| 英文摘要 | In this information explosion era, there are a large amount of digital multimedia content (e.g. text, webpage, picture, audio, video, etc) generated, stored, transferred and transformed, which are still having a explosively growing trend. The fast-growing development of Internet has profoundly promoted the explosive increasing of various multimedia services. With the continuous heterogeneity of Internet developments, different characteristics of network clients (including client device hardware, network accessibility and personalized preferences, etc) imposes a more and more salient contradiction with the identity of multimedia services. For example, through desktop computers users can easily browse a large picture without scrolling or zooming operations, however, the thing becomes more difficult when users browse such a large picture on the small screen of a mobile device, e.g. cell-phone, pocket pc, or smart phone. To meet the demands of mobile clients, the self-adaptive multimedia functions have been proposed. To provide adaptive content delivery, it is necessary to conduct research to analyze and understand multimedia content. This paper has studied several kinds of prevalent multimedia content, including picture, webpage, and video. We implement and present practical and feasible multimedia content adaptation systems. Additionally, based on the extensive pervasiveness of digital images on the Web, we proposed a framework for mining semantic knowledge for Web images, aiming to improve Web image search quality. This paper is targeted to improve multimedia content accessibility and service quality, studying a series of research work as follows: • Introduce multimedia content adaptation concepts, systemically summarize related works in current stage, and demonstrate the importance and essence of this research work. • Propose effective and efficient solutions to multimedia content adaptation according to client capacities, aiming to improve mobile media display and browsing on small screens of handheld devices, such as image, webpage and sports video media. • Propose webpage adaptation solution regarding the limited wireless bandwidth and small display screens on mobile devices. Our experimental results demonstrate that this new solution can comprehensively improve mobile users' web browsing and reduce wireless network consumption. • Propose bran-new solutions to improve sports video access and browsing. Our solutions include sports highlights extraction service on content providers, and effective user interfaces on mobile clients to improve sports video interaction and browsing for users. • Develop web mining technologies to extract high-level semantic knowledge for Web images, such as location, people, time and keywords. Integrating mobile users' contextual information, we apply such semantic knowledge to Web image retrieval, aiming to improve web image search quality for the mobile users. |
修改评论