A quantitative method for graphical representation of H.264 video content dynamics

This paper presents a method for graphical representation of H.264 video content dynamics, based on the extraction of motion information from the video signal in order to define the content's spatiotemporal dynamics levels. For this purpose, the paper introduces a motion compensation and content residual grid, where each video signal is represented as a point with coordinates that are defined by its motion dynamics and content residual, respectively. The proposed method can be effectively integrated as part of the decision process in video adaptation systems, which perform spatial or temporal (or both) adaptation actions, like MPEG-DASH, in order to optimize the video service delivery.