How to interpolate object cropping between object detection model and segmentation model?

Hardware Platform : GPU
DeepStream Version : 5.1
TensorRT Version : 7.2.1
NVIDIA GPU Driver Version : 460.32.03
Issue Type(questions : new requirements

Requirement details :
I am using deepstream-test2 to achieve something like this:

  1. YOLOv5 model plus secondary model U-Net segmentation model;
  2. Add image cropping between pgie and sgie, crop the target of pgie, and then segment the cropped objects separately.

Currently, I have connected pgie and sgie. But I don’t know how to do image segmentation with the target obtained by pgie alone.

If you have add the “U-Net segmentation model” as SGIE, it already supports object cropping inside gst-nvinfer plugin. You don’t need to do anything more.

Please make sure you have set the segmentation model as SGIE correctly.

So how can I get the metadata of multiple objects after segmentation? Is it still in user_meta_data? I can only get the segmentation result of the whole screen from user_meta_data now.

Please refer to Gst-nvinfer — DeepStream 6.1 Release documentation

When the nvinfer is set as SGIE mode correctly, the segmentation user_meta_data is in the object meta data. You need to get the user meta through obj_user_meta_list in NvDsObjectMeta

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.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.