Skip to content

Milestones

List view

  • Delivery of a GitHub repository containing a fully documented python framework for training a deep neural network to skeletonize neural dendrites.

    Overdue by 2 month(s)
    Due by June 30, 2025
    2/2 issues closed
  • Delivery of a GitHub repository containing a fully documented python framework for training a deep neural network to skeletonize neural dendrites

    No due date
  • Creation of a post-processing method to reconnect sub-dendrites that by assumption were “broken” in the inference process.

    No due date
    4/4 issues closed
  • Creation of a production model, trained on 3334^2 pixel images with (a) image tiling and inference on individual tiles, (b) evaluation of tiles-to-image reconstruction methods on analysis results.

    No due date
    3/3 issues closed
  • Creation of a statistical framework to compare results, in python

    No due date
    3/3 issues closed
  • 3. Creation of a PoC model trained on synthetic 1024^2 images 1. no tiling 2. 5 noise levels 3. statistical comparison to PIs results

    No due date
    2/2 issues closed
  • The PI team’s Matlab code analysis results will represent the state-of-the-art analysis for the purposes of this project.

    No due date
    2/2 issues closed
  • Delivery to SSEC of simulated image-skeleton training image pairs in pixel sizes 1024^2 and 3334^2.

    No due date
    1/1 issues closed