Implementing Real-time Vision AI Apps Using NVIDIA DeepStream SDK
Discover how to implement real-time vision AI applications using NVIDIA DeepStream SDK in this 45-minute conference talk. Explore how companies worldwide are modernizing infrastructure with IoT sensors for rich insights, from patient fall detection to traffic flow analysis and customer behavior understanding. Dive into the latest capabilities of DeepStream SDK, witness demonstrations of reference apps built with DeepStream SDK, Transfer Learning Toolkit, and Vision AI pre-trained models. Learn about challenges in video analytics, speeding up intelligent video analytics training and inference, and solving industry-specific problems. Explore the DeepStream SDK overview, graph architecture, key features, and Python bindings. Gain insights into IoT applications, secure communication, over-the-air model updates, and native integration with Triton Inference Server. Discover how to deploy open-source models, unlock greater performance, and implement end-to-end AI using Transfer Learning Toolkit and DeepStream. Examine reference apps for people analytics, social distancing, and face mask detection. Access resources to start building high-performance vision AI applications today. Read more