Optimising task-based video quality

Abstract

Development of techniques for assessing video quality is reviewed. Examples have been provided on the quality of video applications ranging from popular entertainment to new trends such as applications in wide-reaching public systems, not just those used by security forces but also for medical purposes. In particular, two typical usages of task-based video: surveillance video for accurate licence plate recognition, and medical video for credible diagnosis prior to bronchoscopic surgery were introduced by the author. The problem of task-based video quality assessment starting from subjective psychophysiological experiments to objective quality models is discussed. Example test results and models are provided alongside to the descriptions. Finally, a quality optimisation approach, driven by recognition rates is presented.