From rodell at pet.auh.dk Wed May 3 04:43:05 2006 From: rodell at pet.auh.dk (Anders Bertil Rodell) Date: Wed, 03 May 2006 10:43:05 +0200 Subject: [MINC-users] running classify unsupervised Message-ID: <1146645785.28349.6.camel@vega> Does anyone have some examples of how to use the unsupervised -fcm,-hcm, -apriory and -fuzzy options in classify? I have tried a multitude of combinations of the options and the only way I can get fuzzy results seems to be through first creating a training file using a tagfile.. this seems a bit supervised to me Thanks for any help. Anders -- Anders Bertil Rodell PET Centre Aarhus General Hospital From a.janke at gmail.com Thu May 4 02:35:31 2006 From: a.janke at gmail.com (Andrew Janke) Date: Thu, 4 May 2006 02:35:31 -0400 Subject: [MINC-users] running classify unsupervised In-Reply-To: <1146645785.28349.6.camel@vega> References: <1146645785.28349.6.camel@vega> Message-ID: Hi Anders, As far as I am aware classify doesn't do unsupervised classification. In all cases it will train from an input set of tags/classes. I suspect by unsupervised this is merely describing the method of clustering that is used from data generated from the input tag points. I may be wrong and you can do real unsupervised learning with classify, so I shall watch with interest to see if anyone corrects me! :) a On 5/3/06, Anders Bertil Rodell wrote: > Does anyone have some examples of how to use the unsupervised > -fcm,-hcm, -apriory and -fuzzy options in classify? > I have tried a multitude of combinations of the options and the only way > I can get fuzzy results seems to be through first creating a training > file using a tagfile.. this seems a bit supervised to me > > Thanks for any help. > Anders > > -- > Anders Bertil Rodell > PET Centre Aarhus General Hospital > > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From pellegrinis at ninds.nih.gov Thu May 4 15:16:29 2006 From: pellegrinis at ninds.nih.gov (Stefano Pellegrini) Date: Thu, 04 May 2006 15:16:29 -0400 Subject: [MINC-users] real range and pixel range Message-ID: <445A530D.4000307@ninds.nih.gov> Hi all, when I copy a volume using the function (from volume_io) copy_volume_definition, I think I copy also the range of the real value. How can I unset this real range in the copied volume in such a way that the real range match with the pixel range? In general, how can I get rid of the double range? thank you, -stefano From a.janke at gmail.com Thu May 4 18:56:23 2006 From: a.janke at gmail.com (Andrew Janke) Date: Thu, 4 May 2006 18:56:23 -0400 Subject: [MINC-users] real range and pixel range In-Reply-To: <445A530D.4000307@ninds.nih.gov> References: <445A530D.4000307@ninds.nih.gov> Message-ID: Well in this case the default real range of a file is typically 0..1 So in this case your best bet is to first read the voxel range and then set the real range accordingly. ie get_volume_voxel_range() set_volume_real_range() See http://www.bic.mni.mcgill.ca/software/MDP/HTML/VOLUME_IO_prog_ref.html/volume-manipulation.html for more info. a On 5/4/06, Stefano Pellegrini wrote: > Hi all, > > when I copy a volume using the function (from volume_io) > copy_volume_definition, I think I copy also the range of the real value. > How can I unset this real range in the copied volume in such a way that > the real range match with the pixel range? In general, how can I get rid > of the double range? > > thank you, > > -stefano > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From louis.collins at mcgill.ca Sat May 6 14:10:14 2006 From: louis.collins at mcgill.ca (D. Louis Collins) Date: Sat, 06 May 2006 14:10:14 -0400 Subject: [MINC-users] running classify unsupervised In-Reply-To: References: <1146645785.28349.6.camel@vega> Message-ID: <893d17436b821443aeeb105934edccf0@mcgill.ca> Anders, I concur with Andrew. The -mm, -knn, -ann, -hcm, -fcm and -bayes all need estimates of the intensities to build the internal structure required to do the segmentation. So, they all need training data. The training data can come from a training file (containing the intensity info like mean, sd) or a tag file (then the program extracts the mean/sd from the volumes. We do not have an EM technique coded up in classify. However, there is a bit of EM code written by Sylvain Prima. I think that it is in the CVS repository. -Louis On May 4, 2006, at 2:35 AM, Andrew Janke wrote: > Hi Anders, > > As far as I am aware classify doesn't do unsupervised classification. > In all cases it will train from an input set of tags/classes. I > suspect by unsupervised this is merely describing the method of > clustering that is used from data generated from the input tag points. > > I may