OpenCV parallel computing descriptors?

Question

I am developing an object-recognition system. I found that the critical part of my algo is the

extractor.compute();

(After having detector.detect() keypoints)

Is there any way to let compute the feature vector with more core? I can use up to 8 core.

If you build with TBB opencv will use multicores where possible — Martin Beckett, Jun 20 '12 at 23:43
Do you have intel TBB installed? Did you check the USE_TBB box? — Martin Beckett, Jun 20 '12 at 23:46

score 1 · Answer 1 · answered Jun 20 '12 at 23:45

1

Opencv already implements multithread framework for this. Check that you compiled opencv with threading option 'ON'. You should go for an opencv documentation reading, gpu::SURF_GPU may interest you.

You can run cmake again to see the compilation options you used.

answered Jun 20 '12 at 23:45

Eric

2,301
3
23
30

1

note gpu:: uses CUDA (ie an Nvidia graphcis card) it doesn't do multi threading on CPU cores – Martin Beckett Jun 20 '12 at 23:48
@martin Obviously. I thought it was acquired. Thx for exactness. – Eric Jun 20 '12 at 23:51
As Martin said, look for WITH_TBB option. – Eric Jun 20 '12 at 23:52

OpenCV parallel computing descriptors?

1 Answers1