外文翻譯--醫(yī)學(xué)圖像水印基準(zhǔn)

1、外文翻譯-醫(yī)學(xué)圖像水印基準(zhǔn) 江西理工大學(xué)應(yīng)用科學(xué)學(xué)院畢業(yè)設(shè)計(jì)論文外文資料翻譯請(qǐng)注意各個(gè)局部我寫的批注:首頁(yè)請(qǐng)嚴(yán)格按照模板書寫,橫線要上下對(duì)齊 系 : 信息工程系 專 業(yè): 網(wǎng)絡(luò)工程 班 級(jí):071 姓 名:邢星 學(xué) 號(hào):38 附 件: 1.外文資料翻譯譯文;2.外文原文。 指導(dǎo)教師評(píng)語:簽名: 年 月 日 注:請(qǐng)將該封面與附件裝訂成冊(cè)。附件1 外文資料翻譯譯文:中文的譯文在前,英文原文在后,中文要求小四號(hào)字,段落單倍行距,最低要求:大題目、各章節(jié)題目和摘要必須根本正確和通順,英文原文中圖和表、參考文獻(xiàn)可以不翻譯 醫(yī)學(xué)圖像水印基準(zhǔn) 摘要-嵌入EPR的醫(yī)療圖像可以用于傳輸,存儲(chǔ)或者遠(yuǎn)距離的醫(yī)學(xué)應(yīng)用

2、。這需要一個(gè)特定的標(biāo)準(zhǔn)去評(píng)價(jià)嵌入EPR數(shù)據(jù)到醫(yī)療圖像的數(shù)字水印技術(shù)。目前,沒有存在的基準(zhǔn)能去解決這個(gè)問題。水印系統(tǒng)還沒有一個(gè)被普遍接受的性能指標(biāo)。本文提出了一種基準(zhǔn)去評(píng)價(jià)醫(yī)學(xué)圖像水印和數(shù)據(jù)隱藏技術(shù)。 數(shù)字圖像水印的一個(gè)應(yīng)用是在醫(yī)療圖像中隱藏患者的數(shù)據(jù)。病人的數(shù)據(jù)是電子格式的被稱為電子遍歷(EPR)。 所有的工作報(bào)告以數(shù)據(jù)的形式隱藏在醫(yī)療圖像中去驗(yàn)證和隱藏EPR。以不同形式附加EPR的醫(yī)療圖像可以被發(fā)送給世界各地的臨床醫(yī)生去診斷病情。在醫(yī)療圖像中嵌入EPR將會(huì)節(jié)省醫(yī)院信息系統(tǒng)的存儲(chǔ)空間,增強(qiáng)患者數(shù)據(jù)的機(jī)密性并且節(jié)約傳輸所需的帶寬。顯然這個(gè)技術(shù)將會(huì)減少診斷的花費(fèi)。這個(gè)系統(tǒng)可以用數(shù)字水印技術(shù)確保所需

3、要的高等級(jí)平安性。 醫(yī)療圖像數(shù)字水印缺乏系統(tǒng)的標(biāo)準(zhǔn)和條例。醫(yī)療圖像數(shù)字水印團(tuán)體普及世界各地,他們需要一個(gè)標(biāo)準(zhǔn)去交換全局的信息。當(dāng)前流行的基準(zhǔn)集中于細(xì)微評(píng)價(jià)和在傳統(tǒng)非醫(yī)療圖像的退化處理。它們不需要為特定的醫(yī)療圖像類型提供一個(gè)評(píng)價(jià)方案體系或者為了從醫(yī)療圖像處理中產(chǎn)生典型的退化。 早期的醫(yī)療圖像數(shù)字水印技術(shù)幾乎集中于主要兩個(gè)方面:1.篡改檢測(cè)和驗(yàn)證,2.在醫(yī)療圖像中嵌入EPR。篡改檢測(cè)水印用于識(shí)別進(jìn)行操作的醫(yī)療圖像。EPR數(shù)據(jù)可以利用空間域技術(shù)以及改造域技術(shù)嵌入進(jìn)醫(yī)療圖像?？臻g域水印技術(shù)有退化的傾向。嵌入技術(shù)必須無損因?yàn)樵卺t(yī)療應(yīng)用有嚴(yán)格的高質(zhì)量要求。嵌入的比特?cái)?shù)量應(yīng)該足夠大以便于臨床醫(yī)生可以寫他們的

