International Journal of Applied Engineering Research ISSN 0973-4562 Volume 9, Number 22 (2014) pp. 12163-12176 © Research India Publications http://www.ripublication.com
Novel Algorithm for Secure Medical Image Communication using ROI Based Digital Lossless Watermarking and DS A. Umamageswari #1 and G.R. Suresh*2 #
Research Scholar, Department of CSE, Sathyabama University, Chennai, India Professor, Department of ECE, Easwari Engineering College, Chennai, India 1
[email protected] [email protected]
*
Abstract Protection of Medical Image content is very much important in medical image Communication for telediagnosis and telesurgery. Digital lossless watermarking becomes the promising technique to secure the medical content in medical images. To enhance the security this paper proposes SHA-1 with a novel algorithm ACC (Advanced Classical Cipher) for the production of DS (Digital Signature) to achieve high confidentiality and Authentication. An image is compressed using JPEG2000 (DWT) algorithm then medical image content and security information is embedded in RONI (Region of Non Interest) of compressed image using Lossless Watermarking Technique. When medical images are shared through open networks, watermarking also helps to convey integrity and authenticity. Lossless Watermarking is the promising technique to get original image and embedded watermark in the receiver side without any loss of Information. The PSNR (peak Signal to Noise ratio) value is up to 69dBs for all types of DICOM images. Increase in Authentication can be achieved when medical expert’s access secured medical images from the web servers using Kerberos technique. Keyword: Lossless Watermarking, Medical Image Security, medical Image Compression, Authentication and Confidentiality, JPEG2000 Compression, Kerberos, SHA-1, ACC.
INTRODUCTION Sharing of Medical image is used in variety of applications like remote diagnosis, telesurgeries and e-learning [1]. Security can be achieved in HIS (Hospital information System) and PACs environment using DICOM security standards as per HIPAA. Critically injured patients can be treated locally by exchanging of their I.
Paper Code: 27431-IJAER
12164
A. Umamageswari and G.R. Suresh
Medical Images between hospitals located in different countries for giving best treatment to the patients by getting opinions from specialist available in various countries [2][3]. Medical industry expects high level of security algorithms to maintain confidentiality, authenticity, integrity, robustness and information hiding techniques for data transmission [4]. Enforcing content protection using classical access control mechanism is no longer sufficient. It is that necessary to develop security mechanism that guarantee protection of medical image contents in an autonomous way, especially their integrity and traceability. Digital watermarking has been shown as a mechanism to enhance medical image security [5] .Before images is made available into the open network definitely we should compress the medical images to effectively uses the bandwidth. Image compression algorithms can be divided into two main groups, lossless and lossy methods. In lossless compression schemes, only the redundancy is exploited, and the image is recorded in a more efficient manner. All the information is retained and so the reconstructed image is numerically identical to the original image. Lossy compression information deemed irrelevant to the visual perception of the human viewer is discarded and so the compressed image cannot be perfectly reconstructed and distortion is introduced into the reconstructed image [6]. JPEG2000 offers numerous advantages over the JPEG standard. It also offers both lossy and lossless compression. When high quality is concern, JPEG2000 using Discrete Wavelet Transformation promises a high quality final image, even when using lossy compression and also it offers higher compression ratio [7]. Digital Signature algorithms are important on protecting confidential information [8]. To generate the Digital Signature (DS), hash value of the medical image is calculated using SHA-1(Secure Hash Algorithm-1). The algorithm is an iterative, one way hash function that can process image to produce a condensed representation called a message digest. This algorithm enables the integrity of a message to be determined and any change to the message will, with a very high probability, result in a different message digest [9] [10]. This message digest is encrypted using our own algorithm ACC (Advanced Classical Cipher) to produce the Digital Signature. The DS is encrypted hash value [11]. Digital Signature and medical content EMR (Electronic Medical Record) is embedded in the RONI (Region of non interest) using Modified Difference of Expansion Lossless Watermarking technique. The basic principle of our approach is to share a medical image with Knowledge Digest (KD). The proposed KD gives the medical description and interpretation of the image content [12]. To share medical images with some extra header information, unfortunately header files are prone to manipulation and information loss may occur during file format conversion. For Example, most data contained in the header of a DICOM (Digital Imaging and Communication) [13] image file will be lost after conversion into another multimedia format. The combination of Medical image knowledge digest and Digital Signature (DS) of the medical image will be the watermark. The data hiding scheme should have a large embedding capacity to carry more information. Main goal of Digital Lossless watermarking is to protect the copyright and can recover the original image [14]. Lossless watermarking can also be defined on the schemes which can recover the original image form the embedded
