Image processing before preprocessing in nvinfer

[Chereshnya]

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) Jetson TX2
• DeepStream Version 5.1
• JetPack Version (valid for Jetson only)
• TensorRT Version 7.1.3
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs) question

Hi
What kind of image processing does nvinfer do before preprocessing? I know, that nvinfer can do resize, padding and other operations, which could be configure in config.

My question is: are there any image conversations before preprocessing? Can I dump my original image from nvinfer before it goes to preprocessing?

[yuweiw]

Please refer to our nvinfer flowchart FAQ first.
There are also references on how to obtain image data.

[Chereshnya]

Thanks, I’ll check it out
Also, I wonder, how can I get frame from NVMM buffer when using nvvideoconvert ! video/x-raw,(memory:NVMM). I attached pad probe to nvvideoconvert/src, but don’t know, how to copy frame from device memory to host, using received buffer?

This step will help me to determine, what kind of data is received by nvinfer

[yuweiw]

You can try to dump the image data in our open source. Please refer to this FAQ.

[yingliu]

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks