Possible issue with coordinate conversion about brainbrowser HOT 6 CLOSED

Are you sure about this? If you select the functional data in the Volume Viewer demo, you'll see that the steps aren't equal (they're about 2.1:1). For the structural data, it is 1:1, but that's just because that dataset has steps of 1.

@cmadjar care to comment?

from brainbrowser.

The voxel coordinates displayed are the actual coordinates of the Brain Browser pixels and not the voxels of the minc file.

To get the minc voxel coordinates you will need to take into account direction_cosines. See info at: http://en.wikibooks.org/wiki/MINC/Reference/MINC2.0_File_Format_Reference

from brainbrowser.

It looks like this requires two steps:

1. The direction_cosines attribute should be extracted for each dimension from the minc headers in minc2volume-viewer.js.
2. The cosines should be applied in the voxel coordinate conversion in getWorldCoordinates().

from brainbrowser.

@pinelopakik, I'm a little confused by the matrix multiplication shown in the MINC reference you posted a link to. It seems all the matrices are transposed for some reason. If this is the case, that's no problem, but can you verify this with Louis? As a concrete example, can you ask him if the following would be the correct way to calculate the world x coordinate:
w_x = v_x * step_x * c_xx + v_y * step_x * c_xy + v_z * step_x * c_xz + o_x

from brainbrowser.

@pinelopakik, I created the following two matrices, based on the link you sent. Can you verify that they're correct? Also, I found another site giving a different matrix for the conversion. It seems to be older, but the matrix is actually written correctly. Can you make sure that the one we're using is correct?

EDIT: THE BELOW MATRICES ARE INCORRECT

from brainbrowser.

The original link an above matrices are incorrect. The correct conversion can be found here: http://www.bic.mni.mcgill.ca/software/minc/minc2_format/node4.html

from brainbrowser.