NvTracker suppress detection objects

[lqdisme]

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU): dGPU
• DeepStream Version: 6.3
• Issue Type( questions, new requirements, bugs): bugs

Hi,

I have an issue related to NvTracker when running a DeepStream Python application. My pipeline is simple, consisting of Jpegparse → Streammux → NvInfer (YOLO for segmentation) → NvTracker → FakeSink. However, many detections (which I can get from the probe function at NvInfer) are removed after NvTracker. Could you tell me what the problem is and how I can resolve it? The pipeline works well with video instead of JPEG images.

Thank you.

[kesong]

Which type of tracker are you using? Can you share the configure file of the nvtracker? Here is the doc for nvtracker: Gst-nvtracker — DeepStream documentation 6.4 documentation
Can you help a try to set probationAge to 0 in nvtracker configure file?

[lqdisme]

I am using NVDCF. I also tried to set probationAge to 0 but it doesn’t work.
tracker_config.zip (2.1 KB)

[kesong]

Can you share the command line or application to reproduce the issue? If you only have one jpeg image, why you need nvtracker?

[lqdisme]

My pipeline is used to ensure that the model and tracker work well. I tested it with a JPEG image and was surprised that the detections were suppressed after nvtracker. Since nvtracker is not open-source, I can’t find the reason for this.

[kesong]

Yes, you can refer here on why nvtracker suppress several video frames: Gst-nvtracker — DeepStream documentation 6.4 documentation

[yingliu]

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