Parfor loop classification "fix usage of indicated variable" error
5 views (last 30 days)
Show older comments
I'm running a parfor loop and I'm trying to output an array with the columns representing each parfor iteration, and the row elements are converted to a non-zero value at specific locations for each column index. The non-zero element row index is different for every iteration.
The problem: Unable to classify the variable 'Binarizing_Rowsum' in the body of the parfor-loop. For more information, see Parallel for Loops in MATLAB, "Solve Variable Classification Issues in parfor-Loops".
I tried to understand how to fix it by readig the documentation and other answers on the forum but I'm lost. I understand it is considered a sliced variable, however, I don't know how to fix the representation of the variable so that it matchs parfor requirements to get the output I want.
P.S. the correspoding for-loop worked perfectly
Here is the code:
jValues=1:2:length(baseFileNames)-1;
Binarizing_Rowsum=zeros(6100,length(jValues));
LEDdurIndx=NaN(1000,length(jValues));
Light=3000;
parfor idx=1:numel(jValues)
j=jValues(idx);
%Read files
fullFileName=fullfile(thisFolder, baseFileNames{j});
C_data_Array(:,:,idx)= imread(fullFileName,2);
C_data_ArrayMean(:,:,idx)=mean(C_data_Array(:,:,idx),2);
C_data_ArrayStd(:,:,idx)=std(C_data_ArrayMean(:,:,idx));
LEDdetectThresh=(mean(C_data_ArrayMean(:,:,idx),1))+C_data_ArrayStd(:,:,idx);
LEDdurIndx=find(C_data_ArrayMean(:,:,idx)>LEDdetectThresh);
LED_duration_=LEDdurIndx([1,end]); %Updated
LEDinit(idx)=LED_duration(1);%Updated
LEDend(idx)=LED_duration(2);%Updated
Binarizing_Rowsum([LEDdurIndx],idx)=Light; %OUTPUT ERROR: cannot be classified
end
0 Comments
Answers (3)
Walter Roberson
on 17 Aug 2022
BR = Binarizing_Rowsum(:,idx);
BR(LEDdurIdx) = Light;
Binarizing_Rowsum(:, idx) = BR;
5 Comments
Edric Ellis
on 19 Aug 2022
Inside your loop, you are assigning a whole new value to LEDdurIndx. This makes the variable a parfor "temporary", and so any value assigned inside the loop is discarded. I don't see any such problem for Binarizing_Rowsum though...
Bruno Luong
on 19 Aug 2022
Edited: Bruno Luong
on 19 Aug 2022
LEDdetectThresh=(mean(C_data_ArrayMean(:,:,idx),1))+C_data_ArrayStd(:,:,idx);
LEDdurIndx=find(C_data_ArrayMean(:,:,idx)>LEDdetectThresh)
I suspect the code is buggy, the mean can never be bigger than mean + std
LEDdurIndx is therefore empty.
Bruno Luong
on 22 Aug 2022
Edited: Bruno Luong
on 22 Aug 2022
Your code is good to be called "spagetti". If all the image have the same size, you must tell parfor what they are:
jValues=1:2:length(baseFileNames)-1;
n = numel(jValues);
Binarizing_Rowsum=zeros(6100,n);
LEDdurIndx=NaN(1000,n);
Light=3000;
j=jValues(1);
%Read first file to figure out the array sizes
fullFileName=fullfile(thisFolder, baseFileNames{j});
I1= imread(fullFileName,2);
[p,q]=size(I1);
C_data_Array = zeros(p,q,n);
C_data_ArrayMean = zeros(p,1,n);
C_data_ArrayStd = zeros(1,q,n);
LEDdetectThresh = zeros(1,q,n);
Binarizing_Rowsum = zeros(p*q,n);
parfor idx=1:n
j=jValues(idx);
%Read files
fullFileName=fullfile(thisFolder, baseFileNames{j});
C_data_Array(:,:,idx)= imread(fullFileName,2);
C_data_ArrayMean(:,:,idx)=mean(C_data_Array(:,:,idx),2);
C_data_ArrayStd(:,:,idx)=std(C_data_ArrayMean(:,:,idx));
LEDdetectThresh=(mean(C_data_ArrayMean(:,:,idx),1))+C_data_ArrayStd(:,:,idx);
LEDdurIndx = find(C_data_ArrayMean(:,:,idx)>LEDdetectThresh);
LED_duration_=LEDdurIndx([1,end]); %Updated
LEDinit(idx)=LED_duration(1);%Updated
LEDend(idx)=LED_duration(2);%Updated
Temp = zeros(p*q,1);
Temp(LEDdurIndx) = Light;
Binarizing_Rowsum(:,idx)=Temp;
end
0 Comments
See Also
Categories
Find more on Loops and Conditional Statements in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!