A BlindWatermarking Scheme Based on Combined Complex Transforms

In this paper, we investigate a blind watermarking technique via complex signal processing for images. For improving watermark robustness to JPEG compression, our watermarking employs two complex transforms, referred to as complex wavelet transform and complex smear transform. High frequency components are extracted by the complex wavelet transform in order to prevent a quality degradation of host data due to watermarking. The subband signal is spread by an inverse complex smear transform for enhancing robustness to a variety of attacks. These steps correspond to a detecting scheme in the proposed method. From the spreading signal, we derive an optimum impulse signal as a signature. The signature impulse signal is processed by an inverse system to the detection algorithm for generating a watermark, and the watermark is added to the host image. This watermark is detected by a maximum value detection due to the inserted impulse signal. In addition to a payload of the watermark, various rates of detecting the correct key from a decompressed JPEG image are shown for evaluating the performance of the proposed method.