臺灣博碩士論文加值系統

隨著網際網路的迅速發展，人們使用網路日趨增加，但同時也存在著不安全性，部分有心人士利用“全自動機器人程式”竊取或濫用網路資源，造成網路效能及安全性大幅的遞減。為了避免全自動機器人程式的惡意破壞與盜取，多數系統都會套用一種防禦機制，來辨識使用者為人或機器的全自動公開圖靈測試系統（Completely Automated Public Test to tell Computers and Humans Apart ，簡稱CAPTCHA）。目前有各種類型的CAPTCHA廣泛的被使用，當中以文字型態的驗證碼最為普遍，但也隨著光學字元辨識 (Optical Character Recognition，簡稱OCR)的技術漸漸地提升，導致驗證碼面臨遭到破解的威脅，安全性也大幅下降，因此近年來CAPTCHA已逐漸朝向影像、語音等型式發展。因此本篇論文將針對圖像型態CAPTCHA去研究，開發出一個讓使用者可以容易的辨識，又不易被自動化程序惡意攻擊的馬賽克圖像CAPTCHA。根據實驗結果顯示，我們的方法其平均辨識率達93%，平均驗證時間為13.63秒，符合一般驗證碼設計的需求。

With the rapid development of the Internet, people utilize the Internet frequently. However, the Internet is insecure. Some people use "Automatic robot program" to steal or abuse the Internet resources intentionally. This misuse can reduce the network performance and security drastically. In order to avoid this circumstance, a CAPTCHA (Completely Automated Public Test to tell Computers and Humans Apart) mechanism is used to identify whether the user is a person or machine. There are many types of CAPTCHA over the Internet, and the text-based CAPTCHA is widely used. However, the technology of Optical Character Recognition (OCR) has gradually improved. The text-based CAPTCHA confronts the decrypt threat by the new OCR technology. Because of the threat, the image-based CAPTCHA and audio CAPTCHA are proposed. Therefore, the main purpose of this paper is to study the Image-based CAPTCHA. A mosaic image CAPTCHA is proposed that allows users to recognize the image easily and can resist the attacks by automatic robot program. According to the experimental results, the average recognition rate is up to 93% and the average recognition time is 13.63sec. The test results show that the proposed mosaic image CAPTCHA meets the requirements of general CAPTCHA design.

目錄
摘要 I
Abstract II
誌謝 III
目錄 V
圖目錄 VII
表目錄 IX
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機與目的 3
1.3 論文架構 5
第二章 文獻探討 6
2.1 驗證碼的發展 6
2.2驗證碼型態 8
2.2.1文字型態CAPTCHA 8
2.2.2圖像型態CAPTCHA 9
2.2.3語音型態CAPTCHA 17
2.3驗證碼的破解技術 17
2.3.1光學字元辨識技術 18
2.3.2暴力破解法 18
2.3.3資料庫破解法 18
2.3.4字典攻擊法 18
第三章 研究方法 19
3.1系統設計概念 19
3.2系統架構 19
3.3資料庫設計欄位 20
3.4馬賽克圖像驗證碼的處理 21
第四章 實驗結果 24
4.1實驗環境與流程 24
4.2實驗結果與分析 27
4.2.1比較分析 27
4.2.2自動性分析 29
4.2.3可用性分析 30
4.2.4開放性分析 30
4.2.5安全性分析 30
第五章 結論以及未來工作 32
第六章 參考文獻 34


圖目錄
圖1.1：Google網站使用的文字型態CAPTCHA 2
圖1.2：Kitten Auth CAPTCHA 2
圖1.3：Audio CAPTCHA 3
圖1.4：過於扭曲文字型態CAPTCHA 4
圖2.1：Gimpy CAPTCHA 9
圖2.2：Esp-Pix CAPTCHA 10
圖2.3：Assira CAPTCHA 11
圖2.4：What’s up CAPTCHA介面圖 12
圖2.5：吉他的正確方向是非常主觀，並無法說哪一個方向是對的 12
圖2.6：Multi-Modal CAPTCHA 13
圖2.7：Puzzle CAPTCHA 14
圖2.8：Poker為圖像驗證碼 15
圖2.9：Compare CAPTCHA 16
圖2.10：Audio CAPTCHA 17
圖3.1：資料庫各種類名稱 20
圖3.2：馬賽克圖像驗證碼收集分類流程步驟圖 21
圖4.1：實驗問卷流程圖 25
圖4.2：各平均時間 27
圖5.1：加入線條干擾 33



