I know Opencv now supports

// we support 11 types of data:
// video/x-raw, format=BGR -> 8bit, 3 channels
// video/x-raw, format=GRAY8 -> 8bit, 1 channel
// video/x-raw, format=UYVY -> 8bit, 2 channel
// video/x-raw, format=YUY2 -> 8bit, 2 channel
// video/x-raw, format=YVYU -> 8bit, 2 channel
// video/x-raw, format=NV12 -> 8bit, 1 channel (height is 1.5x larger than true height)
// video/x-raw, format=NV21 -> 8bit, 1 channel (height is 1.5x larger than true height)
// video/x-raw, format=YV12 -> 8bit, 1 channel (height is 1.5x larger than true height)
// video/x-raw, format=I420 -> 8bit, 1 channel (height is 1.5x larger than true height)
// video/x-bayer -> 8bit, 1 channel
// image/jpeg -> 8bit, mjpeg: buffer_size x 1 x 1
// bayer data is never decoded, the user is responsible for that
// everything is 8 bit, so we just test the caps for bit depth

but Nvidia HW accelerated converter has this follow formats.

Capabilities:
video/x-raw(memory:NVMM)
format: { (string)I420, (string)NV12, (string)BGRx, (string)RGBA, (string)GRAY8 }
width: [ 1, 2147483647 ]
height: [ 1, 2147483647 ]
framerate: [ 0/1, 2147483647/1 ]
video/x-raw
format: { (string)I420, (string)NV12, (string)BGRx, (string)RGBA, (string)GRAY8 }
width: [ 1, 2147483647 ]
height: [ 1, 2147483647 ]
framerate: [ 0/1, 2147483647/1 ]

Also, most of the Deep Learning models needs RGB or BGR Mat formats...
so maybe it will be a good option to accept RGB.

Also based on video/x-raw(memory:NVMM) It Could be possible to add it?? As CudaMat???