4、診斷報(bào)告。一些有效的水印技術(shù)被用來嵌入文本信息到醫(yī)療圖像中,可以在2,3,4中發(fā)現(xiàn)。流行的基準(zhǔn):非常重要的可靠基準(zhǔn)是Stirmark, Checkmark, Optimark 和 Certimark。所有這些為評(píng)估各種類型平臺(tái)的數(shù)字水印的基準(zhǔn)方法有著共同的低效率。這給所有類型的圖像水印研究設(shè)計(jì)一個(gè)基準(zhǔn)留出了空間。 一個(gè)理想的基準(zhǔn)程序應(yīng)包括審查水印系統(tǒng)參數(shù)的相互依賴關(guān)系,它應(yīng)明確優(yōu)化權(quán)衡水印各方面的約束。使用一個(gè)特定的應(yīng)用來評(píng)估這些參數(shù)多方面的性能度量。 數(shù)字水印的需求比方不可見性,容量和魯棒性,它們互相沖突。因此在這些參數(shù)中必須仔細(xì)權(quán)衡。適宜的評(píng)估必須在一定程度上確保所有的選擇符合需求。醫(yī)療圖

5、像水印的評(píng)估方法不同于其他的基準(zhǔn)因?yàn)橛邢旅娴南拗啤?.1 圖像覆蓋 基準(zhǔn)包含不同大小一定數(shù)量的覆蓋圖像。醫(yī)療圖像可以在不同的形式中使用比方CT,MRI,US和X光。醫(yī)院信息系統(tǒng)包含綜合醫(yī)療圖像數(shù)據(jù)庫(kù)和檢索系統(tǒng),可以讓醫(yī)生在任何時(shí)候去瀏覽患者的圖像。這種系統(tǒng)允許醫(yī)學(xué)圖像以不同的方式被納入到DICOM標(biāo)準(zhǔn)的圖像數(shù)據(jù)庫(kù)效勞器。數(shù)字水印可以在不改變圖像大小或格式的情況下不知不覺的嵌入信息。所以醫(yī)療圖像數(shù)字水印應(yīng)該遵循DICOM格式。3.2 容量 盡管水印的容量是用比特每像素表達(dá)的,但是更方便的單位可以普遍適用于隱藏在醫(yī)療圖像中的EPR文本數(shù)據(jù)稱為最大數(shù)目的嵌入式字符(MNEC)。醫(yī)療圖像數(shù)字水印的容量

6、必須盡可能的大。這就要消除對(duì)注解隱藏,消息驗(yàn)證,初次信息報(bào)告和詳細(xì)診斷報(bào)告可用空間的限制。3.3 不可見性 水印后圖像質(zhì)量評(píng)價(jià)是為了測(cè)量水印之后的圖像失真量。峰值信號(hào)信噪比(PSNR)和均方誤差(MSE)是最廣泛的用于客觀評(píng)估圖像質(zhì)量/失真指標(biāo),但是它們沒有很好的和質(zhì)量評(píng)估進(jìn)行關(guān)聯(lián)。然而,覆蓋圖像的某一局部可以有效的標(biāo)記水印的存在。錯(cuò)誤的信號(hào)可以被人眼當(dāng)做視覺質(zhì)量評(píng)估的噪聲。HVS的重要隱蔽效果將在下面的局部中進(jìn)行解釋。視覺遮蔽:當(dāng)一個(gè)圖像的組件是頻率和方向時(shí),圖像的組件在人眼中將會(huì)不可見。這個(gè)重要的遮蔽效果稱為亮度屏蔽,比照屏蔽和紋理屏蔽。人眼對(duì)平滑區(qū)域的改變比對(duì)紋理結(jié)構(gòu)區(qū)域的改變更為敏感。

7、紋理遮蔽的效果被局部頻率的分發(fā)和紋理方向所決定。一個(gè)參數(shù)用來描述紋理遮蔽效果稱為噪聲可見函數(shù)(NVF)。結(jié)構(gòu)相似測(cè)量:另一個(gè)可感知的度量用來模擬醫(yī)療圖像的水印退化,被稱為結(jié)構(gòu)相似測(cè)量(SSIM)。圖像質(zhì)量評(píng)估以SSIM為根底基于一個(gè)事實(shí),即是高度適應(yīng)的人類視覺系統(tǒng)從視場(chǎng)提取結(jié)構(gòu)信息。SSIM度量是為了測(cè)試醫(yī)療圖像的相似性,因?yàn)镾SIM更集中于局部而不是全局圖像的一致性。沃森度量:覆蓋圖像的非常規(guī)化的區(qū)域和亮度的高度改變會(huì)掩飾水印的存在。這個(gè)現(xiàn)象是被沃森模型給定的。該模型的根本目的是使用圖像塊的相應(yīng)靈敏度閾值來加權(quán)DCT系數(shù)。該閥值是亮度遮蔽和比照遮蔽的混合功能。沃森度量使用最小可絕差(JND)

