ABSTRACT
This chapter presents the technique of combinational watermaking in both the spatial and frequency domains. The idea is to split the watermark image into two parts, which are respectively used for spatial and frequency insertions based on the user preference and data importance. Despite its simplicity, the least significant bits (LSBs) substitution suffers a major drawback. Any noise addition or lossy compression is likely to defeat the watermark. An even simple attack is to simply set the LSB bits of each pixel to one. Watermarking can be applied in the frequency domain by first applying a transform. Similar to spatial-domain watermarking, the values of specific frequencies can be modified from the original. The peak signal-to-noise ratio is often used in engineering to measure the signal ratio between the maximum power and the power of corrupting noise. Because signals possess a largely wide dynamic range, ones’ apply the logarithmic decibel scale to limit its variation.
