From gdevenyi at gmail.com  Fri Sep 11 16:11:53 2015
From: gdevenyi at gmail.com (Gabriel A. Devenyi)
Date: Fri, 11 Sep 2015 16:11:53 -0400
Subject: [MINC-users] Problems with ray_trace and object files
Message-ID: <CAFQOoOoA_RLJOwmqJpKkviBmMqZ6wut__VWRWwdUiP1zhON_OQ@mail.gmail.com>

Hi all,

I?m trying to generate some images of CIVET-generated left and right
hemispheres using ray_trace
like:

> ray_trace -output left_lateral.rgb left_average.obj -size 1000 1000 -crop -sup 3 -bg black -left

As I understand, I would expect this to generate a view of the left
hemisphere, from the left side, so it should be a lateral view.

For some reason, this seems to generate an image that is ?upside down?

If I run:

> convert left_lateral.rgb -flip left_lateral.png

It?s fixed and looks correct.

Am I missing something about ray_trace? Or is this a bug?
?
-- 
Gabriel A. Devenyi B.Eng. Ph.D.
Research Computing Associate
Computational Brain Anatomy Laboratory
Cerebral Imaging Center
Douglas Mental Health University Institute
McGill University
t: 514.761.6131x4781
e: gdevenyi at gmail.com

From sulantha.s at gmail.com  Fri Sep 11 19:00:23 2015
From: sulantha.s at gmail.com (Sulantha Sanjeewa)
Date: Fri, 11 Sep 2015 19:00:23 -0400
Subject: [MINC-users] Problems with ray_trace and object files
In-Reply-To: <CAFQOoOoA_RLJOwmqJpKkviBmMqZ6wut__VWRWwdUiP1zhON_OQ@mail.gmail.com>
References: <CAFQOoOoA_RLJOwmqJpKkviBmMqZ6wut__VWRWwdUiP1zhON_OQ@mail.gmail.com>
Message-ID: <CAPM2tnguD5CXq+pckzePR-O-1DLMHhnMiqPHysV7R8qe1_b9_w@mail.gmail.com>

Hi Gabriel,
I encountered the same problem. This is because of a version problem of ray
trace.
The newer version however has the problem. My solution was to use a older
version of ray trace. Unfortunately I do not have the version umber, but
try with a old version of minc toolkit.
Best regards,
Sulantha.

On Fri, Sep 11, 2015 at 4:11 PM, Gabriel A. Devenyi <gdevenyi at gmail.com>
wrote:

> Hi all,
>
> I?m trying to generate some images of CIVET-generated left and right
> hemispheres using ray_trace
> like:
>
> > ray_trace -output left_lateral.rgb left_average.obj -size 1000 1000
> -crop -sup 3 -bg black -left
>
> As I understand, I would expect this to generate a view of the left
> hemisphere, from the left side, so it should be a lateral view.
>
> For some reason, this seems to generate an image that is ?upside down?
>
> If I run:
>
> > convert left_lateral.rgb -flip left_lateral.png
>
> It?s fixed and looks correct.
>
> Am I missing something about ray_trace? Or is this a bug?
> ?
> --
> Gabriel A. Devenyi B.Eng. Ph.D.
> Research Computing Associate
> Computational Brain Anatomy Laboratory
> Cerebral Imaging Center
> Douglas Mental Health University Institute
> McGill University
> t: 514.761.6131x4781
> e: gdevenyi at gmail.com
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From robert.d.vincent at mcgill.ca  Sat Sep 12 16:01:57 2015
From: robert.d.vincent at mcgill.ca (Robert D. Vincent)
Date: Sat, 12 Sep 2015 16:01:57 -0400
Subject: [MINC-users] Problems with ray_trace and object files
In-Reply-To: <CAFQOoOoA_RLJOwmqJpKkviBmMqZ6wut__VWRWwdUiP1zhON_OQ@mail.gmail.com>
References: <CAFQOoOoA_RLJOwmqJpKkviBmMqZ6wut__VWRWwdUiP1zhON_OQ@mail.gmail.com>
Message-ID: <CAAd5DeQy2y42Z8DfJSx2q6_-b3+gjPj+aUN+RKi-CYAhwp9saA@mail.gmail.com>

Hi Gabriel,

It's a bug. It seems that our bitmap I/O functions changed in the last few
years such that they now flip the Y axis. This affects Display as well as
ray_trace. Thanks for pointing it out, I'll figure out what happened and
fix it.

    -bert


On Fri, Sep 11, 2015 at 4:11 PM, Gabriel A. Devenyi <gdevenyi at gmail.com>
wrote:

> Hi all,
>
> I?m trying to generate some images of CIVET-generated left and right
> hemispheres using ray_trace
> like:
>
> > ray_trace -output left_lateral.rgb left_average.obj -size 1000 1000
> -crop -sup 3 -bg black -left
>
> As I understand, I would expect this to generate a view of the left
> hemisphere, from the left side, so it should be a lateral view.
>
> For some reason, this seems to generate an image that is ?upside down?
>
> If I run:
>
> > convert left_lateral.rgb -flip left_lateral.png
>
> It?s fixed and looks correct.
>
> Am I missing something about ray_trace? Or is this a bug?
> ?
> --
> Gabriel A. Devenyi B.Eng. Ph.D.
> Research Computing Associate
> Computational Brain Anatomy Laboratory
> Cerebral Imaging Center
> Douglas Mental Health University Institute
> McGill University
> t: 514.761.6131x4781
> e: gdevenyi at gmail.com
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From gdevenyi at gmail.com  Tue Sep 15 12:01:42 2015
From: gdevenyi at gmail.com (Gabriel A. Devenyi)
Date: Tue, 15 Sep 2015 12:01:42 -0400
Subject: [MINC-users] Total Intercranial Volume (TIV) in the minc toolchain?
Message-ID: <CAFQOoOpEz0D-e4O_NQLjkh-pvHJehPVGZsxjN-Hcw43wieQ=SA@mail.gmail.com>

Hi all,

I know I can get total brain volume (TBV) with BeAST with excellent
results, but for some groups I think it would be more appropriate to use
TIV to describe the subject.

Does anyone know of a tool/set of commands in minc to achieve this? My
first thought was to use the inverse BeAST mask to remove the brain and
then use some kind of fill-in mask based on the intensity in the empty
space.

-- 
Gabriel A. Devenyi B.Eng. Ph.D.
Research Computing Associate
Computational Brain Anatomy Laboratory
Cerebral Imaging Center
Douglas Mental Health University Institute
McGill University
t: 514.761.6131x4781
e: gdevenyi at gmail.com

From sorench at gmail.com  Sat Sep 19 23:48:39 2015
From: sorench at gmail.com (Soren Christensen)
Date: Sat, 19 Sep 2015 20:48:39 -0700
Subject: [MINC-users] mincreshape -dimsize
Message-ID: <CAHfsrRY2R08Gg94_JrK5tpfrKDtXMEw5ZekqJKW-fHx0qDbEug@mail.gmail.com>

Hi,
 I have an issue with mincreshape I don't understand. I guess it has to do
with trying to interpolate the slowest varying dimension:

file: pwi_bl.mnc
image: signed__ float 0 to 1381.4969482421875
image dimensions: zspace yspace xspace
    dimension name         length         step        start
    --------------         ------         ----        -----
    zspace                     17            5      -38.912
    yspace                    128        -1.75      95.0638
    xspace                    128        -1.75      110.338

u at K450$ mincreshape -dimsize zspace=49 -float -clobber  pwi_bl.mnc
pwi_blIP.mnc
Copying chunks:HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread 0:
  #000: H5Dio.c line 161 in H5Dread(): selection+offset not within extent
    major: Dataspace
    minor: Out of range
(from mivarget): Can't read dataset /minc-2.0/image/0/image-min
(from mivarget): Can't read variable ID# 1
HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread 0:
  #000: H5Dio.c line 161 in H5Dread(): selection+offset not within extent
    major: Dataspace
    minor: Out of range
(from mivarget): Can't read dataset /minc-2.0/image/0/image-max
(from mivarget): Can't read variable ID# 2
.................................................Done.

