06-05-2011, 11:30 AM
Abstract—
Medical image is usually comprised of region ofinterest (ROI) and region of non interest (RONI). ROI is theregion that contains the important information from diagnosispoint of view so it must be stored without any distortion. Wehave proposed a digital watermarking technique which avoidsthe distortion of image in ROI by embedding the watermarkinformation in RONI. The watermark is comprised of patientinformation, hospital logo and message authentication code,computed using hash function. Earlier BCH encryption ofwatermark is performed to ensure inaccessibility of embeddeddata to the adversaries.
Keywords- BCH Encryption, Data authentication, ElectronicPatient Record, Mecdical Images, Spatial domain watermarking
I. INTRODUCTION
DUE to the development of latest technology incommunications and computer networks, exchange ofmedical images between hospitals has become a usualpractice now a days [1]. These medical images areexchanged for number of reasons. For exampleteleconferences among clinicians, interdisciplinary exchangebetween clinicians and radiologists for consultative purposesor to discuss diagnostic and therapeutic measures and fordistant learning of medical personnel [2]. However theseapplications require more attention towards image protection(availability, confidentiality and reliability) [3]. To facilitatesharing and remote handling of medical images in a securemanner watermarking guarantee attractive properties [4,5].It allows permanent association of image content withproofs of its reliability by modifying the image pixel values,independently of the image file format.For protecting the digital images two categories ofwatermarking have been developed: Robust watermarking[6] and Fragile Watermarking [7,8]. Robust watermarks arethose which are difficult to remove from the digital content.Withstand intentional or incidental distortions like compression, scaling, cropping, filtering, A/D or D/Aconversion, etc. and therefore are used for copyrightapplications. Fragile watermarks are those that are easilydestroyed by tampering or modifying the watermarkedcontent hence the absence of watermark to the previouslywatermarked content points to the conclusion that data hasbeen tampered with, and thus are used for dataauthentication applications.One can use the fragile watermarking for authenticationof medical images. In this case, the robustness of watermarkin the image is less concerned, while detection andlocalization of the slight changes of the images are moreimportant [1].In this paper we have extended our work [19] and haveproposed a blind fragile watermarking scheme that does notrequire the original host image during the extraction ofwatermark. First the image has been segmented whichseparates the lung parenchyma from rest of CT scan imagethen three different types of watermarks are embedded in thehost image by replacing the least significant bits (LSBs) ofthe cover segmented image. LSB is a simple non robustembedding technique with a high embedding capacity andsmall embedding distortions. The LSBs of image aregenerally considered as noise caused by the imaging device.So, these bits can be used for secret message embeddingwithout greatly disturbing the image appearance [9]. Thescheme serves for both the purposes of medical imageauthentication and hiding electronic patient record.The portions of an image that contains the significantinformation for diagnosis are called region of interest (ROI)and must be stored without distortion. Usually it is desirableto embed data outside of ROI to give better protection [10]without compromising with the diagnosis information.
II. LITERATURE REVIEW
Different groups of authors have contributed number ofmedical image watermarking techniques. A technique ofembedding electronic patient record (EPR) data in medicalimages is suggested by Acharya et al. [11]. EPR data consistof text file and graphs. Text file is the preliminary reportabout the patient from the radiology department of thehospital and graphs are ECG or EEG. It is an LSB techniqueimplemented in spatial domain. The ASCII characters inEPR data are encrypted before interleaving in medicalimages to improve the security of the data.
Download full report
http://ieeexplore.ieeeiel5/5235874/52671...er=5268167