抗合謀攻擊之數值高程模型浮水印設計

A Collusion-Resistant Watermarking Scheme for Digital Elevation Models

林政毅(Jheng-Yi Lin)

指導教授 ： 蔡展榮

國立成功大學/工學院/測量及空間資訊學系/碩士(2018年)

摘要

本論文主旨在擬定一個實用的數位浮水印方案，並可應用於數值高程模型(DEM)之來源追蹤，來源追蹤與版權保護不同的是，來源追蹤需面臨多名使用者的合謀攻擊。為了抵抗合謀攻擊，本文採用獨立同分布之高斯隨機向量作為浮水印以抵抗合謀攻擊，並以統計機率分析其抵抗能力。為了要權衡浮水印之強健性(Robustness)、不可視性(Imperceptibility)、保真度(Fidelity)與安全性(Security)，本文依內政部千分之一數值高程模型測製規範建立具有保真度之遮罩(fidelity mask)，並以三次樣條內插法內插出合理且平滑的遮罩，而浮水印嵌入強度則是根據遮罩值來決定。本文是以多層面向的評估方法分析浮水印DEM的品質與可應用性，評估方法包括遮罩、高程、坡度、坡向、地表起伏度、地表切割深度，針對原始DEM與浮水印DEM的差值分析其最大最小誤差、平均誤差、均方根誤差。在不可視性測試方面，評估方法包括峰值信號雜訊比、高程影像、梯度影像與等高線。另外，浮水印強健性測試採取的模擬攻擊包括：隨機亂數、均值濾波、中值濾波、裁減、捨位、高程平移、高程縮放，與其他現有文獻不一樣的是，高程平移是正高與橢球高的高程轉換，而非平移一固定值。本文按照DEM的高程與平均坡度將其分類為平原、盆地、台地、丘陵、山地，再從所有6326個DEM中選出具有代表性的8個DEM作為實驗資料。
基於強健性、不可視性、保真度與安全性的實驗成果，總結如下：(1)本文的浮水印設計對DEM常見的信號處理具有強健性。(2)本文的浮水印設計可抵抗平均合謀攻擊與複合型合謀攻擊。(3)人眼難以從浮水印DEM之高程、梯度與等高線之中感知到浮水印的存在。(4)本文以多層面向的評估成果確立浮水印DEM的保真度。
本研究最大貢獻是將Cox等人於1997年提出的獨立同分布之高斯隨機向量，運用於DEM浮水印設計以抵抗合謀攻擊，並檢驗其安全性，且此設計可應用於DEM來源追蹤。

關鍵字

浮水印 ； 數值高程模型 ； 合謀攻擊 ； 來源追蹤

並列摘要

The thesis provides a practical scheme for the application in DEM transaction tracking, which has the task different from application of DEM copyright protection. The significant difference of the task is the resistance to collusion attacks. An idea of independent and identically distributed (i.i.d.) Gaussian random vectors, which has robustness against not only common signal processes but also collusions attacks, is implemented in the thesis, and the Security of the proposed scheme is evaluated by the probability of detection and probability of false positive. The other task, a common issue of watermarking schemes, is the trade-off among the Imperceptibility, Fidelity, Robustness, and Security. To balance the trade-off, the regulation of accuracy requirement for DEM generated in 1/1000 topographic mapping is developed and a cubic spline interpolation is implemented to generate the smooth and rational Human Visual System (HVS) masking called fidelity mask. The differences between original DEMs and marked DEMs are measured to evaluate the fidelity, and they are examined in multidimensional evaluations for the quality of DEMs and their applicability. The Imperceptibility is evaluated by the Peak Signal Noise Ratio (PSNR), contours, images of elevation and gradient. On the other hand, the experiments for Robustness including random noise, mean filter, medium filter, clipping, rounding down, elevation shifting and scaling are executed, especially elevation shifting which is an elevation transformation from the orthometric height to ellipsoidal height. DEMs are classified into plain, basin, tableland, hill and mountain according to its elevation and average of slope. After classifying, the 8 experimental DEMs which are diverse, comprehensive, and representative are selected from all 6326 DEMs.
On the basis of the experiments of DEMs containing Fidelity, Imperceptibility, Robustness and Security, the research can be concluded: (1) The proposed scheme has robustness against common signal processes to DEMs. (2) The proposed scheme is secure against the average collusion attack and complex collusion attack. (3) The watermarks cannot be perceived by human eyes in the marked contours, elevation and gradient images. (4) The results of fidelity are supported by multidimensional evaluations.
The major contribution of the thesis is the proposed watermarking scheme utilizing i.i.d. Gaussian random vectors proposed by Cox et al. in 1997 is genuinely resistant to average collusion attack for DEMs in transaction tracking.

並列關鍵字

Watermarking ； Digital elevation model ； Collusion attack ； Transaction tracking ； Fingerprinting

參考文獻

1. Boneh, D., & Shaw, J. (1995). Collusion-secure fingerprinting for digital data. In Annual International Cryptology Conference, Springer, Berlin, Heidelberg, pp. 452-465.

2. Burrough, P. A., McDonnell, R. A., & Lloyd, C. D. (2015). Principles of geographical information systems. Oxford University Press.

3. Chou, C. H., & Li, Y. C. (1995). A perceptually tuned subband image coder based on the measure of just-noticeable-distortion profile. IEEE Transactions on circuits and systems for video technology, 5(6), pp. 467-476.

4. Cox, I. J., Kilian, J., Leighton, F. T., & Shamoon, T. (1997). Secure spread spectrum watermarking for multimedia. IEEE transactions on image processing, 6(12), pp. 1673-1687.

5. Cox, I., Miller, M., Bloom, J., Fridrich, J., & Kalker, T. (2007). Digital watermarking and steganography. Morgan Kaufmann Publishers.