Have you ever developed an AI-based application that deals with videos and images and just wondered, wouldn’t it be great if there was a toolkit to handle all the I/O streaming bottlenecks of the application? If so, this blog post is for you.
As developers, we always prefer to focus on building core deep learning networks instead of designing end-to-end applications from scratch. Luckily, NVIDIA has a great solution for us: DeepStream SDK.
What is DeepStream SDK?
DeepStream Software Development Kit by NVIDIA is a set of inference plugins and tools for GStreamer, an open-source multimedia framework. DeepStream SDK allows developers to build new AI applications and draw insights from videos in real-time, without having to put too much effort into designing I/O streaming pipelines. It runs on several NVIDIA platforms in the cloud and at the edge and comes with several reference applications that can help you get started with DeepStream quickly.
A DeepStream application is a set of modular plugins connected to form a processing pipeline. These plugins can bring a wide variety of functionalities to your application by leveraging the GPU capabilities of your NVIDIA device. Refer to this link for the DeepStream plugin manual.
DeepStream SDK Support for Smart Distancing App
At Neuralet, we are working on an open-source application to measure social distancing guidelines in real-time using edge devices. This application connects to a video streaming input and identifies cases of social distancing violations. The DeepStream SDK is allowing us to speed up the performance of Smart Social Distancing applications with high accuracy results. You can see the performance results of DeepStream SDK on two NVIDIA platforms in the table below.
Do you need further assistance with Deepstream? Email us at [email protected]
Have a question? Send us a message and we will respond as soon as possible.