Genetic-Algorithm-Based Optimization of Fragile Watermarking in Discrete Hartley Transform Domain

Genetic-Algorithm-Based Optimization of Fragile Watermarking in Discrete Hartley Transform Domain

Sudipta Kr Ghosal (Future Institute of Technology, India) and Jyotsna Kumar Mandal (University of Kalyani, India)
DOI: 10.4018/978-1-5225-0058-2.ch005
OnDemand PDF Download:
No Current Special Offers


In this chapter, a fragile watermarking scheme based on One-Dimensional Discrete Hartley Transform (1D-DHT) has been proposed to verify the authenticity of color images. One-Dimensional Discrete Hartley Transform (1D-DHT) converts each 1 x 2 sub-matrix of pixel components into transform domain. Watermark (along with a message digest MD) bits are embedded into the transformed components in varying proportion. To minimize the quality distortion, genetic algorithm (GA) based optimization is applied which yields the optimized component corresponding to each embedded component. Applying One-Dimensional Inverse Discrete Hartley Transform (1D-IDHT) on 1 x 2 sub-matrices of embedded components re-generates the pixel components in spatial domain. The reverse approach is followed by the recipient to retrieve back the watermark (along with the message digest MD) which in turn is compared against the re-computed Message Digest (MD') for authentication. Simulation results demonstrate that the proposed technique offers variable payload and less distortion as compared to existing schemes.
Chapter Preview


The idea of concealing secret information inside other digital media either in spatial domain or in transform domain is termed as the digital watermarking, which is widely used to authenticate or to verify the owner-ship of a digital content. In spatial domain, robustness and security of the fabricated watermark are two key challenges. In contrast, transform domain techniques fabricates secret bits into the transformed coefficients for better security and robustness. According to the general characteristics, digital watermarking process is classified into three categories namely fragile, semi-fragile and robust respectively. The generic nature of fragile watermarking is to maintain the integrity of digital images and therefore, the embedded watermark is expected to be destroyed while any sort of alterations are made on the watermarked image. Semi-fragile watermarking permits some kinds of alterations on the watermarked images however, severe attacks on the watermarked image are not allowed. The copy-right protection and ownership identification can be accomplished based on the robust watermarking which can resist a designated class of geometrical as well as visual alterations. Fragile watermarking scheme operates the authentication process and checks whether any alteration is occurred. Even a single bit error leads to a different authenticator.

In general, majority of the watermarking schemes are primarily focused to achieve high robustness against common attacks which made watermarking as an effective tool for copy-right protection. Unlikely, limited people had chosen digital watermarking as a way out of authentication. The major objective of this scheme is to design and implement the fragile watermarking in transform domain that can be used to verify the authenticity of color images. Existing schemes are lacking due to the following issues: severe quality degradation, fixed as well as reduced payload and the inability of choosing color image as the cover. To address the above mentioned problems, a novel watermarking scheme in one dimensional Discrete Hartley Transform (1D-DHT) has been proposed that can deal with the authenticity of color images. The scheme provides variable payload that offers a spread from 0.5 to 3 bpB (bits per Byte) with the permissible visual imperceptibility. Moreover, the genetic algorithm (GA) based optimization scheme optimizes the embedded components in transform domain to ensured substantial quality distortion in the watermarked images.

Key Terms in this Chapter

Digital Watermarking: An approach of fabricating watermark into the cover media in an imperceptible way. The privacy of concealed information is maintained among the authorized parties however, the perceptibility of the watermarked image is maintained by the tradeoff between quality and payload.

Peak Signal to Noise Ratio (PSNR): The ratio between power of noisy signal and the power of actual signal is considered to be the Peak signal-to-noise ratio (PSNR) as represented in decibel (dB) unit.

Payload: The metric “payload” is nothing but the capacity of a digital media which denotes the amount of concealed information into the watermarked media. The usual units of payload are bits, Bytes, Kilo Bytes or Mega Bytes respectively. However, the unit “bits per Byte (bpB)” is the latest inclusion into the list.

Authentication: The mechanism of checking the originality of a digital content. An alteration made into the digital media is treated as the violation of authenticity.

Message Digest: A hash value obtained from the secret information which is widely used for integrity verification in this digital era is termed as the message digest.

Cover Media: The media in which the secret watermark is embedded is termed as the cover media. Popular mediums used as the cover includes: image, audio and video etc.

Complete Chapter List

Search this Book: