遍布即時多媒體系統與技術---總計畫

Abstract

鑑於未來網路結構與應用趨於異質結合，本計畫將針對『無所不在』的遍佈即時多媒體 系統與技術進行研究。在異質網路中，每個終端設備接入的網路頻寬及特性並不一致， 且終端設備本身的運算能力也不相同，造成傳統的單一層次多媒體編碼在應用上無法達 到最佳效果。為解決此問題而發展出多層次編碼法與分散式編碼法，以求得整體最佳效 能。此外，建構這類系統需要串流系統的支援與低耗電實作的設計才能實現。結合上述 技術，多媒體內容可以廣為散佈至各種接收環境。然而，必須輔以內容為檢索基礎的搜 尋技術以及智慧財產權保護機制，讓媒體提供者樂於提供，讓媒體使用者方便搜尋，如 此才能達到『無所不在』的多媒體系統概念。 上述多媒體系統相關技術分在六個子計畫中進行: (1) 在P2P 環境下運用分散式視訊編 碼技術的視訊通訊系統設計，(2) 可調視訊編碼之適應性動態精細改進和編碼，(3) 低 耗電之即時性視訊編解碼技術之研究，(4) 多型態多媒體為基礎之內容分析、索引與查 詢，(5) 多媒體傳輸安全機制與可調音訊編碼，以及(6) 有線/無線環境上建構於對等連 網及內容傳播網之可適性串流技術。上述課題均為遍佈即時系統的重要元件，我們將研 究其問題、提出解決方案、模擬實驗、並設計元件與系統之架構。此外，我們亦擬持續 參與MPEG 標準會議，將吾人技術提案到MPEG 標準。本計畫全程三年，其研究涵蓋 異質網路多媒體應用的基礎關鍵技術，預期研究成果將可促進相關係統的研發，並可貢 獻於學術界與產業界。

Due to the heterogeneous nature of future network connections and applications, the goal of this project is to develop the key technologies necessary for a pervasive multimedia system. In a heterogeneous environment, terminals often have different access bandwidth and operating capability. This makes the conventional single-layer coded multimedia difficult to reach every terminal in time with the best available quality. To solve this problem, the scalable coding and distributed coding techniques are invented. In addition to the coding technology, in implementing such a system an overall efficient streaming system design and the low-power circuit design are also required. Combining all the aforementioned technologies, we can deliver scaleable-coded multimedia to various terminals in diversified environments. However, it is still not fully compliant to the 「pervasive」 concept, unless we incorporate the content- based search and digital rights management technologies into the system. With the assistant of these two technologies, the users can easily find what they look for, and the providers can safely contribute their products. This project is divided into six sub-projects: (1) Visual Communication System Design Using Distributed Video Coding Technology in a P2P Environment, (2) Scalable Video Coding – Adaptive Motion Refinement and Coding, (3) A Study on Power-aware Video Coding Technology, (4) Multimedia Multimodal Content Analysis, Indexing and Retrieval, (5) Secure Multimedia Delivery and Scalable Audio Coding, and (6) Adaptive P2P-CDN Layered Streaming over Wired/Wireless Networks. These topics are the necessary components to construct a truly pervasive multimedia system. We will investigate the problems, propose solutions, conduct simulations and design architectures at both system and component levels. In addition, we will continue our MPEG standard activity and submit our technologies to the MPEG standard. The entire duration of this project is 3 years. We expect that the research results will be significant contributions to the academic community and industry.