Agenda&Status 081010: Difference between revisions

From KitwarePublic
Jump to navigationJump to search
Line 81: Line 81:
** Needs
** Needs
*** Large data sets (greater than RAM) - 100GB to 1TB per experiment
*** Large data sets (greater than RAM) - 100GB to 1TB per experiment
*** PixelType may be scalar, vector, or tensor
*** How to deal with multiple channels and time points (e.g. dimensions, vector)
*** How to deal with multiple channels and time points (e.g. dimensions, vector)
*** PixelType may be scalar, vector, or tensor
*** ordering dimensions for iteration (e.g. tracking nuclei needs time before channel whereas segmentation might want channel before time)
*** Subset/crop/resample n dimensional image into m dimensional image (m<n)
*** Crop and resample n dimensional image into m dimensional image (m<n)
*** ordering dimensions for iteration (tracking nuclei need time before channel whereas segmentation might want channel first)
*** Having a multi-resolution representation for images ? ( ala JPEG2000, or wavelets ?)
*** Having a multi-resolution representation for images ? ( ala JPEG2000, or wavelets ?)



Revision as of 16:38, 10 August 2010

How to Join the Tcon

Number to Call

  • 1-800-728-9607 (in the US) or
  • +1 9139049873 (international)
  • access code 6815251

Webex

Project Management

  • Number of Work Assignments?
  • Deadline for sending the Draft of T.O 2?
  • Minutes of Washington meeting from NIH?
  • date of next meeting?

Technical Topics

Stable ITK 3.20

  • How to get it.
  • Maintenance rules (gate keeper ?)

DICOM

Dan and Mathieu to report

Simple ITK

Wrapping

  • Plans for wrapping
  • Tools
    • Pros and Cons
  • Wrapping
  • update from brad on wrapitk test (new swig usage that addresses existing concerns)?

Refactoring Optimization Framework

Modularization

  • Gabe Hart to report of initial Modularization efforts
  • Exploration of Ryppl
    • Interaction with Boost
    • Brad King & Dave Cole Report

Microscopy

  • Scheduling a dedicated tcon for this topic.

Gathering Feedback from the attendees to the Hackathon in Janelia Farm (HHMI)

Management of Large Datasets

  • Sean Megason & Arnaud Gelas
    • Needs
      • Large data sets (greater than RAM) - 100GB to 1TB per experiment
      • PixelType may be scalar, vector, or tensor
      • How to deal with multiple channels and time points (e.g. dimensions, vector)
      • ordering dimensions for iteration (e.g. tracking nuclei needs time before channel whereas segmentation might want channel before time)
      • Crop and resample n dimensional image into m dimensional image (m<n)
      • Having a multi-resolution representation for images ? ( ala JPEG2000, or wavelets ?)
    • Partial solutions
      • meta reader - n-dimensional image set stored on disk (or database) as series of 2-d images. Reader can grab any subset of dimensions and stream. Needs an n-dimensional container (hashtable) that can be sorted.

ITK Readers for Microscopy formats

  • There's a need for readers in the microscopy community in order to be able to use ITK
    • Zeiss (LSM and ZVI)
    • Leica (LIF)
    • Olympus (OIB)
    • Nikon ()
  • BioFormats has readers and example files for these. It is Java with C++ bindings for ITK, BUT it's GPL.

Software Process

Migration to Git

  • Updating Dashboard Machines to Git
  • Contractors to send information to Kitware about the developers who should receive Git write access.
    • (include email and public SSH key)

Before Giving Git Write Access


Clean up Releases

Release Tag Date Major Changes
Alpha-01 August 7 VS6,BCC55, SunCC (5.7?), Metrowerks Removal (based on compiler features), Remove large Images
Alpha-02 August 14 CMake Variables configuration cleanup
Alpha-03 August 21 GDCM 2.0, OpenJpeg-v2, +JPEG2000
Alpha-04 August 28 Third party libraries update: TIFF, JPEG, PNG

ITKv4-TransitionPlanProposal.png


  • Hans Johnson to report on compilers cleanup.
    • and experiences with Git workflow.


Tools

Coding Style

ITKv4_StyleChangeProposal

  • Fixing KWStyle/CDash/CTest for Git : Brad Davis report