The file is written but it appears incorrect with a larger Z FOV for
starters but it also look as if the extreme slices are simply copied
outwards at each end (sorry cannot post images here). Finally the image
appears corrupted in register and freezes up Display and the OS!

file: pwi_blIP.mnc
image: signed__ float 0 to 1381.4969482421875
image dimensions: zspace yspace xspace
    dimension name         length         step        start
    --------------         ------         ----        -----
    zspace                     49          2.5      -57.662
    yspace                    128        -1.75      95.0638
    xspace                    128        -1.75      110.338

-dimsize xspace=64 works without errors.

I have a vague recollection Andrew Janke once explained how to deal with
this but I am unable to locate that post (if it ever existed).

Thanks!
Soren

From a.janke at gmail.com  Sun Sep 20 19:41:37 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Mon, 21 Sep 2015 09:41:37 +1000
Subject: [MINC-users] mincreshape -dimsize
In-Reply-To: <CAHfsrRY2R08Gg94_JrK5tpfrKDtXMEw5ZekqJKW-fHx0qDbEug@mail.gmail.com>
References: <CAHfsrRY2R08Gg94_JrK5tpfrKDtXMEw5ZekqJKW-fHx0qDbEug@mail.gmail.com>
Message-ID: <CAG0CB22HS7GwqSExuG5SYkxZad1tHKfR+qtOvHmg+-rf8X9FLw@mail.gmail.com>

Hi Soren,

This could be one of two things.

1. Is this a MINC1 or MINC2 file? If the former, convert to MINC2
first using mincconvert

2. I think what you are alluding to is "-dimrange zspace=-1" option
that will tell mincreshape to treat a slice dimension as an image
dimension.

There has been a feature request kicking about for a long time to
remove this behaviour as it doesn't make as much sense as it used to
when Peter first implemented slice scaling in MINC1 but I don't think
it's made it into the latest release as of yet.


a



On 20 September 2015 at 13:48, Soren Christensen <sorench at gmail.com> wrote:
> Hi,
>  I have an issue with mincreshape I don't understand. I guess it has to do
> with trying to interpolate the slowest varying dimension:
>
> file: pwi_bl.mnc
> image: signed__ float 0 to 1381.4969482421875
> image dimensions: zspace yspace xspace
>     dimension name         length         step        start
>     --------------         ------         ----        -----
>     zspace                     17            5      -38.912
>     yspace                    128        -1.75      95.0638
>     xspace                    128        -1.75      110.338
>
> u at K450$ mincreshape -dimsize zspace=49 -float -clobber  pwi_bl.mnc
> pwi_blIP.mnc
> Copying chunks:HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread 0:
>   #000: H5Dio.c line 161 in H5Dread(): selection+offset not within extent
>     major: Dataspace
>     minor: Out of range
> (from mivarget): Can't read dataset /minc-2.0/image/0/image-min
> (from mivarget): Can't read variable ID# 1
> HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread 0:
>   #000: H5Dio.c line 161 in H5Dread(): selection+offset not within extent
>     major: Dataspace
>     minor: Out of range
> (from mivarget): Can't read dataset /minc-2.0/image/0/image-max
> (from mivarget): Can't read variable ID# 2
> .................................................Done.
>
> The file is written but it appears incorrect with a larger Z FOV for
> starters but it also look as if the extreme slices are simply copied
> outwards at each end (sorry cannot post images here). Finally the image
> appears corrupted in register and freezes up Display and the OS!
>
> file: pwi_blIP.mnc
> image: signed__ float 0 to 1381.4969482421875
> image dimensions: zspace yspace xspace
>     dimension name         length         step        start
>     --------------         ------         ----        -----
>     zspace                     49          2.5      -57.662
>     yspace                    128        -1.75      95.0638
>     xspace                    128        -1.75      110.338
>
> -dimsize xspace=64 works without errors.
>
> I have a vague recollection Andrew Janke once explained how to deal with
> this but I am unable to locate that post (if it ever existed).
>
> Thanks!
> Soren
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users

From sorench at gmail.com  Sun Sep 20 23:45:40 2015
From: sorench at gmail.com (Soren Christensen)
Date: Sun, 20 Sep 2015 20:45:40 -0700
Subject: [MINC-users] mincreshape -dimsize
In-Reply-To: <CAG0CB22HS7GwqSExuG5SYkxZad1tHKfR+qtOvHmg+-rf8X9FLw@mail.gmail.com>
References: <CAHfsrRY2R08Gg94_JrK5tpfrKDtXMEw5ZekqJKW-fHx0qDbEug@mail.gmail.com>
	<CAG0CB22HS7GwqSExuG5SYkxZad1tHKfR+qtOvHmg+-rf8X9FLw@mail.gmail.com>
Message-ID: <CAHfsrRZg7by5S4SPVv6-Q0x5MU6t2Kg=XEp=BHFf_uV2P4VOUw@mail.gmail.com>

Thanks Andrew!
 Yes file is MINC2 and it was exactly your 2nd point I failed toremember in
detail.
It actually does not seem to do the trick - the image comes out without a
zspace dimension if I add this option.
I used mincresample instead, and using the desired -zsize and -zstep with
no other options achieved the re-slicing I was looking for.
Let me know if you want me to do something to document the mincreshape issue

Thanks!
Soren

On Sun, Sep 20, 2015 at 4:41 PM, Andrew Janke <a.janke at gmail.com> wrote:

> Hi Soren,
>
> This could be one of two things.
>
> 1. Is this a MINC1 or MINC2 file? If the former, convert to MINC2
> first using mincconvert
>
> 2. I think what you are alluding to is "-dimrange zspace=-1" option
> that will tell mincreshape to treat a slice dimension as an image
> dimension.
>
> There has been a feature request kicking about for a long time to
> remove this behaviour as it doesn't make as much sense as it used to
> when Peter first implemented slice scaling in MINC1 but I don't think
> it's made it into the latest release as of yet.
>
>
> a
>
>
>
> On 20 September 2015 at 13:48, Soren Christensen <sorench at gmail.com>
> wrote:
> > Hi,
> >  I have an issue with mincreshape I don't understand. I guess it has to
> do
> > with trying to interpolate the slowest varying dimension:
> >
> > file: pwi_bl.mnc
> > image: signed__ float 0 to 1381.4969482421875
> > image dimensions: zspace yspace xspace
> >     dimension name         length         step        start
> >     --------------         ------         ----        -----
> >     zspace                     17            5      -38.912
> >     yspace                    128        -1.75      95.0638
> >     xspace                    128        -1.75      110.338
> >
> > u at K450$ mincreshape -dimsize zspace=49 -float -clobber  pwi_bl.mnc
> > pwi_blIP.mnc
> > Copying chunks:HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread
> 0:
> >   #000: H5Dio.c line 161 in H5Dread(): selection+offset not within extent
> >     major: Dataspace
> >     minor: Out of range
> > (from mivarget): Can't read dataset /minc-2.0/image/0/image-min
> > (from mivarget): Can't read variable ID# 1
> > HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread 0:
> >   #000: H5Dio.c line 161 in H5Dread(): selection+offset not within extent
> >     major: Dataspace
> >     minor: Out of range
> > (from mivarget): Can't read dataset /minc-2.0/image/0/image-max
> > (from mivarget): Can't read variable ID# 2
> > .................................................Done.
> >
> > The file is written but it appears incorrect with a larger Z FOV for
> > starters but it also look as if the extreme slices are simply copied
> > outwards at each end (sorry cannot post images here). Finally the image
> > appears corrupted in register and freezes up Display and the OS!
> >
> > file: pwi_blIP.mnc
> > image: signed__ float 0 to 1381.4969482421875
> > image dimensions: zspace yspace xspace
> >     dimension name         length         step        start
> >     --------------         ------         ----        -----
> >     zspace                     49          2.5      -57.662
> >     yspace                    128        -1.75      95.0638
> >     xspace                    128        -1.75      110.338
> >
> > -dimsize xspace=64 works without errors.
> >
> > I have a vague recollection Andrew Janke once explained how to deal with
> > this but I am unable to locate that post (if it ever existed).
> >
> > Thanks!
> > Soren
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From robert.d.vincent at mcgill.ca  Mon Sep 21 11:41:27 2015
From: robert.d.vincent at mcgill.ca (Robert D. Vincent)
Date: Mon, 21 Sep 2015 11:41:27 -0400
Subject: [MINC-users] mincreshape -dimsize
In-Reply-To: <CAHfsrRZg7by5S4SPVv6-Q0x5MU6t2Kg=XEp=BHFf_uV2P4VOUw@mail.gmail.com>
References: <CAHfsrRY2R08Gg94_JrK5tpfrKDtXMEw5ZekqJKW-fHx0qDbEug@mail.gmail.com>
	<CAG0CB22HS7GwqSExuG5SYkxZad1tHKfR+qtOvHmg+-rf8X9FLw@mail.gmail.com>
	<CAHfsrRZg7by5S4SPVv6-Q0x5MU6t2Kg=XEp=BHFf_uV2P4VOUw@mail.gmail.com>
Message-ID: <CAAd5DeTWLxajoQabBARdS5Y+Q0QtdpdNkjX4SDPiiDuyvEmnEw@mail.gmail.com>

