English  |  正體中文  |  简体中文  |  Items with full text/Total items : 74010/74010 (100%)
Visitors : 24080889      Online Users : 527
RC Version 7.0 © 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: http://ir.lib.ncu.edu.tw/handle/987654321/65530

    Title: 基於HEVC畫面內編碼特徵之 影像內容檢索技術;Content Based Image Retrieval Utilizing HEVC Intra Coding Features
    Authors: 徐歆茹;Hsu,Hsin-Ju
    Contributors: 通訊工程學系
    Keywords: 壓縮域;基於內容影像檢索;畫面內預測;HEVC;Compression domain;Content-based image retrieval;Intra prediction;HEVC
    Date: 2014-07-31
    Issue Date: 2014-10-15 17:02:54 (UTC+8)
    Publisher: 國立中央大學
    Abstract: 隨著網路的蓬勃發展以及行動裝置的進步,使用者隨手可得影音資訊,這些多媒體資訊不斷地朝向高品質、高解析度發展,使得影音資料量成指數型成長,因此如何管理龐大的資料庫,並根據使用者的需求取得所需之影像便是一個重要的課題。基於影像內容之檢索技術,雖然發展歷史悠久且被廣泛的討論與研究,但多數的研究皆在原始像素域擷取特徵,必須將影像完全解壓縮後再進行處理,會增加許多計算複雜度以及儲存特徵空間。
    本論文針對最新一代視訊壓縮標準HEVC,提出HEVC壓縮域之影像檢索技術,利用HEVC解碼器部分解碼所得之畫面內編碼資訊進行影像檢索。相較於直接使用解碼資訊作直方圖來比對,我們研究HEVC畫面內編碼之特性,將預測模式(intra prediction mode)作濾波處理,並將原本五種不同尺寸的預測單位(Prediction Unit, PU)分為兩組,再搭配其殘餘值(residual)能量之分布,萃取出有效的影像特徵以利檢索。實驗結果顯示,所提出之方法可大幅減少影像特徵的儲存空間,加速系統處理效率,而檢索效能平均準確率(Mean Average Precision, MAP)值為0.247。;With the rapid development of Internet and mobile devices, how to efficiently retrieve multimedia information from huge databases becomes an important issue. In the past decade, numerous algorithms were extensively studied for content based image retrieval. However, most existing works retrieve features in the pixel domain, which requires fully decoding images. It is time consuming and memory wasting to retrieve features in the pixel domain.
    This thesis focuses on the content-based image retrieval in compression domain for the new standard, high efficiency video coding (HEVC). The features are extracted from partially decoded intra frames. Instead of directly taking histogram from the decoded information, the directions of intra mode are filtered and the sizes of prediction unit are grouped into two types. And the histogram of residual power is also adopted as a feature. The retrieval efficiency can greatly benefit by the proposed features.
    The experimental results show that the proposed method can reduce the resource consumption and achieve a good retrieval performance, i.e. Mean Average Precision (MAP) 0.247 in Oxford 5K dataset.
    Appears in Collections:[通訊工程研究所] 博碩士論文

    Files in This Item:

    File Description SizeFormat

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

    社群 sharing

    ::: Copyright National Central University. | 國立中央大學圖書館版權所有 | 收藏本站 | 設為首頁 | 最佳瀏覽畫面: 1024*768 | 建站日期:8-24-2009 :::
    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback  - 隱私權政策聲明