ASIA unversity:Item 310904400/26237
English  |  正體中文  |  简体中文  |  Items with full text/Total items : 90429/105609 (86%)
Visitors : 10283281      Online Users : 89
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version

    Please use this identifier to cite or link to this item:

    Title: 智慧型內容感知視訊調整系統之研究
    Authors: 莊政宏
    Contributors: 資訊學院;資訊工程學系
    Keywords: 影像處理;視訊處理;視訊編輯;視訊調整;視訊雕刻;視訊摘要;細縫裁減;紋理合成;影像變形;image processing;video processing;video editing;video retargeting;video resizing;video carving;seam carving;video summarization;texture synthesis;image warping
    Date: 2011
    Issue Date: 2013-07-18 15:52:19 (UTC+8)
    Abstract: 本計畫為延續前期計畫以空間及時間架構為基礎、具有內容感知之視訊調整技術,預計結合並完 成主要研發事項:(1)空間內容感知之視訊調整技術Part 1、(2)空間內容感知之視訊調整技術Part 2、 (3)時間內容感知之視訊調整技術,並結合(1)(2)(3)項成為智慧型內容感知之視訊調整系統。 在前期計畫中,主要利用物件及背景作為視訊縮放調整的基礎,並且利用基於區域二元圖形(Local Binary Pattern, LBP)的紋理偵測方法找出紋理區域,在紋理區域上以紋理合成及紋理裁切進行放大與 收縮,發展具備紋理感知之視訊調整技術,對於視訊空間內容以物件、背景、紋理等特性區域,利用 其不同的重要性與相關幾何位置之概念,於進行視訊長寬比及大小尺寸調整時,能保持或增強視訊空 間資料的重要內容,完成空間內容感知之視訊調整技術Part 1。 在本年度的計畫中,預計開發多平面圖文資訊萃取技術,擷取視訊內容的文字部分,發展具備文 字感知之視訊調整技術,完成空間內容感知之視訊調整技術Part 2。而後利用視訊分鏡偵測,發展時 間軸、分鏡、及畫面三階層的顯著特徵來評估視訊畫面的重要性,以整張畫面移除或加入的方式進行 視訊時間長度調整工作,達成時間內容感知之視訊調整技術,能保持或增強視訊時間資料的重要內容 與關鍵畫面,最後整合各項技術成為智慧型內容感知之視訊調整系統。本計畫所開發之系統可應用於 調整視訊的長寬比、大小尺寸、及播放時間,接合不同大小尺寸的視訊,因應各種不同長寬比的顯示 需求,及視訊摘要等視訊編輯系統上。

    This project will continue the project in last year and propose spatial and temporal framework-based content-aware video retargeting techniques. The project will be designed to complete the major development issues as follows: (1) part 1 of spatial content-aware video retargeting techniques, (2) part 2 of spatial content-aware video retargeting techniques, and (3) temporal content-aware video retargeting techniques. The three items will be integrated to be smart content-aware video retargeting systems. In the last year, we use objects and backgrounds as the basis of video retargeting. The Local Binary Pattern (LBP) algorithm is applied to detect texture regions. Texture synthesis and cropping are developed for enlarging and shrinking texture regions to complete texture-aware video retargeting techniques. The different importance and geometric positions among these objects, backgrounds, and texture regions can be used for video retargeting to preserve or improve significant content. Then the part 1 of spatial content-aware video retargeting techniques is completed. In this year, the multi-plane document segmentation will be developed for text extraction to complete text-aware video retargeting techniques. And we will develop video shot detection and propose three hierarchical saliency (time-axis, shot, and frame saliency) features to evaluate the importance of video frames. Then a frame removing algorithm based on the hierarchical saliency is developed for temporal content-aware video retargeting. In this way the importance content or key frames will be preserved during video time resizing. Finally, we will integrate all techniques to complete smart content-aware video retargeting systems. Overall speaking, smart content-aware video retargeting systems developed by this project can be applied to the video editing system including adjustment of video aspect ratio, size, time, etc., different video synthesis, in response to the demand of different aspect ratio display, and video summarization.
    Appears in Collections:[Department of Computer Science and Information Engineering] Ministry of Science and Technology Research Project

    Files in This Item:

    File SizeFormat

    All items in ASIAIR are protected by copyright, with all rights reserved.

    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback