Output is not displayed or saved

[slmvsk]

Hi, I was following your manual step-by-step, but this: "After analysis is done, the output table is displayed. The output files of this analysis are saved in the subfolder of “ctFIREout” in the format of overlaid image file, “.CSV” file,
“.mat” file. To be noted, the angles of individual fiber are saved in “HistANG_ctFIRE_2B_D9_ROI2.csv”, overlaid image file name is “OL_ctFIRE_2B_D9_ROI2.tif”."

is not working for some reason.

No errors, MatLab tells that output is saved, but when I am checking the folder I only see 1 csv file with few numbers in a row.
Example image is attached (it is originally a tiff 8 bit stack).

[slmvsk]

Btw I know it is not preprocessed, but I expected to see a mess after first trial and then adjust.

[yuminguw]

Hi slmvsk,

Thanks for your inquiry. Can you tell us which version you were using? What are you going to measure ? What are currently included in the output folder(screenshot?) In addition, it would be very helpful if you can provide a download link to the original stack image such that we can repeat the analysis and do a fast troubleshooting.

Best,
Yuming

[slmvsk]

@yuminguw thank you for the reply! I am using R2022a MatLab and CA5.0, CTF3.0 versions on my Mac. My overall goal is to measure straightness of the dendrites (the image above is the example of straight fibers, I also have more complex curly fiber image), but I was trying also measure only length to see how It works. Because for me the 1st problem is to preprocess and segment them properly, then measure. And It is possible that I will need to look into other measurements to describe straightness after segmentation. Thank you for your time! Here is the link

Best,
Slmvsk

[yuminguw]

Hi Slmvsk,
Just got a chance to look at your image. But I couldn't repeat the error happened to you.
I used the CTF 3.0 MAC 2021a version compiled with MATLAB 2021a (rather than 2022a, maybe a typo in your message above). I only chose the first 2 slices for the test using the default parameters, the analysis went through. To remove some over-segmented fibers, I then changed the background threshold (thresh_im2, 1st parameter in the parameter window) to 10 and the local maximum searching box size (s_xlinkbox, 2nd parameter) to 12. That analysis went through as well with expected output files. A screenshot is shown below:

However, I did notice that there is a bug in the visualization of the stack image. I will release a new build to fix that.(will let you know here when it becomes available)
By the way, as far as your image analysis is concerned, I am not quite sure how to define a fiber in your image. If all signals are from "fibers", then the big variation of the fiber thickness would prevent the algorithm accurately identifying individual fibers. A skeletonization step might be needed as a pre-processing for tracking individual fibers. In addition, for the curvy fiber tracking, it has been a challenging problem to extract intact long curvy fibers in a noise background or complex network. Many tools likely end up with dividing a long curvy fiber to several shorter fiber segments. As a workaround, the straightness of each fiber segment can be first calculated, and then averaged by a weight of their lengths (some applications may also need to consider fiber thickness).

Hope this helps.

Thanks,

Yuming

[slmvsk]

@yuminguw Hi! Thank you for your suggestions! It was very helpful.

Many tools likely end up with dividing a long curvy fiber to several shorter fiber segments.

Yes, that's true. So I want to have a measure to compare this kind of shorter segments and straight short segments in their straightness.) Well, It is next step anyway. Thanks again, I appreciate your time!

[yuminguw]

Hi Slmvsk,

No problem. You are welcome.
Are you able to run the CT-FIRE mac standalone compiled with MATLAB 2021a with correct output? If not, you can try the new Mac Intel 64-bit build out.
If it is still not working I would suggest, if possible, finding a Windows computer and try the windows standalone compiled with MATLAB 2018b built on June 7th 2024.

Please let me know how it goes.

Best,
Yuming