Scalable watermark extraction for real-time authentication of JPEG 2000 images

  • Authors:
  • Charith Abhayaratne;Deepayan Bhowmik

  • Affiliations:
  • Department of Electronic and Electrical Engineering, The University of Sheffield, Sheffield, UK S1 3JD;Department of Electronic and Electrical Engineering, The University of Sheffield, Sheffield, UK S1 3JD

  • Venue:
  • Journal of Real-Time Image Processing
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper proposes a novel scalable authentication scheme that utilizes the progressive enhancement functionality in JPEG 2000 scalable image coding. The proposed method first models the wavelet-based quality scalable coding to identify the effect of the quantization and de-quantization on wavelet coefficient magnitudes and the data embedded within such coefficients as a watermark. A relationship is then established between the watermark extraction rule and the embedding rule, using the magnitudes of the reconstructed and original coefficients. It ranks the wavelet coefficients according to their ability to retain the embedded watermark data intact under various quantization levels corresponding to quality enhancements. Then watermark data is embedded into wavelet coefficients according to their rank followed by JPEG 2000 embedded coding. At the decoder as more and more quality and resolution layers are decoded the authentication metric is improved, thus resulting in gradually increasing complexity of the authentication process according to the number of quality and resolution enhancements. The low complexity authentication is available at low quality low resolution decoding, enabling real-time authentication for resource constrained applications without affecting the authentication metric. Compared to the existing methods, the proposed method results in highly robust scalable authentication of JPEG 2000 coded images.