完整后设资料纪录
DC 栏位 | 值 | 语言 |
---|---|---|
dc.contributor.author | 蔡文锦 | en_US |
dc.contributor.author | Tsai Wen-Jiin | en_US |
dc.date.accessioned | 2014-12-13T10:49:50Z | - |
dc.date.available | 2014-12-13T10:49:50Z | - |
dc.date.issued | 2009 | en_US |
dc.identifier.govdoc | NSC98-2221-E009-085 | zh_TW |
dc.identifier.uri | http://hdl.handle.net/11536/101852 | - |
dc.identifier.uri | https://www.grb.gov.tw/search/planDetail?id=1909330&docId=316604 | en_US |
dc.description.abstract | 同侪网路(Peer-to-Peer, or P2P network)由于其低成本及简单具延展性的特性,已经越 来越被广泛使用在不同的应用上,从早期的档案分享、快速档案下载、到影音媒体串流 等,都是很热门的应用。然而,在影音串流的应用中,由于影音的大频宽需求,使得频 宽不足的Peer 而言,无法下载高频宽的影音资料,或贡献其上载频宽协助影音串流,使 得P2P 的特性无法充分发挥。要解决上述网路与Peer 的异质性问题,方法之一就是将 高频宽的影音资料切割成多个低频宽的子影音档,如此一来,低频宽Peer 也能参与上载 或下载较低频宽的子影音档。而多重描述编码法(Multiple Descriptor Coding)就是一种将 影片切割成多个低频宽子影片的视讯编码法. 多重描述编码法所切割出来的子影片是可以独立被个别解码的。 收到少数子影片 档的Peer 可解出较低品质的影片;而收到愈多子影片档的peer,则可解出愈高品质的影 片。同时,由于各个子影音档在传输时可能走不同路径,封包遗失可以藉由其他子影片 档的内容来加以修复,因此大幅减低封包遗失对影片品质的影响。 多重描述编码法虽然能为网路与Peer 的异质性提供解决之道,然而在实际的应用中 却未被广为使用,主要在于编码效能与即时性的问题。如果单纯将影片档的资料切割成 多份分别压缩,则个别单一的子影片品质难以接受且压缩的效率也差,要让单一的子影 片的品质提升,则必须加入些redundancy,因而更减低了编码效能。要达到最佳的编码 效能往往会增加演算法的复杂度,也影响到编码的即时性。 在此研究计画中,我们预计以三年为期完成以下三个目标:1. 研究利用错误编码 法改善多重描述编码之效能的方法; 2. 研究设计阶层式架构的混合多重描述编码架构; 3. 实现所设计的阶层式多重描述编码于实际的P2P系统中。 计画第一年主要是着重在 研究使用不对等保护(Unequal Error Protection, UEP)的方式改善多重描述编码之编码效 能的方法。我们将根据影片资料的重要性决定分配的方式,来让错误纠正码的使用效率 较好。我们将不预设descriptor 的数目,而是根据视讯档本身的内容、网路状况、或使 用者的需求情形来订定。 第二年的研究重点在于阶层式的混合多重描述编码架构的研 究与设计,主要的想法是希望能结合不同MDC 架构中,对于不同视讯特性切割的优点, 希望兼顾视讯资料的多个特性来达到较佳的编码效能与错误容忍能力。同时考虑多种特 性的混合式MDC 虽也有人提出,但仅限于交错式的混合架构,因此,在video sequence 的每个时间点上,其实都只用了一种特性的切割。在本计画中,我们拟提出阶层式的混 合MDC 架构,使得在sequence 中的每个时间点都是结合多种特性的混合切割,我们预 期如此的切割能有不错的效果。第三年我们将根据前两年的研究结果,实作出所设计的 MDC 方法于P2P streaming 网路中,并分析其在实际应用中的系统效能。 | zh_TW |
dc.description.abstract | Peer-to-Peer (or P2P) network, due to its low-cost and scalability, has been widely used for various applications, such as file sharing, fast downloading, VOIP, IPTV, etc. However, not all the clients can enjoy the media-streaming services over P2P networks due to the problems of heterogeneous device bandwidth and dynamic network congestion. Multiple description coding (MDC) addresses such problems by means of striping a video sequence into two or more descriptions in such a way that each description in independently decodable. The quality of the received video improves with each received description, so devices can determine the number of descriptions to receive according to their available bandwidths. Loss of any one of these descriptions does not cause complete failure. If one of the descriptions is lost, the playback can continue with only a slight reduction in overall quality. There have been a number of proposals for MDC coding each providing their own tradeoff between compression efficiency and error resilience. And very few researches have addressed the problem of coding complexity which is important for streaming applications. In this project, we plan to design a MDC-based coding solution for video streaming over P2P network. There are three goals that we would like to achieve in three years: 1. Study on improving the coding efficiency of MDC by using error correction code, 2. Study on MDC using hierarchical hybrid architectures, 3. Design and implement the proposed MDC solution for P2P applications. In the first year, we will focus on the design of MDC which improves the coding efficiency by using the technique of unequal error protection (UEP) to make the use of error correction code more efficiently. In the second year, we will focus on designing a hierarchical hybrid architecture of MDC which takes advantages of different video features in error concealment. In the third year, we will implement the proposed MDC on a P2P network and evaluate the performance. . | en_US |
dc.description.sponsorship | 行政院国家科学委员会 | zh_TW |
dc.language.iso | zh_TW | en_US |
dc.subject | 多重描述编码 | zh_TW |
dc.subject | 不对等保护 | zh_TW |
dc.subject | 阶层式混合多重描述编码 | zh_TW |
dc.subject | Multiple description coding | en_US |
dc.subject | MDC | en_US |
dc.subject | UEP | en_US |
dc.subject | Peer-to-peer | en_US |
dc.subject | P2P | en_US |
dc.title | 适用于媒体串流之多重描述视讯编码法 | zh_TW |
dc.title | Study on Multiple Description Coding for Video Streaming Applications | en_US |
dc.type | Plan | en_US |
dc.contributor.department | 国立交通大学资讯工程学系(所) | zh_TW |
显示于类别: | Research Plans |