Distributed quality-aware multimedia communications

多媒体通信产业正经历一个快速发展的阶段，新的业务不断涌现，如移动的视频电话和互联网协议电视。此外，多媒体服务越来越多地通过异构网络传输，异构网络生成对终端用户的感知体验质量（QoE）产生不利影响的信号降级组合。传统上，主观测试已经被用于表征多媒体QoE。然而，主观测试，不适合在线应用和客观模型，需要在服务质量监测。有了这些客观模型，就可以系统地调整网络参数以优化速率-质量性能，从而改善多媒体流传输并增强Qoe。我们把这种情况称为“质量感知多媒体通信”。' 在这方面，该项目的主要目标是开发一个客观的视听质量模型，可以在整个网络中分发。利用分布式处理，可以检测到发生不同质量降级的位置。能够检测多媒体流中的特定和相关的质量降低伪像是实现QoE反馈和根本原因分析的重要一步。然而，具体的文物，不同的用户感知到他们的期望和情感状态的水平。 为了在用户级别调整QoE模型，该项目的另一个目标是开发用户期望和情感状态的客观模型。这些模型将基于创新的声学和生理信号监测技术。这里提出的以用户为中心的范例有可能重新定义多媒体通信，并为加拿大和世界各地的多个行业带来重大的经济收益。

The multimedia communications industry is going through a phase of rapid development and new services are emerging continuously, such as mobile video telephony and internet protocol television. Additionally, multimedia services are increasingly being transmitted through heterogeneous networks which generate signal degradation combinations that adversely affect the end-user's perceived quality of experience (QoE). Traditionally, subjective tests have been used to characterize multimedia QoE. Subjective tests, however, are not suitable for online applications and objective models are needed for in-service quality monitoring. With such objective models available, network parameters can be systematically adjusted to optimize rate-quality performance, thus improving multimedia stream delivery and enhancing QoE. We refer to this scenario as 'quality-aware multimedia communications.' In this context, the main goal of this project is to develop an objective audio-visual quality model that can be distributed throughout a network. With distributed processing, locations where different quality degradations occur can be detected. Being able to detect specific and relevant quality-degrading artifacts in the multimedia stream is an important step towards QoE feedback and root cause analysis. Specific artifacts, however, are perceived differently by users depending on their level of expectation and affective state. In order to adjust QoE models at the user level, an additional goal of this project is to develop objective models of user expectation and affective states. The models will be based on innovative acoustic and physiological signal monitoring techniques. The user-centred paradigm proposed here has the potential to redefine multimedia communications and bring to bear significant economic gains to multiple industries in Canada and worldwide.