8、單位來計(jì)算水印圖像的感知錯(cuò)誤。3.4 感興趣區(qū)域(ROI) 一個(gè)重要的考慮因素是使醫(yī)學(xué)圖像嵌入水印讓醫(yī)學(xué)圖像包含感興趣區(qū)域(ROI)。在醫(yī)療圖像中,ROI包括重要的特征信息并且處理不能有任何失真。ROI通常在空間域中被選擇。在空間域水印技術(shù)中,非感興趣局部的像素可以直接被改變。Capacity-NVF-ROI Measure:水印的容量被認(rèn)為是使用低可見性的錯(cuò)誤嵌入進(jìn)特殊覆蓋圖像的比特的數(shù)量。因此容量應(yīng)該與圖像的內(nèi)容聯(lián)系起來。覆蓋圖像的容量被評(píng)估為 CW?21+2/ n22是MWI的方差,n2是方差的噪音,W取決于像素的數(shù)量。如果一個(gè)圖像大小為N×N,WN×N/2。3.5

9、攻擊 系統(tǒng)性能的評(píng)估基準(zhǔn)是在存儲(chǔ)和醫(yī)療圖像的傳輸期間的典型處理操作。各種類型的噪音在傳輸過程中通常會(huì)降低醫(yī)療圖像,并且整體噪聲可以作為高斯建模。噪聲歸因于圖像長(zhǎng)期的存儲(chǔ)被作為斑點(diǎn)模型。3.6 魯棒性 各種醫(yī)療圖像水印魯棒性的處理操作可以在嵌入的信息和萃取的信息中使用比特誤碼率(BER)來評(píng)估。BER的評(píng)估是通過改變每個(gè)降級(jí)過程的強(qiáng)度來實(shí)現(xiàn)的。 為了鑒定覆蓋圖像的噪聲可見區(qū)域,每個(gè)像素會(huì)計(jì)算NVF數(shù)值。首先為了計(jì)算NVF數(shù)值會(huì)使用一個(gè)3×3的鄰域來計(jì)算局部方差。128×128大小,8-bit灰階的MRI圖像心臟磁共振圖像被用作覆蓋圖片。該NVF圖像包含了相應(yīng)的覆蓋圖像,如圖1

10、所示。結(jié)果發(fā)現(xiàn),NVF值在邊緣和紋理局部接近0,而圖像的平滑局部接近于1。 圖.1 覆蓋圖像和它的NVF示意圖 在LSB技術(shù)中,覆蓋圖像中每個(gè)像素的最低有效位可以使用水印來修改。在LSB平面的有效比特總數(shù)為16384比特。如此多數(shù)量的比特足以滿足醫(yī)療圖像的EPR數(shù)據(jù)隱藏的容量需求。這個(gè)覆蓋圖像的LSB平面的分析暴露了LSB平面包含了大量的冗余。EPR數(shù)據(jù)中的每個(gè)字符被7個(gè)比特編碼并且在LSB平面的冗余位上印上水印。圖.2 LSB平面內(nèi)水印的失真 可以進(jìn)一步提升容量通過在高階位平面插入水印。圖2說明了在覆蓋圖像中失真的發(fā)生當(dāng)在6個(gè)平面中插入了3456比特的水印,然后可以看到從第四個(gè)平面開始失真變

11、的可見。測(cè)量不同的細(xì)微變化在表1中顯示。 LSB Plane SSIM PSNR dB Watson metric 1 0.98 49.4 0.028 2 0.92 43.3 0.056 3 0.84 37.4 0.110 4 0.75 31.4 0.210 5 0.65 25.6 0.394 6 0.56 19.8 0.726 4.1 視覺品質(zhì) Vs 容量 視覺品質(zhì) Vs 容量圖表習(xí)慣于用來估計(jì)可隱蔽性的嵌入到圖像覆蓋范圍內(nèi)的最大字符數(shù)量(MNEC)。使用LSB方法,編碼的文本信息以不同的方式嵌入到覆蓋圖像中。通過嵌入進(jìn)不同數(shù)量的字符數(shù)來計(jì)算WPSNR數(shù)值。以不同方式得到的圖像結(jié)果的平均值。