Hi Soren,

Historically, I believe mincreshape really was only meant to resize the two
fastest-varying dimensions (xspace and yspace in this case), at least
according to the man page. However, I'm not sure whether there was some
attempt over the years to make it more flexible, but evidently it is
failing pretty clumsily. It certainly shouldn't issue the errors you are
seeing. I'll look into it.

    -bert

On Sun, Sep 20, 2015 at 11:45 PM, Soren Christensen <sorench at gmail.com>
wrote:

> Thanks Andrew!
>  Yes file is MINC2 and it was exactly your 2nd point I failed toremember in
> detail.
> It actually does not seem to do the trick - the image comes out without a
> zspace dimension if I add this option.
> I used mincresample instead, and using the desired -zsize and -zstep with
> no other options achieved the re-slicing I was looking for.
> Let me know if you want me to do something to document the mincreshape
> issue
>
> Thanks!
> Soren
>
> On Sun, Sep 20, 2015 at 4:41 PM, Andrew Janke <a.janke at gmail.com> wrote:
>
> > Hi Soren,
> >
> > This could be one of two things.
> >
> > 1. Is this a MINC1 or MINC2 file? If the former, convert to MINC2
> > first using mincconvert
> >
> > 2. I think what you are alluding to is "-dimrange zspace=-1" option
> > that will tell mincreshape to treat a slice dimension as an image
> > dimension.
> >
> > There has been a feature request kicking about for a long time to
> > remove this behaviour as it doesn't make as much sense as it used to
> > when Peter first implemented slice scaling in MINC1 but I don't think
> > it's made it into the latest release as of yet.
> >
> >
> > a
> >
> >
> >
> > On 20 September 2015 at 13:48, Soren Christensen <sorench at gmail.com>
> > wrote:
> > > Hi,
> > >  I have an issue with mincreshape I don't understand. I guess it has to
> > do
> > > with trying to interpolate the slowest varying dimension:
> > >
> > > file: pwi_bl.mnc
> > > image: signed__ float 0 to 1381.4969482421875
> > > image dimensions: zspace yspace xspace
> > >     dimension name         length         step        start
> > >     --------------         ------         ----        -----
> > >     zspace                     17            5      -38.912
> > >     yspace                    128        -1.75      95.0638
> > >     xspace                    128        -1.75      110.338
> > >
> > > u at K450$ mincreshape -dimsize zspace=49 -float -clobber  pwi_bl.mnc
> > > pwi_blIP.mnc
> > > Copying chunks:HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread
> > 0:
> > >   #000: H5Dio.c line 161 in H5Dread(): selection+offset not within
> extent
> > >     major: Dataspace
> > >     minor: Out of range
> > > (from mivarget): Can't read dataset /minc-2.0/image/0/image-min
> > > (from mivarget): Can't read variable ID# 1
> > > HDF5-DIAG: Error detected in HDF5 (1.8.15-patch1) thread 0:
> > >   #000: H5Dio.c line 161 in H5Dread(): selection+offset not within
> extent
> > >     major: Dataspace
> > >     minor: Out of range
> > > (from mivarget): Can't read dataset /minc-2.0/image/0/image-max
> > > (from mivarget): Can't read variable ID# 2
> > > .................................................Done.
> > >
> > > The file is written but it appears incorrect with a larger Z FOV for
> > > starters but it also look as if the extreme slices are simply copied
> > > outwards at each end (sorry cannot post images here). Finally the image
> > > appears corrupted in register and freezes up Display and the OS!
> > >
> > > file: pwi_blIP.mnc
> > > image: signed__ float 0 to 1381.4969482421875
> > > image dimensions: zspace yspace xspace
> > >     dimension name         length         step        start
> > >     --------------         ------         ----        -----
> > >     zspace                     49          2.5      -57.662
> > >     yspace                    128        -1.75      95.0638
> > >     xspace                    128        -1.75      110.338
> > >
> > > -dimsize xspace=64 works without errors.
> > >
> > > I have a vague recollection Andrew Janke once explained how to deal
> with
> > > this but I am unable to locate that post (if it ever existed).
> > >
> > > Thanks!
> > > Soren
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From a.janke at gmail.com  Mon Sep 21 18:59:53 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Tue, 22 Sep 2015 08:59:53 +1000
Subject: [MINC-users] mincreshape -dimsize
In-Reply-To: <CAAd5DeTWLxajoQabBARdS5Y+Q0QtdpdNkjX4SDPiiDuyvEmnEw@mail.gmail.com>
References: <CAHfsrRY2R08Gg94_JrK5tpfrKDtXMEw5ZekqJKW-fHx0qDbEug@mail.gmail.com>
	<CAG0CB22HS7GwqSExuG5SYkxZad1tHKfR+qtOvHmg+-rf8X9FLw@mail.gmail.com>
	<CAHfsrRZg7by5S4SPVv6-Q0x5MU6t2Kg=XEp=BHFf_uV2P4VOUw@mail.gmail.com>
	<CAAd5DeTWLxajoQabBARdS5Y+Q0QtdpdNkjX4SDPiiDuyvEmnEw@mail.gmail.com>
Message-ID: <CAG0CB21=B-Rvh51RoFPHdrXz_n+HuArsJMGAW4VUjjKMu=+_Ew@mail.gmail.com>

On 22 September 2015 at 01:41, Robert D. Vincent
<robert.d.vincent at mcgill.ca> wrote:
> Historically, I believe mincreshape really was only meant to resize the two
> fastest-varying dimensions (xspace and yspace in this case), at least
> according to the man page.

Correct, the original reason was due to slice scaling IIRC. If Peter
is still lurking he'll know the answer.

> However, I'm not sure whether there was some
> attempt over the years to make it more flexible, but evidently it is
> failing pretty clumsily. It certainly shouldn't issue the errors you are
> seeing. I'll look into it.

