How to use Floating Point Support in opencv

asked 2015-07-22 11:04:09 -0600

franken
25 ●1 ●5 ●10

updated 2015-07-25 05:42:48 -0600

Hi all,

i'm using raspberry pi 2 with raspbian , i'm working on face and eyes detection using opencv and c++ , i want to improve the performance and reduce time detection in the program and i was wondering if i can use the floating point unit FPU ?? and how to use it in OPENCV .

Any suggestions will be appreciated thanks. :)

NB : the code :https://github.com/Itseez/opencv/tree/master/samples/cpp/tutorial_code/objectDetection

edit retag flag offensive close merge delete

add a comment

1 answer

Sort by » oldest newest most voted

answered 2015-07-24 02:58:50 -0600

mshabunin
1927 ●10 ●31

Try to build OpenCV with NEON support:

cmake  -DCMAKE_TOOLCHAIN_FILE=../opencv/platforms/linux/arm-gnueabi.toolchain.cmake -DENABLE_VFPV3=YES -DENABLE_NEON=YES  ../opencv

edit flag offensive delete link

Comments

i have try to add -march=armv7-a -mfpu=vfp -mfloat-abi=hard in compilation but hasn't improve anything , i will try to re-build opencv with NEON support and let you know thanks :)