Measuring and Managing Picture Quality

This chapter addresses the important area of image and video quality assessment. Methods for conducting subjective trials, and for analyzing the results from them, are first described and a discussion of the properties of some publicly available subjective test databases is provided. Objective measures of video quality are then reviewed and compared in terms of their correlations with subjective opinions. These include structural similarity, the video quality metric (VQM), visual signal-to-noise ratio (VSNR), spatial and temporal most apparent distortion (MAD), motion tuned spatio-temporal quality assessment (aka MOVIE), and the perception-based video metric (PVM). Video metrics also play an important role in the picture compression and delivery processes, for example enabling in-loop rate–quality optimization (RQO). This issue is discussed, describing some of the most common techniques that enable us to select the optimum coding parameters for each spatio-temporal region of a video. Finally, rate control methods are described.