Tuesday, February 19, 2019

Albert Theuwissen Reviews ISSCC Presentations - Part 1

Albert Theuwissen reviews two papers from ISSCC 2019 Image Sensor session held yesterday:

1. Smartsens "A Stacked Global-Shutter CMOS Imager with SC-Type Hybrid-GS Pixel and Self-Knee Point Calibration Single-Frame HDR and On-Chip Binarization Algorithm for Smart Vision Applications" by C. Xu, Y. Mo, G. Ren, W. Ma, X. Wang, W. Shi, J. Hou, K. Shao, H. Wang, P. Xiao, Z. Shao, X. Xie, X. Wang, and C. Yiu:


2. University of Michigan “Energy-efficient low-noise CMOS image sensor with capacitor array-assisted charge-injection SAR ADC for motion-triggered low-power IoT applications” by Kyojin D. Choo, Li Xu, Yejoong Kim, Ji-Hwan Seol, Xiao Wu, Dennis Sylvester, and David Blaauw:

10 comments:

  1. Don't tell me Smartsens infringes these technology patents.

    ReplyDelete
    Replies
    1. from my point of view that their idea to use the first follower as rolling readout is quite new and interesting !

      Delete
  2. 3.5e noise with voltage domain pixels? Seriously, Smartsens? You are beating the laws of physics...

    ReplyDelete
    Replies
    1. there is no measurement condition mentioned for this 3.5e read noise in their table, so I won't take it serious, but i am a bit disappointed that it was presented in such a way on isscc....

      Delete
    2. in the talk was clearly intended in GS mode.

      Delete
  3. Maybe it is the read noise in rolling shutter. The pixel schematic allows rolling shutter read out.

    ReplyDelete
  4. Does the pixel made of 2 wafers without ability to shrink the die size

    ReplyDelete
  5. The ISSCC image sensor papers I heard were ok papers, but not that interesting or inspiring to me. That is the usual nature of ISSCC and I more or less decided on Monday that ISSCC was not a priority in the future for me. The formulaic nature of ISSCC and the very limited time for Q&A, not to mention the difficulty of getting to the microphone queue, makes for a very superficial presentation of new technical work. They should do 10-15 min max presentation and 15-20min Q&A.

    We need to be absolutely sure that IISW never becomes this level of professionally homogenized talks with little discussion. Where is the spark? Where is the spice? Where is the uncovering of important nuances and lessons learned?

    ReplyDelete
    Replies
    1. Albert Theuwissen - Harvest ImagingFebruary 22, 2019 at 12:27 PM

      Eric, I do partly agree, partly disagree.
      ISSCC is running 5 sessions in parallel, to allow people to swap between sessions, so the timing needs to be very strickt. But I do agree that they should shorten the presentation time to benefit from a longer Q&A time. But on the other hand, after the session, there are author interviews. You can have a face-to-face Q&A with the authors while enjoying a glass of wine or a glass of beer (unfortunately not a good Belgian Beer). W.r.t. to access to the microphone : if you choose a seat close to the aisle in the room, access to the microphone is very easy.
      I do fully agree with your statement about the International Image Sensor Workshop ! But that is also why IISW is a WORKSHOP and ISSCC is a CONFERENCE.

      Delete
    2. There were no seats near the aisle that were within convenient viewing distance from the screen. And indeed, I could just email the authors with my questions if I were so inclined, but Q&A often informs everyone at the meeting. Regarding IISW, yes, I remember why we started the workshop. I just worry about it becoming "professionally homogenized" in the future. We did get a lot of great papers to select from for Snowbird so I think IISW will be pretty interesting and informative this June. (btw, preregistration will be announced shortly)

      Delete

All comments are moderated to avoid spam and personal attacks.