There was talk of it but it never happened. As it turns out I even
confused myself and told Soren the wrong thing the correct workaround
is this (to treat in this case zspace as an image dimension):

   -dimsize zspace=-1

It means most of my analysis starts with an incantation such as this
to make everything +ve steps, MINC2 and the same dimension order:

   mincreshape -2 \
      +direction \
      -dimorder zspace,yspace,xspace \
      -dimsize xspace=-1 \
      -dimsize yspace=-1 \
      -dimsize zspace=-1 \
      in.mnc out.mnc

As you rightly point out, this is not all that amenable to new users!


a

From a.janke at gmail.com  Mon Sep 21 21:58:11 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Tue, 22 Sep 2015 11:58:11 +1000
Subject: [MINC-users] Key mappings for register for time
Message-ID: <CAG0CB23wS+qmi5bCONbxHK7nG4HtKhujJPfHdTnSfjyfrL3sHw@mail.gmail.com>

Hi all,

Anyone (Bert?) happen to know if there is a key mapping so that you
can pan through the time dimension in register. Akin to using 2nd
mouse button for the in-plane dimension?

Thanks


a

From robert.d.vincent at mcgill.ca  Mon Sep 21 22:27:37 2015
From: robert.d.vincent at mcgill.ca (Robert D. Vincent)
Date: Mon, 21 Sep 2015 22:27:37 -0400
Subject: [MINC-users] Key mappings for register for time
In-Reply-To: <CAG0CB23wS+qmi5bCONbxHK7nG4HtKhujJPfHdTnSfjyfrL3sHw@mail.gmail.com>
References: <CAG0CB23wS+qmi5bCONbxHK7nG4HtKhujJPfHdTnSfjyfrL3sHw@mail.gmail.com>
Message-ID: <CAAd5DeQC518ie9P2NV9_YBPP__0ipFVPkGn=OSNX9hqLgc7N0w@mail.gmail.com>

Hi Andrew,

Greater than ('>') should step forward and less than ('<') should do the
opposite. I think.

    -bert

On Mon, Sep 21, 2015 at 9:58 PM, Andrew Janke <a.janke at gmail.com> wrote:

> Hi all,
>
> Anyone (Bert?) happen to know if there is a key mapping so that you
> can pan through the time dimension in register. Akin to using 2nd
> mouse button for the in-plane dimension?
>
> Thanks
>
>
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From a.janke at gmail.com  Mon Sep 21 22:35:34 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Tue, 22 Sep 2015 12:35:34 +1000
Subject: [MINC-users] Key mappings for register for time
In-Reply-To: <CAAd5DeQC518ie9P2NV9_YBPP__0ipFVPkGn=OSNX9hqLgc7N0w@mail.gmail.com>
References: <CAG0CB23wS+qmi5bCONbxHK7nG4HtKhujJPfHdTnSfjyfrL3sHw@mail.gmail.com>
	<CAAd5DeQC518ie9P2NV9_YBPP__0ipFVPkGn=OSNX9hqLgc7N0w@mail.gmail.com>
Message-ID: <CAG0CB23Co5ymH+AvpJMrh9ndgjAMDxti2nfHC7o_u9yA58i9Eg@mail.gmail.com>

thanks bert,

I was trying random combinations of shift, control and mouse movements to
no avail.

I'll have a dig in the code and see if I can map it to the mouse somehow
sensibly.

a
On 22/09/2015 12:28 pm, "Robert D. Vincent" <robert.d.vincent at mcgill.ca>
wrote:

> Hi Andrew,
>
> Greater than ('>') should step forward and less than ('<') should do the
> opposite. I think.
>
>     -bert
>
> On Mon, Sep 21, 2015 at 9:58 PM, Andrew Janke <a.janke at gmail.com> wrote:
>
> > Hi all,
> >
> > Anyone (Bert?) happen to know if there is a key mapping so that you
> > can pan through the time dimension in register. Akin to using 2nd
> > mouse button for the in-plane dimension?
> >
> > Thanks
> >
> >
> > a
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From eskild at gmail.com  Tue Sep 22 03:11:29 2015
From: eskild at gmail.com (Simon Eskildsen)
Date: Tue, 22 Sep 2015 09:11:29 +0200
Subject: [MINC-users] Total Intercranial Volume (TIV) in the minc
	toolchain?
In-Reply-To: <CAFQOoOpEz0D-e4O_NQLjkh-pvHJehPVGZsxjN-Hcw43wieQ=SA@mail.gmail.com>
References: <CAFQOoOpEz0D-e4O_NQLjkh-pvHJehPVGZsxjN-Hcw43wieQ=SA@mail.gmail.com>
Message-ID: <CAD3v435qv6_f_nh6faaAA9knimKFCU23AEixN-UF3+moxar+2w@mail.gmail.com>

Hi Gabriel,

It depends on how you define TIV.

If you aim for separating at the inner skull boundary, this is extremely
difficult using "normal" MPRAGE. That is why MEG/EEG/TMS folks use multiple
sequences (T1, T2, FLASH) to generate their boundary element models used
for solving the inverse problem. See for example
http://simnibs.de/version2/mri_sequences

However, if you "just" want brain + CSF, that is possible using normal
MPRAGE only. One way would be to generate a BEaST library with proper TIV
masks. You may also achieve usable masks using surface based tools such as
FSL BET or Freesurfer (did I just say that?!).

I think many would be interested in a library of accurate TIV masks :-)

Simon

On Tue, Sep 15, 2015 at 6:01 PM, Gabriel A. Devenyi <gdevenyi at gmail.com>
wrote:

> Hi all,
>
> I know I can get total brain volume (TBV) with BeAST with excellent
> results, but for some groups I think it would be more appropriate to use
> TIV to describe the subject.
>
> Does anyone know of a tool/set of commands in minc to achieve this? My
> first thought was to use the inverse BeAST mask to remove the brain and
> then use some kind of fill-in mask based on the intensity in the empty
> space.
>
> --
> Gabriel A. Devenyi B.Eng. Ph.D.
> Research Computing Associate
> Computational Brain Anatomy Laboratory
> Cerebral Imaging Center
> Douglas Mental Health University Institute
> McGill University
> t: 514.761.6131x4781
> e: gdevenyi at gmail.com
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From andrew at biospective.com  Tue Sep 22 09:29:49 2015
From: andrew at biospective.com (Andrew Wood)
Date: Tue, 22 Sep 2015 09:29:49 -0400
Subject: [MINC-users] Key mappings for register for time
In-Reply-To: <CAG0CB23Co5ymH+AvpJMrh9ndgjAMDxti2nfHC7o_u9yA58i9Eg@mail.gmail.com>
References: <CAG0CB23wS+qmi5bCONbxHK7nG4HtKhujJPfHdTnSfjyfrL3sHw@mail.gmail.com>
	<CAAd5DeQC518ie9P2NV9_YBPP__0ipFVPkGn=OSNX9hqLgc7N0w@mail.gmail.com>
	<CAG0CB23Co5ymH+AvpJMrh9ndgjAMDxti2nfHC7o_u9yA58i9Eg@mail.gmail.com>
Message-ID: <CABApf9HXosQ_ODLixvh1LeQu_Qdrezt=7PjeV5dCjrLnFdT4vQ@mail.gmail.com>

Hi Andrew,

FYI, I stumbled on an Easter egg that's related: Shift+T opens a window,
plotting voxel intensity over time.

