Hello, recently I started to use OpenCV gpu-enabled class cv::cuda::GpuMat
However, I had to face the reality that no many linear algebra methods have their equivalent for GPU. Methods like
// X,D,V, W, U, Vt are cv::Mat type
cv::eigen(X,D,V) or
cv::SVD::compute(X, W, U, Vt);
This is particularly a blocker for me cause I don’t think those methods are easy to implement for GPU
I was hoping to find in the cuda methods something like
// X,D,V, W, U, Vt are cv::cuda::GpuMat type
cv::cuda::eigen(X,D,V) or
cv::cuda::SVD::compute(X, W, U, Vt);
I would appreciate any pointers to find the equivalent of these methods for cv::cuda::GpuMat type, thank you