Content-adaptive reliable robust lossless data embedding

  • Authors:

  • Lingling An;Xinbo Gao;Yuan Yuan;Dacheng Tao;Cheng Deng;Feng Ji

  • Affiliations:

  • VIPS Lab, School of Electronic Engineering, Xidian University, Xi'an 710071, China;VIPS Lab, School of Electronic Engineering, Xidian University, Xi'an 710071, China and Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education of China, Xidian Un ...;State Key Laboratory of Transient Optics and Photonics, Xi'an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Xi'an 710119, China;Centre for Quantum Computation and Intelligent Systems, Faculty of Engineering and Information Technology, University of Technology, Sydney, Broadway NSW 2007, Australia;VIPS Lab, School of Electronic Engineering, Xidian University, Xi'an 710071, China;VIPS Lab, School of Electronic Engineering, Xidian University, Xi'an 710071, China

  • Venue:
  • Neurocomputing
  • Year:
  • 2012

Quantified Score

Hi-index 0.02

Visualization

Abstract

It is well known that robust lossless data embedding (RLDE) methods can be used to protect copyright of digital images when the intactness of host images is highly demanded and the unintentional attacks may be encountered in data communication. However, the existing RLDE methods cannot be applied satisfactorily to the practical scenarios due to different drawbacks, e.g., serious ''salt-and-pepper'' noise, low capacity and unreliable reversibility. In this paper, we propose an effective solution to RLDE by improving the histogram rotation (HR)-based embedding model. The proposed method is a content-adaptive reliable RLDE or CAR for short. It eliminates the ''salt-and-pepper'' noise in HR by the pixel adjustment mechanism. Therefore, reliable regions for embedding can be well constructed. Furthermore, we basically expect the watermark strengths to be adaptive to different image contents, and thus we have a chance to make an effective tradeoff between invisibility and robustness. The luminance masking together with the threshold strategy is duly adopted in the proposed RLDE method, so the just noticeable distortion thresholds of different local regions can be well utilized to control the watermark strengths. Experimental evidence on 300 test images including natural, medical and synthetic aperture radar (SAR) images demonstrates the effectiveness of the proposed data embedding method.