Cheers,
Andrew

On Mon, Sep 21, 2015 at 10:35 PM, Andrew Janke <a.janke at gmail.com> wrote:

> thanks bert,
>
> I was trying random combinations of shift, control and mouse movements to
> no avail.
>
> I'll have a dig in the code and see if I can map it to the mouse somehow
> sensibly.
>
> a
> On 22/09/2015 12:28 pm, "Robert D. Vincent" <robert.d.vincent at mcgill.ca>
> wrote:
>
> > Hi Andrew,
> >
> > Greater than ('>') should step forward and less than ('<') should do the
> > opposite. I think.
> >
> >     -bert
> >
> > On Mon, Sep 21, 2015 at 9:58 PM, Andrew Janke <a.janke at gmail.com> wrote:
> >
> > > Hi all,
> > >
> > > Anyone (Bert?) happen to know if there is a key mapping so that you
> > > can pan through the time dimension in register. Akin to using 2nd
> > > mouse button for the in-plane dimension?
> > >
> > > Thanks
> > >
> > >
> > > a
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > >
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From robert.d.vincent at mcgill.ca  Tue Sep 22 10:17:15 2015
From: robert.d.vincent at mcgill.ca (Robert D. Vincent)
Date: Tue, 22 Sep 2015 10:17:15 -0400
Subject: [MINC-users] Key mappings for register for time
In-Reply-To: <CABApf9HXosQ_ODLixvh1LeQu_Qdrezt=7PjeV5dCjrLnFdT4vQ@mail.gmail.com>
References: <CAG0CB23wS+qmi5bCONbxHK7nG4HtKhujJPfHdTnSfjyfrL3sHw@mail.gmail.com>
	<CAAd5DeQC518ie9P2NV9_YBPP__0ipFVPkGn=OSNX9hqLgc7N0w@mail.gmail.com>
	<CAG0CB23Co5ymH+AvpJMrh9ndgjAMDxti2nfHC7o_u9yA58i9Eg@mail.gmail.com>
	<CABApf9HXosQ_ODLixvh1LeQu_Qdrezt=7PjeV5dCjrLnFdT4vQ@mail.gmail.com>
Message-ID: <CAAd5DeQiuXg5Q6rCNQ+oVjLVft367GQ8fHmdhFeky3jyYDEVjg@mail.gmail.com>

It wasn't supposed to be an "easter egg", it just never was properly
documented.

     -bert


On Tue, Sep 22, 2015 at 9:29 AM, Andrew Wood <andrew at biospective.com> wrote:

> Hi Andrew,
>
> FYI, I stumbled on an Easter egg that's related: Shift+T opens a window,
> plotting voxel intensity over time.
>
> Cheers,
> Andrew
>
> On Mon, Sep 21, 2015 at 10:35 PM, Andrew Janke <a.janke at gmail.com> wrote:
>
> > thanks bert,
> >
> > I was trying random combinations of shift, control and mouse movements to
> > no avail.
> >
> > I'll have a dig in the code and see if I can map it to the mouse somehow
> > sensibly.
> >
> > a
> > On 22/09/2015 12:28 pm, "Robert D. Vincent" <robert.d.vincent at mcgill.ca>
> > wrote:
> >
> > > Hi Andrew,
> > >
> > > Greater than ('>') should step forward and less than ('<') should do
> the
> > > opposite. I think.
> > >
> > >     -bert
> > >
> > > On Mon, Sep 21, 2015 at 9:58 PM, Andrew Janke <a.janke at gmail.com>
> wrote:
> > >
> > > > Hi all,
> > > >
> > > > Anyone (Bert?) happen to know if there is a key mapping so that you
> > > > can pan through the time dimension in register. Akin to using 2nd
> > > > mouse button for the in-plane dimension?
> > > >
> > > > Thanks
> > > >
> > > >
> > > > a
> > > > _______________________________________________
> > > > MINC-users at bic.mni.mcgill.ca
> > > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > > >
> > > _______________________________________________
> > > MINC-users at bic.mni.mcgill.ca
> > > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> > >
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> >
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From redoute at cermep.fr  Tue Sep 22 10:50:07 2015
From: redoute at cermep.fr (=?UTF-8?B?SsOpcsO0bWUgUmVkb3V0w6k=?=)
Date: Tue, 22 Sep 2015 16:50:07 +0200
Subject: [MINC-users] mincmath
Message-ID: <56016A9F.4030707@cermep.fr>

Hi All,
just a naive question concerning mincmath syntax:

if you want to substract 2 volumes : A - B

the correct syntax is:

mincmath -sub A.mnc B.mnc out.mnc

or

mincmath -sub B.mnc A.mnc out.mnc

(same question for -div  ;)

Thank you
JR

-- 
==================================================================
J?r?me Redout?
Ph.D. - Ing?nieur de Recherche - Universit? Claude Bernard - Lyon1
CERMEP - Imagerie du vivant
59 Bd Pinel. 69677 Bron - FRANCE
tel : 33 (0)4 72 68 86 18 (bureau)
tel : 33 (0)4 72 68 86 00 (standard)
fax : 33 (0)4 72 68 86 10
==================================================================



From a.janke at gmail.com  Tue Sep 22 10:56:46 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Wed, 23 Sep 2015 00:56:46 +1000
Subject: [MINC-users] mincmath
In-Reply-To: <56016A9F.4030707@cermep.fr>
References: <56016A9F.4030707@cermep.fr>
Message-ID: <CAG0CB21i9+OQ=NOdWoQ_wX_ND7S0D_8gDsRx3z5d39cnUviOZQ@mail.gmail.com>

Hi Jerome,

The first is correct (this one):

   mincmath -sub A.mnc B.mnc out.mnc


Be careful of div zero effects when using divide and mincmath.


a


On 23 September 2015 at 00:50, J?r?me Redout? <redoute at cermep.fr> wrote:
> Hi All,
> just a naive question concerning mincmath syntax:
>
> if you want to substract 2 volumes : A - B
>
> the correct syntax is:
>
> mincmath -sub A.mnc B.mnc out.mnc
>
> or
>
> mincmath -sub B.mnc A.mnc out.mnc
>
> (same question for -div  ;)
>
> Thank you
> JR
>
> --
> ==================================================================
> J?r?me Redout?
> Ph.D. - Ing?nieur de Recherche - Universit? Claude Bernard - Lyon1
> CERMEP - Imagerie du vivant
> 59 Bd Pinel. 69677 Bron - FRANCE
> tel : 33 (0)4 72 68 86 18 (bureau)
> tel : 33 (0)4 72 68 86 00 (standard)
> fax : 33 (0)4 72 68 86 10
> ==================================================================
>
>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users

From pvcorrection at yahoo.com  Tue Sep 22 11:13:10 2015
From: pvcorrection at yahoo.com (Olivier Rousset)
Date: Tue, 22 Sep 2015 08:13:10 -0700
Subject: [MINC-users] mincmath
In-Reply-To: <56016A9F.4030707@cermep.fr>
Message-ID: <1442934790.57342.YahooMailAndroidMobile@web160501.mail.bf1.yahoo.com>

The former it is.


