When the four-channel flow data is used to infer the model, Gst-nvmultistreamtiler is used before the nvinfer plug-in to concatenate into a 2×2 big picture. Does nvinfer infer infer directly infer with this big picture?
It depends on your model.
1 Like
The input form of my network is: 4 inputs are stitched together into a 2×2 large graph as input. According to what you said, my understanding is that I can directly use the output of the Gst-nvmultistreamtiler plugin. Can I interpret it that way? Looking forward to your reply，thank you
Yes. Basically, nvinfer can also process the picture size to the model input. If you want process by yourself, please refer to our preprocess plugin.
https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_plugin_gst-nvdspreprocess.html
I want to ask a question: deepstream reads local streams of jpeg images. How to get the name of the image in the probe of the plug-in later.
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
You can get it with the g_object_get (G_OBJECT (src), "location", &location, NULL); API.
Please open a new topic if you have new problem for better reference for others. Thanks
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.