Steganography methods that embed data based on human vision system (HVS) characteristics are called adaptive methods. Based on HVS characteristics, edge regions are more suitable for embedding. The main goal of this dissertation is to study steganography and steganalysis, or in short stegology, of adaptive methods. Therefore, in the first step we analyzed existing adaptive embedding methods and proposed a general structure for them. We introduced a number of parameters that are detrimental in adaptive methods’ security. Thereafter, we proposed a model for evaluating security level of an adaptive method. Most of adaptive methods use LSB Flipping method as means of data embedding. In the second step, we unveiled weaknesses of this adaptive method and proposed an algorithm to defeat an LSB-F based adaptive method. In these attacks, we were able to estimate the amount of payload in a stego image. To ensure the accuracy of our presented stegological assumptions, a new adaptive method is presented. The proposed method uses a local neighborhood analysis for determination of secure locations to use LSB Matching for the embedding purpose. Experiments showed the superiority of our algorithm over well known adaptive methods and LSB Matching based methods. Key Words Steganograghy, Steganalysis, Stegology, Steganology, Adaptive methods, Complexity
