Digital information embedding/extracting

摘要

In an embedding apparatus, an orthogonal transform section (11) divides an original image (1) into a plurality of blocks to perform orthogonal transform every block. An information embedding section (13) embeds a pseudo-random number sequence mapped with inherent digital information in part of a transform coefficients sequence obtained by the orthogonal transform using a weight sequence. In an extracting apparatus, a correlation value calculation section (23) reads sequences of transform coefficients for respective blocks from the data including the pseudo-random number sequence embedded therein in a predetermined order, and calculates an inner product between the sequences of transform coefficients and a pseudo-random number sequence. A pseudo-random number determination section (24) determines that the pseudo-random number sequence has been embedded when the inner product is greater than a predetermine threshold value. An information generation section (25) extracts inherent digital information mapped with the determined pseudo-random number sequence. This method discourages deciphering of the digital information by a third party, suppresses deterioration in image quality by a simple apparatus/method, and improves the rate of detection of watermark data.