-w argument now controls the output tomogram voxel width

Now the voxel width in the output tomogram ("-out" argument) equals the number passed to the "-w" argument. (Earlier, it would always equal the voxel width of the input tomogram ("-in" argument).)

fixed two bugs in the "crop_mrc" program

This small update fixes two bugs:
-fixed a crash which occurs if the xmax, ymax, or zmax arguments are set equal to the size of the image (their maximum allowed values is the size in that direction -1). Now the program handles this user error gracefully and prints a warning message.
-fixed (hopefully) a bug in the way the "origin" was being modified during the cropping process. (The "origin" is one of the fields in the header portion of the MRC file. It will be modified whenever the xmin, ymin, or zmin arguments are non-zero.)

can now search for non-spherical blobs (aligned with the XYZ axes)

I added the "-blob-aspect-ratio" argument to filter_mrc. This new argument allows users to search for blobs that are elongated in the X,Y,Z directions. This also required changing the API for VISFD. The "BlobDog()" function (in the "feature.hpp" file from the VISFD library) now accepts a new argument ("aspect_ratio").

This is not a true anisotropic blob-detector, because the blobs cannot be arbitrarily oriented. It is assumed they will be elongated in the X, Y, or Z directions .

a few small changes and improvements

• The "-best" argument has been renamed to "-tv-best".
• Fixed a minor bug in the -save-progress and -load-progress arguments.
• Fixed and documented the "-spheres-scale" argument.
• Updated some documentation, examples, and default arguments.

fixed bugs in blob detection

corrected Example#2 and fixed a bug in the -discard-blobs argument

new feature: manual control of "-must-link" fragment directions

I added manual control of normal directions when using the "-must-link" command line argument.
(This is a fix for issue #12. This feature is documented here.)
I also added more robust parsing of coordinate strings.

fixed a bug in surface detection

This fixes a typo which would occasionally cause memory errors when attempting to detect surfaces. This bug would either cause crashes (segmentation faults) or cause portions of the image to be ignored during detection. (This was a typo specifically in the code for the "filter_mrc" program. The core VISFD library was unaffected.)

several small bugs fixed and documentation improvements

• fixed bugs in the "-blob", "-must-link", "-draw-spheres" and "-auto-thresh-score" arguments for filter_mrc.
• fixed a bug in the "LabelConnected()" function which affects the "-connect" argument
• added GCC support
• updated the documentation. (I now recommend using the "-tv" argument with 5.0 or higher.)

fixed a bug in the "-mask-rect" and "-mask-sphere" arguments for "filter_mrc"

I fixed a bug in the "-mask-rect", "-mask-sphere", "-mask-rect-subtract", and "-mask-sphere-subtract" arguments for filter_mrc. (They were not working whenever binning occurred, either due to the use of the "-bin" argument, or due to auto-binning.)

added the "-markers" command line argument to "filter_mrc"

This allows "filter_mrc" users to specify their own custom markers for use with watershed segmentation. This was an attempt to replicate the "markers" argument of the skimage.morphology.watershed() function. (Although I have not actually verified that the two codes behave the same way.) I also fixed a bug in the "-find-minima" and "-find-maxima" arguments. I also continued cleaning up the messy VISFD API.