be wrong and you can do real unsupervised learning with > classify, so I shall watch with interest to see if anyone corrects me! > :) > > > a > > > On 5/3/06, Anders Bertil Rodell wrote: >> Does anyone have some examples of how to use the unsupervised >> -fcm,-hcm, -apriory and -fuzzy options in classify? >> I have tried a multitude of combinations of the options and the only >> way >> I can get fuzzy results seems to be through first creating a training >> file using a tagfile.. this seems a bit supervised to me >> >> Thanks for any help. >> Anders >> >> -- >> Anders Bertil Rodell >> PET Centre Aarhus General Hospital >> >> _______________________________________________ >> MINC-users at bic.mni.mcgill.ca >> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users >> > > > -- > Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) > Canada->Montreal Cell: +1 (514) 924 > 2012 > > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > From shahin at bic.mni.mcgill.ca Mon May 8 11:17:34 2006 From: shahin at bic.mni.mcgill.ca (Shahin Zangenehpour) Date: Mon, 08 May 2006 11:17:34 -0400 Subject: [MINC-users] Concerns Re. Strange-looking MNC Output Files Message-ID: Dear All, I hope someone can help me with the following problem. I ran co-variation analyses using DOT on a number of experimental conditions (found in study "SpaceTime2" in /home/bic/shahin/dot/Dbase/ Tables). The operations are labelled as "FCA2", "FCV21" and "FCV214", with their respective voi regressors labelled exactly the same way. The output files are stored in "/data/zlab/shahin/space-time2/ analysis/output/stat/" and they all start with "SpaceTime2_stat_FC" followed by the appropriate strings for each condition (e.g., /data/ zlab/shahin/space-time2/analysis/output/stat/ SpaceTime2_stat_FCA2_FCA2.mnc). As you will note, the outcome of these operations have truncated volumes as opposed to other _stat_ files in the same folder. Sylvain has suggested in passing that there may be problems with the registration process or the transformation files. However, I don't have the slightest clue in how to go about identifying and subsequently solving such problems. Can anyone help directing me to the source? I hope I have provided all the essential information to track this down. Many thanks in advance, SZ ________________________________________________________________________ _______________________________ Shahin Zangenehpour, PhD Postdoctoral Fellow | Montreal Neurological Institute | Cognitive Neuroscience Unit 3801 University Street | Room 276 | Montr?al QC Canada H3A 2B4 | P. 514 398 1717 | F. 514 398 1338 From sylvain at bic.mni.mcgill.ca Mon May 8 12:23:23 2006 From: sylvain at bic.mni.mcgill.ca (Sylvain MILOT) Date: Mon, 8 May 2006 12:23:23 -0400 Subject: [MINC-users] Concerns Re. Strange-looking MNC Output Files In-Reply-To: Message-ID: Hello Shahin, a few things ... If you look at mripet_rsl/guizard_nicolas_200602161008_* with register, you'll notice it is way off. I havent seen many of these over the years since this process is very robust ... but it happens. I would redo this one with 'do_mritopet -mritopet -tracc ...' see mritopet -help for more info. First you should isolate this subject into a separate directory to ease things with do_mritopet and do_pettopet and you can easily do this with symlinks (ln -s). Verify the new results and then redo the do_pettopet part for that subject only. the recipe is: cd /data/zlab/shahin/space-time2/analysis mkdir new cd new mkdir mri pet mrital_xfm (cd mri; ln -s ../../mri/guizard_nicolas_20060127_140219_2_mri.mnc .) (cd pet; ln -s ../../pet/guizard_nicolas* .) (cd mrital_xfm; ln -s ../../mrital_xfm/guizard_nicolas_mrital.xfm .) do_mritopet -mritopet -tracc -batch register mripet_rsl/* # if good batch do_pettopet # if not good call me Also in dot' s setup for this study, the mripet transformation file is wrong for Lauren Klimoff - it should be klimoff_lauren_200601050920_mripet.xfm. The one specified is based on the transmission image. Aside from that, your data is pretty noisy, ie lots of residual uncorrected patient movement ... add to that the fact that your t-tests are not very strong (ie testing normalized CBF with no possibility of accounting for fixed subject effect) and that you probably shouldn't be using a pooled sd measure ... Sylvain On Mon, 8 May 2006, Shahin Zangenehpour wrote: > Dear All, > > I hope someone can help me with the following problem. > > I ran co-variation analyses using DOT on a number of experimental > conditions (found in study "SpaceTime2" in /home/bic/shahin/dot/Dbase/ > Tables). > > The operations are labelled as "FCA2", "FCV21" and "FCV214", with > their respective