Research On Image Information Hiding For Compressed Domain

With the rapid development of computer and network technologies, image, audio, video and other multimedia message are easier to obtain and can be copied arbitrary and dissemination. Internet is an insecure public channel since it appeared. Eavesdrop, modify, copy, and other security issues urgently need to be addressed. Information hiding is an effective means to deal with information security problems which is different from cryptography. It have drawn more and more attention in both academia and industry field. Information hiding is a technology which embeds secret data into digital media without affecting human visual perception. Most images transmitted on the Internet are compressed, therefore, information hiding for the compressed domain has positive significance.In this dissertation, two kinds of compressed images are seen as host objects. Under the research of information hiding, compressive sensing and block truncation coding, I propose an image steganography based on subsampling and compressive sensing and two fragile watermarking schemes for BTC-compressed images. The main contributions of this dissertation are as follows:1. On the basis of detail study of the basic principles of compressed sensing, including sparse representation of signal, linear measurement and sparse reconstruction, I propose an image steganography based on subsampling and compressive sensing by designing a embedding function, embedding intensity and extraction function. Several experimental results show that information hiding schemes in compressed sensing domain is feasibility and security.2. On the basis of detail study of the basic principles of block truncation coding, I propose a novel image authentication scheme with fine image quality for BTC-based compressed images by designing reference matrix and candidate equations to narrow the modify scope of two quantization levels, which could improve the visual quality of watermarked images. On this basis, I further narrow the scope of changes by exploiting modification direction and combining PWLC coding which transforms host image feature into self-recovery watermark. The aim of this proposed scheme is capable of image tamper detection and self-recovery capabilities. Numerous experimental results demonstrate the efficiency and superiority of our proposed two schemes in the datahiding for BTC-compressed images.