How to extract people face features using peoplenet model in deepstream 5.0?

Please provide complete information as applicable to your setup.

**• Hardware Platform (Jetson / GPU)**RTX 1060
• DeepStream Version5.0
• JetPack Version (valid for Jetson only)
• TensorRT Version7.0.0.11
**• NVIDIA GPU Driver Version (valid for GPU only)**450

I am new to Deepstream SDK. CPP preferred and what plugins and dataset I need to use to extract people features using peoplenet model in deepstream SDK.

Are you asking how to retrain the peoplenet model inside deepstream sdk package.
You should create a new topic in TLT forum for better support if yes.

Hi bcao!
I am just learning how to extract face features from the pruned peoplenet model in the face class. It is not about retraining. Do I need to use some plugins or develop an app to make that happen? Thank you!

Got, there is a FaceDetectIR model, you can use that model to detect the face.
Maybe you need to detector in the pipeline pgie(peaplenet) ->sgie(face detector), you can refer https://github.com/NVIDIA-AI-IOT/deepstream_reference_apps/tree/master/back-to-back-detectors for the pipeline

Appreciate it!

Why should he use two detection models ? He can just get the features from peoplenet and slap them into a KNN… Or at the very least use the ROI generated by peoplenet to locate the face and run a classifier on it.

@keionlu you are wasting resources running two detection models.