Palvannan's profile - activity

Hi,

I am using rotationMatrix2D, warpAffine, matchTemplate and minMax opencv API. To convert these functions to use GPU i was able to find gpu equivalent methods for all of them except the "getRotationMatrix2D", is there a GPU equivalent method for "getRotationMatrix2D" function in opencv? If not it is because it doesn't consume significant computation time?

Appreciate the help in advance.

Regards,

Pal

Thanks Steven, i was able to configure with CUDA5.5 after updating the CMake version 3.0( the OpenCVCudaDetect.cmake mentioned the requirement for cmake 2.8.3 minimum) and I was having CMake 2.4.6 before. It was also mentioned in your wiki when i verified later. Even when my system was configured with CUDA 6.0, It automatically detects 5.5. For now i am fine, Will check later on this and update.

Hi Team,

I am using opencv 2.4.8 and i would like to use the gpu apis. I have CUDA5.5/6.0 toolkit in my machine (Red Hat 4.4.7), but i am unable to configure opencv with CUDA. I have used the following commands to configure : cmake -D CMAKE_BUILD_TYPE=RELEASE -D MAKE_INSTALL_PREFIX=/usr/local/home/<customdir> -D WITH_CUDA=ON -D WITH_CUDA=YES -D CUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda ..

Currently the machine points to CUDA6.0(/usr/local/cuda -> CUDA6.0).

Could you please let me know if i am passing required arguments and is opencv 2.4.8 compatible with CUDA 6.0? Thanks for your assistance.

Regards,

Pal