Learn how deep learning algorithms can be used for extracting semantic metadata from video and find out how to take advantage of this metadata for automatic video analysis. We will describe our experience employing deep learning algorithms for face detection, recognition, and anonymization (FaceNet, GANs) and general object recognition (YoloV3). We'll also address practical issues like building and deploying deep learning frameworks on multiple platforms and via GPU-Accelerated Docker containers, both locally and in the cloud. In addition, we'll provide information about several applications for these algorithms such as anonymization of training data, analysis of user-generated content for personalized media services, and automatic camera path generation for 360° video.
Discover how post-production tasks can be accelerated by taking advantage of GPU-based algorithms. In this talk we present computer vision algorithms for corner detection, feature point tracking, image warping and image inpainting, and their efficient implementation on GPUs using CUDA. We also show how to use these algorithms to do real-time stabilization and temporal re-sampling (re-timing) of high definition video sequences, both common tasks in post-production. Benchmarking of the GPU implementations against optimized CPU algorithms demonstrates a speedup of approximately an order of magnitude.