表目錄
表1.1：YAHOO使用之易混淆CAPTCHA 4
表3.1：圖像分類表 22
表4.1：實際畫面說明表圖 26
表4.2：圖像驗證碼比較表 29

[1]巫愛雯(2011)，影像驗證碼技術之研究，碩士論文，台中科技大學多媒體設計系(未出版) ，台中 。
[2]莊凱婷(2013)，以 Poker 為圖像驗證碼之設計研究，碩士論文，交通大學應用藝術研究所，新竹。
[3]郭璟逸(2015)，影像驗證碼應用於行動裝置之研究，碩士論文，臺中科技大學多媒體設計系，台中。
[4]劉晉昇(2011)，一個可自動增加驗證圖庫的圖像CAPTCHA系統，碩士論文，逢甲大學資訊工程所，台中。
[5]賴怡靜(2002)，以錯視為基礎之動態影像驗證碼研究，碩士論文，臺中科技大學多媒體設計系，台中。
[6]L. V. Ahn, M. Blum, N. J. Hopper and J. Langford(2003), “CAPTCHA: Using hard AI problems for security,” In International Conference on the Theory and Applications of Cryptographic Techniques, pp. 294-311.
[7]A. S. Almazyad, Y. Ahmad, and S. A. Kouchay(2011), “Multi-Modal CAPTCHA: A User Verification Scheme,” IEEE International Conference on Information Science and Applications, pp. 1-7.
[8]H. S. Baird, A. L. Coates and R. J. Fateman(2003), “Pessimalprint: A Reverse Turing Test,” Proceedings of 6th International Conferee on Document Analysis and Recognition, pp.154-158.

[9]J. Elson, J. R. Douceur, J. Howell and J. Saul(2007), “Asirra: a CAPTCHA that exploits interest-aligned manual image categorization,” In Proceedings of ACM CCS, pp.1-16.
[10]P. Golle(2008), “Machine learning attacks against the Asirra CAPTCHA,” In Proceedings of the 15th ACM conference on Computer and communications security, pp. 535-542.
[11]R. Gossweiler, M, Kamvar and S. Baluja (2009), “What's up CAPTCHA? a CAPTCHA based on image orientation,” In Proceedings of the 18th international conference on World wide web, pp.841-850.
[12]C. Higgins(2008), “CAPTCHA Variants: KittenAuth,” retrieved from: http://mentalfloss.com/article/18795/captcha-variants-kittenauth(2019/06/18)
[13]C. J. Lee, W. B. Lee, C. S. Liu, K. Ho and C. R. Dow(2012), “A Novel Friendly Image-Based CAPTCHA With Auto-Generation of Test Data,” In Proceedings of the 7th International Conference on Information Warfare and Security: ICIW, pp. 158.
[14]M. Mehrnejad, A. G. Bafghi and A. Harati, E. Toreini(2012), “SEIMCHA: a new semantic image CAPTCHA using geometric transformations,” The ISC International Journal of Information Security, pp.63-76.
[15]G. Mori and J. Malik(2003), “Recognizing objects in adversarial clutter: Breaking a visual CAPTCHA,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp.1-8.
[16]G. Moy, N. Jones, C. Harkless and R. Potter(2004), “Distortion estimation techniques in solving visual CAPTCHAs,” In Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 1-6.
[17]C. Pope and K. Kaur(2005), “Is it human or computer? Defending e-commerce with captchas,” IT professional, pp.43-49.
[18]J. Thatcher(2009), “CAPTCHAs, CAPTCHAs everywhere,” retrieved from: https://jimthatcher.com/captchas.htm(2019/06/18)
[19]A. M. Turing(2009), “Computing machinery and intelligence,” In Parsing the Turing Test. Springer, Dordrecht, pp.23-65.
[20]J. Yan and A. E. Ahmad(2008), “Is cheap labour behind the scene? - Low-cost automated attacks on Yahoo CAPTCHAs,” Proceedings of the 15th ACM conference on Computer and communications security, pp.543-554.