Adaptive Video Watermarking with Robust Pca-based Decoding

Data embedding and recovery in video watermarking, without loss of quality is a challenging one. A new method for video watermarking is presented in this paper. In the proposed method, we use undecimated wavelet transform. The location for data embedding in the LL subband of wavelet coefficients are selected adaptively based on the energy of high frequency subbands. The decoding is performed based on the comparison among the elements of the first principal component resulting from empirical principal component analysis (PCA). Testing and analysis were done by subjecting the watermarked video to various attacks like cropping, filtering and rotation. The results show that the proposed method offers improved performance compared with several conventional methods. Thus this adaptive stationary based data insertion is much suitable for various security multimedia based applications.