Optimization of Least Significant Bit Steganography Using Genetic Algorithm to Improve Data Security

Abstract

Instant message is one of popular application for chatting. This application was already equipped with many security features, but most of these applications saved their chat history in the smartphone database. With so many apps on the internet, recover your chat from the application database is not difficult. Our chat could be recovered using the tools, but most of this application cannot retrieve images with the same quality as what we send in conversation. Steganography using Least Significant Bit (LSB) could embed the message in the picture, but this method is more accessible to detect. One of the techniques that could identify the amount of character inserted using LSB is RS Steganalysis. This paper proposes LSB enhancement using Genetic Algorithm to obscure RS Steganalysis. Genetic Algorithm is used to randomize the embedding character in the pixel so the resulting image (stegoimage) would be hard to analyze. The result showed that the stego-image was difficult to recover or analyze by the general extraction method of Least Significant Bit because Genetic Algorithm has randomly embedded the placement of secret message. The experimental results also demonstrated that the Genetic Algorithm is used to obscure RS Steganalysis while maintaining image quality. Using iteration of the experiment showed fitness value decreasing and get better at 5.855 which only alter 5.3% of 65.536 pixels. The increasing amount of character embedded is also increased RS Steganalysis error.