Registration: Difference between revisions
Jump to navigation
Jump to search
mNo edit summary |
mNo edit summary |
||
| Line 8: | Line 8: | ||
#* Run it from the top level of the subject directory. | #* Run it from the top level of the subject directory. | ||
#* The first time you run it, run with the <code>--crop</code> option, so it imports the data from the <code>crop_values.txt</code> file. | #* The first time you run it, run with the <code>--crop</code> option, so it imports the data from the <code>crop_values.txt</code> file. | ||
#* <code>batchGen.pl</code> | #* The script <code>batchGen.pl</code> creates three files: <code>regbatch.list</code>, <code>subsetbatch.list</code>, and <code>unixbatch.sh</code>. The script <code>unixbatch.sh</code> uses the other two files to dispatch commands in parallel to a processor farm. The <code>subsetbatch.list</code> has a set of commands invoking the <code>subset</code> program. The <code>regbatch.list</code> file contains a list of <code>WristRegistration</code> commands. | ||
#Run <code>unixbatch</code> from the top level of the subject directory. | <pre>E06001_scan> $G/install_linux/bin/wrist/batchGen.pl --crop</pre> | ||
#Run <code>unixbatch.sh</code> from the top level of the subject directory. | |||
#* This should go through and run all the major steps. | #* This should go through and run all the major steps. | ||
#* <code>./unixbatch</code> | #* <code>./unixbatch.sh</code> | ||
#After <code>unixbatch</code> finishes, you need to validate the output of the registration steps | #After <code>unixbatch.sh</code> finishes, you need to validate the output of the registration steps | ||
#* Use <code>WristVis</code> to check the registration | #* Use <code>WristVis</code> to check the registration | ||
#* If there are errors, re-run <code>batchGen.pl</code>, without the <code>--crop</code> option, and choose NOT to re-run all parts. Then let it hopefully help you through the process. | #* If there are errors, re-run <code>batchGen.pl</code>, without the <code>--crop</code> option, and choose NOT to re-run all parts. Then let it hopefully help you through the process. | ||
#Folders <code>collected_results</code> and <code>auto_registr_resuls</code> contain the same registration results in different formats, whereas <code>classified_scans</code> contains the tissue classification results | #Folders <code>collected_results</code> and <code>auto_registr_resuls</code> contain the same registration results in different formats, whereas <code>classified_scans</code> contains the tissue classification results | ||
Revision as of 19:12, 20 May 2009
- Create the subject directory. It must have the format: E#####_scan; where # is substituted with a digit. Import files/folders from Bioengineering:
original_scans/CT_Scans [scans of all wrist positions]original_stacks/[segmented neutral position of wrist]crop_values.txt
- Run
batchGen.plto create the master batch files.- Run it from the top level of the subject directory.
- The first time you run it, run with the
--cropoption, so it imports the data from thecrop_values.txtfile. - The script
batchGen.plcreates three files:regbatch.list,subsetbatch.list, andunixbatch.sh. The scriptunixbatch.shuses the other two files to dispatch commands in parallel to a processor farm. Thesubsetbatch.listhas a set of commands invoking thesubsetprogram. Theregbatch.listfile contains a list ofWristRegistrationcommands.
E06001_scan> $G/install_linux/bin/wrist/batchGen.pl --crop
- Run
unixbatch.shfrom the top level of the subject directory.- This should go through and run all the major steps.
./unixbatch.sh
- After
unixbatch.shfinishes, you need to validate the output of the registration steps- Use
WristVisto check the registration - If there are errors, re-run
batchGen.pl, without the--cropoption, and choose NOT to re-run all parts. Then let it hopefully help you through the process.
- Use
- Folders
collected_resultsandauto_registr_resulscontain the same registration results in different formats, whereasclassified_scanscontains the tissue classification results