From minc-development@bic.mni.mcgill.ca Sun Jul 6 20:15:36 2003 From: minc-development@bic.mni.mcgill.ca (Peter NEELIN) Date: Sun, 6 Jul 2003 15:15:36 -0400 Subject: [MINC-development] Draft API for MINC 2.0 In-Reply-To: <20030613164231.E9376@sickkids.ca> Message-ID: Hi John, I've finally had a chance to quickly read through the draft API - looks good. It would be helpful to have a document to go with it that gave some overview of the basic concepts. The only thing that I noticed to be missing was a set of functions for dealing with spatial voxel-to-world conversions. Either functions for doing voxel-to-world transformations, or for returning a transformation object (perhaps a matrix of some form, although then the programmer would have to deal with non-spatial dimensions that might be intermixed with spatial ones). Another thing that is non-trivial to compute and that might be helpful in the minc library (although it is already in the Proglib library) is a mapping between a world coordinate and the dimension start values. For non-orthogonal axes, the projection onto the axes is not an orthogonal projection and so requires some thought (see convert_origin_to_start.c). This is mostly useful for writing converters when one knows the coordinate of the first voxel and one wants to compute the start values. Apart from that, a very minor point is that MINC 1.0 refers to images and not slices. There are a few new functions that talk about slice max and min, etc. If these refer to minc 1 images (fastest-varying 2 or 3 (for vector) dimensions in the image variable), then they should probably use the same word. Peter ---- Peter Neelin (neelin@bic.mni.mcgill.ca)