ROS vs DeepStream?


I am starting a project and have been doing some research. Admittedly I am not too familiar with some of the subject matter and as such am learning in the process.

My question lies in the difference between using DeepStream.

I am a bit confused on how to interact with it.

Right now my goal is to take some video input and generate some point cloud with it. It seems like this has been done before with ROS and we could use ros-bags or some data set to do some replay back.

So does Deepstream and ROS interact with each other? It seems to me that they are two separate entities that may not mesh well together.

If anyone has some more links of how to use deepstream or anything on deepstream feel free to point me to it. The extent of my knowledge comes from here

Hi micallef, DeepStream is generally geared for multi-stream video analytics whereas ROS is for robotics (like the point cloud processing you mentioned). If you are interested in robotics and point cloud processing, you may want to check out NVIDIA’s Isaac SDK.

You can also find a technical webinar for DeepStream 4.0 and additional DeepStream sample applications at

Thanks dusty! Also your ROS deep learning and hello AI repo’s on github have been such a help in getting me started for a project I am working on. Thank you!