HOGDescriptor::detectMultiScale with TBB on ARM

asked 2013-05-22 04:41:58 -0600

gvi
81 ●2 ●4

Hi all! I have implemented in an ARM a computer vision application that uses OpenCV's HOGDescriptor::detectMultiScale. I have found that the execution of HOGDescriptor::detectMultiScale function is accelerated a lot when I compile OpenCV with TBB option:

Running time without TBB: 317 ms
Running time with TBB: 122.5 ms

Is this the correct behaviour?? I found it very strange, as the ARM I am using has only two cores. How does the TBB acceleration work?? why is so efficient?

I am using a ARM® dual-core Cortex™-A9 MPCore™ / 800 MHz, OpenCV 2.4.3, and TBB 4.1.

Thanks in advance!

edit retag flag offensive close merge delete

add a comment

0

answered 2013-07-10 09:10:27 -0600

nessence
1 ●1 ●2

That is correct behavior. Most of the OpenCV codebase uses Intel TBB and parallel_for() to stripe matrix operations to make use of multiple cores. Most of the algorithms for histogram of gradients can be parallelized which is why it runs so much faster.