Lists

Thursday, October 12, 2017

Compressed Sensing Said to Save Image Sensor Power

Pravir Singh Gupta and Gwan Seong Choi from Texas A&M University publish an open access paper "Image Acquisition System Using On Sensor Compressed Sampling Technique." They say that "Compressed Sensing has the potential to increase the resolution of image sensors for a given technology and die size while significantly decreasing the power consumption and design complexity. We show that it has potential to reduce power consumption by about 23%-65%."

The proposed sensor architecture implementing this claim is given below:


"Now we demonstrate the reconstruction results of our proposed novel system flow. We use both binary and non-binary block diagonal matrix to compressively sample the image. The binary block diagonal(ΦB) and non-binary block diagonal(ΦNB) sampling matrix are mentioned below."

4 comments:

  1. Almost seems silly to comment on this paper, but it does point to value of peer review. At the very least, the authors should have sent the paper to someone who designs image sensors for some helpful comments to improve and clarify the reporting of this effort.

    ReplyDelete
    Replies
    1. Dear Dr Eric,

      Thanks for your suggestion. Our main intention was to come up with simplified use of Compressed Sensing algorithm. In our work, we wanted to show that we can do compressed sensing in a very simplified way instead of putting complicated random num generators, multipliers etc. on sensor(or off sensor) to implement the same. Our intention was not to come up with a commercial or commercial-like implementation of image sensors. But I would like to learn more from you on how our present design can be improved so that we can come up with much clearer and much better implementation of this work. I am really thankful for your time and effort.
      Thanks,
      Pravir

      Delete
    2. Pravir, my comments refer to the paper presentation, not the sensor design per se. What is the impact of the CS on pixel random noise (low light performance)? Also what is the impact on total system power including reconstruction? A table showing key design and simulation features (pixel size, fill factor, array size, frame rate, power breakdown (you have this) etc.) would be helpful. Redrawing circuits in an easy to view format is pretty normal, and delineating what happens in the pixel, column and global sections would be useful. Why not show layout for a smaller pixel? What niche need does such a sensor fulfill? Addressing these things in the paper makes the impact of your hard work easier to assess and appreciate. The input of a sensor designer is to help identify those things that need to be clarified in order to make the paper more useful to that community -- the community you are trying to reach.

      Delete
  2. Thanks a lot Dr. Eric. I will keep all these points in mind for my future work.

    ReplyDelete

All comments are moderated to avoid spam and personal attacks.