Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/19532
Title: Cryptosystems Designs over Random Number Generator and Chaos Theory: Image Encryption Applications
Authors: Saleem Malik, Dania
Keywords: Mathematics
Issue Date: 2021
Publisher: Quaid-i-Azam University Islamabad
Abstract: Due to rapid developments in communicational networks, transmission of massive data by means of them has increased risk of security in multimedia information. For protecting information, data encryption plays a significant role. So, cryptography provides a platform to secure information. Cryptographic modern approaches are mainly focused on block-based encryption techniques which includes “Advance Encryption Standard” and “Data Encryption Standard”, abbreviated as “AES and DES”. But these modern approaches are valuable for textual information. However, owing to massive data transmission in the form of images, these techniques have failed in providing sufficient security due to lack of their efficiency for bulk data and randomness. Therefore, security of images is very challenging. Researchers have been concerned more than ever to secure multimedia information with modern and effective content preservation strategies to face this challenge. From past few decades, many techniques to encrypt images were proposed, and it is found that the encryption techniques based on chaos are most effective. Since the chaotic non-linear systems have some important features like initial and parametric values, sensitivity, randomness, and unpredictability which renders them suitable for encryption of images. In this perspective, the need of hour is to design secure chaos-based cryptosystems which ensure authentication, confidentiality, and integrity of data before transmission. The chaos-based systems offer an appropriate source for abundant pseudorandom sequences generation and are useful to construct encryption nonlinear components. The Substitution boxes commonly known as S-boxes are crucial non-linear components in block ciphers, which plays a key role in multimedia cryptosystems security. The nonlinear S-boxes provide effective security to cryptosystems. For this reason, many encryption schemes for protecting image information use S-boxes in substitution phase where the values of plaintext are replaced with S-boxes to enhance the confusion and make it attack resistant. Moreover, our focus is on confidentiality of content and controlling access that is addressed in encryption, which is only accessible to the parties who have keys for decryption to read the transmitted content (plain-image). The proposed thesis objective is to design cryptosystems that use chaos and random numbers for strong encryption techniques to obtain ciphered images having excellent attack resistant properties. In this perspective, the random numbers and chaos-based techniques have been addressed. Firstly, a hybrid-pseudorandom binary numbers generation (abbreviated as HPRNG’s) is proposed. The HPRNG’s is obtained by feedback shift registers in combination with modified quadratic chaotic map. Since the numbers generated by only feedback shift registers are easily hacked and are not having strong resistance so the modified quadratic chaotic map bits are added to feedback shift register in every cycle. Then the binary streams are used in construction of cryptographically strong block ciphers (S-boxes). The S-boxes are analyzed and tested using different testing techniques commonly used in literature like Nonlinearity, Avalanche criterion, Bit independency etc., which depicts their strong performance. After this, a permutation substitution-based encryption technique is presented. As in literature review, a lot of techniques based on permutation substitution phases have lack of key and image pixels association property, so they are easily hacked. This flaw is overcome in proposed encryption technique by creating a relation between the image pixels and keys. Secondly, in design of strong cryptosystems, key is the main component. Because if the hacker gets any idea about the key, he can easily hack the algorithm. So, the focus of work is on the generation of strong keys. For this, some encryption techniques based on game rules (like chess board game knight movement) and convolution codes along with chaos are presented. Apart from single encryption techniques, cryptosystem algorithms for multiple images encryption have also been designed. Thus, in last a multiple images encryption scheme is proposed. The most overwhelmed feature of these encryption algorithms is that they are not only easy to implement but also offers high level of security that make it resistant against brute force attacks. A detail evaluation of these schemes is done via different analysis that includes Entropy, adjacent correlation between pixels, histograms, histogram variance, differential attacks, key space, and key sensitivity analysis to ensure their robustness. Also, cropping attacks, Noise and pepper attacks are done and there Mean square error (MSE), Peak signal to noise ratio and similarity structural indices are calculated that illustrates the high security level of proposed scheme. Moreover, the differential cryptanalysis is executed for proving the effectiveness of proposed schemes. The entropy values of proposed techniques are near the optimal value 8, which shows high randomness of proposed techniques. Correspondingly, to examine that the proposed encryption scheme has excellent statistical properties that help in the resistance of many attacks, we will analyze it by NIST SP 800-22. The comparison of proposed schemes with some well-defined related schemes is also presented. Comparison results illustrates that the proposed encryption algorithms have better performance than other. Statistical and experimental simulation results depict that the proposed encryption techniques have all desirable characteristics like flexibility, efficiency, and high resistance against attacks like cryptanalysis.
URI: http://hdl.handle.net/123456789/19532
Appears in Collections:Ph.D

Files in This Item:
File Description SizeFormat 
MAT 1719.pdfMAT 17194.7 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.