跳到主要內容

臺灣博碩士論文加值系統

(44.200.169.3) 您好!臺灣時間:2022/12/05 19:19
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

: 
twitterline
研究生:王信雄
研究生(外文):Shing-shoung Wang
論文名稱:植基於邊緣資訊的影像複製偵測方法
論文名稱(外文):An Edge-Based Image Copy Detection Scheme
指導教授:林家禎林家禎引用關係
指導教授(外文):Chia-chen Lin
學位類別:碩士
校院名稱:靜宜大學
系所名稱:資訊管理學系研究所
學門:電算機學門
學類:電算機一般學類
論文種類:學術論文
論文出版年:2006/07/
畢業學年度:94
語文別:中文
論文頁數:57
中文關鍵詞:植基邊緣資訊影像偵測智慧財產權影像特徵向量空間模型
外文關鍵詞:vector space modelintellectual property rightimage signatureEdge-based image copy detection
相關次數:
  • 被引用被引用:5
  • 點閱點閱:257
  • 評分評分:
  • 下載下載:25
  • 收藏至我的研究室書目清單書目收藏:1
隨著網際網路與經濟實惠的數位儲存裝置的廣泛流行,著作權和版權的侵犯問題也日益嚴重,尤其在網際網路的公開環境下,數位影像在使用者間的複製、傳送與散佈,已經成為方便的一件事;如何有效的防止數位影像智慧財產權的不法盜用,是當前最重要的課題之一。目前已經有許多關於影像保護方面的研究被提出來,這些研究可分成以數位浮水印技術與偵測不合法複製的兩種方法。
許多可以偵測原圖經過各種影像處理後複製的方法已經被提出。然而,這些方法無法精確偵測那些經過任意角度1-89度旋轉、截角與平移的篡改圖片。在本篇文章,我們提出了一個植基於影像邊緣資訊複製偵測的方法。主要概念是;首先,一個查詢圖先轉換成一張二元的邊緣圖,僅邊緣圖用以產生影像的特徵值。邊緣圖切割成尺寸為4×4的不重疊區塊,且計算每個區塊內像素值為零的次數。為了建構空間的資訊,我們將每張圖切割為若干區域:中央核心區域、環繞中央核心區域以及外圍區域。最後,每個區域內區塊頻率次數以及各區域之間的關係係數我們應用向量空間模型計算。我們結合兩者成為特徵值。最後藉由我們比對特徵值的方法,我們可以決定哪些是疑似被竄改的圖片。
為了驗證本論文所提出的方法在複製偵測的可行性,本研究針對一般影像最常遇到的影像處理做實驗,實驗結果證明,當影像受到這些類型影像處理竄改時,特別是在任意角度旋轉、影像平移、影像截角等其他參考方法無法精確判斷的處理時,我們的方法依然可以進行偵測。
With the widespread of the Internet and the appearance of lower-priced digital storage devices, the tort of the copyright is more and more serious. Especially under the open environment of the internet network, digital images can be copied, delivered, and distributed more conveniently from user to user. How to protect intellectual property right (IPR) efficiently has become a critical issue. There were a lot of researches about image protection been proposed at present. They can be classified into two different categories: watermarking and copy detection schemes.
Several schemes have been proposed to detect copies which are generated by various image processes. Unfortunately, the existing schemes can not precisely detect tampered images as those are rotated 1 to 89 degrees, cut one part of corner and slightly shifted from the original image. In this thesis, an edge-based digital images copy detection scheme is proposed. The first, a query image is just transformed into a binary edge image, which is the only one used to generate the corresponding image signature. Next, the transformed image is divided into 4×4 non-overlapping blocks. And, the frequencies of zero pixel values in each block are counted. Besides, we divide an image into three regions: center core region, the region surround the center core region and suburb to maintain the spatial information of an image, Finally, the amounts of frequencies in each region’s blocks are counted and the relations among them are calculated by using the vector space model. Above information is combined as a feature vector (also called signature vector) of an image. Later, we can simply compare the feature vector of a query image and those of image database to decide whether the query image is a tampered one or not.
