open-discussion > Shape analysis between non-similar objects?
Showing 1-3 of 3 posts
Results per page:
Sep 23, 2017  10:09 PM | shahrokh nasseri
Shape analysis between non-similar objects?

SPHARM developers and users,


Suppose I have 5 groups with names A, B, C, D and E and in each group there are 8 meat samples such as below:

A1, A2, A3, A4, A5, A6, A7, A8

B1, B2, B3, B4, B5, B6, B7, B8

C1, C2, C3, C4, C5, C6, C7, C8

D1, D2, D3, D4, D5, D6, D7, D8

E1, E2, E3, E4, E5, E6, E7, E8

Each group was subjected to different laboratory procedures (physicochemical methods) and these procedures had a significant effect on the surface of the samples.

I should point out that the cutting of meat samples has been carried out manually. Although I tried to be careful enough, this led to a significant difference in the shape and size of meat samples (due to the non-rigidity of meat samples).

After completing these steps, MRI images with T2 contrast were acquired. Although I have downloaded and installed SlicerSALT, but I've done the Preprocessing, Parameterization, and SPHARM-PDM steps with the Shape Analysis Module in Slicer.

At now, my question is how can I do a surface analysis on them, without considering the overall shape of the meat samples and their volume?

As mentioned above, three steps of Preprocessing, Parameterization and SPHARM-PDM were performed completely with the Shape Analysis Module in Slicer. For example, I've got the output files for one meat sample in the following steps:

in Step1_SegPostProcess: sample1_group04_pp.nii

in Step2_GenParaMesh: sample1_group04_pp_para.vtk,
sample1_group04_pp_surf.vtk, sample1_group05_pp_para.vtk, sample1_group05_pp_surf.vtk, sample1_group06_pp_para.vtk, sample1_group06_pp_surf.vtk

in Step3_ParaToSPHARMMesh:
sample1_group04_pp_surf_MedialAxisScalars.csv, sample1_group04_pp_surf_para.vtk, sample1_group04_pp_surf_paraMix.txt, sample1_group04_pp_surf_paraPhi.txt, sample1_group04_pp_surf_paraPhiHalf.txt, sample1_group04_pp_surf_paraTheta.txt, sample1_group04_pp_surf_SPHARM.coef, sample1_group04_pp_surf_SPHARM.vtk, sample1_group04_pp_surf_SPHARM_ellalign.coef, sample1_group04_pp_surf_SPHARM_ellalign.vtk, sample1_group04_pp_surf_SPHARMMedialAxis.vtk, sample1_group05_pp_surf_MedialAxisScalars.csv, sample1_group05_pp_surf_para.vtk, sample1_group05_pp_surf_paraMix.txt, sample1_group05_pp_surf_paraPhi.txt, sample1_group05_pp_surf_paraPhiHalf.txt, sample1_group05_pp_surf_paraTheta.txt, sample1_group05_pp_surf_SPHARM.coef,
sample1_group05_pp_surf_SPHARM.vtk, sample1_group05_pp_surf_SPHARM_ellalign.coef, sample1_group05_pp_surf_SPHARM_ellalign.vtk, sample1_group05_pp_surf_SPHARMMedialAxis.vtk, sample1_group06_pp_surf_MedialAxisScalars.csv, sample1_group06_pp_surf_para.vtk, sample1_group06_pp_surf_paraMix.txt, sample1_group06_pp_surf_paraPhi.txt, sample1_group06_pp_surf_paraPhiHalf.txt, sample1_group06_pp_surf_paraTheta.txt, sample1_group06_pp_surf_SPHARM.coef, sample1_group06_pp_surf_SPHARM.vtk, sample1_group06_pp_surf_SPHARM_ellalign.coef, sample1_group06_pp_surf_SPHARM_ellalign.vtk, sample1_group06_pp_surf_SPHARMMedialAxis.vtk

The solution to my mind for this analysis is to use Parameterization outputs, especially Parameterization spheres. Is it true? Since I think that mapping the surface mesh of each sample onto a unit sphere causes the  surface analysis between samples to be independent of the radius and geometric shapes of samples, and analysis is done on characteristics and surface differences (such as roughness and smoothness).  Is that true? In  other words, can surface analysis be done using these unit spheres? Because, in my opinion, the quantities on these unit spheres represent the amplitude of the three-dimensional frequency signal of the sample surface.  Is it OK?

My next question starts with the fact that, as I pointed out above, meat samples that are inside a group, although they look different in terms of geometric shape, but they are similar in appearance and surface  characteristics. I must mentioned that if I put all the meat samples (8 * 5 = 40) inside a container, they can be distinguished only by their surface properties. At now, my question is that: Can we create an average  surface using SPHARM for each group and consider it as representative of that group and then compare the surfaces between these representatives?

I am very interested in the topic of Shape Analysis using SPHARM method. I think it's possible to use this method to compare the surfaces of non-similar objects in the Fourier domain.

Please guide me to do it. 

Thanks a lot.


Sep 25, 2017  06:09 AM | Beatriz Paniagua
RE: Shape analysis between non-similar object
Hi Shahrokh,

Could you please ask this question in the new SPHARM (SlicerSALT) user forum?
Thank you,

Sep 26, 2017  07:09 AM | shahrokh nasseri
RE: Shape analysis between non-similar object
Hi Beatriz,

According with your comment, I asked my questions in the SlicerSALT forum ( I hope to receive your guidance as soon as possible.

Thanks a lot.