liu-yikang / sherm-rodentskullstrip Goto Github PK

Hello,

I'm trying to use SHERM on one of my images and I changed Demo.m to read nii.gz:

img_list = dir(fullfile(input_dir, '*.nii.gz'));
animal = {'rat'};
isotropic = [0];

But then, it gives me this error:

Error using <
Matrix dimensions must agree.

Error in sherm (line 42)
[ids_chan, ids_mser] = find((desc_dist >0 & desc_dist) < (min(desc_dist(desc_dist>0)) + 0.05*sum(abs(desc_tmp))));

What is exactly that line and how could I fix it so that it works on my data?

Thanks.

Hello,

when I run the Demo.m file I get this error:

I try to figure out where the problem comes from.
Thanks for your work. It seems pretty great.

This Package needs the: "statistics and machine learning toolbox" (also the "image processing toolbox"). It would be helpful if you could mention these requirements.

PS: the first command in demo.m root_dir = strrep(mfilename('fullpath'),'example',''); does nothing, because you renamed the file.

Dear SHERM team: I read that the default implementation specifies the convexity threshold to 0.85, but I can't find where in the code that's done. Could you please offer a hint where to look for this parameter?

Hi. I got the following error when trying to run sherm on my data (the images are anisotropic and have the correct header). Why does it happen, and what to do to make it work?

Error using get_convexity (line 9)
Error computing the convex hull. The points may be coplanar or collinear.
Error in sherm_parallel_2d (line 108)
parfor i_mser = 1:size(masks_channel, 4)
Error in sherm (line 34)
sherm_parallel_2d(nii,descriptor_filename,drs,brain_vol_range,...

Thanks!

Using my own dataset an anisotroph MRI sigma_rat (T1 & T2):

I get this error:

Hello,

I try to create a shape descriptor template for a mongolian gerbil (very similar to a mouse but not enough to just use the mouse template). I know there's a description about it in the paper but I'm not familiar with all the math you're doing to create the template. Do you have a code for this?

I think the get_shape_descriptor.m calculate the discriptor for the to-be-striped dataset. I tried to use my gerbil template mask to run this function, but the min desc value is always 0 and I don't know why, because the result isn't good at all and far away from perfect ("desc min = 0" means a perfect match?).