Package: ants Version: 1.9+svn532-2 Severity: normal
I tried warping a timeseries image with a warpfield and two consecutive affine transformations, but it fails because the process requires a huge amount of memory, probably due to a leak. This is the timeseries images: m...@head1 /tmp % fslinfo bold.nii.gz data_type FLOAT32 dim1 64 dim2 64 dim3 48 dim4 1101 datatype 16 pixdim1 3.0000000000 pixdim2 3.0000000000 pixdim3 3.0000000000 pixdim4 3.0000000000 cal_max 1462.2584 cal_min 0.0000 file_type NIFTI-1+ and here the reference image: m...@head1 /tmp % fslinfo MNI.nii.gz data_type INT16 dim1 64 dim2 73 dim3 64 dim4 1 datatype 4 pixdim1 3.0000000000 pixdim2 3.0000000000 pixdim3 3.0000000000 pixdim4 1.0000000000 cal_max 0.0000 cal_min 0.0000 file_type NIFTI-1+ I tried to run something like this: WarpTimeSeriesImageMultiTransform 4 bold.nii.gz warped4.nii.gz -R MNI.nii.gz myWarp.nii.gz myAffine.txt otherAffine.txt The process seems to run fine (usual output). However, after about 20% completion (about 20min time) it already consumes more than 20GB and that consumption grows steadily. I assume it processes the image volume by volume and there is a memory leak somewhere. The process eventually dies when it exceeds the machines memory (64GB!). I tried to trim the problem down to be able to debug it and made an image with only the first two volumes of bold.nii.gz. However, running the above command with this image points to another problem: m...@head1 /tmp % WarpTimeSeriesImageMultiTransform 4 example4d.nii.gz warped4d.nii.gz -R MNI.nii.gz myWarp.nii.gz myAffine.txt otherAffine.txt FIELD: myWarp.nii.gz AFFINE: myAffine.txt AFFINE: otherAffine.txt moving_image_filename: example4d.nii.gz output_image_filename: warped4d.nii.gz reference_image_filename: MNI.nii.gz [0/3]: FIELD: myWarp.nii.gz [1/3]: AFFINE: myAffine.txt [2/3]: AFFINE: otherAffine.txt Four-D image size: [64, 64, 48, 2] 4D-In-Spc [3, 3, 3, 3] 4D-In-Org [-66.1627, -88.7711, 84.4337, 0] 4D-In-Size [64, 64, 48, 2] 4D-In-Dir -0 0 1 0 1 -0 0 0 0 -1 -0 0 0 0 0 1 ...... 4D-Out-Spc [3, 3, 3, 3] 4D-Out-Org [-90, 126, -72, 0] 4D-Out-Size [64, 73, 64, 2] 4D-Out-Dir 1 0 0 0 0 -1 0 0 0 0 1 0 0 0 0 1 E: Child terminated by signal ‘Floating point exception’ (+ core dump) Finally, to see whether there is a problem with the volumes, the warpfield or the affines I ran this command on an image with only the first volume of bold.nii.gz: WarpImageMultiTransform 3 example3d.nii.gz warped3d.nii.gz -R MNI.nii.gz myWarp.nii.gz myAffine.txt otherAffine.txt and it behaves exactly like it should and yields the results I wanted to see. So I guess the transformations and input images are fine and the problem lies somewhere in WarpTimeSeriesImageMultiTransform. Thanks in advance, Michael -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
