VISIBLE WATERMARKING WITHIN THE REGION OF NON-INTEREST OF MEDICAL IMAGES BASED ON FUZZY C-MEANS AND HARRIS CORNER DETECTION

David C. Wyld (Eds) : ICCSEA, SPPR, CSIA, WimoA - 2013
pp. 161–168, 2013. © CS & IT-CSCP 2013 DOI : 10.5121/csit.2013.3517
VISIBLE WATERMARKING WITHIN THE
REGION OF NON-INTEREST OF MEDICAL
IMAGES BASED ON FUZZY C-MEANS AND
HARRIS CORNER DETECTION
Debalina Biswas1
, Poulami Das 2
, , Prosenjit Maji3
,
Nilanjan Dey 4
, Sheli Sinha Chaudhuri5
1, 2, 3
Department of CSE, JIS College of Engineering, Kalyani, West Bengal,
India
poulamidas1989@gmail.com,biswas_debalina@rediffmail.com,
maji_katm@rediffmail.com
4
Department of IT, JIS College of Engineering, Kalyani, West Bengal, India
dey.nilanjan@ymail.com
5
ETCE Department, Jadavpur University, Kolkata, India
shelism@rediffmail.com
ABSTRACT
Transfer of medical information amongst various hospitals and diagnostic centers for mutual
availability of diagnostic and therapeutic case studies is a very common process. Watermarking
is adding “ownership” information in multimedia contents to verify signal integrity, prove
authenticity and achieve control over the copy process. Distortion in Region of Interest (ROI) of
a bio-medical image caused by watermarking may lead to wrong diagnosis and treatment.
Therefore, proper selection of Region of Non-Interest (RONI) in a medical image is very crucial
for adding watermark. First part of the present work proposes proper selection of Region of
Non-Interest based on Fuzzy C-Means segmentation and Harris corner detection, to improve
retention of diagnostic value lost in embedding ownership information. The second part of the
work presents watermark embedding in the selected area of RONI based on alpha blending
technique. In this approach, the generated watermarked image having an acceptable level of
imperceptibility and distortion is compared to the original image. The Peak Signal to Noise
Ratio (PSNR) of the original image vs. watermarked image is calculated to prove the efficacy of
the proposed method.
KEYWORDS
Fuzzy C-Means, Harris Corner Detection, ROI, RONI
1. INTRODUCTION
In telemedicine, medical images i.e. electro digital data are continuously sent to and received at
various health care centres and hospitals around the world. The obvious constraint faced in this
situation is to detect the organization from which the particular image has been sent and / or
received. Solution to this problem is watermarking [1, 2, 3, 4] the name of the medical centre on
the digitized image of the referred patient. Apart from this, EPR (Electronic Patient Report)
watermarking [5, 6, 7] is also a common form of medical data exchange with authenticity in

