SlideShare ist ein Scribd-Unternehmen logo
1 von 10
Downloaden Sie, um offline zu lesen
Computer Engineering and Intelligent Systems                                                  www.iiste.org
     ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
     Vol 3, No.3, 2012




       Performance Evaluation of Geometric Active Contour (GAC) and
      Enhanced Geometric Active Contour Segmentation Model (ENGAC)
                     for Medical Image Segmentation

                                        Ajala Funmilola Alaba*, Emuoyibofarhe Justice O

           Department of Computer Science and Engineering, LAUTECH Ogbomoso, Oyo state, Nigeria.

              *E-mail of the corresponding author:    funfaith2003@yahoo.co.uk.

Abstract
         Segmentation is an aspect of computer vision that deals with partitioning of an image into homogeneneous region.
Medical image segmentation is an indispensable tool for medical image diagnoses. Geometric active contour (GAC) segmentation
is one of the outstanding model used in machine learning community to solve the problem of medical image segmentation.
However, It has problem of deviation from the true outline of the target feature and it generates spurious edge caused by noise that
normally stop the evolution of the surface to be extracted.
         In this paper, enhanced Geometric active contour was formulated by using Kernel Principal Component Analysis(KPCA)
with the existing Geometric active contour segmentation model and performance evaluation of the formulated model was carried
out.

Keyword: Geometric active contour, Segmentation, Medical image, Kernel Principal Component Analysis.


1.       Introduction

Image segmentation is an aspect of computer vision that deals with the partitioning an image into regions having homogeneous
meaning. It is used to separate an object from its background and extracts meaningful objects lying in images either by dividing
images into contiguous semantic regions, or by extracting one or more specific objects in images such as medical structures.
However, we have different types of image segmentation methods which are not suitable for medical images since medical images
are diverse, complex and vary from natural images.

Some of the methods used for medical image segmentation fall under deformable models. The use of deformable model was
popularized by Kass et al. (1990), snake model was used to develop active contour model that minimizes energy functional under
the influence of forces which are internal force, image force and external constraint force. Deformable models can be classified
into three categories which are Free-form, Parametric and Geometric active contour model.

Geometric active contour model

This make use of level set method that represent contour or surfaces as Zero level set of a higher dimensional function usually
called a level set function Chunming et al, (2011). Level set can handle topological changes in curve evolution which is not
possible with the classical active contour model. Osher and Sethian (1998) offer a natural and numerical implementation of curve
evolution equation using level set. Readers are refer to Neithammer and Tannenbum (2004) for some previous work on Geometric
active contour model. Xiao et al., (2003) gave the numeric scheme used for GAC model, which is describe in the following
equation.

Let x= (x, y) denote a point in a curve C which is a member in image domain Ω and let the curve be represented by Lipchitz-
continuous function using signed distance function A parameterized differentiable curve C is a differential map
                                                  .

                C : a, b      x( p), y( p)  x( p)
                                                       T
                                                                                                                              (1)
                                                                78
Computer Engineering and Intelligent Systems                                                www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.3, 2012


The rule that determine how each point C moves starting from initial curve C o is
         dC P, t 
                     vC P, t , t 
           dt                                                                                                                (2)
Where V = vector field that refers to as speed, Ignoring C and V in equation 2 gives
         dC
            v
         dt                                                                                                                 (3)
Equation (2) gives a partial differential equation that determines how curve C evolves and often called flow. Minimized the energy
below
                     
             E image  , T  H  G (T )
                                           2
                                                        2
                                                  H x  G T , I x
                                                                                                                            (4)

Where Eimage is the energy that represents distance between the shape of the region defined by the interior of the curve C and the
shape G extracted from the image. Effective minimization of the energy in (4) amount to realizing a trade-off between segmented
object and the background.

Enhanced Geometric Active contour model

The enhanced model (ENGAC) was formulated using Kernel Principal Component Analysis (KPCA) and Geometric active
contour (GAC) segmentation model. KPCA was used to get shape variability within the Geometric active contour segmentation
model. The training set of MRI and CT scan of tran-axial medical images were gotten from University College Hospital, Ibadan,
Oyo state. The medical images were registered using appropriate registration method to pre-process the medical images. Figure 1
gives the overview of the enhanced geometric active contour segmentation model.




