Ben Weiss, Computer Vision Developer at CEVA, presents the "Video Stabilization Using Computer Vision: Techniques for Embedded Devices" tutorial at the May 2016 Embedded Vision Summit.
Today, video streams are increasingly captured by small, moving devices, including action cams, smartphones and drones. These devices enable users to capture video conveniently in a wide range of situations. But, they also pose significant challenges with respect to video quality, including being prone to vibration and shaking. Fortunately, undesired motion can be removed by processing a video stream as it is captured. In this presentation, Weiss surveys video stabilization techniques suitable for embedded platforms, assessing their strengths and weaknesses. He focuses on computer-vision-based video stabilization approaches and explores trade-offs between factors, such as video quality and computation requirements.