In order to verify the detection performance of our proposed scheme, several experiments have been conducted. The experimental results confirm that when the image was been tampered by these image processing listed in this thesis, especially arbitrary angles of rotating, one part of corner cutting and shifted from the original image, the tampered images can be successfully detected by using our proposed scheme.
摘 要 i
ABSTRACT iii
誌 謝 v
目 錄 vi
圖 目 錄 viii
表 目 錄 ix
第一章 緒論 1
1.1研究背景與動機 1
1.2 論文架構 2
第二章 背景知識及相關文獻 3
2.1 Content-Based Image Retrieval 3
2.2 Watermarking 6
2.3 Vector Space Model 8
2.4 Kim’s content-based image copy detection scheme 8
2.5 Wu et al.’s robust content-based image copy detection scheme 10
第三章 植基於邊緣資訊影像複製偵測方法 13
3.1 邊緣偵測 14
3.1.1 尋找邊緣點 14
3.1.2 刪除單一雜點 17
3.2 特徵萃取與表示 19
3.2.1 特徵萃取 19
3.2.2 特徵標準化 20
3.2.2 計算各分割區域間關係係數 21
3.2.3 特徵值表示 22
3.3 特徵比對 22
3.3.1 Measure I:歐式距離加權比較 23
3.3.2 Measure II:絕對值差平均加權比較 24
3.4 範例流程 26
第四章 實驗結果 30
4.1 實驗資料來源 30
4.2 實驗影像處理類型 31
4.3 複製偵測結果 33
4.3.1 排名結果 33
4.3.2 所有處理圖可被複製偵測比較 35
4.3.3 單一處理圖可被複製偵測比較 37
第五章 實驗分析與討論 46
5.1 邊緣偵測門檻值與去除雜點設定 46
5.2 區塊大小設定 47
5.3 區域數目與大小設定 48
5.4 資料標準化 49
5.5 特徵向量長度比較 53
5.6 衡量相似度計算量比較 54
第六章 結論 55
參考文獻 56
[1] C. C. Chang, K. F. Hwang and M. S. Hwang, “A Digital Watermarking Scheme Using Human Visual Effects,” Informatica, Vol. 24, No. 4, pp. 505-511, 2000.
[2] C. C. Chang and C. S. Tsai, “A Technique for Computing Watermarks from Digital Images,” Informatica, Vol. 24, No. 3, June, pp.391-396, 2000.
[3] C. Kim, “Content-based Image Copy Detection,” Signal Processing Image Communication, Vol. 18, pp. 169-184, 2003.
[4] D. N. Bhat and S. K. Nayar, “Ordinal Measures for Image Correspondence,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 20, No. 4, pp. 415-420, 1998.
[5] E. Y. Chang, J. Z. Wang, C. Li, G. Wiederhold, “RIME: A Replicated Image Detector for the World-Wide-Web,” Proceedings of the SPIE Multimedia Storage and Archiving Systems, San Jose, CA, Vol. III, Nov. 1998.
[6] G. Salton, A. Wong, and C. S. Yang, “A Vector Space Model for Automatic Indexing,” Communications of the ACM, Vol. 18, No. 11, pp. 613–620, 1975.
[7] Wu, M. N., Lin, C. C. and Chang, C. C., “A Robust Content-Based Copy Detection Scheme,” Fundamenta Informaticae, Vol. 71, No.1, pp. 351-366, 2006.
[8] Y. K. Chan and C. C. Chang, “A Color Image Retrieval Method Based on Color Moment and Color Variance of Adjacent Pixels,” International Journal on Pattern Recognition and Artificial Intelligence, Vol. 16, No. 1, pp. 113-125, 2002.
[9] SIMPLIcity, WIPE, Virtual Microscope, http://wang.ist.psu.edu/docs/related/
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top