[FEATURE] Does CV-CUDA support the Canny operator?

Does CV-CUDA support the Canny operator? We have a specific requirement for it.