ABSTRACT
Digital watermarking techniques are developed with the intention of protecting multimedia over different threats, by integrating a watermark either visibly or invisibly in a robust manner before publishing or storing the media. This chapter starts with spatial domain–based watermarking approaches, followed by transform domain techniques, and discusses advanced watermarking techniques that rely on the concept of machine learning (ML) algorithms. The watermarking techniques are majorly developed in two processing domains: the spatial domain and transform domain. ML techniques automatically predict output values from input data with high accuracy and efficiency by means of different classification and pattern recognition algorithms, which otherwise would be a time-consuming process. Singular value decomposition-based watermarking schemes in combination with discrete wavelet transform are widely used and provide more transparency and robustness than all other watermarking schemes. Local binary pattern–based watermarking is good at surviving illumination changes and contrast variations.
