No output from SGIE with process-mode=2

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) ->Xavier
• DeepStream Version ->5.0
• TensorRT Version->7.1

I am building a facial recognition pipeline
peoplenet as pgie ->retinaface as sgie1->arcface as sgie2
I need to pass people label output from peoplenet model to retinaface as sgie. How can I achieve this?
The pipeline is working fine with process-mode =1 in sgie config but it’s not working with process-mode=2.config_peoplenet.txt (1.0 KB) dstensor_sgie1_config.txt (1.1 KB)

3

Hi,
Not working in process-mode=2 for sgie, like can not detect? or something else?
if can not detect, can you try below property in sgie configuration with process-mode set to 2
input-object-min-width=0
input-object-min-height=0