國立臺灣大學電機工程學系暨研究所陳良基2006-07-252018-07-062006-07-252018-07-061999-07-31http://ntur.lib.ntu.edu.tw//handle/246246/7721由於多媒體、電腦及電訊傳輸的快速發展，這 三大領域的交集已成為眾所矚目的焦點。而影像資 訊由於其多樣性、高頻寬、以及即時處理的特性， 更成為焦點中的焦點。因此高壓縮比並具有比例調 整功能將是未來的趨勢。在本計畫中，我們將根據 過去的成果，發展出一套具有以上功能的影像編碼 系統。這其中使用了分頻技術、光流演算法、區域 切割技術、隨形轉換以及向量量化的技巧。如何將 以上技術加以改善並做適當的整合將是計畫的關 鍵。在應用層面上，除了高壓縮比之外，影像內容 的分析、物件的比例調整等也將是預期的目標。此 外，此系統的應用範圍也可能涵蓋了影像分析與監 控等。除了系統的模擬之外，硬體的實現也是重要 的一環，本計畫將同時探討實現這些演算法的硬體 架構最佳化。Because of the fast progress of multimedia, computers, and telecommunication, the intersection of the above three fields becomes the focus of technical trends. In addition, digital video attracts much more attention due to its characteristics of diversity, high bandwidth, and real-time processing. High compression ratio and scalability will be the trend in the future. Based on the past research experiences, we will develop a video coding system satisfying the above characteristics in this project. The related techniques include subband coding, optical flow algorithm, region segmentation, arbitrary shape transform, and vector quantization. How to improve and combine all these techniques will be the crucial part of the project. In terms of applications, in addition to high compression ratio, content-based analysis and object scalability are also expected. Furthermore, the applications of the proposed system include video analysis and surveillance. Except the system simulation, hardware implementation is very important. This project will explore the optimal hardware architecture to implement these algorithms.application/pdf130528 bytesapplication/pdfzh-TW國立臺灣大學電機工程學系暨研究所多媒體電腦電訊傳輸影像編碼系統分頻技 術光流演算法區域切割技術隨形轉換向量 量化物件比例調整multimediacomputertelecommunicationvideo coding systemsubband codingoptical flow algorithmregion segmentationarbitrary shape transformvector quantizationobject scalabilityreporthttp://ntur.lib.ntu.edu.tw/bitstream/246246/7721/1/882218E002030.pdf