CrowdSync: User generated videos synchronization using crowdsourcing

摘要

User Generated Videos are contents created by heterogeneous users around an event. Each user films the event with his point of view, and according to his limitations. In this scenario, it is impossible to guarantee that all the videos will be stable, focused on a point of the event or other characteristics that turn the automatic video synchronization process possible. Focused on this scenario we propose the use of crowdsourcing techniques in video synchronization (CrowdSync). The crowd is not affected by heterogeneous videos as the automatic processes are, so it is possible to use them to process videos and find the synchronization points. In order to make this process possible, a structure is described that can manage both crowd and video synchronization: the Dynamic Alignment List (DAL). Therefore, we carried out two experiments to verify that the crowd can perform the proposed approach through two experiments: a crowd simulator and a small task based experiment.