Performance DLStreamer (Gstreamer) vs Pure Openvino (without Gstreamer) for multiple real-time strea

[noleakdefence]

Hi, is there any comparison between using DLstreamer vs Pure Openvino for multiple real-time streams? Do you have any thoughts on this?

I imagine DLstreamer would be better due to the fact of the Gstreamer framework, even though the learning curve is steeper.

We are currently using DLStreamer but with the recent developments of Openvino and oneAPI we are unsure if it is better to continue or migrate.

Thanks! Please let me know if this is not the right place to post this.

[Aznie_Intel]

Hi Noleakdefence,

Thanks for reaching out.

It depends on the objective of your application. Some application works better with DL Streamer than using only OpenVINO and vice versa. DL Streamer also uses OpenVINO™ Runtime inference back-end optimized for Intel hardware platforms. OpenVINO Toolkit includes DL Streamer which provides GStreamer Video Analytics plugin with elements for Deep learning inference using OpenVINO inference engine on Intel CPU, GPU and VPU. For more details on DL Streamer, please refer to the open-source repository here. You may check the samples/demos for both and compare which one is related to your objective.

On another note, there is also Real Time Stream Analysis Demo that demonstrates how to analyze RTSP (Real Time Streaming Protocol) stream using OpenVINO Model Server for inference. This demo is only available for 1 model, but you can learn how to use the RSTP using the OpenVINO Model server from this.

Hope this helps.

Regards,

Aznie

[Aznie_Intel]

Hi Noleakdefence,

This thread will no longer be monitored since we have provided a information. If you need any additional information from Intel, please submit a new question. 

Regards,

Aznie