Perceived quality prediction models for multimedia services vary greatly depending on the type of the data and on the amount of information related to the original signal used. In this research, we have developed machine learning-based reduced-reference bitstream audiovisual quality prediction models by using the parametric version of the publicly available INRS audiovisual quality dataset. As that original INRS dataset did not contain bitstream information but provided both reference and transmitted videos, we have computed its bitstream version to develop the reduced-reference bitstream models. We have compared the performance of the Decision Trees based ensemble methods, Genetic Programming and Deep Learning models on this bitstream version of the dataset and have also compared these results with the results of the no-reference parametric models on the parametric version of the dataset. Decision Trees based ensemble methods outperformed Deep Learning and Genetic Programming based models when reduced-reference bitstream data was used and outperformed all existing no-reference parametric models that were trained and tested on the parametric version of the dataset. Our studies show that Decision Trees based approaches are well suited for no-reference parametric models as well as for reduced-reference bitstream models.

In order to mechanically predict audiovisual quality in interactive multimedia services, we have developed machine learning--based no-reference parametric models. We have compared Decision Trees--based ensemble methods, Genetic Programming and Deep Learning models that have one and more hidden layers. We have used the Institut national de la recherche scientifique (INRS) audiovisual quality dataset specifically designed to include ranges of parameters and degradations typically seen in real-time communications. Decision Trees--based ensemble methods have outperformed both Deep Learning-- and Genetic Programming--based models in terms of Root-Mean-Square Error (RMSE) and Pearson correlation values. We have also trained and developed models on various publicly available datasets and have compared our results with those of these original models. Our studies show that Random Forests--based prediction models achieve high accuracy for both the INRS audiovisual quality dataset and other publicly available comparable datasets.

We have developed reduced reference parametric models for estimating perceived quality in audiovisual multimedia services. We have created 144 unique configurations for audiovisual content including various application and network parameters such as bitrates and distortions in terms of bandwidth, packet loss rate and jitter. To generate the data needed for model training and validation we have tasked 24 subjects, in a controlled environment, to rate the overall audiovisual quality on the absolute category rating (ACR) 5-level quality scale. We have developed models using Random Forest and Neural Network based machine learning methods in order to estimate Mean Opinion Scores (MOS) values. We have used information retrieved from the packet headers and side information provided as network parameters for model training. Random Forest based models have performed better in terms of Root Mean Square Error (RMSE) and Pearson correlation coefficient. The side information proved to be very effective in developing the model. We have found that, while the model performance might be improved by replacing the side information with more accurate bit stream level measurements, they are performing well in estimating perceived quality in audiovisual multimedia services.

We present the INRS audiovisual quality dataset made of 160 unique configurations for audiovisual content including various media compression and network distortion parameters such as video frame rate, quantization and noise reduction parameters, and packet loss rate. The compression and network distortion parameter range values are selected to match real-time communications use cases. The H.264 video codec in 720p resolution and the AMR-WB audio codec are used for encoding video and audio streams. Thirty observers have rated the overall audiovisual quality on the Absolute Category Rating (ACR) 5-level quality scale in a controlled environment. The dataset includes MOS values, packet loss rates measured at bit stream level for both video and audio streams, compression parameters and various packet header information. We have used open source software for producing source audiovisual sequences, end-to-end streaming and a custom video player. These tools and the dataset are free to public access for research and development purposes.

The rapid growth of network services is forcing operators to use of inter-layer coordination between the IP and the transport network management systems in order to optimize resource utilization within the carriers network ecosystem. However, the diversity of technology and the related management systems within the IP and transport networks poses significant challenges in developing a single unified management ecosystem for both IP and transport networks. In this work, we present a basic prototype of a Service Oriented Architecture (SOA) based framework proposed in the EU project ONE for programmable orchestration of IP and transport management processes. The proof-of-concept prototype uses web services and BPEL to facilitate programmable inter-layer coordination, and was tested on a number of use cases. We measured the performance penalty of using the SOA framework against the traditional JAVA based provisioning approaches used in integrated multi-layer management, and show that the penalty incurred in execution times were sufficiently small while bringing several significant advantages during the software design, development, integration and maintenance, which is simple and highly effective.