- Olivier

Sent from Yahoo Mail on Android

From:"J?r?me Redout?" <redoute at cermep.fr>
Date:Tue, Sep 22, 2015 at 10:54 AM
Subject:[MINC-users] mincmath

Hi All,
just a naive question concerning mincmath syntax:

if you want to substract 2 volumes : A - B

the correct syntax is:

mincmath -sub A.mnc B.mnc out.mnc

or

mincmath -sub B.mnc A.mnc out.mnc

(same question for -div? ;)

Thank you
JR

-- 
==================================================================
J?r?me Redout?
Ph.D. - Ing?nieur de Recherche - Universit? Claude Bernard - Lyon1
CERMEP - Imagerie du vivant
59 Bd Pinel. 69677 Bron - FRANCE
tel : 33 (0)4 72 68 86 18 (bureau)
tel : 33 (0)4 72 68 86 00 (standard)
fax : 33 (0)4 72 68 86 10
==================================================================


_______________________________________________
MINC-users at bic.mni.mcgill.ca
http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users


From robert.d.vincent at mcgill.ca  Tue Sep 22 14:12:38 2015
From: robert.d.vincent at mcgill.ca (Robert D. Vincent)
Date: Tue, 22 Sep 2015 14:12:38 -0400
Subject: [MINC-users] mincreshape -dimsize
In-Reply-To: <CAG0CB21=B-Rvh51RoFPHdrXz_n+HuArsJMGAW4VUjjKMu=+_Ew@mail.gmail.com>
References: <CAHfsrRY2R08Gg94_JrK5tpfrKDtXMEw5ZekqJKW-fHx0qDbEug@mail.gmail.com>
	<CAG0CB22HS7GwqSExuG5SYkxZad1tHKfR+qtOvHmg+-rf8X9FLw@mail.gmail.com>
	<CAHfsrRZg7by5S4SPVv6-Q0x5MU6t2Kg=XEp=BHFf_uV2P4VOUw@mail.gmail.com>
	<CAAd5DeTWLxajoQabBARdS5Y+Q0QtdpdNkjX4SDPiiDuyvEmnEw@mail.gmail.com>
	<CAG0CB21=B-Rvh51RoFPHdrXz_n+HuArsJMGAW4VUjjKMu=+_Ew@mail.gmail.com>
Message-ID: <CAAd5DeR-X+Jx_msneMEk-DHouEghXp+_0P2UxO7QB=Y3FsM5BA@mail.gmail.com>

OK, I have a fix for the most serious issue which triggered the bizarre
HDF5 error messages. This in turn causes the image scaling to be wrong, and
led to the inability of Register or Display to load the file properly.

I'm not happy with the behavior of the program in this case even so.
Changing dimension sizes does change the FOV in ways that are fairly
surprising. I assume this is because mincreshape is limited in the ways it
can resample the data, but I would still like to see it act in a more
predictable fashion.

My advice in these situations is often "please use mincresample".

    -bert

On Mon, Sep 21, 2015 at 6:59 PM, Andrew Janke <a.janke at gmail.com> wrote:

> On 22 September 2015 at 01:41, Robert D. Vincent
> <robert.d.vincent at mcgill.ca> wrote:
> > Historically, I believe mincreshape really was only meant to resize the
> two
> > fastest-varying dimensions (xspace and yspace in this case), at least
> > according to the man page.
>
> Correct, the original reason was due to slice scaling IIRC. If Peter
> is still lurking he'll know the answer.
>
> > However, I'm not sure whether there was some
> > attempt over the years to make it more flexible, but evidently it is
> > failing pretty clumsily. It certainly shouldn't issue the errors you are
> > seeing. I'll look into it.
>
> There was talk of it but it never happened. As it turns out I even
> confused myself and told Soren the wrong thing the correct workaround
> is this (to treat in this case zspace as an image dimension):
>
>    -dimsize zspace=-1
>
> It means most of my analysis starts with an incantation such as this
> to make everything +ve steps, MINC2 and the same dimension order:
>
>    mincreshape -2 \
>       +direction \
>       -dimorder zspace,yspace,xspace \
>       -dimsize xspace=-1 \
>       -dimsize yspace=-1 \
>       -dimsize zspace=-1 \
>       in.mnc out.mnc
>
> As you rightly point out, this is not all that amenable to new users!
>
>
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From mishkind at gmail.com  Tue Sep 22 18:36:22 2015
From: mishkind at gmail.com (Mishkin Derakhshan)
Date: Tue, 22 Sep 2015 12:36:22 -1000
Subject: [MINC-users] mincmath
In-Reply-To: <CAG0CB21i9+OQ=NOdWoQ_wX_ND7S0D_8gDsRx3z5d39cnUviOZQ@mail.gmail.com>
References: <56016A9F.4030707@cermep.fr>
	<CAG0CB21i9+OQ=NOdWoQ_wX_ND7S0D_8gDsRx3z5d39cnUviOZQ@mail.gmail.com>
Message-ID: <CAB-GmL3=OcT+Yy2aoQ8rYTnR01apr8bfQmTrbQc_wv0gk=ktbA@mail.gmail.com>

Hi Jerome,
If you haven't discovered it by now, the minccalc -expression command is
also great for both simple and advanced calculations.

Here is an example of subtraction
mincmath -expression "A[0]-A[1]" a.mnc B.mnc out.mnc
A[0] is the first mmc file specified on the command line, so a.mnc
A[1] is the second input, so B.mnc
and out.mc is the output file where the result of a.mnc-B.mnc would go. If
you wanted B.mnc - a.mnc, you could either change the expression to read
"A[1]-A[0]" or you could leave the expression as is, and just change the
ordering of your input files:
mincmath -expression "A[0]-A[1]" B.mnc a.mnc out.mnc

There are lots of other types of expressions possible, but one of the most
useful ones I've used is the if/else
mincmath -expr "A[1] != 0 ? A[0]/A[1]:0" a.mnc b.mnc out.mnc

The general syntax of the expression is first evaluate the stuff before the
?, if it is true, then in the output file you put the stuff on the left of
the colon, else, write out the stuff on the right of the :

So the above would read, if a voxel in A[1] is not equal to zero, then in
the output file put in A[0] divided by A[1]. And if A[1] is equal to zero,
then instead of outputting the division, just output 0. Like in the above
a.mnc maps to A[0] and b.mnc maps to A[1].

See the minccalc man page for lots of documentation.

hth,
mishkin


On Tue, Sep 22, 2015 at 4:56 AM, Andrew Janke <a.janke at gmail.com> wrote:

> Hi Jerome,
>
> The first is correct (this one):
>
>    mincmath -sub A.mnc B.mnc out.mnc
>
>
> Be careful of div zero effects when using divide and mincmath.
>
>
> a
>
>
> On 23 September 2015 at 00:50, J?r?me Redout? <redoute at cermep.fr> wrote:
> > Hi All,
> > just a naive question concerning mincmath syntax:
> >
> > if you want to substract 2 volumes : A - B
> >
> > the correct syntax is:
> >
> > mincmath -sub A.mnc B.mnc out.mnc
> >
> > or
> >
> > mincmath -sub B.mnc A.mnc out.mnc
> >
> > (same question for -div  ;)
> >
> > Thank you
> > JR
> >
> > --
> > ==================================================================
> > J?r?me Redout?
> > Ph.D. - Ing?nieur de Recherche - Universit? Claude Bernard - Lyon1
> > CERMEP - Imagerie du vivant
> > 59 Bd Pinel. 69677 Bron - FRANCE
> > tel : 33 (0)4 72 68 86 18 (bureau)
> > tel : 33 (0)4 72 68 86 00 (standard)
> > fax : 33 (0)4 72 68 86 10
> > ==================================================================
> >
> >
> > _______________________________________________
> > MINC-users at bic.mni.mcgill.ca
> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>

From a.janke at gmail.com  Tue Sep 22 22:56:09 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Wed, 23 Sep 2015 12:56:09 +1000
Subject: [MINC-users] Quick way to flip data in a MINC file
Message-ID: <CAG0CB217RQF4_6B91jciP_B=YOsK2Ey+yNqRjzbGxGWVbbwPrw@mail.gmail.com>

Hi all,

Does anyone have any bright ideas/hackery of how to flip data (for a
dimension) in a MINC file? Here is what I currently do:

   https://github.com/BIC-MNI/minc-widgets/blob/master/volflip/volflip#L84

This takes about 2.5 minutes for a average sized volume (700x272x187).
This has been more than fine for a long time, but now I have a use
case that requires more speed.

