Soft Computing Methods for the Design, Deployment, and Reliability of Networks and Network Applications
Talk: Automatic measuring of perceptual quality, a major QoE component of video- and audio-based services
Presenter: G. Rubino, INRIA, France
The ultimate goal when designing an application or a service running on the Internet is the satisfaction of the final user. For a multimedia application or service concerning the transportation of video, voice or audio content, this mainly translates into obtaining a high enough Perceptual Quality (PQ), the subjective view of the user about how much satisfactory is that content when he/she accesses it. Measuring PQ is current business in today’s industry: the definitive way of doing it is by means of a panel of human users that react to the content in a controlled experiment called a subjective test. Many norms exist for these tests, on the type of media used and on the objectives of the assessment.
We are interested in automatically measuring PQ, and in obtaining values close to those coming from subjective testing. This talk describes our solution to this specification, the PSQA project, its current status and the ongoing work. PSQA stands for Pseudo-Subjective Quality Assessment. It is a technology developed at INRIA capable of providing a numerical (that is, quantitative) and accurate (meaning highly correlated with the results of subjective tests) estimation of the quality of a video, voice or audio flow as perceived by the end user, automatically and in real time if necessary. The flow can be either one-way (for instance, video treaming) or interactive (for instance, in some IP telephony service). PSQA has been already used by industry in several collaborative projects in different contexts (mobile TV, P2P-based video streaming, VoIP services, etc.).
The talk will describe the main components of our technology and its main potential applications, some past and present research activities around it, and some of the perspectives concerning our future efforts. In particular, we will describe the case of HTTP-based video streaming, because of its importance in today applications. We will also underline associated current open problems, both concerning the theoretical foundations of PSQA and important practical aspects of it. PSQA is based on specific statistical learning techniques that will be briefly described in the talk.
Bio: Gerardo Rubino is a senior researcher at INRIA (the French National Institute for Research in Computer Science and Control) where he is the leader of the DIONYSOS group, working on the analysis and design of networking technologies. He is Board Member of the Media & Networks Cluster, Brittany, France, and INRIA’s representative at the SISCom Brittany Research Cluster. Among past responsibilities, he has been Scientific Delegate for the Rennes unit of INRIA for 5 years, responsible of research in networking at the Telecom Bretagne engineering school for 5 years, Associate Editor of the Operations Research international journal “Naval Research Logistics” for 9 years, former member of the Steering Board of the European Network of Excellence EuroFGI and responsible of the relationships between the network and European industry. He has also been the head of the International Partnership Office at INRIA Rennes for 5 years. He is a member of the IFIP WG 7.3. He belongs to the Steering Committee of QEST (www.qest.org).
He is interested in quantitative analysis of complex systems using probabilistic models. He presently works on performance and dependability analysis, and on perceptual quality assessment of audio and video applications and services. In particular, he is the author of the PSQA (Pseudo-Subjective Quality Assessment) technology for automatic perceptual quality real-time evaluation. He also works on rare event analysis, he is a member of the Steering Committee of RESIM, the only workshop dedicated to the topic, and co-author of the book entitled “Rare Event Simulation Using Monte Carlo Methods” (Wiley, 2009). He is the author of more than 200 scientific works in applied mathematics and computer science.