[雙語(yǔ)翻譯]外文翻譯-醫(yī)學(xué)圖像水印基準(zhǔn)（英文）

1、A Benchmark for Medical Image Watermarking Navas K. A, Sasikumar M and Sreevidya S Electronics and Communication Engineering Department College of Engineering, Trivandrum, Kerala, India, 695016 Phone: (91) 471-2515658

2、 Fax: (91) 471-2598370 E-mail: kanavas@rediffmail.com Keywords: EPR, Data hiding, Benchmark Abstract - The medical images with EPR embedded in it can be used for transmission, storage or telemedicine application

3、s. There is a need of specific standards for the evaluation of watermarking techniques used for embedding EPR data on medical images. No existing benchmark addresses this issue. There are no universally accepted perf

4、ormance measures applicable for every watermarking system. In this paper a benchmark is proposed for the evaluation of medical image watermarking and data hiding techniques. 1. INTRODUCTIONHiding patient data in the m

5、edical 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 wate

6、rmarking for authentication and EPR hiding. 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 wi

7、th medical images will save storage space of the Hospital Information System, enhance confidentiality of the patient data and save the bandwidth required for transmission. Obviously this will reduce the cost of diagn

8、osis. 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 medical images.

9、 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

10、typical non-medical image degradation processes. They do not provide an evaluation scheme applicable for specific medical image types, or for typical degradations arising from medical image processing. 2. MEDICAL IM

11、AGE WATERMARKING TECHNIQUESAlmost all the earlier works in medical image watermarking have focused mainly on two areas: 1. Tamper detection and authentication and 2. Embedding EPR in medical images. Tamper detection w

12、atermarks 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 wate

13、rmarking techniques are prone 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 larg

14、e enough for the clinicians to write their diagnosis report. Some of the available watermarking techniques used for embedding text information into medical images can be found in [2,3,4]. Popular Benchmarks: The imp

15、ortant 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 methods. This

16、 makes a room for research on devising a benchmark for all kinds of image watermarking. 3. A NOVEL BENCHMARK FOR MEDICAL IMAGE WATERMARKINGAn ideal benchmarking procedure should involve examining the set of mutually d

17、ependent parameters of the watermarking system and it should clearly optimize the trade off between various constraints of watermarking. Various performance metrics are used to evaluate these parameters based on a sp

18、ecific application. The requirements of watermarking such as imperceptibility, capacity and robustness are hampering each other. Therefore, a trade off is essential between these parameters. A proper evaluation h

19、as 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 followin

20、g constraints. 3.1 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 Informatio

21、n System 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

22、 image database server with the DICOM standard. Digital watermarking can imperceptibly embed messages without changing image size or format. So the watermarked medical image can conform to the DICOM format. 3.2 Capa

23、city Though the capacity of watermark is expressed in bits per pixel, more convenient unit that can be generally 237reveals that the LSB plane contains a large amount of redundancy. Each character in the EPR data is en

24、coded using 7-bits and watermarked into the redundant bits of LSB plane. Fig.(2) Distortion due to watermarking in LSB planes The capacity can be further improved by inserting the watermark into the higher order bit p

25、lanes. Fig. 2 illustrates the distortion occurred in the cover image when 3456 bits of the watermark were inserted into six planes and it could be seen that, fourth plane onwards the distortion became visible. Variat

26、ion in the different imperceptibility measures is tabulated in Table 1. 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

27、 6 0.56 19.8 0.726 Table 1. Variation in imperceptibility measures 4.1 Visual quality Vs Capacity The Visual quality Vs capacity graph is used to estimate the maximum number of characters (MNEC) that can be embedded

28、 into the cover image within the imperceptibility limits. Using LSB method, encoded text information was embedded into cover images of different modalities. The WPSNR values were calculated for various amounts of emb

29、edded characters. The results obtained for images of different modalities were averaged. The WPSNR value that ensures imperceptibility of watermark was found to be 40 dB. It was observed that LSB techniques ensure mi

30、nimum degradation to cover image. For evaluating WPSNR, error (difference between cover image and watermarked image) was scaled by the corresponding NVF values evaluated at each pixel. It was found that CT images pro

31、vide the highest value of imperceptibility for the given number of embedded characters compared to images in other modalities. This is due to the high contrast between adjacent regions in CT images. Capacity of the

32、ROI mapped to horizontal details was calculated. Finally, the capacity of horizontal details without ROI was calculated as per (1). In the blind watermarking scenario, two different pseudorandom sequences were embedde

33、d in the ROI and non-ROI regions so that the watermark detector can correctly identify the ROI. As expected, the number of bits that could be embedded into the horizontal details decreased with the increase in the si

34、ze of ROI. Fig.(2) Degradation in visual quality with capacity 4.2 Visual quality Vs Attack strength The degradation in the visual quality of the watermark is illustrated using Visual quality Vs Attack strength graph.

35、WPSNR decreased with increase in the variance of Gaussian noise. The WPSNR used to measure visual degradation of medical images with noise uses Contrast Sensitivity Function (CSF) as the weighting factor. The frequenc

36、y response of CSF is modeled as a band pass filter and the error signal is filtered by this BPF. Fig.(2) Degradation in visual quality with attack4.3 Bit error rate Vs Attack strength Bit error rate Vs Attack strength

37、graph is used to find out the robustness of the watermark against various attacks. The bit error rate between the original and extracted watermark increased with the increase in the variance of speckle noise. It was