Please provide complete information as applicable to your setup.
**• Hardware Platform (Jetson / GPU):ORIN
**• DeepStream Version:6.3
**• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
**• Issue Type( questions, new requirements, bugs):New requirements
**• I am currently using YOLO and SORT for object detection and tracking, but my purpose for using SORT is not to continuously track the same ID of an object. My goal is to use YOLO for frame-by-frame detection. For example, in the first frame, I perform detection and obtain the bounding boxes for the objects. Then, in the second frame, I do not perform detection, but instead use SORT to base the bounding boxes on the predictions obtained from the first frame. I then visualize the detection results from the first frame and the predicted results from the second frame using DeepStream OSD.
**•
Do you mean you want to set PGIE inference interval? NvDCF tracker can support the feature. Please here: Gst-nvtracker — DeepStream documentation
Yes, I have followed this tutorial to use NvDCF tracker technology to do it, set the interval to 2s, but during the test of MP4, the bounding box of the target object of detection and tracking has been flickering discontinuously, how to solve this problem? In the first frame, I perform the detection and can get the bounding box of the object, can I use SORT to build the bounding box based on the predictions obtained from the first frame in the second frame, instead of performing the detection, and finally use DeepStream OSD to visualize the detection results of the first frame and the predictions of the second frame?
Do I need to create a custom nvdstracker plugin now and implement my requirements in that tracker? How can I create a custom nvdstracker plugin, and could you provide me with an example and relevant tutorial?
For custom nvtracker implementation, please refer: Gst-nvtracker — DeepStream documentation
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