12、在40分貝下WPSNR數(shù)值確保水印的不可見性。研究發(fā)現(xiàn),LSB的技術(shù)可確保覆蓋圖像的最低退化。 為了評(píng)估WPSNR,誤差(覆蓋圖像和水印圖像的差異)按比例由每一個(gè)像素相應(yīng)的NVF數(shù)值來評(píng)估。結(jié)果發(fā)現(xiàn),與其他方式相比,在嵌入進(jìn)一定數(shù)目字符的情況下,CT圖像提供了最高等級(jí)的隱蔽性。這是由于CT圖像臨近的區(qū)域有鮮明的比照。 ROI的容量映射為水平清晰度是可以被計(jì)算出來的。最后,按照per1計(jì)算沒有ROI的水平清晰度的容量.在盲水印方案中,兩個(gè)不同的偽隨機(jī)序列號(hào)被嵌入進(jìn)ROI和非ROI區(qū)域,以便于水印檢測(cè)器可以正確的區(qū)分ROI。正如預(yù)期的那樣,嵌入進(jìn)水平清晰度的比特減少的數(shù)量對(duì)應(yīng)于ROI增加的大小,如

13、圖(3)所示。 圖.3 視覺品質(zhì)與容量的退化4.2 視覺品質(zhì) Vs 攻擊強(qiáng)度 使用視覺品質(zhì) Vs 攻擊強(qiáng)度圖表來說明水印視覺品質(zhì)的退化。WPSNR的減少伴隨著高斯噪聲方差的增加。 用于測(cè)量噪聲醫(yī)學(xué)圖像視覺退化的WPSNR使用比照靈敏度(CSF)作為權(quán)重因數(shù)。CSF的響應(yīng)頻率作為帶通濾光器和被帶通濾光器過濾的錯(cuò)誤信號(hào)的參考,如圖(4)所示。 圖.4 視覺品質(zhì)與攻擊的退化4.3 比特錯(cuò)誤率 Vs 攻擊強(qiáng)度 比特錯(cuò)誤率 Vs 攻擊強(qiáng)度圖表被用來發(fā)現(xiàn)水印對(duì)各種攻擊的魯棒性。在原始的水印和提取的水印中誤碼率隨著斑點(diǎn)噪聲方差的增加而增加。 很顯然微波域中水印的容量明顯小于空間域的,但是魯棒性強(qiáng)于空間域。這

14、是因?yàn)槭褂每臻g域水印嵌入技術(shù)對(duì)像素的操作很敏感,然而在嵌入進(jìn)微波次能帶的時(shí)候它剩余了未改變的局部,如圖(5)所示。 圖.5 比特錯(cuò)誤率和攻擊的退化 本文為醫(yī)療圖像的文本隱藏提出了一個(gè)基準(zhǔn)。討論了容量的界限,不可見性和魯棒性。這些基準(zhǔn)標(biāo)準(zhǔn)將會(huì)非常好的被全球的ROI圖像數(shù)據(jù)隱藏團(tuán)體用來設(shè)計(jì)和評(píng)估算法。作者參與了為這個(gè)領(lǐng)域的研究人員制定一個(gè)完整的基準(zhǔn)數(shù)據(jù)表的工作。 參 考 文 獻(xiàn)1 M. Kutter and F. A. P. Petitcolas, “A fair benchmark for image watermarking systems, Electronic Imaging 99. Se

15、curity and Watermarking of MultimediaContents, USA, vol. 3657, January 1999.2 S. Dandapat, Opas Chutatape and S. M. Krishnan,“Perceptual model based Data Embedding in a Medical Image, Proc. Int. Conf. on Image Processing, vol.4,pp. 2315-2318, October 2004.3 Xuanwen Luo, Qiang Cheng and Joseph Tan, “

16、A Lossless Data Embedding Scheme for Medical Images in Application of e-Diagnosis, Proc. 25th Annual Int.Conf. of the IEEE EMBS, Mexico, vol.1, pp.i-c, September 2003.4 A. Nikolaidis S. Tsekeridou A. Tefas V Solachidis, “A Survey on Watermarking Application Scenarios and Related Attacks, Proc. Int.

