Self-embedding watermarking scheme against JPEG compression with superior imperceptibility

Abstract

To improve the imperceptibility, security and tamper detection performance, a self-embedding watermarking scheme against JPEG compression is proposed in this work. The recovery watermark of all blocks in the host image, which consists of 6-bit DC-code and 5-bit AC-code and simultaneously used to both tamper detection and tamper recovery, is recombined based on the secret key and embedded in the quantized DCT coefficients. This strategy not only enhances the ability against the known forgery attacks due to introducing multi-blocks independency, but also improves the tamper detection performance by designing the tamper detection method based on the multi-neighbor characteristic and multi-threshold optimization. To achieve the better imperceptibility, the 7 middle frequency DCT coefficients of an 8?×?8 block are chosen to hide the 11-bit watermark by adopting the weight-function modulo based embedding method. Robustness against JPEG compression is enhanced by setting the quantization step of the chosen DCT coefficients according to the standard JPEG quantization table and an adjustable scaling factor which balances the robustness with imperceptibility. We also discuss the effects of different scaling factors on the imperceptibility and robustness in terms of peak signal to noise ratio (PSNR) and survival quality factor (SQF) of watermarked image generated by a given scaling factor. Experimental results demonstrate that the proposed method outperforms conventional semi-fragile restorable watermarking schemes in imperceptibility, tamper detection and tamper recovery in various forgery attacks especially for JPEG compression with low QFs.