Professor George Irfan Essa, gave a lecture on December 11, 2014 titled “Video Cameras Everywhere: Data-Driven Methods for Video Analysis and Enhancement ".
Professor Irfan Essa started with describing the pervasiveness of image and video content, and how such content is growing with the ubiquity of cameras. He used this to motivate the need for better tools for analysis and enhancement of video content. He started with some of the earlier work on temporal modeling of video, then led up to some of the current work and described two main projects: (1) The approach for a video stabilizer, currently implemented and running on YouTube, and its extensions; (2) A robust and scalable method for video segmentation.
Professor Irfan Essa described, in some details, the Video stabilization method, which generates stabilized videos and is in wide use. The method allows for video stabilization beyond the conventional filtering that only suppresses high frequency jitter. This method also supports removal of rolling shutter distortions common in modern CMOS cameras that capture the frame one scan-line at a time resulting in non-rigid image distortions such as shear and wobble. The method does not rely on a-priori knowledge and works on video from any camera or on legacy footage. He showcased examples of this approach and also discussed how this method is launched and running on YouTube, with Millions of users.
Professor Irfan Essa also described an efficient and scalable technique for spatio-temporal segmentation of long video sequences using a hierarchical graph-based algorithm. This hierarchical approach generates high quality segmentations and he demonstrated the use of this segmentation as users interact with the video, enabling efficient annotation of objects within the video. He also showed some recent work on how this segmentation and annotation could be used to do dynamic scene understanding.