AFNI and SUMA new features - April 2005

April 2005:

• Add 'Pleg(m,x)' (Legendre polynomial) function to 3dcalc. Also 'stdev(...)' and 'sem(...)' functions for standard deviation and standard error of the mean, each function taking an arbitrary list of arguments.
• AFNI programs can now read NIfTI-1 '.nii' (or '.nii.gz') files directly. The goal is to make interoperation with other software packages (e.g., FSL, SPM) simpler on the pitiful users.
• Add '.nii' file output to THD_writedset() function, with the result that most AFNI programs will now write NIfTI-1.1 formatted files if the output -prefix ends in '.nii' (or in '.nii.gz').
  • This specifically includes to3d (converting image files to AFNI or NIfTI-1.1 datasets).
  • Program 3dcopy can be used to convert NIfTI-1.1 datasets to AFNI format, if so desired.
  • Program 3dAFNItoNIFTI can be used to convert AFNI format datasets to NIfTI-1.1 format.
  • NIfTI-1.1 and AFNI datasets do not completely overlap in their conceptual space. As a result, not all meta-data is preservable when writing a .nii file from AFNI. And not all NIfTI-1.1 files can be read into AFNI (for example, AFNI doesn't support 64-bit float datasets, or vector-valued time series -- 5 dimensional NIfTI-1.1 files).
• Add -automask option to 3dDeconvolve; however, using a precomputed mask (via -mask) is more efficient.
• Add function mri_write_jpg() function; now, if an output 2D image filename ends in '.jpg', a JPEG file will be written (via program cjpeg) rather than a PPM file. Applies in particular to program imupsam (for interpolating an image to a finer matrix).
  • Also modified mri_write() functions to output to stdout if the filename is "-"; in this way, a program like imupsam can write to a pipe.
• Modified plug_tag.c to allow tags to be applied to "functional" datasets, as well as "anatomicals". A few other plugins were also fixed up in this regard.
• Enable the gamma 'g' arrows in an image viewer to adjust the gamma factor for RGB images; previously, the user couldn't do much to modify the way RGB images (from RGB datasets, or in the volume renderer) were displayed.
  • Also, if AFNI_STROKE_THRESHOLD is enabled, click-and-drag adjustment of the gamma factor will work (much as click-and-drag works for the grayscale adjustment; see README.environment).
  • The 'Norm' button will reset the color gamma factor to 1.0, when you really screw it up and can't see anything any more!
  • Most people don't know it, but the 'Sharpen' button on the 'Disp' control panel will sharpen the underlay images, both grayscale and RGB.
• 3dDeconvolve's -nodata option will now work with the -stim_times option. For more details, go to http://afni.nimh.nih.gov/afni/doc/misc/3dDeconvolveSummer2004/ and scroll to the bottom.
• The "transforms" options on the Disp image viewer control panel will now directly affect the viewing of RGB underlay datasets.
  • And the 'g' (gamma) control arrows can be used to change the appearance of such colored datasets.
  • These controls do not affect the overlay image. However, you can 'transform' the overlay image from the hidden right-click popup available under the 'Inten' label on the 'Define Overlay' control panel.
  • The diffusion directionality image below was created from an RGB dataset, which was filtered via the 'Median21' and 'Sharpen' controls to produce the pleasing watercolor-ish effect. Followed by a 'Save:agif' to make the animated GIF.

In the comments below, Daniel Glen mentions his new 3dDWItoDT program for diffusion tensor calculations. Here is an image produced by colorizing the principal eigenvector of the D matrix: