Abstract:

We present a robust digital watermarking embedding strategy with blind detection based on Tetrolet transform and singular value decomposition (SVD). Firstly, the binary watermarking image is pretreated by two-factor chaotic encrypting processing. The original host image is transformed by Tetrolet transform, and then, the low-frequency coefficients in two-level are decomposed by the means of the blocked SVD. Lastly, the encrypted watermark information is embedded into those largest coefficients. The proposed strategy can extract the watermarking image without reference to the original host image or the original watermarking image, which achieves blind detection and improves the practicability of the digital watermarking system. Simulation results show that the proposed watermarking algorithm is secure, invisible and robust to common image processing and geometric attacks.

Key words: Tetrolet transform, chaos mapping, singular value decomposition, blind detection