Failing this I'm going to have to resort to writing something in C/C++
or perhaps even use some HDF5 tool.



a

From jon at pipitone.ca  Wed Sep 23 05:46:44 2015
From: jon at pipitone.ca (Jon Pipitone)
Date: Wed, 23 Sep 2015 05:46:44 -0400
Subject: [MINC-users] Quick way to flip data in a MINC file
In-Reply-To: <CAG0CB217RQF4_6B91jciP_B=YOsK2Ey+yNqRjzbGxGWVbbwPrw@mail.gmail.com>
References: <CAG0CB217RQF4_6B91jciP_B=YOsK2Ey+yNqRjzbGxGWVbbwPrw@mail.gmail.com>
Message-ID: <20150923094641.GA4009@brick>

Hey Andrew,

My go-to these days for minc/hdf5 file manipulation is python/numpy, so
you could try that. Using Jason's pyminc interface, do something like: 

    from pyminc.volumes.factory import *
    import numpy as np

    infile = volumeFromFile('input.mnc')
    outfile = volumeLikeFile('input.mnc', 'output.mnc')

    outfile.data = np.copy(infile.data[::-1,...]) # flips first dim

    outfile.writeFile()
    outfile.closeVolume()

Untested, but should save you time on file I/O at least. I used np.copy
there because IIRC pyminc doesn't like numpy views (which is what you
get when you do unfancy indexing operations like this) but I could be
wrong. 

Jon. 

p.s. I've used volflip many many times. Thank you!

On 09/23, Andrew Janke wrote:
> Does anyone have any bright ideas/hackery of how to flip data (for a
> dimension) in a MINC file? Here is what I currently do:
> 
>    https://github.com/BIC-MNI/minc-widgets/blob/master/volflip/volflip#L84
> 
> This takes about 2.5 minutes for a average sized volume (700x272x187).

From zijdenbos at gmail.com  Wed Sep 23 06:33:31 2015
From: zijdenbos at gmail.com (Alex Zijdenbos)
Date: Wed, 23 Sep 2015 06:33:31 -0400
Subject: [MINC-users] Quick way to flip data in a MINC file
In-Reply-To: <CAG0CB217RQF4_6B91jciP_B=YOsK2Ey+yNqRjzbGxGWVbbwPrw@mail.gmail.com>
References: <CAG0CB217RQF4_6B91jciP_B=YOsK2Ey+yNqRjzbGxGWVbbwPrw@mail.gmail.com>
Message-ID: <CABEEk17pT50qqABLqfyQ_v8RbJePgys7omr0xk95HLPAXZdCzg@mail.gmail.com>

You are running through the volume 4 times, that does seem a bit excessive.

I use mincresample for this with a flipping xfm, using the code snippet
below. Draws on a perl lib that I can't (yet) share, but I'd be happy to
provide the relevant subs (i.e., a working script) if anybody is
interested. But the code below is probably understandable as-is. It
supports flipping about the center-of-mass (com), start=0, or (what you are
after) the volume center (center).

-- A

-----

my %Dirs = ('xspace', 0, 'yspace', 1, 'zspace', 2);

# Generate the flipping xfm
my $Flip = &TmpFile('flip.xfm');

my @Center = (0, 0, 0);

if ($Center eq 'com') {
    my @dimi = @Dirs{@Dir};
    $Center[@dimi] = (&CoM($In))[@dimi];
}
elsif (($Center eq 'center') || ($Center eq 'start')) {
    my $s = &Sampling($In);

    foreach my $space (@Dir) {
    my $dimi = $Dirs{$space};

    die "Volume $In has no $space dimension?!\n" if (! $s->{n}[$dimi]);

    $Center[$dimi] = ($Center eq 'start') ?
        $s->{start}[$dimi] : $s->{start}[$dimi] + ($s->{n}[$dimi] - 1)/2.0
* $s->{step}[$dimi];
    }
}

my @Scales = (1, 1, 1);

foreach my $space (@Dir) {
    $Scales[$Dirs{$space}] = -1;
}

&Run(['param2xfm', '-clobber', '-scales', @Scales, '-center', @Center,
$Flip]) && die;

# Resample the input volume
my @MR = ('mincresample', '-clobber', $Interpolation, '-transformation',
$Flip);
push(@MR, ($Center eq 'start') ? '-tfm_input_sampling' : ('-like', $In));
&Run([@MR, $In, $Out]) && die;


On Tue, Sep 22, 2015 at 10:56 PM, Andrew Janke <a.janke at gmail.com> wrote:

> Hi all,
>
> Does anyone have any bright ideas/hackery of how to flip data (for a
> dimension) in a MINC file? Here is what I currently do:
>
>    https://github.com/BIC-MNI/minc-widgets/blob/master/volflip/volflip#L84
>
> This takes about 2.5 minutes for a average sized volume (700x272x187).
> This has been more than fine for a long time, but now I have a use
> case that requires more speed.
>
> Failing this I'm going to have to resort to writing something in C/C++
> or perhaps even use some HDF5 tool.
>
>
>
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>
>

From redoute at cermep.fr  Wed Sep 23 07:41:04 2015
From: redoute at cermep.fr (=?UTF-8?B?SsOpcsO0bWUgUmVkb3V0w6k=?=)
Date: Wed, 23 Sep 2015 13:41:04 +0200
Subject: [MINC-users] mincmath
In-Reply-To: <CAB-GmL3=OcT+Yy2aoQ8rYTnR01apr8bfQmTrbQc_wv0gk=ktbA@mail.gmail.com>
References: <56016A9F.4030707@cermep.fr>
	<CAG0CB21i9+OQ=NOdWoQ_wX_ND7S0D_8gDsRx3z5d39cnUviOZQ@mail.gmail.com>
	<CAB-GmL3=OcT+Yy2aoQ8rYTnR01apr8bfQmTrbQc_wv0gk=ktbA@mail.gmail.com>
Message-ID: <56028FD0.8070009@cermep.fr>

Thank you all!

Jerome

Le 23/09/2015 00:36, Mishkin Derakhshan a ?crit :
> Hi Jerome,
> If you haven't discovered it by now, the minccalc -expression command is
> also great for both simple and advanced calculations.
>
> Here is an example of subtraction
> mincmath -expression "A[0]-A[1]" a.mnc B.mnc out.mnc
> A[0] is the first mmc file specified on the command line, so a.mnc
> A[1] is the second input, so B.mnc
> and out.mc is the output file where the result of a.mnc-B.mnc would go. If
> you wanted B.mnc - a.mnc, you could either change the expression to read
> "A[1]-A[0]" or you could leave the expression as is, and just change the
> ordering of your input files:
> mincmath -expression "A[0]-A[1]" B.mnc a.mnc out.mnc
>
> There are lots of other types of expressions possible, but one of the most
> useful ones I've used is the if/else
> mincmath -expr "A[1] != 0 ? A[0]/A[1]:0" a.mnc b.mnc out.mnc
>
> The general syntax of the expression is first evaluate the stuff before the
> ?, if it is true, then in the output file you put the stuff on the left of
> the colon, else, write out the stuff on the right of the :
>
> So the above would read, if a voxel in A[1] is not equal to zero, then in
> the output file put in A[0] divided by A[1]. And if A[1] is equal to zero,
> then instead of outputting the division, just output 0. Like in the above
> a.mnc maps to A[0] and b.mnc maps to A[1].
>
> See the minccalc man page for lots of documentation.
>
> hth,
> mishkin
>
>
> On Tue, Sep 22, 2015 at 4:56 AM, Andrew Janke <a.janke at gmail.com> wrote:
>
>> Hi Jerome,
>>
>> The first is correct (this one):
>>
>>     mincmath -sub A.mnc B.mnc out.mnc
>>
>>
>> Be careful of div zero effects when using divide and mincmath.
>>
>>
>> a
>>
>>
>> On 23 September 2015 at 00:50, J?r?me Redout? <redoute at cermep.fr> wrote:
>>> Hi All,
>>> just a naive question concerning mincmath syntax:
>>>
>>> if you want to substract 2 volumes : A - B
>>>
>>> the correct syntax is:
>>>
>>> mincmath -sub A.mnc B.mnc out.mnc
>>>
>>> or
>>>
>>> mincmath -sub B.mnc A.mnc out.mnc
>>>
>>> (same question for -div  ;)
>>>
>>> Thank you
>>> JR
>>>
>>> --
>>> ==================================================================
>>> J?r?me Redout?
>>> Ph.D. - Ing?nieur de Recherche - Universit? Claude Bernard - Lyon1
>>> CERMEP - Imagerie du vivant
>>> 59 Bd Pinel. 69677 Bron - FRANCE
>>> tel : 33 (0)4 72 68 86 18 (bureau)
>>> tel : 33 (0)4 72 68 86 00 (standard)
>>> fax : 33 (0)4 72 68 86 10
>>> ==================================================================
>>>
>>>
>>> _______________________________________________
>>> MINC-users at bic.mni.mcgill.ca
>>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>> _______________________________________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>


