IEEM Programming Procedure For Detecting Boundary Of Carotid Artery

Document Sample
IEEM Programming Procedure For Detecting Boundary Of Carotid Artery Powered By Docstoc
					                                                    (IJCSIS) International Journal of Computer Science and Information Security,
                                                    Vol. 8, No. 5, August 2010




  IEEM PROGRAMMING PROCEDURE FOR
   DETECTING BOUNDARY OF CAROTID
              ARTERY
                 V.Savithri                                                    Dr.S.Purushothaman
   Professor,Department of Computer Science                        Principal, Sun College of Engineering and
      Mother Teresa Women’s University                                            Technology
            Kodaikanal, INDIA                                                   Kanyakumari, INDIA
           Savi3_8@yahoo.co.in                                              dr.purushothaman.s@gmail.com


Abstract--This paper presents an IEEM programming
Abstract--                                                         represents, in fact, the most powerful instrument today
procedure for use on noisy B-mode ultrasound images
                            B-                                     available for predictingcoronary disorders in people over
of the carotid artery. This programming procedure is               fifty .Images are taken by ultrasound equipment working at
based on Image Enhancement,            Edge detection and          frequencies ranging from 1 to 20 MHz, and are obtained via
Morphological operations in Boundary detection. This               a probe that has to be positioned on the patient’s neck.
procedure may simplify the job of the practitioner
                                      of                           [3]There are two phases to the measurement process: i)
for analyzing accuracy and variability of segmentation             carotid image capturingand ii) thickness measurement.
results. Possible plaque regions are also highlighted. A                 Ultrasound plays an important role in diagnosis and
thorough evaluation of the method in the clinical                  illness and injury. The noninvasive imaging of different
environment shows that inter       observer variability is         parts of the body has other applications in medical
evidently
evidently decreased      and so is the overall analysis            diagnosis, such as in tissue characterization and
time. The       results   demonstrate that it has the              measurement of tissue motion.The measurement of tissue
potential to      perform    qualitatively better     than         motion[9] is a broad category. It can include the analysis of
applying existing methods             in    intima     and         the motion of physically active organs such as the heart and
adventitial layer detection on B-mode images.
adventitial                     B-                                 discrete structures such as cardiac valves and arterial walls.
                                                                   It can refer to the analysis of motion induced in passive
                                                                   tissues, such as liver or lung, due to active organs such as
Keywords- Artery, boundary detection , imaging,
Keywords-                                                          the heart or by external sources such as low frequency
Ultrasonic, parallel programming.                                  vibration or compression. The response of the tissue is a
                                                                   function its elasticity, which is directly related to the
                                                                   healthiness of the tissue. [12]The measurement of blood
                   I. Introduction                                 flow velocity is an important parameter in diagnosing
                                                                   vascular diseases such as venous thrombosis. Regardless of
           Over the last few years, image processing has           the particular medical application of ultrasound, all
been playing an increasingly important role in many                applications require the transmission and capture of radio-
scientific areas. This is due, among other reasons, to the         frequency (RF) ultrasonic signals. The signals must be
ever-improving performance of computers that are now               processed in some way to extract the desired information. In
capable of quickly processing the characteristically large         the case of assessing tissue motion, Doppler ultrasound has
amounts of data produced by images. This processing is             been very popular, particularly in the measurement of blood
mostly oriented toward extracting either qualitative or            flow. Practically all commercial ultrasound blood flow
quantitative information from object images. In particular,        measurement systems utilize the frequency-domain Doppler
[1][2]precise dimensional characterization of objects              technique. Doppler techniques have been around for a long
through contact-less measurement techniques is a very              time, and have been extensively covered in the literature. In
important task in several environments such as industrial          addition to Doppler, however, time-domain methods of
quality and process control and medical diagnosis.[4] A            measuring tissue motion also exits, which have not been
typical application field of medical image processing is in        comprehensively reviewed.[17]Time-domain methods have
the diagnosis of atherosclerosis.[5]The atherosclerosis            potential advantages over Doppler techniques in many
process is strongly linked to carotid thickening and               applications, and the use of time-domain based methods is
plaques, whose presence can be clearly detected in                 becoming more and more widespread.
artery longitudinal section images provided by ultrasound                Due to[6] the huge amount of information
                                                                                (
                                                                   intravascular(IVUS) images are increasing their role in the
techniques.[6]The analysisof   the   carotid    ultrasound
                                                                   diagnosis and treatment of several diseases. Manual




