search

awhite40 / MIC-Modeling-Polymer-Composites

Developing structure property linkages for glass fiber reinforced polymer systems.
awhite40.github.io/MIC-Modeling-Polymer-Composites
0 stars 0 forks source link

Potentially important statisitics #11

Open awhite40 opened 9 years ago

awhite40 commented 9 years ago

From Image: Fiber orientation distribution Fiber length Volume Fraction (directly from segmentation) Fiber Diameter (shouldn't vary according to person who made sample; should be 30um)

tonyfast commented 9 years ago

What are the scales of your voxels then and how do they relate to the diameter?

Do you know anything about the average fiber length?

awhite40 commented 9 years ago

Our voxels are 10um x 10um x 10 um. The diameter is 30um. No we do not know the average fiber length because:

  1. Our sample is cut out of a larger part so the fibers will be cut off at the edges
  2. The processing conditions will lead to breakage of the fibers (reference is in the processing conditions post)

However, we do know that as part of the processing, the fibers were cut to half an inch. So our fibers will be no longer than that.

tonyfast commented 9 years ago

Okay. So they what is the max fiber volume and length in voxels? We can use these. On Oct 10, 2014 4:09 PM, "Alicia White" notifications@github.com wrote:

Our voxels are 10um x 10um x 10 um. The diameter is 30um. No we do not know the average fiber length because:

  1. Our sample is cut out of a larger part so the fibers will be cut off at the edges
  2. The processing conditions will lead to breakage of the fibers (reference is in the processing conditions post)

However, we do know that as part of the processing, the fibers were cut to half an inch. So our fibers will be no longer than that.

— Reply to this email directly or view it on GitHub https://github.com/awhite40/MIC-Modeling-Polymer-Composites/issues/11#issuecomment-58709451 .

glahoti6 commented 9 years ago

@tonyfast @awhite40 http://loci.wisc.edu/software/ctfire While I was searching for something, I found this. This is for extracting collagen fibers in an image. Its output includes distributions for fiber angle, fiber length, fiber straightness, and fiber width. In our case, it seems to be working well for a single layer (a single image) but thats not what we want. There is an option for inputing multiple images but I am not sure whether the interconnectedness between the images is accounted in the final output or not. Interesting things are algorithms that this software is using. It uses Curvelet Transform and FIRE. Out of this two, FIRE seems to be helpful for fiber extraction in our case. I have tried the software. Now I will be trying FIRE matlab code separately. Let me know your thoughts.

awhite40 commented 9 years ago

I was looking at the documentation for Fire and the paper they published had an interesting section about how they tested their algorithm for extraction of fibers on a comutationally generated "sample" network with known fibers. It might be interesting to look at for the idea of generating microstructures. The title of the paper is "An algorithm for extracting the network geometry of three-dimensional collagen gels" and it was published in the journal of microscopy.

glahoti6 commented 9 years ago

@awhite40 http://www.mathworks.com/matlabcentral/fileexchange/43400-skeleton3d http://www.mathworks.com/matlabcentral/fileexchange/43527-skel2graph-3d