臺灣博碩士論文加值系統

近年來，由於數位化的時代，數位資料廣泛的透過網際網路傳播，任何人都可以輕易擷取資料並進一步修改做私人商業行為，數位資料相關智慧財產權的問題也漸漸地浮現於檯面上，而其中可以藉由數位浮水印來進行檢測著作權的相關權益問題，本論文將著重利用浮水印來保護影像本身並且偵測竄改位置進行還原。
本論文提出一種方法來偵測影像是否遭受到竄改，若偵測到竄改則對影像進行回復。首先將保護的影像利用SPIHT(set partitioning in hierarchical trees)進行壓縮，產生SPIHT位元流，而SPIHT若沒有錯誤發生時壓縮特性非常良好，只需少數位元流進行解碼，即可還原成一張近似原圖的影像，但考慮在影像傳輸過程中，可能會造成SPIHT位元流產生錯誤，使還原影像品質非常差。為了保護SPIHT位元流，我們使用QR碼的錯誤更正技術，將SPIHT位元流編碼為QR碼，而QR碼有四種錯誤更正級別，如何進行最有效率的分配，藉由本論文所提出的錯誤更正級別配置(error correction level allocation)演算法，將有限的QR碼個數使用合適的錯誤更正分配，並將編碼完成的QR碼當作浮水印植入原始影像之第一、二層位元層(LSB bit plane)。若要復原影像時，萃取出浮水印後再使用標準QR碼解碼器解碼，還原成SPIHT位元流，最後透過SPIHT解碼器產生還原影像，與接收到影像作比對，一旦偵測到疑似錯誤像素，將遭到竄改位置利用還原影像資料覆蓋來達到回復的效果。
實驗結果證明我們所提出的方法可以偵測數位影像遭到竄改以及回復，並且可達到期望的影像品質，此外還可以抵抗胡椒鹽雜訊並進行回復，平均回復的效果比傳統中值濾波器、自適性中值濾波器好。

Nowadays, digital information transmits widely through the Internet. Anyone could retrieve information easily, and modify information for commercial purposes. In light of these situations, more and more intellectual property issues occurs. One of the solutions to this issue is applying digital watermark, which is used to examine whether the copyright of an object is infringed. In this thesis, we focus on using watermark to protect image, such that the tampered part is detected and recovered.
In this thesis, we propose a method to examine whether an image is tampered or not. If an image is tampered, watermark will recover the tampered image. At first, we use SPIHT (set partitioning in hierarchical trees) technology to compress image. The compression property of SPIHT is very good when no error occurs in the SPIHT bit stream. Using only a small percentage of the SPIHT bit stream in decoding, the decoded image is highly similar to the original image. However, when there is error in the SPIHT bit stream, the recovered image can be bad. A SPIHT bit stream can be divided data packets of some length. In this thesis, we use QR-code error correction technology to protect SPIHT bit stream. In other words, we encode SPIHT data packets into QR codes. There are four error correction levels (H,Q,M,L) in QR-code. How to use error correction level efficiently is a key issue in this work. In our scheme, we make suitable error correction level the assignment to the QR-coded SPIHT packets. Then, serving as a watermark, those QR-coded SPIHT packets are embedded into the LSB (least significant bit) bit1 and bit2 of the original image. When we want to recover image, we extract the watermark. We used standard QR-decoder to decode and get back the SPIHT bit stream. Finally, the SPIHT-decoded image is compared to the received image. Big-difference locations indicate tampering. The tampered part will be replaced by the SPIHT counterpart.
Experimental results show that the method we proposed can detect digital image which is tampered and we can recover it. Recovered image can achieve expected image quality. Moreover, our scheme removes salt and pepper noise, and is better than median filter and adaptive median filter.

