How to merge multiple GST batched buffer into one buffer

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Please provide complete information as applicable to your setup.

**• Hardware Platform (Jetson / GPU): GPU 1080
**• DeepStream Version: 5.0.1
**• Issue Type( questions, new requirements, bugs): questions

Hi, Suppose I want to create a pipeline with 4 input streams, 2 streams to a people net and 2 to vehicle net.
After predicting using nvinfer module, I draw bounding boxes and thus have 2 batched GST buffers, how can I merge these two batched buffers into one and tilted then render them onto screen?
Thanks for your help.

Edit: I’ve read the deepstream muxer and it only supports RGBA/NV12 input buffers

3

No. Deepstream can not support such case now. You may consider to combine the two models together so that you can infer all streams with one batch.

4

Do you have another idea to tackle this problem. If I combine all streams together, it may slow down the inference process.

5

For current deepstream, no other suggestion.

6

What about using nvstreamdemux on both streams and nvstreammux again ?

7

It is not supported now. Recombining nvstreamdemux output with nvstreammux fails - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums

9

Oh thanks alot, I’ve thought about implementing that but if it not working I try to find another way.