Prediction-based watermarking schemes using ahead/post AC prediction

Authors:
Jing-Ming Guo;Chia-Hao Chang
Affiliations:
Department of Electrical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan;Department of Electrical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan
Venue:
Signal Processing
Year:
2010

Citing 8
Cited 2

Blind image data hiding based on self reference

Pattern Recognition Letters
Variance-Classified Capacity Watermarking Using Discrete Cosine Transform

IIH-MSP '08 Proceedings of the 2008 International Conference on Intelligent Information Hiding and Multimedia Signal Processing
Adaptive Visible Watermarking of Images

ICMCS '99 Proceedings of the IEEE International Conference on Multimedia Computing and Systems - Volume 2
Watermarking in conjugate ordered dither block truncation coding images

Signal Processing
General blind watermark schemes

WEDELMUSIC'02 Proceedings of the Second international conference on Web delivering of music
A variable step size LMS algorithm

IEEE Transactions on Signal Processing
A robust variable step-size LMS-type algorithm: analysis andsimulations

IEEE Transactions on Signal Processing
Denoising and copy attacks resilient watermarking by exploiting prior knowledge at detector

IEEE Transactions on Image Processing

Joint fingerprinting and decryption with noise-resistant for vector quantization images

Signal Processing
Reversible joint fingerprinting and decryption based on side match vector quantization

Signal Processing

Quantified Score

Hi-index	0.08

Visualization

Abstract

This study presents two predicted-based watermarking schemes, namely ahead AC-predicted watermarking (AAPW) and post AC-predicted watermarking (PAPW), by embedding information into low frequency AC coefficients of discrete cosine transform (DCT). The proposed methods utilize the DC values of the neighboring blocks to predict the AC coefficients of the center block. The low frequency AC coefficients are modified to carry watermark information. The least mean squares (LMS) is employed to yield the intermediate filters for cooperating with the neighboring DC coefficients to predict the original AC coefficients. During the LMS filter training, the training blocks are classified into different categories according to their texture angles and variances. The classified trained filter sets are then used to predict the AC coefficients even more precisely. As documented in the experimental results, the image quality and the embedded capacity of the proposed schemes are superior to former methods in the literature. Moreover, many attacks are addressed to show the robustness of the proposed methods.