Abstract
The content of the integrity of digital image needs copyright protection; the process of copyright protection used the methods of digital watermarking techniques. The strength and robustness of watermarking algorithms faced a problem of geometrical attacks. In this paper proposed novel algorithms for digital image watermarking using BP neural networks using features transform. The SIFT transform function is a core feature transform function of a digital image, and the BP neural network is multi-layer neural network models. The BP neural network generates the pattern for the embedding. The BP neural network and RBF neural network cascaded and generates dynamic patterns for embedding. The processing of RBF neural network models is very efficient due to a single layer network. The proposed algorithms implemented in MATLAB software and perform well know image dataset for the processing of image watermarking. For the validation of algorithms measure some standard parameters in concern of quality measure peak to signal-noise ratio. In concern of security strength measure number of correlations of a pixel, and in terms of complexity measure time of encoding and decoding of watermarking algorithms.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have