How do you access the original frame for processing?

[andrew.qian]

Please provide complete information as applicable to your setup.

Deepstream 7.1 triton multi arch Jetson

When we are given the arguments passed into the CustomTensorPreparation function, how do we access the original surface/frame at the top of the metadata hierarchy?

I have not found documentation on how to get the original frame passed into the first pgie, even though that would be useful for processing later down the line.

Thanks.

[junshengy]

@andrew.qian

When we are given the arguments passed into the CustomTensorPreparation function, how do we access the original surface/frame at the top of the metadata hierarchy?

You can refer to this link, it show how to get nvbufsurface/meta from NvDsPreProcessBatch.

apps/tao_others/deepstream-pose-classification/nvdspreprocess_lib/nvdspreprocess_lib.cpp

@andrew.qian

I have not found documentation on how to get the original frame passed into the first pgie, even though that would be useful for processing later down the line.

What does this mean? The data sent to nvinfer is actually a batch formed by nvstreammux, not a frame. gst_nvinfer_submit_input_buffer in /opt/nvidia/deepstream/deepstream/sources/gst-plugins/gst-nvinfer/gstnvinfer.cpp shows how to process batches.

[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.