                   Figure 1: Overview of enhanced Geometric Active Contour model

The enhancement process started for the conventional Geometric active contour segmentation model by building Kernel Principal
Component Analysis (KPCA) feature space for medical shapes to be extracted and segmented. The steps involves were;

                                                                        79
Computer Engineering and Intelligent Systems                                                                  www.iiste.org
      ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
      Vol 3, No.3, 2012


 (i)     Get a training set of registered medical images
 (ii)    Represent the contour of the registered medical images with high dimensional function using signed distance function, such that
         a point was represented by (Xi ,Xj) where i = 1……N and j = 1…..N
(iii)    Calculation of mean offset map represented by φ for the points using

                                            
                                                 1 N
                                                     X i , X j
                                                 N i 1
                                                                             
                                                                                                                             (5)
(iv)     The centered kernel was formed for the points this equation
                     K (i,j)=             X i    ,  X j       X i ,  X j   K X i X j                             (6)
                     X i   X i   
             where
(v)      The centered kernel was decomposed by getting
         (a) Diagonal matrix that represent Eigen values represented by (α1……αn)
         (b) Orthonormal matrix that contain Eigen vectors represented by (M i1……MiN)

         (c) Covariance matrix calculated using
                                                              C 
                                                                         1 N
                                                                                        
                                                                             Xi , X j  Xi , X j
                                                                         N i 1
                                                                                                      T



(vi)         The Eigenvector of Covariance matrix was computed for the decomposed kernel denoted by

                                                              
                                              N M ki
                                        Vk           Xi
                                             i 1 
                                                    k                                                                                        (7)

(vii)         Any subsequent projection of element X of the input space I on the KPCA space was denoted by
                                                                                                                                    P' X  and the co-

              ordinate of
                            P'  X  on kth component Vk , was calculated by
                                               1 N
                                       Bk        U ki K  X , X i 
                                               k i 1                                                                                       (8)

(viii)        The projection of
                                    X  onto the subspace          spanned by the first l eigenvectors was calculated by
                                                       l
                                          P '   Bk Vk  
                                                      k 1                                                                                   (9)

                 The square distance
                                        d between   X  and its projection on the KPCA space was denoted by
                                          2
                                          F
      (ix)

                                                  d F   X , P'  X     X   P'  X 
                                                    2                                                      2

                                                                                                                                             (10)

      To hybridize GAC and Kernel PCA, these are the steps;

  (i)        Perform nonlinear PCA on the constraint SDF using Polynomial kernel function.
                                        K o i, j   i ,  j
                                                                          c

                                                                                                                                              (11)
      From equation (11), when c = 2 and the square bracket is opened,
                                                  c                  2                    2
                                       i ,  j        i   j           i   j
                                                                                                                                              (12)
        where
                            c=       variance parameter.
                                                                                     80
Computer Engineering and Intelligent Systems                                                        www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.3, 2012


             2
 i   j        = square distance between two SDFs Φi and Φj
o     = nonlinear mapping corresponding to Polynomial kernel

(2)     Minimize the energy functional equation below;
                                           Eenergy ( x)   xi , Z ' xi 
                                            F
                                                                                                                                (13)
where
 F
Eenergy
          denotes the facts that the shape knowledge is expressed as a distance in feature space.
x = test shape represented by Signed Distance Function (SDF).
 using
                                                                               dx
                                                     Eenergy   x Eenergy.
                                                          F              F

                                                t                               d                                             (14)
where

                                    g i  ko i ,  j 
                                       N
                   E   F                                      c
                         energy
                                      i 1
2.          Methods of Performance Evaluation

            The performance evaluations of the models were carried out using the following metrics by Lauren (2001). These metrics
     were used to evaluate the performance of enhanced Phasewire image segmentation on Livewire image segmentation.

     (a) Segmentation/Running time

              This is the time taken to compute the segmentation including the running time of the algorithm or the time spent to
     perform interactive segmentation. The research was implemented on Compaq Presario CQ60 notebook PC with AMD Athlon
     Dual-core QL-62 2.00 GHz processor.

     (b) Segmentation Accuracy

             The accuracy of the segmentation algorithms is generally based on Volume measurement. This is the volume of
     segmented feature in the image domain. It is calculated by multiplying number of voxel by the volume of each voxel.

     (c) Hausdsorff distance

                This evaluating metric measures the degree of mismatch between the initial constraint and the boundary of the
     extracted feature. It identifies the point in one set that is the farthest from the other and outputs the distance to the nearest point
     in the other set. For two sets of points A  a1 ,....., a m and B  b1 ,....., bn the hausdorff distance H(A,B) is defined as;


                                         H ( A, B) = max( h( A, B), h(B, A))
                 Where     h(A, B) = max min || a- b ||
                    aε A bεB
(d) Assessment by the Medical expert.