摘　要 i
Abstract ii
致謝 iv
目　錄 v
圖索引 vii
表索引 ix
中英文對照表 x
符號索引 xii
第一章 緒論 1
1.1前言 1
1.2研究動機 1
1.3論文架構 2
第二章 相關技術介紹 3
2.1二維離散小波轉換 3
2.2 SPIHT演算法 4
2.2.1 SPIHT編碼 5
2.2.2 SPIHT解碼 8
2.2.3 SPIHT壓縮效能 9
2.2.4 SPIHT漸進式解碼 9
2.2.5 SPIHT缺點 10
2.3 QR碼簡介 12
第三章 系統架構 16
3.1結合SPIHT與QR碼 17
3.1.1 QR碼錯誤更正能力 17
3.1.2問題公式化 18
3.2錯誤更正級別配置(全域搜尋) 19
3.3嵌入QR碼浮水印方式 25
3.4萃取QR碼浮水印方式 34
3.5圖示錯誤率(pattern error rate PER) 37
3.6加速錯誤更正級別配置(次佳解suboptimal)分段收尋 39
3.7竄改位置偵測機制與影像復原 45
3.7.1竄改位置偵測機制 45
3.7.2影像復原機制 49
第四章 實驗結果與討論 51
4.1參數設定 51
4.2 SNR與PSNR轉換 52
4.3理論PSNR值與模擬PSNR值誤差 53
4.4影像還原與去除胡椒鹽雜訊流程 55
4.5實際模擬數據 58
第五章 結論與未來展望 67
參考文獻 69

[1] A. Said and W. A. Pearlman, “A new, fast and efficient image code based on set partitioning in hierarchical trees,” IEEE Trans. On Circuits and Systems for Video Technology, vol. 18,no. 6,pp. 891-898,June 2000.
[2] K.T. Lay and L.J. Wang, “Packetization and unequal erasure protection for transmission of SPIHT-encoded images,” IEICE Trans. On Communications, vol. E97-B,no. 1,pp. 226-237, January 2014.
[3] 周佳怡, ‘‘用於SPIHT影像封包之不等錯誤保護傳輸技術,’’ 國立台灣科技大學電子工程研究所, 2011.
[4] K.C Liu, “Self-embedding watermarking scheme for colour images by bi-level moment-preserving technique,” IET image Processing, vol. 8, iss. 6,pp. 363-372, June 2014.
[5] T.Y. Lee and S.D. Lin, “Dual watermark for image tamper detection and recovery,” ELSEVIER Pattern Recognition, vol. 41, iss. 11,pp. 3497-3506, November 2008.
[6] P.L. Lin,C.K. Hsieh and P.W. Huang, “A hierarchical digital watermarking method for image tamper detection and recovery,” ELSEVIER Pattern Recognition, vol. 38, iss. 12, pp. 2519-2529, December 2005.
[7] S.B. Chaluvadi and M.V.N.K. Prasad, “Efficient image tamper detection and recovery technique using dual watermark,” Proc. IEEE Int. Conf. Nature & Biologically Inspired Computing, 2009. NaBIC 2009. World Congress on, vol. 1, pp. 993-998, Dec. 2009.

[8] N. Wang and C.H. Kim, “Tamper detection and self-recovery algorithm of color image based on robust embedding of dual visual watermarks using DWT-SVD,” Proc. IEEE Int. Conf. Communications and Information Technology, 2009. ISCIT 2009. 9th International Symposium on, vol. 1, pp. 157-162, Sept. 2009.
[9] Q. Song and H.B. Zhang, “Image tamper detection and recovery using dual watermark,” Proc. IEEE Int. Conf. Wireless Communications Networking and Mobile Computing (WiCOM), 2010 6th International Conference on, vol. 1,pp. 1-4 , Sept. 2010.
[10] M.J. Chen and X.B. Sun, “A digital image watermarking of self-recovery base on the SPIHT algorithm,” Proc. IEEE Int. Conf. Signal Processing Systems (ICSPS), 2010 2nd International Conference on, vol. 2,pp. 621-624, July 2010.
[11] R.C. Gonzalez and R.E. Woods, “Digital Image Processing third edition,” Prentice Hall, 2008
[12] “Information technology – automatic identification and data capture techniques – bar code symbology – QR code, ”International Standard ISO/IEC 18004, ISO/IEC, 2006