Robust image watermarking using multiresolution analysis

The paper proposes an image watermarking scheme using wavelet domain. Wavelet transform, due to its sub bands decomposition, offers better trade-off between embedding strength and robustness performance compared to other unitary transforms such as DCT, DFT, Fourier-Mellin, Walsh-Hadamard etc. Data embedding in different sub bands of the cover image offers resiliency against various types of image impairments. Degree of resiliency is increased further by embedding the same watermark information in multi-level wavelet coefficients of the cover image. Experiment results show that the scheme is resilient to various linear and non linear filtering, image sharpening, cropping, rescaling, noise addition, JPEG and JPEG 2000 compression operations etc. Compression adaptive negative modulation increases robustness against low quality compression. The proposed algorithm shows better visual as well as statistical imperceptibility and resiliency against JPEG and JPEG 2000 compression compared to two popular DCT domain algorithms and data recovery process does not require the cover image, watermarked image and/or watermark message unlike the latter two schemes.