This session introduces building a scalable real-time video analytics system using open-source technologies. The talk includes an example that shows object detection on a Kubernetes-based platform that records, ingests, stores, inferences, aggregates, visualizes, and plays the video data. We will finish the session by demonstrating object detection with GPU-accelerated AI inference using NVIDIA DeepStream plugins for GStreamer running with Pravega, a CNCF project.
