【作者】 赵刚；

【导师】 杨宗凯；

【作者基本信息】 华中科技大学 ， 通信与信息系统， 2005， 博士

【摘要】 近几年来,移动互联网已成为当今全球信息产业的热点,基于移动互联网的多媒体业务无疑具有广阔的应用前景。多媒体传输质量是多媒体应用成功与否的关键。然而,复杂的网络传输条件(低带宽、高误码、时变特性)、有限的终端处理能力以及不断变化的移动用户需求给多媒体传输带来了巨大的挑战。无线多媒体的传输必须能够自适应地满足传输网络、终端和用户需求,才能有效保证传输质量。针对这一问题,本文将研究自适应的无线多媒体传输技术,以使多媒体资源能适应不同的网络带宽,有效对抗各种网络差错,充分满足无线多媒体业务的需求,从而优化多媒体传输质量。本文主要围绕图像和视频两种媒体的无线传输问题展开研究,包括四个方面的内容:(1)JPEG2000 图像传输特性模型及其误码保护机制研究; (2)H.264/AVC 视频特性分析模型及其误码保护机制研究; (3)H.264/AVC 视频无线传输带宽自适应技术研究; (4)多媒体教育资源自适应传输系统框架研究。本文得到国家自然科学基金项目“基于媒体传输特性的无线多媒体技术的研究”(项目编号60202005)和国际学术交流(新加坡南洋理工大学)合作项目“面向分布式视音频内容获取终端的推拉式提取工具”,以及国家十五科技攻关计划“网络教育关键技术及其示范工程”,“教育资源管理系统”(项目编号2001BA101A07)项目的资助。本文取得的研究成果包括如下几个方面: JPEG2000 图像无线传输特性模型及其误码保护机制:现有对JPEG2000 的率失真优化及抗误码方案的研究主要通过仿真方法获得,计算量大且十分耗时,对非均匀误码保护方案的设计效率不高。本文提出了一个理论分析模型来预测各种无线信道条件下传输的JPEG2000 图像失真统计特性。基于该分析模型,我们为在无线信道上传输JPEG2000 图像设计了高效、准确的非均匀误码保护方案,达到了较好的图像传输效果。H.264/AVC 视频特性模型及其误码保护机制:现有针对视频的非均匀误码保护机制或导致编码效率降低; 或基于启发式方法,缺乏理论依据; 较少考虑到人眼感知系统的特点。本文依据人眼感知系统对视频运动敏感程度的特性,更多还原

【Abstract】 In recent years mobile Internet has emerged as one of the largest sector of the global information industry and the demand of multimedia services in the mobile Internet is increasing. The key to the success of wireless multimedia services is the transmission quality. However, the transmission quality of multimedia stream will degrade greatly in low bandwidth, high error rate and time-variant wireless channel.In additional, limited terminal capability and various mobile user requirements also bring challenges to wireless multimedia transmission. Wireless multimedia should be adapted to the transmission network, the terminal and mobile users’requirements. To solve the above problems, this thesis addresses on adaptive multimedia transmission in mobile Internet. The objective of this work is to enable multimedia resources adapted to different network bandwidth, effectively fight against all kinds of transmission errors, and optimize the transmission quality of wireless multimedia. The research of this thesis mainly focuses on wireless image and video transmission, and includes four issues: (1) transmission property model of JPEG2000 encoded image over wireless channel and design of error protection schema; (2) property model of H.264/AVC video and design of error protection schema; (3) network bandwidth adaptation technology for wireless H.264/AVC video transmission; (4) multimedia learning resource adaptation and delivery framework. The work in this thesis has been supported by the National Science Foundation of China “Investigation of Wireless Multimedia Techniques based on Multimedia Transmission Property ”(No.60202005) and international academic (Nanyang Technological University in Singapore) cooperation project “Push-Pull Extraction Tool For DistributEd AudioVIsual Content Terminal ”respectively, and also supported by the project “Learning Resource Management System”(No.2001BA101A07), which is a sub-project of the “Key Technology and Demonstration Project of Network Learning ”project in the 10th National Five Year Plan for Key Problem Tackling in Science and Technology. The contributions of this thesis include: 1. The transmission property model of JPEG2000 image over wireless channel and error protection schema: Existing researches on rate-distortion optimization and error-resilient schema are mainly based on simulation approaches requiring large amounts of computation time. This thesis proposed an analytical model to estimate statistical distortion for the transmission of JPEG2000 encoded image over wireless channel. Based on this model, we design effective and accurate unequal error protection schema for JPEG2000 image transmission over wireless channel with better transmission quality. 2. Property Model of H.264/AVC video and error protection schema: Current unequal error protection schemas are based on heuristic approaches and take no account of the characteristics of human visual system. According to the sensitive characteristics to video motions of human eyes, this thesis utilized perceived motion energy model by taking account to the new encoding features of H.264/AVC standard to analyze and model the motions in H.264/AVC encoded video. The bit stream of the video is divided into several quality layers and an unequal error protection has also been designed to protect this layered bit stream for the transmission over wireless channels. Experiment results show that higher video transmission quality is obtained. 3. H.264/AVC video adaptation model and system: Towards the limited bandwidth and time-variant characteristics of wireless channel, we proposed an adaptation model based on MPEG-21 Digital Item Adaptation framework for wireless H.264/AVC video transmission. The proposed system not only ensures the viewers’perceived qualities as high as possible but also realizes format-independent video adaptation mechanism and enhances the flexibility and scalability of multimedia adaptation system. 4. Multimedia learning resource adaptation and delivery framework: For the terminal adaptation and learner requirements adaptation in mobile learning, this thesis proposedan adaptation and delivery framework for multimedia learning resource. By integrating the adaptation of learning object and the adaptation of multimedia content together, this framework ensures higher general transmission quality of learning resource than other approaches which only consider one type of the adaptation.更多还原