請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86569
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 吳家麟(Ja-Ling Wu) | |
dc.contributor.author | Wen-Cheng Tsou | en |
dc.contributor.author | 鄒汶橙 | zh_TW |
dc.date.accessioned | 2023-03-20T00:03:47Z | - |
dc.date.copyright | 2022-08-31 | |
dc.date.issued | 2022 | |
dc.date.submitted | 2022-08-11 | |
dc.identifier.citation | [1] J Fridrich. Symmetric ciphers based on two-dimensional chaotic maps. International Journal of Bifurcation and Chaos, 8:1259–1284, 1998. [2] OLCAY TANER YILDIZ ERCAN SOLAK, CAHIT ÇOKAL and TÜRKER BIYIKOĞLU. Cryptanalysis of fridrich’s chaotic image encryption. International Journal of Bifurcation and Chaos, 20(5):1405–1413, 2010. [3] Eric Yong Xie, Chengqing Li, Simin Yu, and Jinhu Lü. On the cryptanalysis of fridrich's chaotic image encryption scheme. Signal Processing, 132:150–154, mar 2017. [4] Chengqing Li, Dongdong Lin, Bingbing Feng, Jinhu Lu, and Feng Hao. Cryptanalysis of a chaotic image encryption algorithm based on information entropy. IEEE Access, 6:75834–75842, 2018. [5] Guodong Ye, Chen Pan, Xiaoling Huang, Zhenyu Zhao, and Jianqing He. A Chaotic Image Encryption Algorithm Based on Information Entropy. International Journal of Bifurcation and Chaos, 28(1):1850010–750, January 2018. [6] Sun Fu-Yan, Liu Shu-Tang, and Lü Zong-Wang. Image encryption using highdimension chaotic system. Chinese Physics, 16:3616, 12 2007. [7] Yueping Li, Chunhua Wang, and Hua Chen. A hyper-chaos-based image encryption algorithm using pixel-level permutation and bit-level permutation. Optics and Lasers in Engineering, 90:238–246, 2017. [8] Karim H. Moussa, Ahmed I. El Naggary, and Heba G. Mohamed. Non-linear hopped chaos parameters-based image encryption algorithm using histogram equalization. Entropy, 23(5), 2021. [9] Philippe M. Binder and Roderick V. Jensen. Simulating chaotic behavior with finitestate machines. Phys. Rev. A, 34:4460–4463, Nov 1986. [10] SHUJUN LI, GUANRONG CHEN, and XUANQIN MOU. On the dynamical degradation of digital piecewise linear chaotic maps. International Journal of Bifurcation and Chaos, 15(10):3119–3151, 2005. [11] Zbigniew Galias. The dangers of rounding errors for simulations and analysis of nonlinear circuits and systems?and how to avoid them. IEEE Circuits and Systems Magazine, 13(3):35–52, 2013. [12] Chengqing Li, Dongdong Lin, Bingbing Feng, Jinhu Lü, and Feng Hao. Cryptanalysis of a chaotic image encryption algorithm based on information entropy. IEEE Access, 6:75834–75842, 2018. [13] Chengqing Li, Bingbing Feng, Shujun Li, Jüergen Kurths, and Guanrong Chen. Dynamic analysis of digital chaotic maps via state-mapping networks. IEEE Transactions on Circuits and Systems I: Regular Papers, 66(6):2322–2335, 2019. [14] Zbigniew Galias. Periodic orbits of the logistic map in single and double precision implementations. IEEE Transactions on Circuits and Systems II: Express Briefs, 68(11):3471–3475, 2021. [15] Chunlei Fan and Qun Ding. Analysing the dynamics of digital chaotic maps via a new period search algorithm. Nonlinear Dynamics, 97, 07 2019. [16] Martin Hanik, Mehmet Arif Demirtaş, Mohammed Amine Gharsallaoui, and Islem Rekik. Predicting cognitive scores with graph neural networks through sample selection learning. Brain Imaging and Behavior, 16(3):1123–1138, nov 2021. [17] Thomas N. Kipf and Max Welling. Semi-supervised classification with graph convolutional networks, 2016. [18] William L. Hamilton, Rex Ying, and Jure Leskovec. Inductive representation learning on large graphs, 2017. [19] Rasmus Rothe, Radu Timofte, and Luc Van Gool. Dex: Deep expectation of apparent age from a single image. In 2015 IEEE International Conference on Computer Vision Workshop (ICCVW), pages 252–257, 2015. | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86569 | - |
dc.description.abstract | 混沌映射性近幾年被廣泛應用在密碼學上,主要仰賴其確定性、非週期性、及對初始值的敏感性。而在影像加密的領域中,混沌映射更是被套用為在空間域中的加密元件,使其成為一種偽隨機亂數生成器(PRNG)。跟傳統的影像加密方法相比,這種加密方式提供的最大好處是不會讓密文檔案大小過度膨脹,並且能提供極高速的加密速度,在現今要求即時多媒體應用的需求下,混沌映射是非常好的解藥。在此論文中,我們提出了一種以圖神經網路為基底的攻擊手法,透過針對特定混沌映射,在有限精度下設計的先備知識訓練集,預訓練出一個能夠迴歸參數的圖神經網路模型。攻擊時我們假設攻擊者能夠竊聽到一部分的混沌序列,並將該序列輸入模型進行迴歸,進而推導出出該混沌序列所對應到的參數,此參數在影像加密演算法中即相當於金鑰。換言之,若此法順利迴歸出混沌序列之參數,則表示攻擊(或破密)成功 | zh_TW |
