A novel batch image encryption algorithm using parallel computing (Q2662717): Difference between revisions

This paper consideres how to encrypt multiple images fast and proposes (with current multi-core CPU-s; GPU-s are not mentioned) a very natural ``parallel encryption scheme that encrypts a batch of images in a preemptive manner, which makes full use of processors of a modern computer'' (p. 13); to support the idea are presented results of some experiments. Unfortunately nothing could be reproduced, implementation is described with four lines of linear pseudocode (referring self-destructing threads) and some shemas; instead of used programming language and compiler is indicated only IDE: ``The developed platform is codeblocks 17.12'' (p.12); indication of used in experiments computer's CPU is erronous ``computer with Inter(R) Core(TM) i7-6700 CPU(4 cores, 8 threads)'' and the clock speed is not indicated, Intel CPU i7-6700 for desktops comes with different clock speeds: 2.6 GHz..4.2 GHz. Organization of preemtive multitasking is missing, e.g. in page 7 the overhead from organizing parallel processing of a shared resource is not considered at all ``time for execution on single processor is equal to sum of times for execution in parallel processors'' -- with large number of small images the overhead could be significant. Many sentences are difficult or impossible to understand: ``two threads are encrypting the same image simultaneously'' thus ``we don't know what the encrypted data will be'' (p.3) and neither would reader, ``Different threads compete with each other to encrypt one of the batch images'' (p.3), ``The time of our proposed is the maximum running time of the thread among multiple threads'' (p.12); authors claim : ``our scheme is flexible to encrypt different sizes of images'' (p.3), but do not explain if/how the chaos-based encryption parameters depend on image size. This is a description of an idea, not a solution.