Understanding video retrieval

With the improving capabilities of current hardware systems, there are ever growing possibilities to store and manipulate videos in a digital format, leading to a growing number of video archives. People build their own digital libraries from materials created through digital cameras and camcorders, and use systems such as YouTube to place this material on the web. Unfortunately, this data creation prowess is not matched by any comparable tools to organise and retrieve video information. There is a need to create new retrieval engines to assist the users in searching and finding video scenes they would like to see from many different video files. Unlike text retrieval systems, retrieval on digital video datasets is facing a serious problem: The Semantic Gap. This is the difference between low-level data representation of videos and the higher level concepts user associates with video. This book introduces several approaches to bridge this semantic gap, explains different evaluation strategies and presents state-of-the-art video retrieval tools. The target audience is everyone who is interested in getting to know the research approaches that led to the popular video retrieval tools.