• P-ISSN 0974-6846 E-ISSN 0974-5645

Indian Journal of Science and Technology


Indian Journal of Science and Technology

Year: 2016, Volume: 9, Issue: 6, Pages: 1-6

Original Article

1D Chaos-based Image Encryption Acceleration by using GPU


Chaos-based image encryption algorithm is one of the most important methods that are considered as the main part of many structuring encryption systems. In this paper, a new implementation of One-Dimension (1D) chaos-based image encryption algorithm is presented using the parallelism features of GPU and CPU. In order to use the parallelism power of CPU, the parallel computing toolbox of MATLAB, provides efficient methods for Parallel Task Processing (PARFOR) and Parallel Data Processing (SPMD). For further improving the execution time of the algorithm, sequential partitions are performed on CPU and the parallel pieces are executed on the GPU. The results of serial and parallel implementation on the color images with different resolutions, using MATLAB parallelism methods show when the size of the pictures increase, the performance of the 1D chaos-based image encryption algorithm in parallel implementation by the both parallel task “PARFOR” and data processing “SPMD” methods, becomes better. Also, the results of the implementation illustrate that the execution time when PARFOR scheme is used becomes better when the image size is higher than a threshold. Furthermore, the results of Cuda and Visual C++ implementation on the color images with different resolutions show that the simulation time using Cuda C++ is almost three times better than visual C++. Total results of the comparison show that when a combination of CPU and GPU is used, the execution speed reached its best state. Because of accelerating the image encryption process using the power of CPU and GPU, the proposed implementation is suitable for the multimedia application systems.

Keywords: Accelerating Image Encryption Process, Graphic Processing Unit (GPU), One-Dimension (1D) Chaos-based Image Encryption Algorithm, Parallel Data Processing (SPMD), Parallel Task Processing (PARFOR)


Subscribe now for latest articles and news.