A multi-scale image watermarking based on integer wavelet transform and singular value decomposition

Image watermarking technique is one of effective solutions to protect copyright, and it applied a variety information security application domains. It needs meet four requirements imperceptibility, robustness, capacity security. A multi-scale secure image method proposed in this work, which based on the Integer Wavelet Transform (IWT) Singular Value Decomposition (SVD). Four IWT sub-bands are firstly obtained after 1-level host image, corresponding singular diagonal matrices can be using SVD. Then, each matrix divided into non-overlapping sections terms size embedding watermark. Particularly, upper left part same as The watermark directly embedded parts afterwards by multiplying different scaling factors complete final operation. Especially, novel optimized authentication mechanism designed resolve false positive problem, exists SVD-based algorithms. In addition, three-dimensional optimal mapping algorithm search through objective evaluation function, significantly improve imperceptibility robustness. experimental test comparison analysis illustrate that scheme demonstrates high with peak signal noise ratio values 45 dB strong robustness average normalized correlation 0.92.