Graph Based Multimedia Analysis

Chapter Four: Multi-view exocentric video summarization

Abstract

In this chapter, the problem of multi-view exocentric video summarization is addressed. The goal is to efficiently represent the most significant information from a set of videos captured for a certain period by multiple cameras. The problem is highly challenging because of the huge size of the data, presence of many unimportant frames with low activity, inter-view dependencies, and significant variations in illumination. We propose a graph-based solution to the above problem. We first detect shots from the videos and use different features (e.g. color, texture, and shape) for characterizing shot representative frames. Then, only the informative shot representative frames are retained ...