On Perception-Based Error Protection for Mobile Multimedia

Abstract: Representation of multimedia requires large amounts of data. Therefore, multimedia source signals are compressed to reduce the data rate that would otherwise be required for its transmission. However, compression renders multimedia data more vulnerable to transmission errors. Therefore, error protection is needed when transmitting multimedia to mobile users over wireless channels. As channel coding adds parity symbols to the data for error protection, which in turn increases transmission data rate, efficient utilization of this added parity budget becomes a key for insuring quality of service for mobile multimedia. Computing the optimal parity allocation among packets of a multimedia stream may not be possible in realtime due to the huge number of ways the parity can be assigned to these packets. Therefore, it is important to design powerful error protection schemes while imposing low computational and memory demands on the system. This is particularly true for systems having limited resources, such as mobile handheld devices. Given that human beings are the final judges on the quality of multimedia services, it will also be beneficial to use perceptual quality metrics that correlate well with human perception when evaluating different error protection schemes. In this work, we have considered the above aspects when designing error protection for mobile multimedia: Providing near optimal performance while keeping complexity low; designing and evaluating of error protection schemes based on perceptual quality metrics. The thesis is divided into five parts. In the first part, region of interest (ROI) identification, coding and advantages of ROI coding are investigated. In addition, a framework is proposed for using ROI coding in wireless imaging. The second part provides a number of unequal error protection (UEP) schemes for wireless imaging ranging from very simple and less efficient equal error protection to very complex yet optimal error protection. The proposed error protection schemes provide an excellent trade-off between performance and omplexity. In the third part, a framework for perception-based error protection of images and video streams is proposed. This framework is based on the fact that the human visual system is more sensitive to quality degradation in the ROI compared to the background. The proposed error protection scheme provides excellent performance. Part four deals with realtime systems and/or systems having limited memory and processing resources. Specifically, an UEP is proposed for wireless imaging systems which provides near optimal performance and puts very low processing/memory demands on the systems. Finally, in the fifth part, optimal UEP is analyzed. Based on the analysis, optimal UEP is modeled using several mathematical models. Each of the considered models is evaluated for its fitness in providing error protection. The simulation results show that most of the investigated models give excellent performance.