                                                              85                                http://sites.google.com/site/ijcsis/
                                                                                                ISSN 1947-5500
                                                          (IJCSIS) International Journal of Computer Science and Information Security,
                                                          Vol. 8, No. 5, August 2010




segmentation is slow and lacks of objectivity.[10]                        Image brightness[11] may be improved by modifying
Consequently, automatic segmentation and tracking of the                  the histogram of the image.
vessel inner wall in IVUS images has been approached in
several recent works.[8] The poor quality of the images                   Image Denoising:
suggests the use of techniques such as probabilities or fuzzy
logic guiding an active contour to adjust the inner wall.                 An image is often corrupted by noise in its acquisition or
                                                                          transmission. Noise is any undesired information that
       II. Materials and Methods                                          contaminates an image. Noise appears in images from a
Problem Definition                                                        variety of sources. The goal of denoising is to remove the
                                                                          noise while retaining as much as possible the important
                                                                          signal features. Denoising can be done through filtering.
          The noise created during ultrasound scanning
                                                                          Filters reduces noise. Gaussian highpass filter helps to
leads to difficulty in defining the boundary of the vessel.
                                                                          reduces a noise.
The image is further deteriorated by the occurrence of lipid
rich plaque a poorly angled transducer during image
                                                                          % H = guassian_filter(co,ro,fo);
acquisition. Difficult in highlighting plaque region.
                                                                          H = 1 – H;
                                                                          Out = Zeros(co,ro);
BOUNDARY DETECTION                                                        Outf = imf. * H;
Ultrasonic Artery Images                                                  Out = abs ( ifft2(outf));
                                                                          Imshow(im), title (‘Original image’), figure,
                                                                           Imshow((out)), title (‘Filtered Image’) figure,
                                                                                         title(‘2D
                                                                           Imshow (H), title(‘2D view of B’), figure, surf(H),
                                                                           Title (‘3D view of H’)

                                                                          Edge Detection:

                                                                          The IEEM defines edges as Zero-crossings of second
       1(a)
Figure 1(a) Carotid artery image. (b) Definition of echo zones and        derivatives in the direction of the greatest first derivate. This
                            interfaces.                                   works in multistage process (i) image is smoothed by
                                                                          Gaussian convolution (ii) 2D first derivate operator is
                                                                          applied to the smoothed image to highlight region of the
Figure. 1(a) shows a representative image of a carotid artery.            image with high spatial derivatives. The effectiveness of
The femoral artery has a similar appearance. The echoes in                this algorithm is determined by three parameters (i) width of
the region of interest can be schematically grouped into                  the Gaussian kernel (ii) upper threshold (iii) lower threshold
seven echo zones Z1–Z7 [Figure. 1(b)]. Previous studies                   used by tracker.
[12], [13] have shown that the leading edge (upper side) of
Z3, Z5, and Z7, denoted as I3, I5, and I7, can be mapped to               Morphological operations for Boundary detection:
the near-wall intima–lumen interface, the far-wall lumen–
intima interface, and the far-wall media–adventitia interface,            Morphological operations are very effective in the detection
respectively. Consequently, the distancebetween I3 and I5                 of boundaries in a binary image X. The following boundary
represents the LD and the distance between I5 and I7 is the               detectors are widely used:
far-wall IMT.With this understanding, the determination of
ultrasonic measurement of the artery becomes equivalent to                  Y=X–(X B)
accurately detecting the echo boundaries I3, I5, and I7.                    Y=(X         B ) = X or
          [14]The femoral artery has a similar appearance.
                                                                            Y = (X       B) - ( X    B)
The echoes in the region of interest can be schematically
                                                                          where Y is the boundary image, operator denotes erosion
grouped into seven echo zones Z1–Z7 .[15][16] Previous
studies have shown that the leading edge (upper side) of Z3,              operator     denotes dilation ‘ – ‘ denotes set theoretical
Z5, and Z7, denoted as I3, I5, and I7, can be mapped to the               subtraction.
near-wall intima–lumen interface, the far-wall lumen–intima
interface, and the far-wall media–adventitia interface,                   %Boundary detector
respectively. Consequently, the distance between I3 and I5                Close all;
represents the LD and the distance between I5 and I7 is the               Clear all;
far-wall IMT. With this understanding,[14 the determination               Clc;
of ultrasonic measurement of the artery becomes equivalent                a=imread(‘carotid.jpg’);
to accurately detecting the echo boundaries.                              b=[010;111;010];
                                                                          a1=imdilate(a,b);
                                                                          a2=imerode(a,b);
                                                                          a3=a-a2;
     III. IEEM programming procedure                                      a4=a1-a;
Image Enhancement:                                                        a5=a1-a2;
                                                                          imshow(a)
Histogram equalization provides more                    visually          figure,imshow(a1),title(‘Dilated Image’)
pleasing results across a wider range of                images.           figure, imshow(a2),title(‘Eroded Image’)