17、Conf. Image Processing,vol.3, pp.991-994, Oct. 2001.5 Rajendra Acharya U., P. Subhanna Bhat, Sathish Kumar and Lim Choo Min, “Transmission and storage of medical images with patient information, Journal ofComputers in Biology and Medicine, vol. 33, pp.303-310, 2003.6 Deepthi Anand and U.C. Niranjan,

18、 “Watermarking medical images with patient information, Proceedings of the 20th Annual International Conference of theIEEE Engineering in Medicine and Biology Society, vol. 2, pp.703-706, Nov.1998.7 G. Coatrieux, H. Maitre, B. Sankur, Y. Rolland and R.Collorec, “Relevance of Watermarking in Medical

19、Imaging,IEEE-EMBS Information Technology Applications in Biomedicine, pp.250-255, 2001.8 Xuanwen Luo and Qiang Cheng, “Health Information Integrating and Size Reducing, Proc. IEEE Nuclear Science Symposium, Medical Imaging Conference and Workshop of Room-Temperature Semiconductor Detectors, 2003.外文原

20、文:英文原文在后,要求從PDF格式的原文中粘貼形成word格式文章,字體新羅馬小四號(hào)字,英文每個(gè)段落字體要兩端對(duì)齊A Benchmark for Medical Image WatermarkingAbstract - The medical images with EPR embedded in it can be used for transmission, storage or telemedicine applications. There is a need of specific standards for the evaluation of watermarking techni

21、ques used for embedding EPR data on medical images. No existing benchmark addresses this issue. There are no universally accepted performance measures applicable for every watermarking system. In this paper a benchmark is proposed for the evaluation of medical image watermarking and data hiding tech

22、niques1. INTRODUCTION Hiding patient data in the medical image is one of the applications of digital image watermarking. The patient data in the electronic format is called Electronic patient record EPR. All works reported in data hiding in medical image are watermarking for authentication and EPR h

23、iding. The medical images of different modalities with EPR attached to them can be sent to the clinicians residing at any corner of the globe for the diagnosis. Embedding of EPR with medical images will save storage space of the Hospital Information System, enhance confidentiality of the patient dat

24、a and save the bandwidth required for transmission. Obviously this will reduce the cost of diagnosis. This kind of a system requires a high level of security, which can be ensured by using digital watermarking techniques Literature is devoid of a systematic norms or regulations for watermarking medi

25、cal images. Medical image watermarking communities around the world need a standard benchmark for the exchange of information globally. The currently popular benchmarks focus on evaluating imperceptibility and robustness under typical non-medical image degradation processes. They do not provide an e

26、valuation scheme applicable for specific medical image types, or for typical degradations arising from medical image processing2. MEDICAL IMAGE WATERMARKING TECHNIQUES Almost all the earlier works in medical image watermarking have focused mainly on two areas: 1. Tamper detection and authentication

27、and 2. Embedding EPR in medical images. Tamper detection watermarks are used for identifying manipulations done on medical images. EPR data can be embedded into the medical image using spatial domain techniques as well as transform domain techniques. Spatial domain watermarking techniques are prone

28、to degradations. The embedding technique must be lossless because of the stringent requirements on high quality in medical applications; however the number of embedded bits should be large enough for the clinicians to write their diagnosis report. Some of the available watermarking techniques used f

29、or embedding text information into medical images can be found in 2,3,4Popular Benchmarks: The important available benchmarks are Stirmark, Checkmark, Optimark and Certimark. All the these benchmarks share the common inefficiency of providing a platform for evaluating all kinds of image watermarking

30、 methods. This makes a room for research on devising a benchmark for all kinds of image watermarking3. A NOVEL BENCHMARK FOR MEDICAL IMAGE WATERMARKINGAn ideal benchmarking procedure should involve examining the set of mutually dependent parameters of the watermarking system and it should clearly op

31、timize the trade off between various constraints of watermarking. Various performance metrics are used to evaluate these parameters based on a specific applicationThe requirements of watermarking such as imperceptibility, capacity and robustness are hampering each other. Therefore, a trade off is es

32、sential between these parameters. A proper evaluation has to ensure that all the selected requirements are met to a certain level of assurance. The evaluation method for medical image watermarking techniques differs from the other benchmarks because of the following constraints The requirements of w

33、atermarking such as imperceptibility, capacity and robustness are hampering each other. Therefore, a trade off is essential between these parameters. A proper evaluation has to ensure that all the selected requirements are met to a certain level of assurance. The evaluation method for medical image

34、watermarking techniques differs from the other benchmarks because of the following constraints3.1 Cover Image Set The benchmark incorporates a number of cover images of varying size. The medical images are available in different modalities such as CT, MRI, US, and X-ray. The Hospital Information Sys

