A Blind Image-Adaptive Data Hiding Method in Wavelet Domain with Robustness against Compression

Abstract

In this paper we introduce a robust image-adaptive data hiding in wavelet domain that hide data in images while imposing minimal perceptual degradation in original image. As it is well accepted, data embedding in the low and mid frequency parts of an image have better robustness under compression attacks than high frequency parts. So our method embeds more data in low frequency coefficients than high frequency coefficients with use of scaler QIM method. Also the encoder employs local criteria for selecting the coefficients for embedding. The hidden data can be recovered reliably under attacks, such as jpeg or jpeg2000 compression.