Parallelization of Least Significant Bit Algorithm for Image Steganography in GPU using CUDA

Abstract

This paper presents a method for accelerating the performance of LSB (Least Significant Bit) Algorithm by the concept of parallelization through GPU (Graphics Processing Unit) of the system using CUDA(Compute Unified Device Architecture).The paper also illustrates Image Steganography as an application of LSB algorithm for the encryption of audio file format within a higher resolution digital image file. The principal idea behind of the system is to perform optimal programing of the LSB algorithm for GPU compatibility to parallelize the computation by using M*N threads of the GPU for an image of M*N resolution in which the audio bit encryption is applied. The results of CPU and GPU aided image steganography are calculated separately and compared to estimate performance in gain achieved in parallelization.