A secure framework for remote diagnosis in health care: A high capacity reversible data hiding technique for medical images

https://doi.org/10.1016/j.compeleceng.2020.106933Get rights and content

Highlights

  • A high capacity data hiding algorithm which is reversible in nature.

  • Utilization of the original pixels along with the interpolated pixels boosts the payload capacity.

  • A novel weighted image interpolation technique that enhances visual quality of the stego image.

Abstract

The health care industry involves the processing of large-scale images for various applications. Remote diagnosis is one such significant area where medical images are sent across vulnerable communication media. Hence, a secure and robust framework is necessary to hide and retrieve patient information in medical images. Here, we propose a high capacity reversible data hiding technique that can be used to embed patient data using a new weighted interpolation technique. In this approach, to improve the payload capacity, interpolated pixels are effectively utilized for the data embedding process using modular arithmetic. The original cover pixels are also employed to embed data using the difference expansion method. The framework developed is tested with standard benchmark images and medical images. The experimental results prove that the proposed method proffers a better output in comparison with the other contemporary methods in this domain.

Introduction

In the digital era, the protection of confidential data from unauthorized access is highly demanded. Reversible data hiding (RDH) plays an integral role in secure communication because the embedding process is done in such a way that the hidden data is unnoticeable to attackers. It has a variety of applications in the area of medical imagery, military imagery, etc. Clinical information can be embedded in a medical image and transmitted to a doctor who is sitting anywhere across the globe. Even a minute alteration in the original medical image is unacceptable for correct diagnosis. One can use the data hiding algorithm, which reversible [1, 2], so that the receiver can extricate the hidden data and also construct the original image.

Usually, the two main classifications of data hiding algorithms are reversible and irreversible. In the irreversible category, the restoration of the original image is not possible, but it renders high embedding rate. The effectiveness of a data-hiding scheme is evaluated mainly using parameters like embedding capacity and visual quality. Embedding capacity or payload capacity gives the total number of secret bits, which can be embedded into a cover image using the given algorithm. Succeeding the data embedding process, the cover image is called the stego image and is needed to compute the visual similarity between these two images. Normally, there is a trade-off between these two parameters. Thus, it is a challenging research area. Recently, a substantial bulk of work has been done in the field of RDH. Some researchers developed an algorithm in the spatial domain, while some have utilized frequency domain.

Image-based RDH schemes are appropriate for applications like medical imagery because we can hide large amounts of electronic patient record (EPR) data into a medical image. Interpolation is a technique used to find out unknown values using known information. Interpolation is used for up-sampling and the interpolated pixels are used for data embedding. A few research works are based on this idea and the original pixels are untouched in the process of data embedding. A novel approach has been put into use in this paper where along with the interpolated pixels, the original pixels were also utilized for data embedding so that the payload capacity of the algorithm is very high and we could reconstruct the original image as such. Recently remote diagnosis plays a key role in health care [3], and the security aspect of medical images in this context has gained significance.

The main stimulus for this work is the desire to implement a high-capacity RDH scheme suitable for medical applications. The considerable offerings of the proposed work are:

  • 1

    A new image interpolation technique that outperforms the recently published interpolation techniques.

  • 2

    Implementation of a high-capacity data-hiding algorithm, which is reversible in nature.

  • 3

    On the basis of the analysis of the latest literature in accordance to the up-sampling and data embedding, it was observed that the original pixels were untouched. To boost the payload capacity of the proposed algorithm, the original pixels have also been utilized.

  • 4

    The outcome of the experiments authenticate that the prospective algorithm produces excellent visual quality stego images.

The remaining paper is structured in this way; Sect. 2 examines the analogous work in this area. Sect. 3 discusses the recommended method, and experimental results and analysis are given in Section 4. Finally, the concluding remarks are given.

