Index patients, including enhanced patient autonomy, improved clinical

Index Terms—DICOM, Image Encryption,
Initial state, High speed scrambling, Medical image,                              Pixel adaptive diffusion

 

1) INTRODUCTION

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

 

1.1BACKGROUND

 

 

 

      

Medical imaging is
the technique and process of   creating visual
representations of the interior of a body called medical images. These medical
images are used for clinical analysis and medical intervention. Medical images
are used to diagnose and treat various kinds of diseases. The
availability of electronic data within the modern health information
infrastructure presents significant benefits for medical providers and
patients, including enhanced patient autonomy, improved clinical treatment,
advances in health research and public health surveillance. Advances in information
and communication technologies provide new means to access, handle and move
medical records, but at the same time, the medical records compromise their security
due to their ease of accessing. Medical images involve a lot of privacy of the
patients which are very confidential and sensitive Disastrous events may occur if
these private images are stolen, viewed or used by unauthorized accesses. A
hacker or a malicious attacker may use the images for their personal benefits
like medical marketing and fraudulent insurance claims, which may cause life-threatening
risks. Therefore, protecting medical images is quite important. Many ways have
been devised to protect the medical images but out of which the most efficient
one is encrypting the image since encryption provides integrity,
confidentiality, non-repudiation and authentication. Only with the correct key
one can recover the original image. The Digital Imaging and Communication in
Medicine (DICOM) Standard is used globally to store, exchange, and
transmit medical images.    Recently
various encryption schemes based on PN  
sequence, scrambling, neural networks, rotation matrix etc are proposed
to provide medical images with high level of security1-3.In 4 Weijia Cao et
al. presented a medical image encryption algorithm using edge maps derived from
source image. Many kinds of edge maps with various edge detectors can be
applied as the keys.In 5 the medical image is
encrypted by performing image partition and encryption. Yuling Liu
proposes a ROI based reversible data hiding scheme in this paper. Finally the
encrypted medical images are combined with LSB to recover ROI losslessly. In
6 Eric Yong et al.
proposed an encryption scheme based on properties of Fridrich’s chaotic image      encryption scheme. In 7 the authors propose
an encryption algorithm based on quaternion rotation.

9Performs two phase operations. In first
phase to perform rotations based on binary key and in second phase to generate
pseudo-random numbers to perform pixel value permutation to encrypt medical
image.. 10 Proposes an encryption algorithm using the matrix product and
exclusive addition.
11 First compresses the medical image in
the discrete wavelet transform (DWT) domain and then encrypts the image with an
algorithm based on basic pixel permutation and randomness. Although there are
various encryption schemes as mentioned above to protect medical images some of
them have weaknesses in different aspects. For e.g.In 4 the
techniques are not suitable for bulky data, e.g. medical images, and they are
vulnerable to the differential attack because of their low security level. The
method in 5 is applicable only in the case of a cover image with a
small ROI and also the quality of the stego-image gets reduced when the data
size is increased.6 Uses cryptanalyzed image encryption schemes and survey
reveals that most of these cryptanalyzed image encryption schemes employ permutation
and/or substitution stages and weaknesses in them could be identified using
simple known/chosen plain text attacks. This is largely because of their weak
design and blind dependency only on chaotic properties of key-stream for
security without incorporation of operations that can establish intricate confusion
and diffusion characteristics in the encryption process. In 9 Experimental analysis shows proposed technique
does not provide any statistical information through histogram analysis. In
11 the main drawbacks are: 1)for fine analysis, it becomes computationally
intensive 2)its discretization, the discrete wavelet transform is less
efficient and natural.

 

1.2CONTRIBUTIONS

 

In
order to overcome the problems in existing encryption algorithms an encryption
scheme based on substitution permutation is proposed. In this encryption scheme
first random data is inserted into the plain image and two rounds of high speed
scrambling and pixel adaptive diffusion is done in order to shuffle the pixel
positions and to spread the inserted data over the entire range. In this way it
provides a high level of security of the medical images. This method can be
applied to medical images of any representation format. In this proposed
encryption scheme pixel adaptive diffusion is implemented using two operations
namely Bitwise-Xor(BX) and Modulo-Arithmetic(MA).Some key advantages of the proposed
encryption scheme are as follows: (1)Key size is variable and can be as long as
256 bits.(2)It is robust against data loss and impulse noise.(3)It can encrypt
identical images into different cipher 
images even using same key and this is possible because of the random
data insertion.(4)provides two operations to implement pixel adaptive diffusion
: BX and MA in which the former has high efficiency in hardware platforms and
latter can achieve higher speed in software platforms.

 

2)
PROPOSED ENCRYPTION SCHEME

Secret
key K has a length of 256 bits and this key is decomposed to provide keys for
scrambling and diffusion operation respectively.

2.1KEY
STRUCTURE

The
256 bit key is decomposed into 6 parts namely x0, r, R1, R2 of 52
bits each and d1, d2 of each 24 bits. In this encryption
scheme the pseudo random numbers for high speed scrambling and pixel adaptive
diffusion is generated by using the Logistic Sine System (LSS) proposed in 8
which is given below:

 

,r€0,4,(iteration variable)€0,1.Using initial state (x0,r)
and LSS pseudo random numbers are generated for the two rounds of high speed
scrambling and pixel adaptive diffusion. The variables x0, r, R1, and
R2 are float numbers that can be converted from a 52-bit stream by

                             FN=

Two
integers d1
and d2 can be obtained from a 24 bit stream by

                    

The
initial states(x0, r)
for two rounds of encryption can be obtained as follows

Where
i= {1, 2}.

 

2.2RANDOM
DATA INSERTION

Randomly
generated data is inserted into the plain image. The data inserted can be
regarded as a noise that is added to the plain image. The inserted data will be
different for each execution and hence a high level of security can be provided.
Consider that the plain image is of size HW.Generate two random vectors R and 0