Presentation is loading. Please wait.

Presentation is loading. Please wait.

Outline  BatchMake Overview  BatchMake and XNAT Integration  XNAT Overview  GridWizard.

Similar presentations


Presentation on theme: "Outline  BatchMake Overview  BatchMake and XNAT Integration  XNAT Overview  GridWizard."— Presentation transcript:

1

2 Outline  BatchMake Overview  BatchMake and XNAT Integration  XNAT Overview  GridWizard

3 BatchMake Overview  Goals  BatchMake  Slicer BatchMake Modules  Use-case High-throughput mouse brain morphology  Funding NIH/NLM A2D2 Award, 2004-2005 NAMIC, 2007-2008 NIH/NIMH STTR Phase I, High Throughput Web- Based Image Analysis of Mouse Brain MR Imaging Studies, UNC, PI: M. Styner, 2007-2008

4 Goals  Facilitate the development and deployment of scripts for high-throughput image processing  Simplify the specification and reporting of statistical experiments  Use cases Develop the script using 10 images on a local PC, and deploy using 1,000s of images on a Grid Conduct and report a Taguchi experiment to determine the significant parameters of a method Conduct leave-on-out analyses to compare two different methods

5 BatchMake  Scripting language  Editor  Processing engines  Report Server

6 Scripting Language  Modeled after CMake Set( VAR value ) ListFilesInDir( VAR dir ) ForEach( VAR list ) ForNFold( ITERATION TRAININGSET i TESTINGSET i fullSet numberOfFolds )  Extensible

7 Scripting Language  Applications that use Slicer execution model are available as commands # Find RegisterImages and determine # its command-line parameters Set( @RI RegisterImage ) Set( RI.FixedImage filename) Set( RI.MovingImage filename) Set( RI.Registration “Bspline”) Run( RI )

8 Editor  Syntax highlighting  Define non-Slicer apps as commands  Launch a processing engine  Monitor results Loops are unrolled ○ Jobs ○ Data ○ Steps

9 Processing Engines  Launched from Editor  Launched from command-line  Local processing (bash / bat)  Condor Grid  Sun Grid  KWGrid  Globus (limited support)

10 Report Server: BatchBoards  Script declares an experiment design, independent vars, dependent vars, thumbnails, and a BatchBoard server  BatchBoard server (website and mysql database) receives data from scripts, provides a database browser, provides experiment reports.

11 BatchBoard Reports  Taguchi table webpage  Graphs of independent vs. dependent vars.  Color highlighting when dependent vars passes a threshold

12

13 Slicer Modules (approx.)  Enable BatchMake (automatic with GetBuildTest2)  “Batch Processing” filter group Applications/CLI/BatchMake  EMSegmenter  GaussianBlur (“hello world”)

14 High-throughput mouse brain morphology (Phase II)  The proposed web-based system with the server side modules will be evaluated directly by end-users in the following studies: Fragile X mouse behavioral phenotype from Phase 1, Neuregulin modulated schizophrenia mouse phenotype, Postnatal alcohol exposure in adolescent mice, Intra-uterine cocaine exposure in rats.

15 BatchMake and XNAT  Explore data via XNAT  Export pointers to interesting data as an XNAT Catalog  BatchMake commands exist for creating lists from XNAT Catalogs  BatchMake commands exist for downloading data from XNAT (status?)

16 XNAT  Server  Desktop  File format for meta-data and data collections

17 Grid Wizard  Command-line tool for distributing a program to a variety of compute services


Download ppt "Outline  BatchMake Overview  BatchMake and XNAT Integration  XNAT Overview  GridWizard."

Similar presentations


Ads by Google