35、tem contains Integrated Medical Image Database and Retrieval System that enables doctors to browse patient images at any time. Such a system allows medical images in different modalities to be integrated into an image database server with the DICOM standard. Digital watermarking can imperceptibly em

36、bed messages without changing image size or format. So the watermarked medical image can conform to the DICOM format3.2 Capacity Though the capacity of watermark is expressed in bits per pixel, more convenient unit that can be generally applied to EPR text data hiding in Medical images is imum Numbe

37、r of Embedded Characters MNEC. For medical image watermarking, the capacity must be as high as possible. This is to remove a constraint of available space for hiding annotations, authentication message, first information report and detailed diagnosis report.3.3 Imperceptibility Measures The quality

38、assessment of an image after watermarking is done to measure the amount of distortion due to the watermarking. Peak Signal-to-Noise Ratio PSNR and Mean Squared Error MSE are the most widely used objective image quality/distortion metrics, but they are not correlating well with perceived quality meas

39、urement. However, certain portions of the cover image can effectively mask the presence of the watermark. The error signals that are visible to human eye need to be taken as noise for visual quality assessment. The important masking effects of HVS are explained in the following section1 Visual Maski

40、ng: When an image component is in the frequency and orientation, that image component becomes less conspicuous to the human eye. The important masking effects are Luminance Masking, Contrast masking and Texture masking. Human eye is less sensitive to changes in textured areas than in smooth areas. T

41、he texture masking effect is determined by local frequency distribution and texture direction. The texture masking effect is described using a parameter called Noise Visibility Function NVF2. Structural Similarity Measure: Another perceptual metric used to model the degradation of watermarked medica

42、l images is the Structural Similarity Measure SSIM. Image quality assessment based on SSIM is based on the fact that the HVS is highly adapted to extract structural information from the viewing field. SSIM metric is ideal for testing of similarities in medical images because it focuses on local rath

43、er than global image similarity3 .Watson Metric: Regions of non-regular and highly changing luminance in the cover image are able to mask the presence of watermark. This phenomenon is given by Watson model. The basic aim of the model is to weight the DCT coefficients in an image block by its corresp

44、onding sensitivity threshold. The threshold is a compound function of luminance masking and contrast masking. Watson metric is used to calculate the perceptual error in the watermarked image in Just Noticeable Difference JND units.3.4 Region of Interest ROI An important factor to be considered while

45、 watermarking medical images is that medical images contain Region of Interest ROI. In medical images, ROI is an area that contains diagnostically important information and must be processed without any distortion. The ROI is usually selected in the spatial domain. In spatial domain watermarking tec

46、hniques, the pixels in non-ROI parts can be modified directly1. Capacity-NVF-ROI Measure: The watermark capacity is considered as the number of bits that can be embedded into the particular cover image with low error visibility. Therefore the capacity measure must be associated with the content of i

47、mage. The capacity of the cover image is evaluated as, C W log21+2/ n2 1 Where2 is the variance of MWI andn2 is the noise variance and W depends upon the number of pixels. For an image of size N × N, W N × N/2.3.5 Attacks The benchmark evaluates the performance of the system under typical

48、processing operations during storage and transmission of medical images. Various types of noises usually degrade medical images during transmission and the overall noise can be modeled as Gaussian. The noise due to long-term storage of the image is modeled as speckle noise3.6 Robustness Measure The

49、robustness of the watermark to various medical image-processing operations can be evaluated using the Bit Error Rate BER between the embedded message and the extracted message. The BER is evaluated by varying the strength of each degradation process.4. RESULTS AND DISCUSSION In order to identify the

50、 regions of noise visibility in the cover images, NVF values were calculated at each pixel. First local variance was measured using a 3×3 neighborhood in order to calculate NVF values. 128×128 size, 8-bit gray scale MRI image of heart was used as the cover image. The NVF image obtained cor

51、responds to the cover image is shown in Fig. 1. It was found that NVF values were close to 0 in edges and textured portions, whereas it was close to 1 in flat portions of the image. Fig.1 Cover image and its NVF map In the LSB technique, the least significant bit of each pixel in the cover image is

52、modified using the watermark. The total number of bits available in the LSB plane was 16384 bits. This much amount of bits is sufficient to meet the capacity requirements of EPR data hiding in medical images. The analysis of LSB plane of the cover image reveals that the LSB plane contains a large amount of redundancy. Each character in the EPR data is encoded using 7-bits and watermarked into the redundant bits of LSB planeFig.2 Distortion due to watermarking in LSB planes The capacity ca