voi regressors labelled exactly the same way. > > The output files are stored in "/data/zlab/shahin/space-time2/ > analysis/output/stat/" and they all start with "SpaceTime2_stat_FC" > followed by the appropriate strings for each condition (e.g., /data/ > zlab/shahin/space-time2/analysis/output/stat/ > SpaceTime2_stat_FCA2_FCA2.mnc). > > As you will note, the outcome of these operations have truncated > volumes as opposed to other _stat_ files in the same folder. Sylvain > has suggested in passing that there may be problems with the > registration process or the transformation files. However, I don't > have the slightest clue in how to go about identifying and > subsequently solving such problems. Can anyone help directing me to > the source? I hope I have provided all the essential information to > track this down. > > Many thanks in advance, > SZ > > ________________________________________________________________________ > _______________________________ > Shahin Zangenehpour, PhD Postdoctoral Fellow | Montreal > Neurological Institute | Cognitive Neuroscience Unit > 3801 University Street | Room 276 | Montr?al QC Canada > H3A 2B4 | P. 514 398 1717 | F. 514 398 1338 > > > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > --- Sylvain Milot (sylvain at bic.mni.mcgill.ca) (trinity at bic.mni.mcgill.ca) Brain Imaging Centre Montreal Neurological Institute Webster 2B, Room 208 Montreal, Qc., Canada, H3A 2B4 Phone : (514) 398-4965, Fax: 398-8948 Mobile : (514) 712-1768 Office : 527 Av Des Pins O., Room 204 Montreal, Qc., H2W 1S4 From pellegrinis at ninds.nih.gov Tue May 9 13:01:47 2006 From: pellegrinis at ninds.nih.gov (Stefano Pellegrini) Date: Tue, 09 May 2006 13:01:47 -0400 Subject: [MINC-users] keep_real_range Message-ID: <4460CAFB.80108@ninds.nih.gov> Hi all, I am tring to resample a b1 map (b1_map.mnc) to another image (emt_off.mnc) but I have a problem with the range of the output image. the input image, has this range file: b1_map.mnc image: signed__ double -100 to 50 while the output iamge has file: reg_rig_off_b1_map.mnc image: signed__ double 0 to 1 Instead I don't want the output range to change. the command that I have used is mincresample -clobber -transformation transformation.xfm -like emt_off.mnc -keep_real_range b1_map.mnc res_b1_map.mnc When I try to open it with AFNI, it recognize the original range. I think there is something that is not clear to me. Can anybody help me? thanks, -stefano From citrog at gmail.com Tue May 9 17:44:50 2006 From: citrog at gmail.com (Gil Citro) Date: Tue, 9 May 2006 17:44:50 -0400 Subject: [MINC-users] Same Problem - Error compiling conglomerate 1.6 Message-ID: <9c80f12d0605091444p6563a62ycbe71901e6b84c43@mail.gmail.com> I posted a very similar problem to this before, but unfortunately the previous solution doesn't seem to be working. The problem is that when I try to compile conglomerate 1.6, I get an error. The following are the last several lines of compiler output, and I've attached the while thing as compile-error. /usr/local/lib/libbicpl.a(rgb_io.o)(.text+0x226):/usr/local/mni/bicpl-1.4.2/Images/rgb_io_ppm.c:100: undefined reference to `pm_allocrow' /usr/local/lib/libbicpl.a(rgb_io.o)(.text+0x2d8):/usr/local/mni/bicpl-1.4.2/Images/rgb_io_ppm.c:118: undefined reference to `ppm_writeppmrow' /usr/local/lib/libbicpl.a(rgb_io.o)(.text+0x2e9):/usr/local/mni/bicpl-1.4.2/Images/rgb_io_ppm.c:121: undefined reference to `pm_freerow' collect2: ld returned 1 exit status make[1]: *** [composite_images] Error 1 make[1]: Leaving directory `/usr/local/mni/conglomerate-1.6' make: *** [all] Error 2 enter the text in that file's own buffer. The previous solution was to use the --with-image-ppm option with ./configure for bicpl 1.4.2 and then to manually make the following change in the Makefile Old: LIBS = -lppm -lvolume_io -lminc -lnetcdf -lm New: LIBS = -lppm -lvolume_io -lminc -lnetcdf -lm -lppm and to make sure libnetpbm9-dev was installed, which it is. If I configure bicpl with --with-image-ppm but don't change the Makefile, bicpl won't compile. If I configure bicpl without --with-image-ppm , bicpl will compile, as will congolmerate, but then ray_trace will not. Between this time and last time, I have reinstalled Debian Sarge from scratch. It's an Athlon64 3000+ box, and last time I was using a 64 bit kernel, but this time I'm using the regular Debian 32 bit kernel to improve compatibility with some software. Other than that, I think everything is the same. I think I reinstalled all the packages that were previously installed when it worked. If anyone has any suggestions for what to try next, I'd appreciate it. Thanks! Gil From a.janke at gmail.com