How to preprocessing(splitting to 2 image) cropped image from primary detector before classification?

Accelerated Computing Intelligent Video Analytics DeepStream SDK
#1

• Jetson Xavier NX
• DeepStream Version:5
• JetPack Version: 4.4

I have a primary detector that detects an object, then I wish to use dsexample to split the cropped image from primary detector, is that possible? The pipeline I wish to do is as follows.

Full image --> nvinfer(primary detector) --> cropped image --> dsexample(split the cropped image to 2) --> nvinfer(classifier)

Because I wish to process multiple vehicle’s multiple lines number plate…

#5

You also need to customized nvinfer code for your pipeline.

If you are using nvinferserver, you can refer Issue with adding custom preprocessing step

#7