Exploit the outputs of Yolo


I would like to be able to use the outputs of any Yolo with the deepstream framework. I mean the video with the bounding boxes, their coordinates, the detected classes, etc. Where do I need to look on deepstream to be able to exploit this for another program?


You can refer to callback “osd_sink_pad_buffer_probe” in sample_apps/deepstream-test1

Detection model in test1 can be replaced by yolo model.