We are working on semantic segmentation using DeepStream. We found sample analytics apps in git and default test app with DS for the same, but they are not having complete pipeline implementation like Test5 App.
Is there any sample app available for the same, which has pipeline from segmentation to IoTHub messaging?
please find in NVIDIA-AI-IOT directory. please refer to GitHub - NVIDIA-AI-IOT/deepstream_reference_apps: Samples for TensorRT/Deepstream for Tesla & Jetson
you can refer to deepstream-test4.
@fanzh , thank for responding. The git repo doesn’t have example for semantic segmentation pipeline. We need a model and pipeline which can give us more than one class like person, car, road, trees, signal, etc using single model and generate the text meta to send it to IoTHub.
1 please refer to /opt/nvidia/deepstream/deepstream-6.0/sources/apps/sample_apps/deepstream-segmentation-test, it is segmentation demo, start command is like this : ./deepstream-segmentation-app -t infer dstest_segmentation_config_semantic.txt /opt/nvidia/deepstream/deepstream/samples/streams/sample_720p.jpg
you can combine deepstream-segmentation-app and deepstrem-test5.
2 you can refer to GitHub - NVIDIA-AI-IOT/deepstream-segmentation-analytics: A project demonstration to do the industrial defect segmentation based on loading the image from directory and generate the output ground truth.
@fanzh Thanks for providing the semantic segmentation example link, we will look into and let you know.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.