[openvl-devel] SUBVOLUME HANDLER:

[Susan F. <sf...@cs...>]

Much of what I am using OpenVl for
involves combining and splitting massive volumes up. Due to the large
size of the data, it is very inefficient, and sometimes not possible, to
read in parts I don't need. I have put subvolume and slice handling into
vlvolume rather than create volume processing plugins because some object
that are private to vlvolume were necessary.
In particular, in vlvolume I have the following:

READING PART OF A VOLUME:
readImages reads slices. I have implemented it to allow for reading one 
datatype into another so that I don't need to create a separate volume when 
these don't match (see cropREDtoRGBA, cropRGBtoRGBA, cropRGBtoRed). This 
occurs in much of my work since I am adding the alpha channel to rgb data. It 
is also useful for grabbing only one channel for regiongrowing. I also can 
read only part of the slice. I am now changing this so that the fileio 
plugins automatically detect if only part of a file is desired based on 
info.origFilePos. So far. I have put this as slab only into vox (not checked 
yet). This may take a while since it affects code in many plugins, but when 
it's done the above crop___ routines will change to copy___ instead, which 
seems like cleaner code.

I think that a separate SubVolume class and slice class, which would derive 
from vlvolume may be a good idea. This would include most of the code I put 
in vlvolume so far, and would return vlvolume to it's original level of 
purity. Any comments?