162 Computer Science & Information Technology (CS & IT)
telemedicine. However, medical watermarking [8, 9, 10, 11] has its own limitations and requires a
sophisticated execution such that minimal data is tampered.
The most informative area of the medical image is called as Region of Interest (ROI). ROI [12,
13] is the area used for diagnosis by the physician. Any distortion in ROI may lead to wrong
diagnosis and treatment. Hence, for watermarking medical images, area outside ROI in the image
should be chosen. This part of the image is called as Region of Non Interest (RONI).
With the onset of modern medical science, use of several imaging techniques has become the
basement for diagnosis of diseases among the medical professionals. Some of the imaging
techniques are, Magnetic Resonance Imaging (MRI), Computed Tomography (CT), Ultrasound
(US), Skiagram, Positron Emission Tomography (PET), Mammogram, Intravascular Ultrasound
(IVUS) etc.
Intravascular ultrasound is a medical imaging technique, which uses an ultrasound probe
positioned at the tip of a coronary catheter. Intravascular ultrasound is a tomographic imaging
modality performed during coronary angiography. In case of IVUS, ultrasound signal reflected
from the surrounding tissue is reconstructed into a real-time tomographic gray scale image [14].
IVUS can identify the boundaries of different layers of blood vessels such as internal elastic
intima and external elastic intima. Borders between tissues with different composition and
acoustic properties cause strong ultrasound reflections. Two such borders in the coronary artery
walls are the lumen/intimal border and external elastic membrane. The strong signal originating at
these tissue interfaces allows recognition and manual or automated planimetry. Thus, IVUS
becomes capable of measuring lumen area by identified borders of lumen and intima. By
measuring maximum and minimum diameter of External Elastic Lamina, IVUS makes it possible
to diagnose atherosclerotic arteries. IVUS can calculate atheroma through EEM and Lumen CSA
measurements i.e. plaque plus media area [15]. Comprehensively, it can be said that IVUS is
capable of measuring maximum and minimum thickness of plaque along with media, plaque plus
media eccentricity and plaque burden. Length of plaque can be measured by the motorized
transducer pullback function.
MRI (Magnetic Resonance Imaging) is a noninvasive medical technique, which helps the
physicians to diagnose the patients, detect abnormalities and treat medical conditions. It is used to
detect structural abnormalities of the body. MRI is capable of giving images of different organs
and structures inside the body in more detail than X-Ray, ultrasound and CT- Scan. MRI is a
radiology technique that uses magnetism, radio waves and a computer for imaging different parts
of the body [16, 17]. MRI of different parts of the body is suggested for different reasons. MRI of
head is suggested for detecting brain tumors, injuries in the brain, nerve injury, bleeding in the
brain. Spine MRI can check the discs and nerves of the spine for conditions such as spinal
stenosis, disc bulges, and spinal tumors. Chest MRI is done to examine the heart, valves and
lungs. Chest MRI can detect abnormalities in heart and lungs functioning, lung cancer and breast
cancer too. MRI is also capable of imaging blood vessels to detect any abnormalities in the flow
of blood through the vessels. Gall bladder, kidneys, liver, pancreas can also be examined by MRI
for detecting abnormalities [18]. MRI is often suggested for fetal assessment in pregnant women.
While doing MRI the patient is placed on a moveable bed. Then the bed is inserted into a magnet,
which creates a strong magnetic field. The magnetic field aligns the protons of hydrogen atoms,
which are then exposed to a beam of radio waves. This spins the various protons of the body, and
they produce a faint signal that is detected by the receiver portion of the MRI scanner. The
receiver information is processed by a computer, and an image is produced. The image and
resolution produced by MRI is quite detailed and can detect tiny changes of structures within the
body. For some procedures, contrast agents, such as gadolinium, are used to increase the accuracy
of the images.

Computer Science & Information Technology (CS & IT) 163
ROI and RONI marked IVUS and MRI image is shown in the figure 1.
(a)
(b)
Figure 1. (a) ROI and RONI marked IVUS image, (b) ROI and RONI
marked MRI image
2. METHODOLOGY
2.1. Fuzzy C-Means
In pattern recognition, a clustering method known as Fuzzy c-means (FCM) is widely used.
Bezdek proposed FCM in 1973[19, 20]. FCM is also known as Fuzzy ISODATA. In this
clustering technique, one piece of data belongs to two or more clusters. FCM based segmentation
is fuzzy pixel classification. FCM allows data points or pixels to belong to multiple classes with
varying degree of membership function between 0 and 1.
FCM possesses unique advantage of grading linguistic variables to fit for appropriate analysis in
discrete domain on pro-rata basis.
FCM computes cluster centers or Centroids by minimizing the dissimilarity function with the help
of iterative approach. By updating the cluster centers and the membership grades for individual
pixel, FCM [21] shifts the cluster centers to the "right" location within set of pixels.
To accommodate the introduction of fuzzy partitioning, the membership matrix (U) = [uij] is
randomly initialized according to Equation 1, where uij being the degree of membership function
of the data point of ith
cluster xi.
∑=
=∀=
c
i
ij nju
1
,...,1,1
(1)
The performance index (PI) for membership matrix U and ci’s used in FCM is given Equation 2.

uij is between 0 and 1.
ci is the centroid of cluster i.
dij is the Euclidian distance between ith
centroid (ci) and jth
data point.
m є [1, ∞] is a weighting exponent.
To reach a minimum of dissimilarity function there are two conditions. These are given in
Equation 3 and Equation 4.
∑
∑
=
=
= n
j
m
ij
n
j j
m
ij
i
u
xu
c
1
1
(3)
∑ =
−








=
c
k
m
kj
ij
ij
d
d
u
1
)1/(2
1
(4)
3.1.1. Algorithm of FCM
Step 1. The membership matrix (U) that has constraints in Equation 1 is randomly
initiated.
Step 2. Centroids (ci) are calculated by using Equation 3
Step 3. Dissimilarity between centroids and data points is computed using Equation 2.
Stop if its improvement over previous iteration is below a threshold.
Step 4. A new U is computed using Equation 4. Go to Step 2.
3.2. Alpha Blending Technique
Alpha Blending [22, 23] is a well known technique in computer graphics for blending each pixel
of the first source image with the corresponding pixel in the second source image. The equation
for executing alpha blending is as follows,
Final pixel = alpha * (First image's source pixel) + (1.0-alpha) * (Second image's source pixel)
The blending factor or percentage of colors from the first source image used in the blended image
is called the "alpha". The alpha used in algebra is in the range 0.0 to 1.0, instead of 0 to 100%.
Alpha-blending blind Image hiding technique to generate Watermark image is given by
Watermark Embedding:
WII=alpha*(CI) + (1.0-alpha)*(WI) (8)
where, WII=watermarked image, CI=selected sub-band of the cover image, WI= selected
corresponding sub-band of the watermark image.

3. PROPOSED METHOD
Watermark embedding process is explained in Figure 2.
Step 1. Fuzzy C-means Segmentation is applied on gray medical image.
Step 2. Segmented Image is binarized.
Step 3. Binary area open is used to remove small objects from binarized image.
Step 4. Sobel Edge detection is applied on the binarized image.
Step 5. Harris Corner Detection Algorithm [24, 25] is applied.
Step 6. Maximum Harris Diameter is calculated and based on the Harris points the centre
is determined.
Step 7. A circle is drawn based on the centre point and diameter on the sample image to
mark the boundary between Region of Interest (ROI) and Region of Non-Interest
(RONI).
Step 8. An area from the RONI is selected based on the watermark image size.
Step 9. Watermark image is embedded on the selected area based on alpha blending
technique.
The watermarked image is overlaid on the selected position of the original image to generate the
Final watermarked Image.
4. RESULT AND DISCUSSION
MATLAB 7.0.1 Software is extensively used for the study of the medical image watermarking
embedding technique. Concerned images obtained in the result are shown in Figure 2 through 6.
(a) (b) (c) (d)
(e) (f) (g) (h)
(i) (j)
Figure 2. (a) Original IVUS Image, (b) Fuzzy C-Means Segmented Image, (c) Edge detected Image, (d)
Marked ROI, (e) Harris Corner Detected Image, (f) Maximum Harris Diameter, (g) Marked ROI using
Harris Corner detection, (h) ROI, (i) RONI, (j) Watermark Embedded Image

Figure 3. (a) Original MRI Image, (b) Fuzzy C-Means Segmented Image, (c) Edge detected Image, (d)
Marked ROI, (e) Harris Corner Detected Image, (f) Maximum Harris Diameter, (g) Marked ROI using
Harris Corner detection, (h) ROI, (i) RONI, (j) Watermark Embedded Image
Figure 4. (a) 3D Reconstruction of the detected boundary between ROI and RONI of IVUS image, (b) 3D
Reconstruction of the detected boundary between ROI and RONI of MRI image.
Figure 5. Watermark Embedding Using Alpha-Blending Technique
Figure 6. Watermark Embedding into the RONI of Medical Image