Thu May 11 00:00:07 2006 From: a.janke at gmail.com (Andrew Janke) Date: Thu, 11 May 2006 00:00:07 -0400 Subject: [MINC-users] Same Problem - Error compiling conglomerate 1.6 In-Reply-To: <9c80f12d0605091444p6563a62ycbe71901e6b84c43@mail.gmail.com> References: <9c80f12d0605091444p6563a62ycbe71901e6b84c43@mail.gmail.com> Message-ID: Hi Gil, This seems rather strange given that you have described an almost identical system to what we use here. Do you happen to have the approriate bit from the config.log? It may yield some answers. a On 5/9/06, Gil Citro wrote: > I posted a very similar problem to this before, but unfortunately the > previous solution doesn't seem to be working. The problem is that > when I try to compile conglomerate 1.6, I get an error. The following > are the last several lines of compiler output, and I've attached the > while thing as compile-error. > > /usr/local/lib/libbicpl.a(rgb_io.o)(.text+0x226):/usr/local/mni/bicpl-1.4.2/Images/rgb_io_ppm.c:100: > undefined reference to `pm_allocrow' > /usr/local/lib/libbicpl.a(rgb_io.o)(.text+0x2d8):/usr/local/mni/bicpl-1.4.2/Images/rgb_io_ppm.c:118: > undefined reference to `ppm_writeppmrow' > /usr/local/lib/libbicpl.a(rgb_io.o)(.text+0x2e9):/usr/local/mni/bicpl-1.4.2/Images/rgb_io_ppm.c:121: > undefined reference to `pm_freerow' > collect2: ld returned 1 exit status > make[1]: *** [composite_images] Error 1 > make[1]: Leaving directory `/usr/local/mni/conglomerate-1.6' > make: *** [all] Error 2 > enter the text in that file's own buffer. > > The previous solution was to use the --with-image-ppm option with > ./configure for bicpl 1.4.2 and then to manually make the following > change in the Makefile > > Old: LIBS = -lppm -lvolume_io -lminc -lnetcdf -lm > New: LIBS = -lppm -lvolume_io -lminc -lnetcdf -lm -lppm > > and to make sure libnetpbm9-dev was installed, which it is. > > If I configure bicpl with --with-image-ppm but don't change the > Makefile, bicpl won't compile. If I configure bicpl without > --with-image-ppm , bicpl will compile, as will congolmerate, but then > ray_trace will not. > > Between this time and last time, I have reinstalled Debian Sarge from > scratch. It's an Athlon64 3000+ box, and last time I was using a 64 > bit kernel, but this time I'm using the regular Debian 32 bit kernel > to improve compatibility with some software. > Other than that, I think everything is the same. I think I > reinstalled all the packages that were previously installed when it > worked. > > If anyone has any suggestions for what to try next, I'd appreciate it. Thanks! > > Gil > > > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > > > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From a.janke at gmail.com Thu May 11 00:03:11 2006 From: a.janke at gmail.com (Andrew Janke) Date: Thu, 11 May 2006 00:03:11 -0400 Subject: [MINC-users] keep_real_range In-Reply-To: <4460CAFB.80108@ninds.nih.gov> References: <4460CAFB.80108@ninds.nih.gov> Message-ID: Hi Stefano, If I understand correctly, I think what you need to do is just drop the --keep_real_range argument altogether. Mind you if you are transferring in and out of afni/nifti/spm all the time you would probably be best served to use -float or -double for all your files. This will definitely reduce the number of conversion problems you will have with ranges. a On 5/9/06, Stefano Pellegrini wrote: > Hi all, > > I am tring to resample a b1 map (b1_map.mnc) to another image > (emt_off.mnc) but I have a problem with the range of the output image. > the input image, has this range > > file: b1_map.mnc > image: signed__ double -100 to 50 > > while the output iamge has > > file: reg_rig_off_b1_map.mnc > image: signed__ double 0 to 1 > > Instead I don't want the output range to change. the command that I have > used is > > mincresample -clobber -transformation transformation.xfm -like > emt_off.mnc -keep_real_range b1_map.mnc res_b1_map.mnc > > When I try to open it with AFNI, it recognize the original range. I > think there is something that is not clear to me. > > Can anybody help me? > thanks, > > -stefano > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From o.winz at fz-juelich.de Thu May 11 09:19:19 2006 From: o.winz at fz-juelich.de (Oliver Winz) Date: Thu, 11 May 2006 15:19:19 +0200 Subject: [MINC-users] time and time-with values with volume_io library Message-ID: <446339D7.5090107@fz-juelich.de> Hi everyone, i have a question about the volume_io library, is there a possibility to determine the time and time-width values from a dynamic image file with this library. i want to write a little program that subtracts the contribution of vascular radioactivity from dynamic PET images (interpolating blood values with frame times....). i have a few problems doing it with the minc2 library, therefore i use the volume_io lib. thank' a lot, oliver -- Oliver Winz, M.Sc. Molecular Neuroimaging Group Institute of Medicine Research Centre Juelich D-52425 Juelich Germany Phone: +49 2461 61-6493 Web: http://www.fz-juelich.de/ime/mni Location: Building 15.9, Room 3010 From citrog at gmail.com Thu May 11 11:56:21 2006 From: citrog at gmail.com (Gil Citro) Date: Thu, 11 May 2006 11:56:21 -0400 Subject: [MINC-users] Same Problem - Error compiling conglomerate 1.6 In-Reply-To: References: <9c80f12d0605091444p6563a62ycbe71901e6b84c43@mail.gmail.com> Message-ID: <9c80f12d0605110856u44d11f47h5ace27b1c78a41a0@mail.gmail.com> Meanwhile, I found a solution. 1. Downloaded bicpl-1.4.3, which, which configured with --with-image-ppm does not require manual editing of the Makefile. 2. Tried conglomerate 1.6.2, which still will not compile without manual tweaking, but making the following change in the Makefile after configuration works. The same change also works in the ray_trace 1.0 Makefile. < LIBS = -lbicpl -lvolume_io -lminc -lnetcdf -lm > LIBS = -lbicpl -lvolume_io -lminc -lnetcdf -lm -lppm As for why the conglomerate and ray_trace Makefiles require manual editing, I'm not sure but have attached the config.log files for both. The same problem seems to have existed for bicpl 1.4.2, but has been fixed in 1.4.3. Thanks. Gil On 5/11/06, Andrew Janke wrote: > > This seems rather strange given that you have described an almost > identical system to what we use here. Do you happen to have the > approriate bit from the config.log? It may yield some answers. > From o.winz at fz-juelich.de Thu May 11 12:54:50 2006 From: o.winz at fz-juelich.de (o.winz@fz-juelich.de) Date: Thu, 11 May 2006 18:54:50 +0200 Subject: [MINC-users] time and time-with values with volume_io library Message-ID: <525a80524eb7.524eb7525a80@fz-juelich.de> hi everyone, i think i've solve the problem. first i check the input volume for irregular volume dimension 'is_volume_dimension_irregular(input,0)', when true, i use 'get_volume_irregular_widths(input,0,(number_of_frames -1),widths)' and 'get_volume_irregular_starts(input,0,(number_of_frames-1),starts)', where 'widths' and 'starts' arrays of type Real (double). greetings, oliver > Message: 3 > Date: Thu, 11 May 2006 15:19:19 +0200 > From: Oliver Winz > Subject: [MINC-users] time and time-with values with volume_io library > To: minc-users at bic.mni.mcgill.ca > Message-ID: <446339D7.5090107 at fz-juelich.de> > Content-Type: text/plain; charset=ISO-8859-15; format=flowed > > Hi everyone, > > i have a question about the volume_io library, is there a > possibility to > determine the time and time-width values from a dynamic image file > with > this library. > i want to write a little program that subtracts the contribution of > vascular radioactivity from dynamic PET images (interpolating blood > values with frame times....). > i have a few problems doing it with the minc2 library, therefore i > use > the volume_io lib. > > thank' a lot, > > oliver From citrog at gmail.com Thu May 11 13:36:07 2006 From: citrog at gmail.com (Gil Citro) Date: Thu, 11 May 2006 13:36:07 -0400 Subject: [MINC-users] Problem Buliding MNI Models Message-ID: <9c80f12d0605111036g56b5558i6b3b239bade6a326@mail.gmail.com> I'm having trouble building and installing mni-models_average305-lin-1.0 mni-models_colin27-lin-1.0 mni-models_icbm152-lin-1.0 Below is an example of the errors I'm getting. They're all similar errors, so I suspect there's a common cause. I'm running Debian Sarge on an Athlon64 but running the 32 bit i386 kernel rather than the x86_64 kernel. NetCDF 3.5.0 and all related Debian packages I could find are installed. I thought it might be a problem with the version of NetCDF, so I installed 3.6.1 from source under /usr/local but that had no effect. If anyone has a suggestion, I'd appreciate it. Thanks. Gil ** Making icbm_avg_152_t1_tal_lin_pad.mnc from icbm_avg_152_t1_tal_lin.mnc by resampling on 2mm grid and padding by 16mm [autocrop] [root at bayes:/usr/local/mni/mni-models_icbm152-lin-1.0] [2006-05-11 13:31:55] /usr/local/bin/mincresample -clobber icbm_avg_152_t1_tal_lin.mnc icbm_avg_152_t1_tal_lin_pad.mnc -start -106 -142 -88 -step 2 2 2 -nelements 107 125 107Transforming slices:...........................................................................................................Done (from miopen): Can't write compressed file ncredef: ncid -1: Not a netCDF id ncattput: ncid -1: Not a netCDF id ncclose: ncid -1: Not a netCDF id autocrop: crashed while running minc_modify_header (termination