Multimedia API camera_v4l2_cuda image to VPI [cuda to cuda] image mapping

I need to perform following steps:

  1. Capture camera frame using multimedia api in cuda memory.
  2. Use that cuda memory buffer image for furhter processing such as pyramid generation using VPI.

To do so, I would like to decode camera frame using camera_v4l2_cuda (Jetson Linux API Reference: 12_camera_v4l2_cuda (camera capture CUDA processing)) and pass to VPIImage (VPI - Vision Programming Interface: Image)

Can anyone suggest me the fastest way to do so?