QUALITY OF EXPERIENCE OF MULTIMEDIA SERVICES

Abstract

Nowadays, multimedia services are becoming a part of the individual's life; and thus, many ongoing efforts have been given to develop and optimize the technology for multimedia content creation, distribution, and consumption. As a result of advancement in the technology, there is a rivalry among the standards, services, and thus, the service providers. In this context, assurance of the consumer satisfaction by providing a required level of quality for the provided service is of crucial importance for the service providers. From another viewpoint, providing a better Quality of Experience (QoE) to the provided service could be a differentiation strategy to achieve the competitive advantage. To provide the best QoE, all the involving entities in the multimedia communication chain should pose a defined quality level. Moreover, the QoE centric operations and optimization techniques should be incorporated in the multimedia communication system. For QoE centric operation and optimization, the assessment of the QoE of the service is an important step, and thus, the QoE assessment is a scope of this dissertation. Among multimedia services, the video communication is becoming a dominant service. Moreover, the latest advancement in imaging technology-light field imaging, is expected to be a next generation imaging technology. Due to the possibilities of a wide range of applications provided by the light field imaging, lots of attention is pulled from industry and academia by this technology. Therefore, this dissertation is mainly focused on two types of multimedia content: video and light field image. Every step involved in multimedia communication such as acquisition, representation, encoding, network, and rendering/presentation, produces the artifacts, and ultimately degrades the quality. The main aim of this dissertation is to propose the novel theories and frameworks for assessing the QoE of video and light field image. First part of the dissertation is focused on video QoE assessment. Before designing a quality assessment metric, and for optimizing the multimedia communication networks, the knowledge of the impact of transmission impairments on video QoE is crucial. In communication environment, the spatial-temporal sensitive video content need to be transmitted over a noisy and bandwidth limited wired or wireless channel. The network produced impairments introduce the artifacts, and the artifacts ultimately degrade the video QoE. The achieved results show that there is a significant impact of the impairments: packet loss, jitter, and bandwidth on video QoE. However, the initial delay introduced by the network does not adversely affect the QoE. From the analysis results it is noticed that, the video QoE and the level of quality distortion introduced by the network impairments is also depending on the video scene i.e. video content itself. Therefore, a study is performed to understand the impact of visual content on the QoE. The dataset designed to study the impact of transmission impairments in video QoE, is also used to benchmark the state-of-the art quality measures. The achieved results strengthen the need of a No-Reference metric for video communication services. Therefore, a blind QoE assessment metric for video communication services is proposed. The performance evaluation of the proposed metric evidences its superiority over existing metrics. Next, a big revolution in image acquisition systems is given by the introduction of the light field imaging. The basic idea behind the light field imaging systems is the use of a micro-lenses array positioned at the focal point of the camera lens, in front of the image sensor. In this way, it is possible to record multiple views of a scene by using a single camera in a single shot, thus avoiding problems related to calibration and camera synchronization. The micro lens array records information on the incident light direction at different positions, i.e. it records the light field. The availability of low cost acquisition devices turned the light field theory into practice, thus allowing novel applications of the imaging systems. Currently, many ongoing efforts have been given towards the optimization and standardization of the system. In this context, the knowledge of the quality of the processed images is important. The work has been started by defining the quality related issues, and quality distortion model of the light field technology; which is significantly different compared to the traditional 2D/3D visual technology. Next, the increasing interest towards this media calls for methods to protect these data from manipulations and unauthorized reproduction or diffusion. For this purpose, a watermarking scheme is designed by exploiting a tradeoff between quality, robustness, and capacity for copyright protection. Evaluating the quality issue, that is the imperceptibility of the watermark or its impact on the cover data, is a challenging task. This becomes more challenging for light field data, because there is no validated or standard quality assessment protocol available for light field data. Therefore, our next aim is to devise a quality assessment framework for light field image. Towards, this aim a Reduced Reference light field image quality assessment framework is proposed. The achieved analysis results show that the predicted quality scores are very close to the corresponding subjective opinion scores. The lack of a validated subjected quality assessment framework for light field image pushes us to work in this topic-to propose a quality assessment framework for light field images, particularly for the images captured by using the commercially available light field cameras such as Lytro Illum and Raytrix. Therefore, an extensive study was performed by considering the Lytro Illum images. During the study, four encoding methods, including standard image compression methods and recently the proposed light field image compression methods are considered. As a result of the study, a new light field image quality dataset is available for the research community. Moreover, the dataset is used to study the impact of compression artifacts on light field image QoE and to benchmark the existing quality metric, when applied for light field image. Following, this study, a subjective quality assessment framework is proposed for light field imaging.