Peak Signal to Noise Ratio (PSNR)
It measures the quality of a watermarked signal. This performance metric is used to determine
perceptual transparency of the watermarked signal with respect to original signal [26]:
( )
2
, ,
2
,,,
maxx y x y
x yx yx y
MN P
PSNR
P P
=
−∑ (9)
where, M and N are number of rows and columns in the input signal,
is the original signal and is the watermarked signal.
PSNR between the Original MRI Image Watermark MRI Image is 28.8255 and PSNR between
the Original IVUS Image Watermark IVUS Image is 23.5042 as shown in Table 1.
Table 1.
PSNR
Original MRI Image Vs. Watermark MRI Image 28.8255
Original IVUS Image Vs. Watermark IVUS Image 23.5042
5. CONCLUSION
It is quite common to exchange medical information between various hospitals and diagnostic
centres for mutual availability of diagnostic and therapeutic case studies. EPR (Electronic patient
record) or Hospital Logo can be hidden within a bio medical image for high security than
transferring EPR/Logo through the internet. As the bio medical images like MRI, CT Scan, USG
images convey information for detecting and diagnosis of diseases, exchanging patients’ records
hidden within these images of patients’ diagnostic results requires high reliability. That is why it
is necessary to find out such regions of these images, which do not contain diagnostic
information. In this paper, region of non-interest (RONI) is found out by using Fuzzy C-means
algorithm, which is a clustering method. After identifying the RONI, data is hidden within that
region using watermark embedding. As data is hidden within less informative region, distortion of
that region of the image due to watermark embedding does not cause any loss in the diagnostic
value of the medical image.
REFERENCES
[1] Jessica Fridrich, Miroslav Goljan, Rui Du,”Invertible authentication” In Proc. SPIE, Security and
Watermarking of Multimedia Contents III, Vol. 3971, pp. 197–208, San Jose, USA. (Jan. 2001).
[2] Gouenou Coatrieux, Main, H., Sankur, B., Rolland, Y., & Collorec. R.”Relevance of watermarking in
medical imaging” In IEEE-embs Information Technology Applications in Biomedicine, pp. 250–255,
Arlington, USA. (Nov. 2000).
[3] Nilanjan Dey, Moumita Pal,Achintya Das,”A Session Based Watermarking technique Within the NROI of
Retinal Fundus Images for Authencation Using DWT,Spread Spectrum and Harris Corner Detection”,
International Journal of Modern Engineering Research, Vol.2, Issue 3, May-June 2012, pp-749-757 ( ISSN:
2249-6645) .
[4] Nilanjan Dey, Poulami Das, Achintya Das, Sheli Sinha Chaudhuri, “Feature Analysis for the Blind–
Watermarked Electroencephalogram Signal in Wireless Telemonitoring Using Alattar’s Method”, 5th
International Conference on Security of Information and Networks (SIN 2012), In Technical Cooperation

