Imperceptibility—Robustness tradeoff studies for ECG steganography using Continuous Ant Colony Optimization

Highlights

• ECG steganography is performed using DWT–SVD and quantization watermarking scheme.

• Imperceptibility-robustness tradeoff is investigated.

• Continuous Ant Colony Optimization provides optimized Multiple Scaling Factors.

• MSFs are superior to SSF in providing better imperceptibility-robustness tradeoff.

Abstract

ECG Steganography ensures protection of patient data when ECG signals embedded with patient data are transmitted over the internet. Steganography algorithms strive to recover the embedded patient data entirely and to minimize the deterioration in the cover signal caused by the embedding. This paper presents a Continuous Ant Colony Optimization (CACO) based ECG Steganography scheme using Discrete Wavelet Transform and Singular Value Decomposition. Quantization techniques allow embedding the patient data into the ECG signal. The scaling factor in the quantization techniques governs the tradeoff between imperceptibility and robustness. The novelty of the proposed approach is to use CACO in ECG Steganography, to identify Multiple Scaling Factors (MSFs) that will provide a better tradeoff compared to uniform Single Scaling Factor (SSF). The optimal MSFs significantly improve the performance of ECG steganography which is measured by metrics such as Peak Signal to Noise Ratio, Percentage Residual Difference, Kullback–Leibler distance and Bit Error Rate. Performance of the proposed approach is demonstrated on the MIT-BIH database and the results validate that the tradeoff curve obtained through MSFs is better than the tradeoff curve obtained for any SSF. The results also advocate appropriate SSFs for target imperceptibility or robustness.

Introduction

Recent advances in medical devices enable ubiquitous patient monitoring. In order to achieve this, the acquired medical information along with patient data is sent to the physician/care giver through internet. It is essential to ensure the protection of patient data in such transfer (Al Ameen et al., 2012, Law, 1996). Data hiding schemes such as steganography can be used in such situations to shield the identity of the medical information. Watermarking techniques such as least significant bit and substitution schemes are used for hiding the data. The data to be protected is the watermark and the information that carries the watermark is referred to as cover signal. The success of steganography lies in maintaining the embedding induced deterioration of the signal to be minimal and also be robust to external attacks (Katzenbeisser and Petitcolas, 2000, Shih, 2007, chap. 4; Tareef and Al-Ani, 2015, Ziou and Jafari, 2014). In medical domain, steganography schemes protect patient data by hiding it inside their medical information (Zielinska, Mazurczyk, & Szczypiorski, 2014). In this paper, we are interested in hiding patient data into their ECG signals using steganography. Since watermarking leads to cover signal deterioration, steganography algorithms strive to modify the cover signal less such that diagnosability is not affected.

Among other methods, transform domain techniques are widely used in steganography. Steganography in transform domain consists of decomposing the cover signal and embedding the watermark in one or more frequency sub-bands. In ECG steganography, research in the past has focused on different transforms and watermarking techniques. Ibaida and Khalil (2013) performed ECG steganography using Discrete Wavelet Transform (DWT) and Least Significant Bits (LSB) algorithm. In addition, encryption and scrambling techniques were used to improve the security. Chen et al. (2014) compared the performance of Discrete Cosine Transform (DCT), Discrete Fourier Transform (DFT) and DWT using quantization based watermarking algorithm on ECG steganography. It was observed that DCT and DWT provide better results than DFT. A DWT and Singular Value Decomposition (SVD) based ECG steganography was proposed by Edward Jero, Ramu, and Ramakrishnan (2014). They converted the ECG signal into 2D matrix and the patient data was embedded into one of the DWT frequency sub-bands using SVD watermarking algorithm. Moreover, researchers have used steganography in medical images as well (Acharya et al., 2004, Giakoumaki et al., 2006, Lei et al., 2014, Raul et al., 2007)

One of the widely used watermarking techniques is the quantization approach. In SVD based quantization, scaling factors play a vital role in minimizing the deterioration of the cover signal (Mishra et al., 2014, Run et al., 2012). In the context of using a single scaling factor, a low scaling factor provides better imperceptibility while high scaling factor provides better robustness against external attacks. Hence, scaling factor governs the tradeoff between imperceptibility and robustness. Thus, it plays a vital role in determining the quality of the watermarked signal. However, it is easier to decipher the uniform scaling factor in the event of hack. Multiple Scaling Factors (MSFs) (Mishra et al., 2014, Run et al., 2012) are a better choice in such situations but require optimization techniques to find them. Ali and Ahn (2015) show that the algorithm can fail in Mishra et al. (2014). But they use a very less scaling factor range (0.005–0.06) and compare the extracted watermark using correlation. Patient data retrieval which is the focus of this paper is measured using Bit Error Rate.

Loukhaoukha, Chouinard, and Taieb (2011) proposed Lifted Wavelet Transform (LWT) and SVD based image watermarking method. They used Multi Objective Ant Colony Optimization (MOACO) to determine the MSFs. It was observed that the robustness of the watermark improved without losing its transparency. Run et al. (2012) compared the performance of SVD based watermarking technique with DCT and DWT for copyright protection. They embedded the watermark into the principle components of cover image to improve the reliability. Then, the optimized MSFs were computed using Particle Swarm Optimization algorithm. DWT–SVD image watermarking using Firefly algorithm (FA) was introduced by Mishra et al. (2014). The fitness function of FA algorithm was a linear combination of imperceptibility and robustness. They report that the proposed approach is capable of identifying optimal MSFs such that the performance is better than existing methods. Ali and Ahn (2014) developed DWT–SVD watermarking algorithm with self-adaptive differential evolution (DE) to improve the performance of image watermarking. The scaling factors were optimized using self adaptive DE algorithm to yield highest possible robustness and better imperceptibility.

In this study, the focus is to hide patient data in ECG signal with minimal deterioration to the signal. We use DWT to decompose the cover signal. SVD and quantization approach is used for watermark embedding. An optimization problem is formulated and solved using Continuous Ant Colony Optimization (CACO) algorithm to obtain MSFs. These MSFs when used in quantization provide superior imperceptibility and robustness of watermark than any Single Scaling Factor (SSF). One of the important outcomes of the study is a tradeoff curve between imperceptibility and robustness as a function of watermark size for different SSFs. If one wants to use a SSF for ECG type signals, this tradeoff curve can be utilized to pick up the appropriate scaling factor to achieve the respective imperceptibility and robustness.

Rest of the paper is organized as follows: In Section 2, the materials and methods discuss the database, CACO, DWT–SVD based ECG steganography and the proposed method. The results are discussed in detail in Section 3. The overall summary of the proposed research work is presented in Section 4.