• Hardware Platform (Jetson / GPU) GPU
• DeepStream Version 6.2
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only) 525
• Issue Type( questions, new requirements, bugs) questions
Hi, I’m trying to do output parsing in Python for object detection and pose estimation simultaneously and then do Gst-nvtracker. My approach is to create pyds.NvDsObjectMeta given rect_params like this and 17 keypoints for each object. However, when I refer pyds.NvDsObjectMeta, I haven’t any variable for such additional variables for such key points.
The reason is that I want to check if key points between frames of the same ID after tracking move too much or not. Therefore, I need a way to store keypoints in the object. Do you have any suggestions?
you can use nvds_add_user_meta_to_obj to add user meta to object meta, please refer to function attach_tensor_output_meta in deepstream SDK, the path is opt\nvidia\deepstream\deepstream\sources\gst-plugins\gst-nvinfer\gstnvinfer_meta_utils.cpp
Thank @fanzh, how to create user_meta given a numpy array so that I can run nvds_add_user_meta_to_obj?
There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks
here is a numpy sample Appsrc with numpy input in Python
first you might see how attach_tensor_output_meta is used in C function attach_tensor_output_meta. then you might refer to pyds.nvds_add_display_meta_to_frame in deepstream_python_apps/deepstream_test_1.py at master · NVIDIA-AI-IOT/deepstream_python_apps · GitHub, nvds_add_user_meta_to_obj is similar to nvds_add_display_meta_to_frame .
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.