An Investigation into Steganography Algorithms

Abstract

Steganography is a Greek word that means hiding the existence of the secret message. The Least Significant Bit (LSB) and Bit Plane Complexity Segmentation (BPCS) are two main mechanisms of the steganography. This project investigates the two mechanisms in terms of their resistance against statistical attacks, the hiding capacity and stego images quality. The two algorithms were implemented on two images of the size 512 x 512. One image was complex, while the other was not. The LSB algorithm was implemented by substituting the least significant bits (1LSB), the least two significant bits (2LSB) and the least three significant bits (3LSB). The least two and three significant bits were substituted to discover if this is an effective way to increase the hiding capacity in the LSB technique. Furthermore, the BPCS algorithm was implemented by using three threshold values, which are 0.3, 0.4 and 0.48. The reason behind using multiple values of the threshold is to study its effect on the histogram of the stego image, the hiding capacity, and the stego images quality. The results of this implementation showed that both algorithms are vulnerable to the histogram attacks. Furthermore, the carrier image in the 3LSB technique should be chosen carefully because the quality of some images is affected if the least three significant bits are substituted. In terms of hiding capacity, the BPCS technique provides more capacity than the LSB (1LSB,2LSB,3LSB) algorithm does. However, using the full capacity in the BPCS algorithm could cause distortion in the stego image. It has been also found that the hiding capacity in the BPCS mechanism depends on the complexity of the image and the threshold value. In addition, increasing the threshold value can contribute to enhancing the quality of the stego image and reducing the noise in the histograms of the stego images.