Section snippets

Related works

Various internet of things (IoT) based electronic health architectures have been proposed but securing EPR data still needs research attention. Tian [4] developed a difference expansion technique for data hiding, and their maximum embedding capacity is 0.5 bit per pixel (bpp) only. Lou et al. [5] suggested an RDH scheme, which was based on adaptive difference expansion and their results are better than the previous methods. Pawar et al. [6] suggested a method based on shifting of the histograms

Proposed method

Fig. 1 exemplifies the block diagram of the propounded algorithm. Initially, the original input image is up-sampled using the proposed interpolation technique discussed in subSection 3.1 to produce the cover image. After the cover image generation, the data embedding process is performed. Secret data is embedded into the interpolated pixel using a modular arithmetic operation discussed in subSection 3.2. To boost the payload capacity of the algorithm original reference pixels are also utilized

Experimental results and analysis

To prove the efficiency of the proposed technique, experiments were carried out on nine standard images (shown in Fig. 4) of the general category taken from literature. The technique was then applied to six medical images (shown in Fig. 5) obtained from the DICOM library to demonstrate how efficiently the electronic health records can be embedded in the image. For embedding purposes, we have used data generated by a pseudo-random function. Experiments were conducted using MATLAB R2015a.

Eq. (18)

Conclusions

A high-capacity RDH technique mastered for disclosing clinical information via medical images has been proposed. The developed scheme exploits the concept of image interpolation. A new weighted interpolation has been designed which assigns additional weightage to the nearest pixels and our proposed image interpolation dominates few interpolation techniques of recent times. Also, the interpolation in the most recent works consumed more computational time for generating the cover image because it

CRediT authorship contribution statement

P.V. Sabeen Govind: Conceptualization, Methodology, Data curation, Writing - original draft. M.V. Judy: Supervision, Validation, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Sabeen Govind P V is currently pursuing Ph.D. from Cochin University of Science and Technology, Kerala, India. He completed his M. Tech in Computer Science from University of Kerala. He is working as an Assistant Professor in Rajagiri College of Social Sciences (Autonomous), Kerala. His research interests include Image processing and Data hiding.

References (25)

  • T. Chuen et al.

    High capacity reversible hiding scheme based on interpolation, difference expansion and histogram shifting

    Multimed Tools Appl, Springer

    (2014)
  • J. Tian

    Reversible data embedding using a difference expansion

    IEEE Trans Circu Syst Video Technol

    (2003)
  • Cited by (12)

    • Reversible data hiding with enhancing contrast and preserving brightness in medical image

      2022, Journal of Information Security and Applications
      Citation Excerpt :

      And these methods are superior when large amounts of patient-identifying information need to be embedded. Govind et al. [17] proposed a high capacity reversible data hiding technique that can be used to embed patient data using a new weighted interpolation technique, the payload capacity has been improved in the method. Bhardwaj et al. [18] proposed a novel encrypted dual image reversible data hiding, this method gave very high embedding rates on both natural and medical image.

    • A high-capacity and reversible patient data hiding scheme for telemedicine

      2022, Biomedical Signal Processing and Control
      Citation Excerpt :

      But the cover image generated via this method has an identical weakness (i.e. block effect) with the nearest neighbor interpolated image. Afterward, a novel image scaled-up scheme [29] is developed to form a cover image, and modular arithmetic and difference expansion is applied for data embedding. This algorithm achieves an embedding rate of 2.3 bit per pixel (bpp) for peak signal noise ratio (PSNR) of 36 dB.

    View all citing articles on Scopus

    Sabeen Govind P V is currently pursuing Ph.D. from Cochin University of Science and Technology, Kerala, India. He completed his M. Tech in Computer Science from University of Kerala. He is working as an Assistant Professor in Rajagiri College of Social Sciences (Autonomous), Kerala. His research interests include Image processing and Data hiding.

    Judy M V, Ph.D. in Computer Science, and currently working as an Associate Professor in the Department of Computer Applications, Cochin University of Science and Technology, Kerala, India. Her main research interests include Computational biology, Evolutionary algorithms, Data mining, Big data analytics and Image processing.

    This paper is for CAEE special section SI-ipsa Reviews processed and recommended for publication to the Editor-in-Chief by Area Editor Dr. E. Cabal-Yepez.

    View full text