Hello Ruopeng ! First of all thank you very much for all the work you've done with this software, it's been a pleasure working with it !

I think I just have a little problem with track points coordinates. I want to compute the mean value of an image on a track (the b0 image for example), using matlab. As I saw in previous posts, coordinates in the track files are in mm, so they have to be divided by the voxel_size specified in the header to obtain the voxel coordinates. Then I floor the values so I get whole voxels coordinates (and also add 1 to each coordinate because matlab indexes don't start at zero). I add the values of the corresponding voxels in the b0 image, and divide the total by the number of points in the track to get the mean b0 value.

I've tried on only one fiber to see the result. Trackvis gives me a mean of 0.418886 but my script gives me a mean of 0.4039. If I do the calculation by taking only one trackpoint per voxel, the result I get is 0.4069. I also tried to round the voxel coordinate instead of flooring it, but the result I get is shorter…

But the more important thing is that I have also some trouble also with selecting fibers that go trought an ROI. I take each fiber's point, convert the coordinates as described above and see if the corresponding voxel belongs to the ROI to select the fiber. I'm able to select only 15 tracks whereas trackvis selects 38 (using the Any Part selection of the ROI filter).

Do I miss some point with the coordinate transformation ? I read that one should apply the vox_to_ras matrix (http://trackvis.org/blog/forum…..onvention/), but how can you do it as it's a 4x4 matrix ?

Sorry for the inconveniance, I hope you can help me on this :-)

David