Multiple ROIs processing per frame

Accelerated Computing Intelligent Video Analytics DeepStream SDK
#1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
T4
• DeepStream Version
5.0
• JetPack Version (valid for Jetson only)
NA
• TensorRT Version
7.0.0.11
• NVIDIA GPU Driver Version (valid for GPU only)
440.64.00
• Issue Type( questions, new requirements, bugs)

Is it possible to use nvvideoconvert to crop a single frame at two different ROIs and put them into a single batch of nvstreammux? Generally speaking, if a frame has two ROIs, and each frame needs to generate two separate labels (after classification), what is the canonical way of building such a DeepStream pipeline?

#3

Hey customer, We will check it internally

#4

This is not supported and user will need to customize this. Suggested approach - User can add a T element after decoder and feed same frame to multiple nvvideoconvert elements which will operate on given ROI and then the buffers can be fed to nvstreammux component. Support for multiple ROI is added to our roadmap and will be part of one of the upcoming releases

#6

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.