Spread Spectrum Watermark in DC: A View from the Embedding Processing

Spread spectrum watermarking is a robust methodology for image copyright identification. Embedding location and weight are two critical factors that influence watermark performance. As a rule of thumb, direct current (DC) coefficients are generally excluded from embedding the spread spectrum watermark due to the introduction of block artifacts. We investigate into this issue from the view of watermark embedding and reveal the fact that a desired embedding location should have small module and high stability. The DC location, while exhibiting high stability, is constrained by its large module. To address this limitation, we present a method for adjusting the DC coefficients (ADC). First, the discrete cosine transform (DCT) is performed on a matrix comprised of DC coefficients from all blocks of the host image. Second, the coefficients at the end of the high-frequency band, with the smallest module, are selected to form the embedding location. To adapt different real scenes, an attack-based weight (ABW) method is presented. It incorporates prior attack information into the weight determination. Experimental results demonstrate that the performance of the state-of-the-art algorithms is improved using the adjusted embedding location, and the integration of ADC and ABW methods further enhances the performance of the watermark system, which achieves an average PSNR of 48 dB with the bit error ratio (BER) approaching zero.