                                                                     86                                 http://sites.google.com/site/ijcsis/
                                                                                                        ISSN 1947-5500
                                               (IJCSIS) International Journal of Computer Science and Information Security,
                                               Vol. 8, No. 5, August 2010




Segmentation of the intima-media region:
                    intima-

We introduce a new method for the segmentation of
the imtima-media region in ultrasound images, which
combines splines(for the adventitia detection),
dynamic       programming(dp),       smooth  intensity
thresholding surfaces and a successful geometric
active contour model and known for its accuracy,                                         G-canny



flexibility and robustness. Several image features are
used in the segmentation. Human interaction is
minimal. It is able to segment both near-end and far-
end carotid walls; it supports to detect plaques of
different sizes, shapes and classes.

         IV. Results and discussion

                    TABLE 1
       READING VARIABILITY(%) WHEN
   MEASUREMENTS WERE PERFORMED BY
  THREE READERS BEFORE APPLYING IEEM
                PROCEDURE.                                                            morphed
READER1 READER2 READER3 AVERAGE
(accuracy)  (accuracy) (accuracy) (accuracy %)
    82%        80%         84%        82%



                                                                                     TABLE 2
                                                                    READING VARIABILITY(%) WHEN
                                                                 MEASUREMENTS WERE PERFORMED BY
                                                                 THREE READERS AFTER APPLYING IEEM
                 Original image
                                                                            PROCEDURE.

                                                              READER1         READER2              READER3        AVERAGE
                                                              (accuracy)      (accuracy)           (accuracy)     (accuracy %)
                                                                  94%           94.5%                  96%           94.5%




                                                                                    V. Conclusion

                     gray
                                                                        In conclusion, We have proposed a method based
                                                              on IEEM programming procedure to automatically measure
                                                              ultrasonic artery images. The human knowledge of the
                                                              artery image is incorporated in the system, which makes the
                                                              system capable in processing images of different quality.
                                                              Human factors in the determination of the boundaries are
                                                              reduced. Evaluation of the system shows reduced inter
                                                              observer variability as well as overall analysis time. The
                                                              automated artery boundary detection system can replace the
                                                              old manual system in a clinical application environment.

                      indexed




                                                         87                                 http://sites.google.com/site/ijcsis/
                                                                                            ISSN 1947-5500
                                                      (IJCSIS) International Journal of Computer Science and Information Security,
                                                      Vol. 8, No. 5, August 2010




