error

[Jackywei1]

While I was running your program, the following error occurred:
'Incorrect use of.*
The dimensions of the matrix must be consistent.'

[wqshmzh]

Which line produces the mistake ? Please tell me the line number. The code in Matlab R2018b is working and there's no error on my PC.

[Jackywei1]

I don't know which line of props is. The range is roughly in the FCM clustering cycle, but it may also be the problem with my matlab version, which is 2016a

[wqshmzh]

Well, I probably find the error. Please wait for my corrections. Thanks for your advise.

[Jackywei1]

Thank you for your code, which is very helpful to me

[wqshmzh]

Well, I have corrected the error. Please redownload the FCM_SICM_Main.m file. If you can run this code then, please let me know. :-)

[wqshmzh]

If you cannot acquire the desired segmentation result in one run, please give it another run.

[Jackywei1]

OK this time, but there is something wrong with the code in your first project

[Jackywei1]

In falrcm algorithm, region_ level_ On line 14 of the information function, the fspecial function has one less input parameter

[wqshmzh]

Try this "mask = fspecial('gaussian', [l l], sigma);"

[wqshmzh]

I forgot to delete the variable "flag" in the code. Sorry. Also I have uploaded a newer version of "region_level_information.m" file. Please redownload.

[Jackywei1]

I think there's a problem between lines 95-107 in the main function,
Misuse ./
Matrix dimensions must be consistent.

[wqshmzh]

I originally want to call GPUs to run the entire code. The variable "flag" is used to determine the number of callable GPUs on your PC. It is welcomed for you to use GPU to boost if you transfer all the matrices to type "gpuArray".

[wqshmzh]

Try to add "U_col_sum = repmat(U_col_sum, [1 cluster_num]);" below line 106. R2016a doesn't have automatic broadcast mechanism (Already incorporated in Numpy for Python) used in R2018b for matrices and this is the reason causing this ./ problem.

[Jackywei1]

According to your method, the above problems will also appear in the fuzzy clustering cycle

[wqshmzh]

Try to redownload the package now. If you can run, please let me know.