Custom Machine Learning Model in deepstream-imagedata-multistream.py

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

I have a question about the deepstream python app, the deepstream-imagedata-multistream sample for specific. I want to use the Face Recognition library of adam_geitgey in this sample. And the problem is that I can’t find the appropriate line to directly interact with the frame and then draw bounding box (with CV2) on it. I have a hard time in understanding the code, hope somebody can help me with it.
Thanks.

HARDWARE
**• Hardware Platform (Jetson / GPU): Jetson Nano **
**• DeepStream Version: 6.0 **
• JetPack Version (valid for Jetson only): 4.4
• Python Version: 3.6
**• Issue Type: Question **

3

DeepStream use gst-nvdsosd to draw the bboxes on the images and videos. Please study DeepStream SDK first before you start with DeepStream python samples. Welcome to the DeepStream Documentation — DeepStream 6.2 Release documentation

DeepStream pyds only provide to export the frame as numpy array (just as the sample deepstream_python_apps/deepstream_imagedata-multistream.py at master · NVIDIA-AI-IOT/deepstream_python_apps · GitHub ), please make sure the cv2 code you want to use is working with numpy array before you try to integrate into DeepStream python app.