Where should frame data be placed for accelerated functions if we're not using CSI?

We’re considering the following camera link frame grabber for use in combination with Jetson cards and want to ensure that we can use the accelerated functions the jetson architecture offers.
https://www.epixinc.com/products/pixci_minih2.htm

Epix says PIXCI frame grabber either streams video data directly into system memory or, optionally, into GPU memory. Is system memory/GPU memory where video data needs to be for all their acceleration functions? Are we missing out on capabilities of the Jetson Architecture by using this approach vs the CSI?

Thanks,
Josh

Suppose yes,
You can reference to v4l2cuda sample code of MMAPI.

Thanks

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.