A Fast Semi-Fragile Watermarking Scheme Based on Quantizing the Weighted Mean of Integer Haar Wavelet Coefficients

Abstract

Most semi-fragile watermarking algorithm based on Haar wavelet needs to repeatedly switch between the wavelet domain and spatial domain to avoid the overflow of pixel gray value after embedding watermark. In order to improve the efficiency of watermarking embedding algorithm, a fast semi-fragile watermarking scheme is proposed. According to the correspondence of position between spatial pixels and wavelet coefficients as well as the correspondence of adjusting amount between spatial pixels and wavelet coefficients in 1-level integer Haar wavelet transform matrix, the proposed scheme can avoid the overflow checking of pixel gray value by controlling the adjusting amount of wavelet coefficients before watermark embedding. Meanwhile, a cumulative weighted voting method is suggested to reduce the false alarm pixels in the original binary watermark error map. Experiment shows that the proposed scheme is feasible and effective. It is sensitive to malicious tampering and can resist a certain degree of JPEG lossy compression, JPEG2000 lossy compression and other conventional image processing.