Embodying information into images by an MMI-based independent component analysis algorithm
Liu Ju, Jiande Sun, Zhengfeng Du, Yong Wan · 2003
The signals measured by multi-sensors are always the mixtures of several independent sources. Therefore, it is necessary to separate them from each other for practical applications. Independent component analysis (ICA) is a novel signal processing method presented in dealing with such problems. It is also found very useful in many other problems such as biomedical signal separation, communication and multimedia information processing. We gave a minimizing mutual information based ICA algorithm and then applied it into information hiding or digital watermarks. We consider the original image as a mixture of some independent character images. We can embody the watermark into the separated images and then remix them as a watermarked image. Simulations results show the validity in embodying information into images.