A Novel Image Steganography and Steganalysis Technique Based on Pattern Searching

Abstract

Image steganography is a method, which hides a message into a cover image and generates a stego-image. A secret message is embedded into the image as noise, which is nearly impossible to be detected by human eyes. The stego-image is then sent to the receiver who can extract the message with or without stego-key, which depends on the hidden scheme. Most of the hiding techniques hide the message straight away into the significant bits of the image pixels therefore disturbing the resolution, which degrades the quality of the image and makes the image prone to various types of attacks. A new approach that can make the message more secure and enhance the image quality is proposed here. This approach uses LSB substitution method along with the proposed algorithm. The suggested method hides the secret message based on searching the pattern for {00, 01, 10, 11} and matching the above pattern with last two bits of each pixel in RGB format. The secret message along with two RGB images is then taken as an input where the message is converted into binary form first, and the complete algorithm is applied to hide the secret message within the images. Inversely, using the concept of a dictionary file and applying the algorithm accordingly can retrieve the secret message. The proposed method hides and retrieves the message based on a combination of the two images. The two images work as lock and key and both are necessary for proper retrieval of the message. The quality of the image from which the original message is extracted is not affected at all.