A Real-Time Ownership Identification Approach for Semi-Blind Invisible Watermarking Using Statistical Information of Watermark Bit Stream

Nowadays, due to the advanced technologies in telecommunication and broadcasting networks, there are ever-increasing business demands for immediate watermarking schemes in many real-time applications to protect their digital properties against misuse or piracy. Studies revealed that the spatial-domain watermarking schemes such as the EISB (Enhanced Intermediate Significant Bit) require lower computational complexity than the transform-domain techniques. Therefore, they could be more appropriate compared to the transform-domain techniques for real-time applications such as medical and satellite image protection applications. However, they are not as robust as the transform-domain techniques. Hence, there is still a great challenge for emerging effective watermarking schemes for real-time ownership identification applications. To take up this challenge, we proposed an approximation technique for the ownership identification applications using the ratio of the number of Is to the number of both Is and Os in the bit stream of the watermark. We were successfully implemented the proposed approach in the EISB scheme. The results of experiment revealed that the proposed approach identified the owner of the watermarked image, even after the embedded watermark was totally corrupted by the JPEG2000 image compression attack.