Wisdom of the Crowd in Egocentric Video Curation

摘要: Videos recorded by wearable egocentric cameras can suffer from quality degradations that cannot always be fixed current methods. When several video are viewing the same scene, each having highly variable quality, it is possible to combine them into a single high-quality video. Current techniques select for point in time highest stream, but may not relevant. E.g. best come person happen look sideways main attraction. We propose curation of stream multiple videos requiring selected will also view most interesting region scene. Importance determined "wisdom crowd", i.e. number looking at region. The resulting more and higher than any individual streams possibly obtain. Several examples presented demonstrating effectiveness this technique.