Abstract
Information hiding is the science of communication in such way that the hidden message into cover media can not to be known or detected. The requirement of steganography with efficient performance, such as high embedding capacity, high image quality, and undetectability, has become an important topic. In this dissertation, information hiding techniques with high performance in spatial domain are proposed, which contains two parts. First part is used the concept of LSB as a basis and adopted an Adaptive LSB approach considered that the pixel differencing falls into lower-level or higher-level. Second part is used the concept of EMD as a basis and extend original methodology, which one pixel is changed by 1 since the embedded algorithm is applied, into a group with multi-pixel and each pixel is changed by k. Thus, a series of hiding techniques are proposed, including FoEMD, HoEMD, and AdEMD. Furthermore, for hiding data more flexibly and completely, a strategy of modulus adjusting is proposed and combined with above techniques of FoEMD, HoEMD, and AdEMD. As shown in the experimental results, our proposed approaches have higher capacity than that of previous researches, and they own the characteristic of efficient performance.