How do I allocate a 4-channel floats container float4 for vx_image

It’s a simple question as the title described.


An image usually ranges from [0,255], it is more efficient to use U8/U16… format.
The F32/2F32 type is designed for depth rather than a standard color image.

Maybe you can check if nvxcu_point4f_t can be used for your use case.
Or could you share more information about which API do you want to use?