Robust Visual Hashing for Image Authentication
In this paper, secure and robust image (visual) hashing based on the Discrete Wavelet Transform (DWT) and Non-negative Matrix Factorization (NMF) is proposed. In the proposed method, the considered image is undergone preprocessing manipulations. Then, the DWT is applied on the preprocessed image to generate image features that are largely invariant under perceptually insignificant manipulations. Consequently, the image feature data are reduced using NMF. The generated NMF coefficient matrix is interleaved, quantized and encoded to produce a binary string that is interleaved to generate the image hash.