-- 
==================================================================
J?r?me Redout?
Ph.D. - Ing?nieur de Recherche - Universit? Claude Bernard - Lyon1
CERMEP - Imagerie du vivant
59 Bd Pinel. 69677 Bron - FRANCE
tel : 33 (0)4 72 68 86 18 (bureau)
tel : 33 (0)4 72 68 86 00 (standard)
fax : 33 (0)4 72 68 86 10
==================================================================



From a.janke at gmail.com  Wed Sep 23 21:44:51 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Thu, 24 Sep 2015 11:44:51 +1000
Subject: [MINC-users] Quick way to flip data in a MINC file
In-Reply-To: <20150923094641.GA4009@brick>
References: <CAG0CB217RQF4_6B91jciP_B=YOsK2Ey+yNqRjzbGxGWVbbwPrw@mail.gmail.com>
	<20150923094641.GA4009@brick>
Message-ID: <CAG0CB20NK6f1fR45Z2WBkKNa75Mpprs99XXtT8hkhtv72+v4+Q@mail.gmail.com>

Hi Jon,

Thanks for the pyminc snippet, I've given it a try and the speed is
(as expected) better, still not as fast as I'd like but getting there.

a


On 23 September 2015 at 19:46, Jon Pipitone <jon at pipitone.ca> wrote:
> Hey Andrew,
>
> My go-to these days for minc/hdf5 file manipulation is python/numpy, so
> you could try that. Using Jason's pyminc interface, do something like:
>
>     from pyminc.volumes.factory import *
>     import numpy as np
>
>     infile = volumeFromFile('input.mnc')
>     outfile = volumeLikeFile('input.mnc', 'output.mnc')
>
>     outfile.data = np.copy(infile.data[::-1,...]) # flips first dim
>
>     outfile.writeFile()
>     outfile.closeVolume()
>
> Untested, but should save you time on file I/O at least. I used np.copy
> there because IIRC pyminc doesn't like numpy views (which is what you
> get when you do unfancy indexing operations like this) but I could be
> wrong.
>
> Jon.
>
> p.s. I've used volflip many many times. Thank you!
>
> On 09/23, Andrew Janke wrote:
>> Does anyone have any bright ideas/hackery of how to flip data (for a
>> dimension) in a MINC file? Here is what I currently do:
>>
>>    https://github.com/BIC-MNI/minc-widgets/blob/master/volflip/volflip#L84
>>
>> This takes about 2.5 minutes for a average sized volume (700x272x187).
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users

From a.janke at gmail.com  Wed Sep 23 22:02:13 2015
From: a.janke at gmail.com (Andrew Janke)
Date: Thu, 24 Sep 2015 12:02:13 +1000
Subject: [MINC-users] Quick way to flip data in a MINC file
In-Reply-To: <CABEEk17pT50qqABLqfyQ_v8RbJePgys7omr0xk95HLPAXZdCzg@mail.gmail.com>
References: <CAG0CB217RQF4_6B91jciP_B=YOsK2Ey+yNqRjzbGxGWVbbwPrw@mail.gmail.com>
	<CABEEk17pT50qqABLqfyQ_v8RbJePgys7omr0xk95HLPAXZdCzg@mail.gmail.com>
Message-ID: <CAG0CB22vxgJk=37sQ58Bq6U2p+4H4byuRh8XrkZK3NkVtEx37g@mail.gmail.com>

Hi Alex,

The mincresample approach (with -nearest_neighbour) is what is used in
volgemodel for MDA's and works well, especially if you have other
transforms to concatenate in, but I was after something faster and
data based. Beware also that from a look at your chunk of code that I
think this will not do what you want if a volume has direction
cosines.

To get around this you could use minc_modify_header to set the starts
and direction cosines to identity, mincresample with your flip xfm and
then set the direction cosines and starts back to what they were.


a


On 23 September 2015 at 20:33, Alex Zijdenbos <zijdenbos at gmail.com> wrote:
> You are running through the volume 4 times, that does seem a bit excessive.
>
> I use mincresample for this with a flipping xfm, using the code snippet
> below. Draws on a perl lib that I can't (yet) share, but I'd be happy to
> provide the relevant subs (i.e., a working script) if anybody is
> interested. But the code below is probably understandable as-is. It
> supports flipping about the center-of-mass (com), start=0, or (what you are
> after) the volume center (center).
>
> -- A
>
> -----
>
> my %Dirs = ('xspace', 0, 'yspace', 1, 'zspace', 2);
>
> # Generate the flipping xfm
> my $Flip = &TmpFile('flip.xfm');
>
> my @Center = (0, 0, 0);
>
> if ($Center eq 'com') {
>     my @dimi = @Dirs{@Dir};
>     $Center[@dimi] = (&CoM($In))[@dimi];
> }
> elsif (($Center eq 'center') || ($Center eq 'start')) {
>     my $s = &Sampling($In);
>
>     foreach my $space (@Dir) {
>     my $dimi = $Dirs{$space};
>
>     die "Volume $In has no $space dimension?!\n" if (! $s->{n}[$dimi]);
>
>     $Center[$dimi] = ($Center eq 'start') ?
>         $s->{start}[$dimi] : $s->{start}[$dimi] + ($s->{n}[$dimi] - 1)/2.0
> * $s->{step}[$dimi];
>     }
> }
>
> my @Scales = (1, 1, 1);
>
> foreach my $space (@Dir) {
>     $Scales[$Dirs{$space}] = -1;
> }
>
> &Run(['param2xfm', '-clobber', '-scales', @Scales, '-center', @Center,
> $Flip]) && die;
>
> # Resample the input volume
> my @MR = ('mincresample', '-clobber', $Interpolation, '-transformation',
> $Flip);
> push(@MR, ($Center eq 'start') ? '-tfm_input_sampling' : ('-like', $In));
> &Run([@MR, $In, $Out]) && die;
>
>
> On Tue, Sep 22, 2015 at 10:56 PM, Andrew Janke <a.janke at gmail.com> wrote:
>
>> Hi all,
>>
>> Does anyone have any bright ideas/hackery of how to flip data (for a
>> dimension) in a MINC file? Here is what I currently do:
>>
>>    https://github.com/BIC-MNI/minc-widgets/blob/master/volflip/volflip#L84
>>
>> This takes about 2.5 minutes for a average sized volume (700x272x187).
>> This has been more than fine for a long time, but now I have a use
>> case that requires more speed.
>>
>> Failing this I'm going to have to resort to writing something in C/C++
>> or perhaps even use some HDF5 tool.
>>
>>
>>
>> a
>> _______________________________________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>>
>>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users