Fix #8177

In the previous version only the default stream was/could be used, i.e.
cv::cuda::Stream::Null().

With this change, HOG::compute() will now run in parallel over different
cuda::Streams.

The code has been reordered so that all data allocation is completed
first, then all the kernels are run in parallel over streams.

Fix #8177

See https://github.com/Itseez/opencv/issues/5721

COMMENTS:
* The second __syncthreads() is necessary, I am sure of that.
* The code works without the first __syncthreads() too, but I have however added it for symmetry. Anyway it doesn't affect time performances, I have checked it with some profiling with nvvp

changed hog to work with variable parameters and changed the hog sample to test it with more options

added comments and tests

use abstract interface with hidden implementation

move GPU constants initialization from constructor to compute method