於視覺密碼學中詐欺攻擊防禦機制之研究 On the Cheating Prevention Schemes in the k-out-of-n Visual Cryptography
- 發布日期：
- 最後更新日期：109-05-13
- 資料點閱次數：242
:::
Naor 和 Shamir兩位學者於1994年提出k-out-of-n的視覺密碼學，在此架構中唯有疊和k張以上的投影片才能以視覺回復機密影像。由於視覺密碼學的安全性及不需運算解密的特性，應用於許多方面如：視覺認證、資訊隱藏及影像加密。
這篇論文提出於視覺密碼學中詐欺攻擊的明確定義，定義包含了以人類視覺為方向的可信條件、亮度條件及安全條件。也提出了詐欺攻擊的三種攻擊分類，分別為已知加密法、已知投影片及已知機密影像三種分類。論文中以一個合理的情況及演算法說明，詐欺攻擊確實會危害其它的參與者。在詐欺攻擊活動的分析中，CA-1及CA-2詐欺攻擊活動並不適用於k-out-of-n的門檻。而論文所提出的詐欺攻擊活動被證明為有效的攻擊。
論文最後提出於視覺密碼學中詐欺攻擊的四種防禦機制，分別為零位元增加防禦機制、認證防禦機制、2-out-of-(n+l) VC 防禦機制及多重機密影像防禦機制。這四種防禦機制都被證實能有效防止詐欺攻擊，而另一HTCP防禦機制於論文所提出的過濾攻擊下並無法提供有效的防禦。論文所提出的四種防禦機制也被證實能逹到和視覺密碼學相同的安全等級。
In 1994, Naor and Shamir proposed the k-out-of-n Visual Cryptography(VC) scheme such that only more than or equal to k participants can visually recover the secret through superimposing their transparencies. Due to the unconditional security and the property of decoding without computation, VC has been applied to numerous applications such as visual authentication and identification, steganography, and image encryption.
The dissertation proposed an explicit definition of cheating in the k-out-of-n VC. In VC, the property of applying of Human Visual System (HVS) for decoding distinguishes VC from secret sharing schemes. According to the property, the definition consists of authentic conditions, brightness condition, and security condition. Moreover, three cheating attack models, encryption algorithm only, known transparency, and known secret, are proposed based on the amount of the information known to the cheater. A reasonable scenario and the algorithm of the proposed cheating activity are used to show that cheating is possible in VC. According to the definition of cheating, Hu and Tzeng’s cheating activities, CA-1 and CA-2, are not considered as a successful cheating activity in the k-out-of-n VC.
The dissertation then proposed four cheating prevention schemes. The zero-bit appended cheating prevention scheme, the authentication based cheating prevention scheme (ABCP), the 2-out-of-(n+l) VC cheating prevention scheme, and the homogenous secret images based cheating prevention scheme (HBCP). The proposed four cheating prevention schemes are shown to provide well solutions to deal with cheating attacks. With respect to Hu and Tzeng’s cheating prevention scheme (HTCP), it suffers cheating activities after the proposed basis matrix filtering attack (BMFA). On the contrary, the proposed four schemes, ZACP, 2-out-of-(n+l) VC, ABCP, and HBCP, are proven as secure as the VC scheme.
資料來源：http://www.airitilibrary.com/Publication/alDetailedMesh1?DocID=U0005-0607200714073300