The increasing popularity of multimedia network technology, data transmission on the Internet becomes very easy and convenient. However, the pubic channel does not secure enough because digital data transmission on the public network is very easy to be modified, intercepted, and copied. Even though traditional cryptography schemes can provides very good security, however as an encrypted data normally reveals the importance of its content. The encrypted data might also attract the interest by hackers. To provide a secure transmission, data hiding becomes an important issue. Data hiding or called Steganography is a kind of camouflage approach which can provide a secure transmission. Data hiding differs from traditional cryptography schemes, in that it embeds secret information into normal host media such as images, music, and texts. It can decrease the suspect of the interest by the hackers because the embedded media does not look like it has valuable.
In this thesis, we have proposed three data-hiding schemes. The first proposed scheme is based on the VQ encoding and the principal component analysis. The principal component analysis is used to reorder VQ codewords, and then to spilt them into two cluster groups. To decrease the reconstructed image quality, the codewords uniform selection is also provided. The second scheme uses the concept of the table lookup which partitions gray-level colors into several groups or called clusters. Each cluster is given a unique value by a pseudo random number generator with a known seed. Secret bits are then embedded into the selected groups. To increase the data security, a seed of the pseudo random number generator must be changed at each time for decreasing the possible detection. The third data hiding scheme is based on block truncation coding (BTC) scheme and the concept of pixels clustering. The proposed scheme utilities two reconstructed values and one bitmap of each BTC block to hide secret data. Two reconstructed values for each block uses the second proposed data hiding scheme to hide secret data On the other hand, for the binary bitmap, the Hamming code is applied to each binary bitmap of BTC blocks for increasing the hiding capacity.
