Failed to accessing segmentation data for PeopleSegNet(mrcnn)

[kotreshrr17]

Hi, I was trying to run PeopleSegNet(mrcnn) model from deepstream_tao_apps in deep stream. I was able to generate model engine file and I was able to get the bounding box coordinates. I am not able to get the mask details. I was trying to access those with NvDsObjectMeta. I am getting the following error. could you please help us how to access the segmentation mask.

TypeError: Unable to convert function return value to a Python type! The signature was
(self: pyds.NvDsObjectMeta) → _NvOSD_MaskParams

• Config file used: deepstream_tao_apps/pgie_peopleSegNetv2_tao_config.txt at master · NVIDIA-AI-IOT/deepstream_tao_apps · GitHub
• Hardware Platform (GPU) : Tesla T4
• DeepStream Version : 6.0
• TensorRT Version : 8.0.1

Thanks

[Morganh]

@kotreshrr17

I was able to generate model engine file and I was able to get the bounding box coordinates. I am not able to get the mask details.

Can you observe the mask when you run the application?

[kotreshrr17]

I am not able to access the metadata of the mask. Deepstream pipeline was running fine without any errors. But when we tried to read meta data using probe with NvDsObjectMeta.
We could access the rectangular parameters of the objects but not the mask parameters.
I was getting errors with respect to NvOSD_MaskParams. like this: TypeError: Unable to convert function return value to a Python type! The signature was
(self: pyds.NvDsObjectMeta) → _NvOSD_MaskParams

[Morganh]

@kotreshrr17

Deepstream pipeline was running fine without any errors. But when we tried to read meta data using probe with NvDsObjectMeta.

Could you create a new topic in deepstream forum? Seems that the model is running well.

[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