Search

Abstract: Color edge detection is much more efficient than gray scale detection when edges exist at the boundary between regions of different colors with no change in intensity. This paper presents adaptive templates, which are capable of detecting various color and intensity changes in color image. To avoid conception of multilayer proposed in literatures, modification has been done to the CNN structure. This modified structure allows a matrix C, which carries the change information of pixels, to replace the control parts in the basic CNN equation. This modification is necessary because in multilayer structure, it faces the challenge of how to represent the intrinsic relationship among each primary layer. Additionally, in order to enhance the accuracy of edge detection, adaptive detection threshold is employed. The adaptive thresholds are considered to be alterable criteria in designing matrix C. The proposed synthetic system not only avoids the problem which is engendered by multi-layers but also exploits full information of pixels themselves. Experimental results prove that the proposed method is efficient. [Copyright &y& Elsevier]

Abstract: The security of digital image attracts much attention recently. A hash-based digital image encryption algorithm has been proposed in Ref. . But both the theoretical analysis and computer simulation show the characteristic of diffusion is too weak to resist Chosen Plaintext Attack and Known Plaintext Attack. Besides, one bit difference of the plain pixel will lead to only one corresponding bit change of the cipher pixel. In our improved algorithm, coupled with self-adaptive algorithm, only one pixel difference of the plain-image will cause changes of almost all the pixels in the cipher-image (NPCR>98.77%), and the unified average changing intensity is high (UACI>30.96%). Both theoretical analysis and computer simulation indicate that the improved algorithm can overcome these flaws and maintain all the merits of the original one. [Copyright &y& Elsevier]

Abstract: The construction of a new Hash function attracts much attention recently. In Kwok and Tang (2005) [Kwok HS, Tang WKS. A chaos-based cryptographic Hash function for message authentication. Int J Bifurcat Chaos 2005;15:4043–50], a chaos-based Hash function has been proposed. In this paper, the potential flaws in the original algorithm are analyzed in detail, and then the corresponding improving measures are proposed. We enhance the influence that each bit of the final Hash value is closely related to all the bits of the message or key and a single bit change in message or key results in great changes in the final Hash value. Simulation results show that the proposed improving algorithm has strong diffusion and confusion capability, good collision resistance, extreme sensitivity to message and secret key. [Copyright &y& Elsevier]

Abstract: A novel combined cryptographic and hash algorithm based on chaotic control character is proposed in this paper. The control character is generated by chaotic iteration. The plaintext is pre-processed in terms of control character, and then encrypted by the look-up index table. At the same time, the chaotic trajectory is changed continuously according to the control character, which can avoid the dynamical degradation of chaos. Besides, the look-up index table is updated by utilizing the control character continuously, and the index item of the final look-up index table can be considered as the hash value of the whole paragraph of plaintext. Therefore, the proposed algorithm can perform both encryption/decryption and hash in a combined manner. Compared with Wong’s and our former algorithms, the proposed one has decreased the ratio of ciphertext/plaintext to 1, and improved the efficiency greatly. Theoretical analysis and computer simulation verify the validity of the proposed algorithm. [Copyright &y& Elsevier]