with 12% contrast detection threshold and 3.5% contrast matching; the APS readout has 57dB dynamic range with 1% FPN.

Technology Session 2 too has a nice paper on organic film sensor by Panasonic and Fujifilm:

Thin Organic Photoconductive Film Image Sensors with Extremely High Saturation of 8500 Electrons/μm2
M. Mori*, Y. Hirose*, M. Segawa*, I. Miyanaga*, R. Miyagawa*, T. Ueda*, H. Nara*, H. Masuda*, S. Kishimura*, T. Sasaki*, Y. Kato*, Y. Imada**, H. Asano**, H. Inomata**, H. Koguchi**, M. Ihama** and Y. Mishima**, *Panasonic Corporation and **FUJIFILM Corporation, Japan
We have developed an image sensor with thin organic photoconductive film (OPF) laminated on CMOS circuits. Owing to high capacity of a charge storage node, the saturation level is 12 dB higher than those of conventional image sensors. Because of the very thinness of the laminated film, i.e. 0.5 μm, the device is crosstalk-free and an incident light angle of over 30 degrees is realized.

The VLSI Symposium tip sheet gives a cross section of the 3um and 0.9um organic pixels in the paper:

The Technology Symposium has Samsung ToF paper in Session 10:

Time of Flight Image Sensor with 7μm Pixel and 640x480 Resolution
S. Kim, S. Cha, H. Park, J. Gong, Y. Noh, W. Kim, S. Lee, D.-K. Min, W. Kim, T.-C. Kim and E. Jung, Samsung Electronics Co., Ltd., Korea
Time of flight (ToF) sensor with pixel size of 7x7μm and VGA resolution is developed using a backside illumination (BSI) structure. Quantum efficiency (QE) of near infrared (NIR) light is improved dramatically by applying thick epitaxial layer, reflection metal and anti-reflection layer. The depth error ranges are 2cm and 10cm at 90% and 10% reflection condition at the distance of 7m, respectively.

Also, there is an hour-long Advanced CMOS Image Sensors lecture by Y. Egawa, Toshiba in VLSI Technology Short Course a day before the Symposia start.


  1. Almost at the same time when Aptina presents the 14M at this VLSI Symposium, they also present its little sister (10M) at the IISW. Would have loved to see it the other way.

  2. Albert, 10MP is only one of the working modes of this sensor and you will see more papers on other working modes very soon. :)

  3. "An Ultra-Low Noise Photoconductive Film Image Sensor with a High-Speed Column Feedback Amplifier Noise Canceller" What is the principle of this low noise readout?? Is this similar to what proposed Fowler before??

  4. Albert, we would love to see the IISW program announced.

    1. Albert TheuwissenApril 22, 2013 at 6:52 PM

      I quickly checked and I was told that there are some last-minute changes in the program and that the technical program chair is working to fit the last pieces of the puzzle. On the short term the program will become public.

  5. Seeing not one but two presentations of organic film sensors, tough both by the same corporations, it makes me wonder what the current state of them are and what are the main weaknesses when it comes to implementing them in SLRs.

    Additionally I wonder why those sensors have micro-lenses at all? Could they cause more issues than solve?


All comments are moderated to avoid spam.