Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-10827

Work on GLMM paper example

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: meas_deblender
    • Labels:
      None
    • Story Points:
      11
    • Sprint:
      DRP F17-1, DRP F17-2
    • Team:
      Data Release Production

      Description

      Peter Melchior and I have been working on a paper to publish the GLMM algorithm that we developed as part of the deblender. This ticket exists to create a more simplistic example than the deblender (hyperspectral imaging) for the paper.

        Attachments

          Activity

          Hide
          fred3m Fred Moolekamp added a comment -

          The main work in this ticket was to create an example using hyperspectral data to be included in the paper. There is still some writing that will need to be done, which can be scheduled in a separate (shorter) ticket.

          Show
          fred3m Fred Moolekamp added a comment - The main work in this ticket was to create an example using hyperspectral data to be included in the paper. There is still some writing that will need to be done, which can be scheduled in a separate (shorter) ticket.
          Hide
          fred3m Fred Moolekamp added a comment -
          Show
          fred3m Fred Moolekamp added a comment - See the following notebooks for demonstration of the algorithm: https://github.com/fred3m/hyperspectral/blob/master/washington_dc.ipynb (straight NMF with normalization) https://github.com/fred3m/hyperspectral/blob/master/washington_dc-sparsity.ipynb (sparsity) https://github.com/fred3m/hyperspectral/blob/master/washington_dc-entropy.ipynb (entropy, not currently working)
          Hide
          fred3m Fred Moolekamp added a comment -

          Hi Peter,
          Can you take a look at the hyperspectral results? I still can't get the entropy solution to work, so after you take a look at these we can discuss whether or not it's worth spending any more time on this.

          Show
          fred3m Fred Moolekamp added a comment - Hi Peter, Can you take a look at the hyperspectral results? I still can't get the entropy solution to work, so after you take a look at these we can discuss whether or not it's worth spending any more time on this.
          Hide
          pmelchior Peter Melchior added a comment -

          The examples is well set up. Sparsity does seem to do some good, but I don't like the normalization of the per-pixel component amplitudes. It necessitates the use of a component for shadows just because it is dark, not because of any inherent spectrum it may have. That's against the spirit of spectral unmixing...

          Show
          pmelchior Peter Melchior added a comment - The examples is well set up. Sparsity does seem to do some good, but I don't like the normalization of the per-pixel component amplitudes. It necessitates the use of a component for shadows just because it is dark, not because of any inherent spectrum it may have. That's against the spirit of spectral unmixing...
          Hide
          fred3m Fred Moolekamp added a comment -

          Increased story points since the majority of this ticket was discovering bugs/misconceptions about the algorithm and work on the actual paper only began in earnest after the first 6 story points were completed.

          Show
          fred3m Fred Moolekamp added a comment - Increased story points since the majority of this ticket was discovering bugs/misconceptions about the algorithm and work on the actual paper only began in earnest after the first 6 story points were completed.
          Hide
          fred3m Fred Moolekamp added a comment -

          Peter Melchior, is it ok to close this ticket now?

          Show
          fred3m Fred Moolekamp added a comment - Peter Melchior , is it ok to close this ticket now?
          Hide
          pmelchior Peter Melchior added a comment -

          Please add the convergence plots for A,S and the relevant Z variables.

          Show
          pmelchior Peter Melchior added a comment - Please add the convergence plots for A,S and the relevant Z variables.
          Hide
          pmelchior Peter Melchior added a comment -

          Example is convincing

          Show
          pmelchior Peter Melchior added a comment - Example is convincing

            People

            • Assignee:
              fred3m Fred Moolekamp
              Reporter:
              fred3m Fred Moolekamp
              Reviewers:
              Peter Melchior
              Watchers:
              Fred Moolekamp, Peter Melchior
            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Summary Panel