Automatic video editing of geotagged mobile videos: Master's Thesis

Demand for efficient ways to represent vast amount of video data has grown rapidly in recent years. This is due the ever growing amount of videos recorded and shared by people. Enablers for this evolution can be found from increased capacity of storages, improved processing power of mobile devices and faster network connections. In this thesis an automatic video editing system for geotagged mobile videos is designed and implemented. The system creates a video summary from a set of unedited video clips. Geotags are used to connect videos from the same context together. The system finally creates the video summary where videos from the same context are represented as video scenes. This work represents some principles of video editing and how they can be applied in video abstraction. Also previous works done in the research field of video abstraction are presented. This work is significant for the research field as abstraction of mobile videos has not been extensively researched before. Secondly the use of geotags in video abstraction is introduced in this work. Evaluations of the system prove the concept useful as it clearly improves coherence and enjoyability of the video summaries