Watermarking In Digital Audio Or Video And Their Capacity Improving

This paper analyzes the characteristic of digital watermarking in audio or video and the application of apperception of human auditory system(HAS) and human visual system(HVS) on digital watermarking . After systematically comparing every current method, by which digital watermarking is embedded in audio or video, with others, and analyzing current fields that watermarking embedded in, it presents a digital watermaking method which is based on correlation detection and can be draw out without original signal, and forms a data-hiding algorithm based on correlation detection and a robust watermarking algorithm based on MPEG-4 video object encoding, by respectively applying the method in audio and video .A data-hiding algorithm based on correlation detection can transform watermaking information into new sequence by spreading spectrum from every bit to hide according to the positive or negative correlation to known module, and adjust the sub-vectors coefficient of wavelet conversion according to the new sequence to hide data. Considering of the variable data size, the algorithm defines a kind of sign to end the data-extracting process, which advances the algorithmic efficiency. The algorithm is still some resistance to noise while it has hidden some text data, which can depress the misinformation probability, and its capacity to hide data is bigger, and can achieve the target to transmit important information efficiently by hiding them in a mult-media signal on public network.A robust watermarking algorithm for object-based MPEG-4 encoding detailedly analyses the structure of object-based MPEG-4 encoding bitstream and its technique of re-synchronization , and select GOP ,which can be visited random, as based group to embed digital watermarking .By defining subsection number and some special marks for start and end, this algorithm can draw out digital watermarking subsection from other visual sequence or other video object to replace missing digital watermark , and form integrated digital watermarking after recomposition while frame-skipping or block-skipping is occurred ,or while video is cut and rebuilt. This algorithm is very simple and instantaneous for it only decodes some header-information and does some VLC decoding, inverse- scanning, re-scanning, and VLC re-coding on chrominance blocks of B-VOP .