Acknowledgement                                                           principles and description of a computerized analyzing
                                                                          system,” Clin. Physiol., vol. 6, no. 11, pp. 565–577,
                                                                          1991.
         We would like to thank members of                           [13] P. Pignoli, E. Tremoli, A. Poli, and R. Paoletti, “Intimal
Madras Medical College and New hope Scan center                           plus medial thickness of the arterial wall: A direct
And Hospital , in the Department of Radiologist                           measurement with ultrasound imaging,” Circulation,
and Neurologist ,and Dr.P.Kamalakannan, offered                           vol. 74, pp. 1399–1406, 1986.
valuable comments advices and for providing                          [14] R. E. Bellman and S. Dreyfus, Appled Dynamic
patients personal reports                                                 Programming. Princeton, NJ: Princeton University
                                                                          Press, 1962.
References                                                           [15] A. A. Amini, T. E.Weymouth, and R. C. Jain, “Using
                                                                          dynamic programming for solving variational problems
[1] R. Chelappa et al., “The past, present, and future of
                                                                          in vision,” IEEE Trans. Pattern Anal. Machine Intell.,
     image and multidimensional signal processing,” IEEE
                                                                          vol. 12, pp. 855–867, Sept. 1990.
     Signal Processing Mag., vol. 15, pp. 21–58, Mar. 1998.
                                                                     [16] R. J. Kozick, “Detectinh interfaces on ultrasound
[2] D. H. Ballard and C. M. Brown, Computer V ision.
                                                                          images of the carotid artery by dynamic programming,”
     Englewood Cliffs, NJ: Prentice-Hall, 1982.
[3] L. Angrisani, P. Daponte, C. Liguori, and A. Pietrosanto,
                                                                          SPIE, vol. 2666, pp. 233–241, 1996.
                                                                     [17] W. Liang, R. Briwning, R. Lauer, and M. Sonka,
     “An image based measurement system for the
                                                                          “Automated analysis of brachial ultrasound time
     characterization of automotive gaskets,” Measurement,
                                                                          series,” in Proc. SPIE Conf. Physiol. Function
     vol. 25, pp. 169–181, 1999.
[4] M. G. Bond and S. K. Wilmoth et al., “Detection and
                                                                          Multidimensional Images, vol. SPIE 3337, San Diego,
                                                                          CA, Feb. 1998, pp. 108–118.
     Monitoring of Asymptomatic Atherosclerosisi in
     Clinical Trials,” Amer. J. Med., vol.86, (suppl 4A), pp.
     33–36, 1989.
[5] N. M. El-Barghouty, T. Levine, S. Ladva, A. Flanagan,
     and A. Nicoladeis, “Histological verification of                                          Savithri Vedachalam , working
     computerized carotid plaque characterization,” Eur. J.                                    as professor in Department of
     V ascular Endovascular Surg., vol. 11, pp. 414–416,                                       Computer Application in Hindu
     1996.                                                                                     College, Pattabiram, Chennai,.
[6] F. De Man, I. De Scheerder, M.C. Herregods, J. Piessens                                    She received her degrees,
     and H. De Geest Role of Intravascular Ultrasound in                                       M.Phil.(C.S.) from Alagappa
     Coronary Artery Disease: A new gold standart?                                             University and M.C.A. from
     Beyond Angiography. Intravascular Ultrasound State-             Annamalai University, India.Her area of research includes
     Of-The-Art XX Congres of the ESC, Vol 1 (August                 Medical Imaging, Image Processing, Object tracking, 3-D
     1998)                                                           Image analysis. She published more than 10 papers in
[7] D. Hausmann, Andre J.S. Lundkvist, Guy Friedrich,                national , International Conferences and Journals.
     Krishnankutty Sudhir, Peter J. Fitzgerald and Paul G.
     Yock Lumen and Plaque Shape in Atherosclerotic
     Coronary Arteries Assesed by In EVO Intracoronary               .
     Ultrasound Beyond Angiography. Intravascular
     Ultrasound: State-Of-The-Art XX Congres of the ESC,
     Vol 1 (August 1998)                                                                    . Dr. S. Purushothaman is working as
[8] F. Escolano, M. Cazorla, D. Gallardo and R. Rizo                                        Principal and professor in Sun College
     Deformable Templates for Plaque Thickness                                              of Engineering and Technology,
     Estimation of Intravascular Ultrasound Sequences                                       Nagerkoil, India. He received his Ph.D
     Pattern Recognition and Image Analysis. Preprints of                                   from IIT Madras, M.E from Anna
     the VI1 National Symp. On Patt. Recog. and Im.                                         University Chennai and B.E from PSG
     An.Vol 1 (April 1997)                                                                  College of Technology, Coimbatore
[9] M.A.Bottalico, A.Starita, “EcoStudio:A computer tool                                    His area of research includes Artificial
     to      support      carotid     ultrasound      images         Neural Networks, Image Processing and signal processing.
     analysis,Engineering in Medicine and Biology                    He published more than 50 research papers in national and
     Soc.,IEEE,pp.2428-2430,2000.                                    international journals.
[10]    Song Chun          Zhu, Alan        Yuille, Region
     Competition:Unzfiing Snakes, Region Growing, and
     BayesIMDL for Multiband Image Segmentation. IEEE
     Trans. Pattern An. Mach. Intelligence, Vol. 18, No 9 , (
     September 1996).
[11] Nobuyuki Otsu A Threshold Selection Method from
     Gray-Level Histograms. IEEE Trans. on Sys. Man and
     Cybernetics, Vol. SMC-9,Na 1, pp 62-65, (January
     1979)
[12] I. Wendelhag, T. Gustavsson, M. Suurkula, G.
     Berglund, and J. Wikstrand, “Ultrasound measurement
     of wall thickness in the carotid artery. Fundamental

                                                                88                                http://sites.google.com/site/ijcsis/
                                                                                                  ISSN 1947-5500

				
DOCUMENT INFO
Description: IJCSIS invites authors to submit their original and unpublished work that communicates current research on information assurance and security regarding both the theoretical and methodological aspects, as well as various applications in solving real world information security problems.