Novel Algorithm for Secure Medical Image Communication
12165
image [15] [16]. The watermarked images are shared through the web servers. The medical experts who are accessing the images should be registered with the web servers through their user id and password [17]. The strict authentication can be provided to these medical experts by using Kerberos. Kerberos introduces intermediate server which has the database all the medical experts should register their user id and password with this database. The intermediate authentication server produces ticket to access the medical images which are available in the websites, so the doctors registered properly with the websites through this Kerberos only can able to access the message [18][19]. After embedding the watermark inside into an image, image quality can be calculated by Peak Signal to Noise Ratio (PSNR) using Root Mean Square Error (RMSE) and compression Ratio. Compression Ratio and PSNR should be high for better quality image [20] [21]. Compression Ratio can be calculated by the ratio between the size of the image before compression and size of the image after compression [22] [23]. (1) The quality of the watermarked image is measured by PSNR. Bigger in PSNR better in quality of watermarked image. PSNR for image with size M×N is given by (2) (3) Where f (m, n) is pixel gray values of the original image. f1 (m, n) is pixel gray values of watermarked image. The PSNR value based on payload or capacity rate (watermark size). The rest of this paper is organized as follows. In section 2 algorithms used for proposed is described [A.JPEG2000 Image compression algorithm, B. SHA-1 Secure Hash algorithm, C. Advanced Classical Cipher-ACC, D. Lossless WatermarkingModified Difference of Expansion, E. Algorithm for Kerberos] and then results and discussion for this proposed work is discussed in section 3. The paper concludes in section 4.
II. METHODOLOGY USED A. JPEG2000 Image Compression
The JPEG 2000 image compression consists of four basic steps in the algorithm-preprocess, transformation. In our work we implemented JPEG2000 compression without quantization because medical images contains sensitive information, these information should not get lost during compression. JPEG2000 utilizes a new coding method called Embedded Block Coding with Optimized Truncation (EBCOT). Step 1: Pre-processing: Image is decomposed to components to maximum of 256. These components are decomposed into rectangular tiles. Step 2: Transformation: JPEG2000 uses discrete wavelet Transformation (DWT). Each tile is decomposed into different resolution levels, these levels are made up of sub bands of coefficients.
12166
A. Umamageswari and G.R. Suresh
Step 3: Quantization: Sub bands of coefficients are quantized and collected as blocks. Step 4: Entropy Encoding: The bit planes of the coefficients in a code block are entropy encoded. Encoding can be done in such a way that certain ROI can be coded at a higher quality than the background.
Fig 1: Image before Compression
Fig 2: Image after Compression
B. Secure Hash Algorithm-SHA I This Hash algorithm accepts first row of the pixel mapped table of the original image as input and do some confusion and diffusion mathematically to produce the fixed length of output as a message digest value. The output message digest size will be only 160 bits. The following algorithm explains the entire step by step procedure of SHA 1 Step1: Appending Padding Bits. The original message is "padded" (extended) so that its length (in bits) is congruent to 448, modulo 512. The padding rules are: The original message is always padded with one bit "1" first. Then zero or more bits "0" are padded to bring the length of the message up to 64 bits less than a multiple of 512. Step 2: Appending Length. 64 bits are appended to the end of the padded message to indicate the length of the original message in bytes. The rules of appending length are: The length of the original message in bytes is converted to its binary format of 64 bits. If overflow happens, only the low-order 64 bits are used.
Novel Algorithm for Secure Medical Image Communication
12167
Break the 64-bit length into 2 words (32 bits each). The low-order word is appended first and followed by the high-order word. Step 3: Preparing Processing Functions. SHA1 requires 80 processing functions Step 4: Preparing Processing Constants. SHA1 requires 80 processing constant words defined as: K(t) = 0x5A827999 ( 0
