Multi primary detecter inference for different cameras, is it possible?

Accelerated Computing Intelligent Video Analytics DeepStream SDK
,
1

• Hardware Platform: Jetson NX
• DeepStream Version: 5.0
• JetPack Version: 4.4
• TensorRT Version: 7.1.0

Hi Sir/Madam:

This is a general question:

From the deepstream pipeline, normally video stream will go through a primary detecter (gst-nvinfer), then to tracker (optional).
So, what if I installed several cameras to one Deepstream pipeline, but for different cameras, I want to use different models.
Is it possible in this way?

Best
Yichao

3

No. With nvstreammux, the multiple streams will be bundled into the batch, either PGIE or SGIE should work on batch but not single stream. If different streams need different models, please put them in different pipelines.

4

An extension of the above question, what if for each stream of a few different models will be used.

E.g. use pretrained vehicle detect, human detect on the same video stream , how will it be given that is only one pgie? Thanks

5

Hi Chunleonglee:

I will open a new topic with your question, since I am also curious about it.
Here’s the link: Multi primary detecter inference for one camera, is it possible?

Best
Yichao