status=768) ** Building 16mm data for icbm_avg_152_t1_tal_lin From rtoro at snv.jussieu.fr Fri May 12 12:18:26 2006 From: rtoro at snv.jussieu.fr (Roberto Toro) Date: Fri, 12 May 2006 17:18:26 +0100 Subject: [MINC-users] nii2mnc In-Reply-To: References: Message-ID: <67EFDB45-45EC-477C-8E99-3A242B09996D@snv.jussieu.fr> Hi! is there a way of specifying a rotation matrix when converting a nii volume to mnc? (I'm trying to convert a normalized nii volume and it appears upside down) thanks in advance! roberto This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation. From a.janke at gmail.com Fri May 12 16:31:07 2006 From: a.janke at gmail.com (Andrew Janke) Date: Fri, 12 May 2006 16:31:07 -0400 Subject: [MINC-users] nii2mnc In-Reply-To: <67EFDB45-45EC-477C-8E99-3A242B09996D@snv.jussieu.fr> References: <67EFDB45-45EC-477C-8E99-3A242B09996D@snv.jussieu.fr> Message-ID: Not specifically, you can flip and change the direction order though. -flipx -yzx for example. a On 5/12/06, Roberto Toro wrote: > Hi! > is there a way of specifying a rotation matrix when converting a nii > volume to mnc? > (I'm trying to convert a normalized nii volume and it appears upside > down) > > thanks in advance! > > roberto > > This message has been checked for viruses but the contents of an attachment > may still contain software viruses, which could damage your computer system: > you are advised to perform your own checks. Email communications with the > University of Nottingham may be monitored as permitted by UK legislation. > > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From a.janke at gmail.com Wed May 17 02:37:16 2006 From: a.janke at gmail.com (Andrew Janke) Date: Wed, 17 May 2006 02:37:16 -0400 Subject: [MINC-users] Problem Buliding MNI Models In-Reply-To: <9c80f12d0605111036g56b5558i6b3b239bade6a326@mail.gmail.com> References: <9c80f12d0605111036g56b5558i6b3b239bade6a326@mail.gmail.com> Message-ID: Hi Gil, Looks like netcdf cannot find the gzip runtime libraries (to uncompress the input file). If this minc install came from a binary package this would be suprising as they are supposed to be compiled static. a On 5/11/06, Gil Citro wrote: > I'm having trouble building and installing > > mni-models_average305-lin-1.0 > mni-models_colin27-lin-1.0 > mni-models_icbm152-lin-1.0 > > Below is an example of the errors I'm getting. They're all similar > errors, so I suspect there's a common cause. I'm running Debian Sarge > on an Athlon64 but running the 32 bit i386 kernel rather than the > x86_64 kernel. > > NetCDF 3.5.0 and all related Debian packages I could find are > installed. I thought it might be a problem with the version of > NetCDF, so I installed 3.6.1 from source under /usr/local but that had > no effect. > > If anyone has a suggestion, I'd appreciate it. Thanks. > > Gil > > ** Making icbm_avg_152_t1_tal_lin_pad.mnc from > icbm_avg_152_t1_tal_lin.mnc by resampling on 2mm grid and padding by > 16mm > [autocrop] [root at bayes:/usr/local/mni/mni-models_icbm152-lin-1.0] > [2006-05-11 13:31:55] /usr/local/bin/mincresample -clobber > icbm_avg_152_t1_tal_lin.mnc icbm_avg_152_t1_tal_lin_pad.mnc -start > -106 -142 -88 -step 2 2 2 -nelements 107 125 107Transforming > slices:...........................................................................................................Done > (from miopen): Can't write compressed file > ncredef: ncid -1: Not a netCDF id > ncattput: ncid -1: Not a netCDF id > ncclose: ncid -1: Not a netCDF id > autocrop: crashed while running minc_modify_header (termination status=768) > ** Building 16mm data for icbm_avg_152_t1_tal_lin > > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From citrog at gmail.com Thu May 18 11:39:55 2006 From: citrog at gmail.com (Gil Citro) Date: Thu, 18 May 2006 11:39:55 -0400 Subject: [MINC-users] Problem Buliding MNI Models In-Reply-To: References: <9c80f12d0605111036g56b5558i6b3b239bade6a326@mail.gmail.com> Message-ID: <9c80f12d0605180839q20dd972fve6c16461cb392e96@mail.gmail.com> Actually no, I compiled MINC 2.0.10 myself from the source. I also compiled NetCDF 3.6.1 from source, because NetCDF 3.5.0 would not compile on my system. Any suggestions for how to make NetCDF find gzip? Thanks again. Gil On 5/17/06, Andrew Janke wrote: > > Looks like netcdf cannot find the gzip runtime libraries (to > uncompress the input file). If this minc install came from a binary > package this would be suprising as they are supposed to be compiled > static. > > On 5/11/06, Gil Citro wrote: > > I'm having trouble building and installing > > > > mni-models_average305-lin-1.0 > > mni-models_colin27-lin-1.0 > > mni-models_icbm152-lin-1.0 > > From nikelski at bic.mni.mcgill.ca