VPI do not support conversion between VPI_IMAGE_FORMAT_YUYV and VPI_IMAGE_FORMAT_RGB8

[zhuce_cgf]

In our system, the image format from camera is YUYV. We plan to use VPI to convert into RGB before transmitting it into neural network. But vpiSubmitConvertImageFormat returned an error message “VPI_ERROR_INVALID_IMAGE_FORMAT: CUDA: Conversion not implemented between VPI_IMAGE_FORMAT_YUYV and VPI_IMAGE_FORMAT_RGB8”.

Does it mean VPI do not support conversation between YUV and RGB for now? If it’s true, is there an effeicient way to convert the image from YUV to RGB since CPU is so slow.

Thanks.

[AastaLLL]

Hi,

There are some options in the VIC backend for the YUV format.
Please check the document below to see if it can meet your requirement:

https://docs.nvidia.com/vpi/algo_imageconv.html

Thanks.

[zhuce_cgf]

But I can’t find yuyv in column “input type”.

[AastaLLL]

Hi,

Is MMAPI an option for you?
If yes, you can use it to convert the YUYV format into RGBA8.

Thanks.

[zhuce_cgf]

What is MMAPI? Do you have any documents for it? Thanks.

[AastaLLL]

Hi,

You can find the sample in the below folder:

/usr/src/jetson_multimedia_api/samples

The corresponding document can be found in the below link:

https://docs.nvidia.com/jetson/archives/r35.1/ApiReference/l4t_mm_07_video_convert.html

Thanks.