Deepstream5.0 SDK Application

Hi,
Ask the following questions:

  1. Using deepstream for inference, we need to support the following recording strategies: start recording after detecting the target, the recording time is 30 seconds; start recording after detecting the target, and stop recording after the target disappears. Can it be achieved? Can you provide a demo?

  2. Use deepstream for inference. After detecting the target, save the image data in jpg format and implement it in C language. Can you provide a demo?

  3. Use deepstream for inference. Can the inferred data stream be pushed to the html page for playback? Can you provide a solution?

  4. Use deepstream to push rtsp stream. Can the process of pushing stream control the start and stop? Can you provide a demo?

  5. How to calculate the computing power required by the algorithm model? The relationship between the size of the algorithm model and GPU, CPU, memory?

Hi @1018724818
you could start with the sample code we provided now with the doc - https://docs.nvidia.com/metropolis/deepstream/dev-guide/index.html .

Thanks!