with ACM Special Interest Group on Security, Audit and Control (SIGSAC). Proceedings by ACM Press
and Digital Library, 22-26 October 2012, Jaipur, India.
[5] Nilanjan Dey, Sayantan Mukhopadhyay, Achintya Das, Sheli Sinha Chaudhuri “Analysis of P-QRS-T
Components Modified by Blind Watermarking Technique Within the Electrocardiogram Signal for
Authentication in Wireless Telecardiology Using DWT”, International Journal of Image, Graphics and
Signal Processing, vol. 4,2012.
[6] Nilanjan Dey, Anamitra Bardhan Roy, Achintya Das, Sheli Sinha Chaudhuri,” Stationary Wavelet
Transformation Based Self-Recovery of Blind-Watermark from Electrocardiogram Signal in the Wireless
Telecardiology”, International Workshop on Intelligence and Security Informatics for International
Security (IIS'12) , 11-12 October 2012, Trivandrum, India, [ Springer in Communications in Computer and
Information Science Series(CCIS), ISSN: 1865:0929. CCIS is abstracted/indexed in ISI Proceedings,
Scopus, EI and DBLP ].
[7] Nilanjan Dey, Prosenjit Maji , Poulami Das, Achintya Das , Sheli Sinha Chaudhuri, “An Edge Based
Watermarking Technique of Medical Images without Devalorizing Diagnostic Parameters”, International
Conference on Advances in technology and Engineering, NMIMS University, Mumbai, India , January 23-
25, 2013.
[8] Arijit Kumar Pal, Poulami Das, Nilanjan Dey “Odd-Even Embedding Scheme Based Modified Reversible
Watermarking Technique using Blueprint”, published in arXiv preprint arXiv: 1303.5972,2013.
[9] Nilanjan Dey, Prasenjit Maji , Poulami Das, Shouvik Biswas, Achintya Das, Sheli Sinha Chaudhuri,
“Embedding of Blink Frequency in Electrooculography Signal using Difference Expansion based
Reversible Watermarking Technique”, Scientific Bulletin of the Politehnica University of Timisoara -
Transactions on Electronics and Communications p-ISSN 1583-3380, vol. 57(71), no. 2, 2012
[10] Nilanjan Dey, Shouvik Biswas , Poulami Das, Achintya Das, Sheli Sinha Chaudhuri,” Lifting Wavelet
Transformation Based Blind Watermarking Technique of Photoplethysmographic Signals in Wireless
Telecardiology”, Second World Congress on Information and Communication Technologies (WICT 2012),
Trivandrum, India: October 30-November 02, 2012.
[11] Nilanjan Dey, Poulami Das, Achintya Das, Sheli Sinha Chaudhuri, “DWT-DCT-SVD Based Intravascular
Ultrasound Video Watermarking “,Second World Congress on Information and Communication
Technologies (WICT 2012), Trivandrum, India: October 30-November 02, 2012.
[12] Baisa L. Gunjal, Suresh N. Mali,”ROI Based Embedded Watermarking of Medical Images for Secured
Communication in Telemedicine” In International Journal of Computer and Communication Engineering.
Vol. 6 (2012).
[13] Malay Kumar Kundu, Sudeb Das, ”Lossless ROI Medical Image Watermarking Technique with Enhanced
Security and High Payload Embedding” In International conference on Pattern Recognition. (1457-1460).
[14] Lin, C. P., Honye, J., Chang, C. J., & Kuo, C. T.”Clinical Application of Intravascular Ultrasound in
Coronary Artery Disease” An Update. ACTA Cardiol Sin. Received. (August 21, 2010).
[15] http://bme240.eng.uci.edu/students/06s/yuhsianh/ Quantitative%20measurement.htm
[16] http://www.medicinenet.com/mri_scan/article.htm
[17] http://www.nlm.nih.gov/medlineplus/ency/article/003335.htm
[18] http://www.webmd.com/a-to-z-guides/magnetic-resonance-imaging-mri
[19] Jacob C Dunn,” A Fuzzy Relative of the ISODATA Process and Its Use in Detecting Compact Well
Separated Clusters” In J. Cyber. Vol. 3. pp. 32-57. (1973)
[20] James C. Bezdek, “Pattern Recognition with Fuzzy Objective Function Algorithms. In Plenum” New York,
USA. (1981).
[21] Nilanjan Dey, Anamitra Bardhan Roy ,Moumita Pal, Achintya Das, ”FCM Based Blood Vessel
Segmentation Method For Retinal Images”, International Journal of Computer Science and Network
(IJCSN) (ISSN 2277– 5420), June 2012,Volume 1, Issue 3.
[22] Akhil pratap singh, Agya Mishra, “Wavelet Based Watermarking on Digital Image” In Indian Journal of
Computer Science and Engineering. Vol. 1(2). pp. 86-91.
[23] Nilanjan Dey, Anamitra Bardhan Roy ,Sayantan Dey, ”A Novel Approach of Color Image Hiding using
RGB Color planes and DWT”, International Journal of Computer Applications Volume 36– No.5,
December 2011, (ISSN: 0975 – 8887),
[24] Chris Harris, Mike Stephens, “A Combined Corner and Edge Detector” In Proceedings of 4th Alvey
Vision Conference. (1988).
[25] Konstantinos G. Derpanis, The Harris Corner Detector”, 2004.
[26] http://ncalculators.com/statistics/signal-noise-ratio-calculation.htm.

VISIBLE WATERMARKING WITHIN THE REGION OF NON-INTEREST OF MEDICAL IMAGES BASED ON FUZZY C-MEANS AND HARRIS CORNER DETECTION

More Related Content

What's hot (18)

Similar to VISIBLE WATERMARKING WITHIN THE REGION OF NON-INTEREST OF MEDICAL IMAGES BASED ON FUZZY C-MEANS AND HARRIS CORNER DETECTION (20)

More from cscpconf (20)

Recently uploaded (20)

VISIBLE WATERMARKING WITHIN THE REGION OF NON-INTEREST OF MEDICAL IMAGES BASED ON FUZZY C-MEANS AND HARRIS CORNER DETECTION