How to get Position Information of recognition Object

hello, I run the sample deepstream-test1,It recognizes people and draws a frame.I read the code,But I don’t know where to draw the frame.I want to get the location information of the object and not give it a picture frame. Is there any way?

nvdsosd plugin draws the frame.
In deepstream_test1_app.c -> osd_sink_pad_buffer_probe() “NvDsObjectMeta” has the location information.