UC Irvine

Previous researches have shown that image encryption could be done using techniques like DES, IDEA, RES but the new and effective way for fast and secure encryption using chaos-based cryptography is the most preferred encryption technique. Chaos-based encryption algorithms are a combination of multiple chaotic maps and the same process can be repeated for multiple cycles for higher security. But, as the number of steps to process an image increases the processing time increase too. Another, reason for increased processing time is the number of pixels being encrypted. Since, chaos-based algorithms are considered as ideal for encrypting images in real time applications, higher security and lesser response time are essential.

This research primarily focusses on examining encryption techniques using two-dimensional chaotic maps and comparing the encryption algorithms security level and response time for different image sizes. After serial implementation, parts of encryption and decryption process that can be parallelized are evaluated and implemented using GPU and CUDA programming and then quantitative results are compared with the serial implementation.