dc.description.abstract | Owing to their deterministic, nonperiodic, and sensitive to initial state characteristics, chaotic maps are widely applied to cryptography in recent years. In chaos-based image encryption schemes, chaotic maps behave like a pseudo random number generator for encrypting digital images in the spatial domain. Contrary to traditional methods, the major advantage of chaos-based image encryption algorithm is its high encrypt speed and limited size expansion of the cipherimage. Thanks to this advantages, chaos-based image encryption algorithm has become a great remedy in real-time secure multimedia application areas nowadays. In this work, by designing a flexible and novel algorithm to generate appropriate training set based on the domain knowledge of specific chaotic maps in finite precision and pre-training a Graph Neural Network model for conducting parameters regressions, a GNN based parameter regression attack is proposed. To launch the attack, we assume that attacker has ability to eavesdrop a portion of the chaotic sequence, and the captured sequence segment is then input to the pre-tarined model we mentioned above for getting the corresponding parameter. The obtained parameter is equivalent to the key in a chaos-based image encryption algorithm. If the parameter can be found by the network, we say the attack is successful. | en |
dc.description.provenance | Made available in DSpace on 2023-03-20T00:03:47Z (GMT). No. of bitstreams: 1 U0001-2607202216052200.pdf: 2312347 bytes, checksum: 56257ebff518babd461f1b6a594f1c87 (MD5) Previous issue date: 2022 | en |
dc.description.tableofcontents | 目錄 口試委員審定書 i 致謝 iii 摘要 v Abstract vii 目錄 ix 圖目錄 xiii 表目錄 xv 第一章 簡介 1 第二章 背景介紹和攻擊者設定 3 2.1 混沌映射影像加密演算法 . . . . . . . . . . . . . . . . . . . . . . . . 3 2.2 單峰映射 (logistic map) . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.3 有限精度下的混沌映射 . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.4 攻擊者模型 (Adversary model) . . . . . . . . . . . . . . . . . . . . . 6 第三章 訓練集生成演算法 7 3.1 可接受區間 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3.2 訓練資料集生成 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.2.1 GenerateSMN 函數 . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.2.2 CalculateMPL 函數 . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.2.3 GetMeanofMPL 函數 . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.2.4 GetMaxofMPL 函數 . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.2.5 CalCleanPro 函數 . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.2.6 VisitfromCand 函數 . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.2.7 分配參數值 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.3 測試資料集生成 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 第四章 模型與實驗設計 13 4.1 圖神經網路模型 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 4.1.1 版本一 (Model 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 4.1.2 版本二 (Model 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 第五章 實驗結果 17 5.1 資料集設定 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 5.2 模型一實驗結果 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 5.2.1 不同精度下的學習曲線 (learning curve) . . . . . . . . . . . . . . 19 5.2.2 訓練趨於穩定後損失函數之比較 . . . . . . . . . . . . . . . . . . 21 5.3 模型二實驗結果 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 5.3.1 不同精度下的學習曲線 (learning curve) . . . . . . . . . . . . . . 23 5.3.2 訓練趨於穩定後損失函數之比較 . . . . . . . . . . . . . . . . . . 26 5.4 模型二在資訊受限之表現 . . . . . . . . . . . . . . . . . . . . . . . . 27 5.5 gv 和 cp 在模型二對實驗各項指標的影響 . . . . . . . . . . . . . . . 30 5.5.1 固定 cp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 5.5.2 固定 gv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 5.6 實際攻擊案例及分析 . . . . . . . . . . . . . . . . . . . . . . . . . . 34 5.6.1 簡易影像加密演算法 . . . . . . . . . . . . . . . . . . . . . . . . 34 5.6.1.1 GenerateCS 函數 . . . . . . . . . . . . . . . . . . . . 35 5.6.1.2 Confusion 函數 . . . . . . . . . . . . . . . . . . . . . 35 5.6.1.3 Diffusion 函數 . . . . . . . . . . . . . . . . . . . . . . 35 5.6.2 攻擊流程 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 5.6.3 金鑰空間縮減分析 . . . . . . . . . . . . . . . . . . . . . . . . . . 36 第六章 結論 39 參考文獻 41 | |
dc.language.iso | zh-TW | |
dc.title | 針對有限精度單峰映射影像加密演算法之圖神經網路參數迴歸攻擊 | zh_TW |
dc.title | A GNN Based Parameter Regression Attack on Image Encryption Algorithm Using Logistic Map under Finite Precision Constraints | en |
dc.type | Thesis | |
dc.date.schoolyear | 110-2 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 許超雲(Chau-Yun Hsu),陳駿丞(Jun-Cheng Chen),陳文進(Wen-Chin Chen) | |
dc.subject.keyword | 混沌映射,有限精度,圖神經網路,參數迴歸, | zh_TW |
dc.subject.keyword | Chaotic maps,Finite precision,Graph neural network,Parameter regression, | en |
dc.relation.page | 59 | |
dc.identifier.doi | 10.6342/NTU202201736 | |
dc.rights.note | 同意授權(全球公開) | |
dc.date.accepted | 2022-08-12 | |
dc.contributor.author-college | 電機資訊學院 | zh_TW |
dc.contributor.author-dept | 資訊工程學研究所 | zh_TW |
dc.date.embargo-lift | 2022-08-31 | - |
顯示於系所單位: | 資訊工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-2607202216052200.pdf | 2.26 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。