Accessing Original RTSP Frames Before Resize in Deepstream Parallel Inference App

[tclark4]

Please provide complete information as applicable to your setup.

Please provide complete information as applicable to your setup.
• Hardware Platform (Jetson / GPU)
GPU
• DeepStream Version
6.2
• JetPack Version (valid for Jetson only)
n/a
• TensorRT Version
8.5.2
• NVIDIA GPU Driver Version (valid for GPU only)
520.61.05
• Issue Type( questions, new requirements, bugs)
Question

Hi all,
I’m working with Deepstream parallel reference app to perform object detection on multiple RTSP streams (currently 12 of varying resolutions). I’d like to access the original, un-resized frames before Deepstream performs any pre-processing.

I understand Deepstream utilizes GStreamer for multimedia handling and as part of the GStreamer pipeline, all frames from RTSP feeds are resized to a uniform size for batch inferencing. I’m currently looking at the deepstream_parallel_infer_app.cpp file to understand how RTSP frames are ingested.

My goal is to access the raw image before they are resized to the same dimension.
Could someone point me in the right direction within Deepstream to achieve this? Is deepstream_parallel_infer_app.cpp the relevant file, or is there some other location I should be looking?

Additionally, any links to Deepstream documentation or references related to my question would be greatly appreciated.

Thanks in advance.

[yuweiw]

You can refer to tee plugin. The pipeline is as follows

                 -> <dump the original image>
rtsp source->tee
                 -> nvstreammux->pgie->...

[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