Thu May 18 13:12:39 2006 From: nikelski at bic.mni.mcgill.ca (EJ Nikelski) Date: Thu, 18 May 2006 13:12:39 -0400 Subject: [MINC-users] xfm2param problem Message-ID: <446CAB07.5080202@bic.mni.mcgill.ca> Hi List, I have a need to modify xfm files by extracting or retaining various transform types (transform, rescale, etc). For example, taking an xfm that transforms to Talairach, but keeping the rescale set to 1.0, providing me with an xfm that puts my volume into Talaiarch-ish orientation, while keeping the native brain dimensions (this can be useful for brain painting). Up until now I've used xfm2param/param2xfm to modify the xfms as needed -- without a problem. However, while working on a new project, I noticed this rather strange phenomenon.If you create an xfm as follows ... param2xfm -transform 10 10 10 -rotation 20 20 20 -scale 30 30 30 x0.xfm ... you get as expected. Next, use the individual transformation type info, and then combine with xfmtool, i.e., param2xfm -transform 10 10 10 xT.xfm param2xfm -rotation 20 20 20 xR.xfm param2xfm -scale 30 30 30 xS.xfm xfmtool xT.xfm xR.xfm xS.xfm xTRS.xfm ... "cat xTRS.xfm" shows the 3 appended xfms - OK. But an xfm2param of the file shows strange translation values ... jnikelski@[~/xxx]: xfm2param xTRS.xfm after parameter extraction -center 0.00000 0.00000 0.00000 -translation 327.16223 309.88623 258.71875 -rotation 20.00000 20.00000 20.00000 -scale 30.00000 30.00000 30.00000 -shear -0.00000 -0.00000 0.00000 Should this not work? Or am I missing something fundamental here? .. which is, of course, all too likely :( Thanks, -Jim -- E. James Nikelski, Ph.D. Postdoctoral Research Fellow Bloomfield Centre for Research in Aging Lady Davis Institute for Medical Research Sir Mortimer B. Davis - Jewish General Hospital McGill University Tel: (514) 340-8222 x 2298 Fax: (514) 340-8295 From a.janke at gmail.com Thu May 18 14:33:19 2006 From: a.janke at gmail.com (Andrew Janke) Date: Thu, 18 May 2006 14:33:19 -0400 Subject: [MINC-users] xfm2param problem In-Reply-To: <446CAB07.5080202@bic.mni.mcgill.ca> References: <446CAB07.5080202@bic.mni.mcgill.ca> Message-ID: Hi Ed, First I presume you mean: param2xfm -translation (not -transform) What you have to remember here is that xfm2param and param2xfm assume a certain order of operations. An affine transformation doesn't care (or know) about translation, rotation and scale, it is just a matrix operation. So this means that if you wish to do the below, you just need to make sure that you apply the operations in the same order as param2xfm does. ie: Instead of: xfmtool xT.xfm xR.xfm xS.xfm xTRS.xfm use: xfmtool xS.xfm xR.xfm xT.xfm out.xfm Your translation of ~300 came from a translation of 10 and then a rotation (moves the initial translation out of plane) and then a scale of 30 (effectively multiplies the translation by 30). Andrew On 5/18/06, EJ Nikelski wrote: > Hi List, > > I have a need to modify xfm files by extracting or retaining various > transform types (transform, rescale, etc). For example, taking an xfm > that transforms to Talairach, but keeping the rescale set to 1.0, > providing me with an xfm that puts my volume into Talaiarch-ish > orientation, while keeping the native brain dimensions (this can be > useful for brain painting). Up until now I've used xfm2param/param2xfm > to modify the xfms as needed -- without a problem. > > However, while working on a new project, I noticed this rather > strange phenomenon.If you create an xfm as follows ... > > param2xfm -transform 10 10 10 -rotation 20 20 20 -scale 30 30 30 x0.xfm > > ... you get as expected. Next, use the individual transformation type > info, and then combine with xfmtool, i.e., > > param2xfm -transform 10 10 10 xT.xfm > param2xfm -rotation 20 20 20 xR.xfm > param2xfm -scale 30 30 30 xS.xfm > xfmtool xT.xfm xR.xfm xS.xfm xTRS.xfm > > ... "cat xTRS.xfm" shows the 3 appended xfms - OK. But an xfm2param of > the file shows strange translation values ... > > jnikelski@[~/xxx]: xfm2param xTRS.xfm > after parameter extraction > -center 0.00000 0.00000 0.00000 > -translation 327.16223 309.88623 258.71875 > -rotation 20.00000 20.00000 20.00000 > -scale 30.00000 30.00000 30.00000 > -shear -0.00000 -0.00000 0.00000 > > > Should this not work? Or am I missing something fundamental here? .. > which is, of course, all too likely :( > > > Thanks, > > -Jim > > > > -- > E. James Nikelski, Ph.D. > Postdoctoral Research Fellow > Bloomfield Centre for Research in Aging > Lady Davis Institute for Medical Research > Sir Mortimer B. Davis - Jewish General Hospital > McGill University > Tel: (514) 340-8222 x 2298 > Fax: (514) 340-8295 > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From nikelski at bic.mni.mcgill.ca Thu May 25 15:34:29 2006 From: nikelski at bic.mni.mcgill.ca (EJ Nikelski) Date: Thu, 25 May 2006 15:34:29 -0400 Subject: [MINC-users] Mac OS X (Tiger) and register Message-ID: <447606C5.3040400@bic.mni.mcgill.ca> Hi List, I just moved from Panther to Tiger. I installed the netcdf/minc/register/Display packages available at http://packages.bic.mni.mcgill.ca/osx-10.4/ ... but now find a curious behaviour. When starting register, I get some GLUT error messages ... >>2006-05-25 15:13:07.488 register[538] *** _NSAutoreleaseNoPool(): >>Object 0x328ff0 of class NSCFString autoreleased with no pool in place >>- just leaking >>2006-05-25 15:13:07.488 register[538] GLUT Warning: glutUseLayer: >>unknown layer, 0. ... and while in register, the ctrl and alt keys only allow translating the image, I can no longer zoom. Now, while I'm certain that this problem is somehow related to my setup, I'm not sure what it might be. Is it GLUT related? All worked fine under Panther. Any ideas? Thanks, -Jim -- E. James Nikelski, Ph.D. Postdoctoral Research Fellow Bloomfield Centre for Research in Aging Lady Davis Institute for Medical Research Sir Mortimer B. Davis - Jewish General Hospital McGill University Tel: (514) 340-8222 x 2298 Fax: (514) 340-8295 From a.janke at gmail.com Wed May 31 02:49:24 2006 From: a.janke at gmail.com (Andrew Janke) Date: Wed, 31 May 2006 02:49:24 -0400 Subject: [MINC-users] Mac OS X (Tiger) and register In-Reply-To: <447606C5.3040400@bic.mni.mcgill.ca> References: <447606C5.3040400@bic.mni.mcgill.ca> Message-ID: Hi Jim, This is a tad wierd but I suspect this is a GLUT issue. The warnings you have posted below are "normal'. I have no clue where they come from (and no impetus to hunt them down as they appear to be harmless). If you use a "real" 3 button mouse does the problem persist? I am unsure if you are trying to mix the apple X11 hacks for 2nd and 3rd mouse button with the register control key stuff. Andrew On 5/25/06, EJ Nikelski wrote: > Hi List, > > I just moved from Panther to Tiger. I installed the > netcdf/minc/register/Display packages available at > > http://packages.bic.mni.mcgill.ca/osx-10.4/ > > ... but now find a curious behaviour. When starting register, I get some > GLUT error messages ... > > >>2006-05-25 15:13:07.488 register[538] *** _NSAutoreleaseNoPool(): > >>Object 0x328ff0 of class NSCFString autoreleased with no pool in > place >>- just leaking > >>2006-05-25 15:13:07.488 register[538] GLUT Warning: glutUseLayer: > >>unknown layer, 0. > > ... and while in register, the ctrl and alt keys only allow > translating the image, I can no longer zoom. Now, while I'm certain that > this problem is somehow related to my setup, I'm not sure what it might > be. Is it GLUT related? All worked fine under Panther. Any ideas? > > Thanks, > > > -Jim > > > > > -- > E. James Nikelski, Ph.D. > Postdoctoral Research Fellow > Bloomfield Centre for Research in Aging > Lady Davis Institute for Medical Research > Sir Mortimer B. Davis - Jewish General Hospital > McGill University > Tel: (514) 340-8222 x 2298 > Fax: (514) 340-8295 > _______________________________________________ > MINC-users at bic.mni.mcgill.ca > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users > -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From a.janke at gmail.com Wed May 31 02:52:48 2006 From: a.janke at gmail.com (Andrew Janke) Date: Wed, 31 May 2006 02:52:48 -0400 Subject: [MINC-users] New versions of a number of packages Message-ID: Hi all, A number of new releases of MINC tools have happened over the past few weeks on: http://packages.bic.mni.mcgill.ca/tgz/ Most of these now have debian packages but not all of them. These updates are primarily small changes needed to ensure that all of these packages compile cleanly with minc2. A big thanks to Claude (LEPAGE) of the BIC for putting a lot effort into this in the last weeks. -- Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/) Canada->Montreal Cell: +1 (514) 924 2012 From jp.mailhot at UMontreal.CA Tue May 30 14:50:29 2006 From: jp.mailhot at UMontreal.CA (Jean-Philippe Mailhot) Date: Tue, 30 May 2006 14:50:29 -0400 Subject: [MINC-users] Display problems using Display Message-ID: <1149015029.447c93f52ead3@www.courrier.umontreal.ca> Hi all, I installed Display on Linux Fedora Core 5 but I have some difficulties when I try to increase the size of the slice window where there are the 3 orthogonal views of slices. In fact, the size of the window increase but not the 3 slices in that window ! Can anybody help me? thanks, Jean-Philippe Mailhot BRAMS (Brain,Music and Sound) LNMCA,Universit? de Montr?al