Parallelization of a Loop with 'parfor - Unsuccessfully
1 view (last 30 days)
Show older comments
Hello, I created an algorithm which work on Image Patches. I go through matrix of an image and apply a function of a neighborhood predefined.
For some reason MATLAB won't allow me using the 'parfor' command.
Here is the code:
mInputFrames = padarray(mInputFrames, [searchWinEffRadius searchWinEffRadius], 'replicate');
parfor iRowIdx = 1:videoFramesNumRows
for jColIdx = 1:videoFramesNumCols
paddedImageRowIdx = iRowIdx + searchWinEffRadius;
paddedImageColIdx = jColIdx + searchWinEffRadius;
vSearchWindowRowIdx = [(paddedImageRowIdx - searchWinEffRadius):(paddedImageRowIdx + searchWinEffRadius)];
vSearchWindowColIdx = [(paddedImageColIdx - searchWinEffRadius):(paddedImageColIdx + searchWinEffRadius)];
for kFrameIdx = 1:videoFramesNumFrames
vPatches = ((kFrameIdx - 1) * singleFrameSearchWindowNumPixels) + ...
[1:singleFrameSearchWindowNumPixels];
mCurrSearchWindow = mInputFrames(vSearchWindowRowIdx, vSearchWindowColIdx, kFrameIdx);
mCurrSearchWindowPatches(:, vPatches) = Im2ColSliding(mCurrSearchWindow, [localWinSize, localWinSize]);
end
mOutputFrame(iRowIdx, jColIdx) = PatchFilter(mCurrSearchWindowPatches, refPixelRowIdx, refPixelColIdx, vGaussianKernel, weightsStd);
end
end
The error I'm getting: "The PARFOR loop cannot run due to the way variable 'mCurrSearchWindowPatches' is used.
The work on the loops is per pixel which is completely indepndent (In the result) from its neighborhood. What can I do to allow the parallelization to kick in?
Thank You.
0 Comments
Accepted Answer
More Answers (0)
See Also
Categories
Find more on Loops and Conditional Statements in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!