Estimation of QoE of video traffic using a fuzzy expert system

Abstract

Quality of experience (QoE) in multimedia traffic has been the focus of extensive research in the last decade. The estimation of the QoE provides valuable input in order to measure the user satisfaction of a particular service. QoE estimation is challenging as it tries to measure a subjective metric where the user experience depends on a number of factors that cannot simply be measured. In this work, we present a methodology and a system based on fuzzy expert system to estimate the impact of network conditions (QoS) on the QoE of video traffic. At first, we conducted subjective tests to correlate network QoS metrics with participants' perceived QoE of video traffic. Second, we propose a No Reference method based on fuzzy expert system to estimate the network impact on the video QoE. The membership functions of the proposed fuzzy system are derived from normalized probability distributions correlating the QoS metrics with QoE. We propose a simple methodology to build the fuzzy inference rules. We evaluated our system in two different sets of experiments. The estimated video quality showed high correlation with the subjective QoE obtained from the participants in a controlled test. We integrated our system as part of a monitoring tool in an industrial IPTV test bed and compared its output with standard Video Quality Monitoring (VQM). The evaluation results show that the proposed video quality estimation method based on fuzzy expert system can effectively measure the network impact on the QoE.

A QoE-oriented scheduling scheme for HTTP streaming service in LTE system

“…Authors in [10] pick the QoE-content type, sender bit rate, block error rate and mean burst length as four key parameters that can impact video quality, followed by constructing a model between MOS (Mean Opinion Score) and these four parameters. In [11], the authors proposed a video quality estimation method based on a fuzzy expert system to measure the impact of network condition (packet loss rate, packet loss burstiness and jitter) on the QoE of video services. The QoE optimization problem can be translated to a multiple objectives QoS optimization problem, thus traditional QoS optimization techniques can be applied to QoE optimization scenarios.…”

Prediction of perceptual quality for mobile video using fuzzy inference systems

“…Again, either machine learning was used to assess QoS/QoE correlation [9], or a fuzzy expert system was used for QoE estimation [14], [15]. The majority employed the mean opinion score (MOS) as a quality measure.…”

A hybrid prediction model for video quality by QoS/QoE mapping in wireless streaming

“…While the work in [14] proposed a nonmachine learning parametric model that combined only application layer parameters for different video contents. Other research studies examined QoS parameters from the network layer solely, such as packet loss, delay, and jitter [12], [15], [16]. Mushtaq et al [12] evaluated six different machine learning techniques to assess QoS/QoE correlation of video streaming over cloud networks.…”

“…Mushtaq et al [12] evaluated six different machine learning techniques to assess QoS/QoE correlation of video streaming over cloud networks. A fuzzy expert system was used for QoE estimation by Pokhrel et al [15] and Alreshoodi et al [16]. All three employed the mean opinion score (MOS) as a quality measure.…”