A New Information Hiding Method In Low Frequency Domain Of A Photo

The digitalization of information medium provides much convenience for information store and access and also improves the efficiency and accuracy of information expression. Especially, with the development of computer and network communication technology, data exchange and transmission become comparably simple. However, it brings forth some new problems, such as information security of the digital media, intellectual property protection and authentication, etc. Cryptographic techniques can't solve the problem completely, for encrypted information is prone to attract curiosity and observation of the aggressor, thus it may be cracked.What's more, when received information is decrypted, encrypted contents are no longer protected, that is to say, cryptography can just protect media contents in transmission. There is a need for security technique as the supplement of cryptography, i. e., information hiding.This dissertation puts forward a new method of information hiding and its basic idea is hiding information in the low frequency domain of a photo. As we know, hiding information in the high frequency can probably be affected by the attack of JPEG compressing and low-pass filter and so on, however, hiding data in the low frequency domain may make much and even unacceptable chang of source photo. Two kinds of method to divide the photo into blocks are presented in the dissertation, then the techniques of appending zero, DCT transform and changing low frequency coefficents are utilized to hide information (in fact, information is masked in DC components). So information is hided in the low frequency domain of a photo. Furthermore, good quantity evaluation is got after compared with original photos. The software of stirmark is used to attack test on hided results in the dissertation and good experimental results are obtained. It shows that the proposed technique has fair robustness in JPEG compression and medium filtering, while has no robustness in some geometry transform such as replacement, rotation and scaling. Nevertheless, if the same transform is made to the original photo which matchs the attacked photo during the test process, good results can also be got. Furthermore, the comparision with some classical hiding techniques such as the LSB method and DCT-transform method is presented, which proves that the proposed technique in the dissertation gains much advantage over JPEG compression. Finally, the shortage of the proposed technique and further research direction are pointed out.