This is the evaluation based on Expert comment on the output of the system. The questionnaire and the system were made
available for the Medical expert on the ease of usage and segmentation quality of the output


3           Discussion of Results


                                                                                81
Computer Engineering and Intelligent Systems                                               www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.3, 2012


          The training sets used for the evaluation were obtained from University College Hospital Ibadan (UCH). Trans-axial
Human Brain medical images for both Computed tomography (CT) and Magnetic Resonance Imaging (MRI) were used. Figure
2(a) is a MRI scan of a human brain that serves as the test image for the registration, while Figure 2(b) is a CT scan of a human
brain that was used as the refrence image for the registration. Figure 2(c) is a registered image for both CT and MRI




                          (a)                       (b)                        (c)                         (d)


Figure 2: Medical image Registration (a) MRI scan of trans-axial Human Brain (b) CT scan of trans-axial Human Brain (c)
           Registered image (d) Region of interest for segmentation




        (a)                         (b)                   (c)                        (d)                         (e)




          (f)                          (g)                       (h)                              (i)
   Figure 3: Segmentation Results for segmenting Caudate nucleus from trans-axial Human Brain
        (GAC)     (a) contour at iteration = 100 (b) contour at iteration = 250 (c) Final segmentation contour at iteration =350
                        (d) magnitude of Eigenvalue (e) 3D view of segmented area.
                                                                82
Computer Engineering and Intelligent Systems                                                  www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.3, 2012



        (ENGAC) (f) contour at iteration = 100 (g) contour at iteration = 250 (h) Final segmentation contour at iteration =350
                     (i) 3D view of segmented area.

Figure 2(d) shows a region of interest or an initial constraint for the region to be segmented. The constraint then evolved according
to specified time steps (0.005s). During the process, embedding functions continues to change the topology little by little according
to the given time steps. The embedding function then reconstructed after each set of the time-steps using incremental solver in
GAC. The evolving energy stops when the initial constraint minimized the energy functional to reach the shape in kernel space in
ENGAC.

Figure 3(a) and (b) depict intermediate results of ENGAC Segmentation model at iteration equal 100 and 250 respectively. Figure
3(c) gives the final segmentation result at iteration equal 350. The caudate nucleus in trans-axial human brain was fully extracted
using the model. Figure 3(d) gives the magnitude of Eigen value and isovalues for ENGAC image segmentation that results in a
set of shape parameter vectors. From this set, the mean parameter vector and the covariance matrix as well as the Eigen vector and
Eigenvalues were interpreted as mode of shape variation. At different points in the feature space on the co-ordinates x and y, x
represents 0.002 and y represents 0.039. The nine (9) vectors found were represented as shown in the graph and the non-linearity
of these Eigen vectors was what kernel identifies by keeping the Eigen vectors higher than tolerance of 1e -18 and by normalizing
these Eigen vectors to get the exact shape. Figure 3(e) gives the 3D view of the final segmentation of the specified region of
interest. This could help the physician to locate the position of caudate nucleus in the entire image domain.

Figure 3(f) and (g) depict intermediate results of GAC Segmentation model at iteration equal 100 and 250 respectively. Figure
3(h) gives the final segmentation result at iteration equal 350. The caudate nucleus in trans-axial human brain was fully extracted
using the model. Figure 3(e) is the 3D view of the final segmentation of the specified region of interest

4        Performance Evaluation

The results of the performance evaluation were given below;


Segmentation time
        The segmentation time was evaluated for Geometric active contour (GAC) and enhanced Geometric active contour
(ENGAC). Table 1 gives the result for the evaluation. Figure 4 shows that as the number of iteration increases the segmentation
time increases for both segmentation methods but ENGAC model’segmentation time is 16.03% slower to segmentation time in
Geometric Active Contour (GAC) model.

Table 1: Segmentation time for GAC model and ENGAC model for caudate nucleus segmentation in trans-axial Human Brain


                                     No of       (ENGAC)              (GAC)
                                   iteration
                                               Segmentation      Segmentation
                                                 time(sec)         time(sec)
                                      30           58.65             45.42
                                      60            67.29             60.17
                                      90            79.31             78.49
                                      120           91.18             80.35
                                      150          100.49             84.38




                                                                 83
Computer Engineering and Intelligent Systems                                                    www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.3, 2012



                                           Segmentation time (ENGAC) Versus
                                               Segmentation time (GAC)
                                           120
                                           100
                              Time (sec)    80                                  Segmentation
                                            60                                  time(sec)
                                                                                ENGAC
                                            40
                                                                                Segmentation
                                            20                                  time(sec) GAC
                                             0
                                                 0           100       200
                                                     No of iteration

                            Figure 4: Segmentation Time (ENGAC Vs GAC)

Segmentation accuracy

The segmented area volume that was used to measure the accuracy was calculated by multiplying number of voxel by the volume
of each voxel. Table 2 gives the results obtained for GAC and ENGAC segmentation model. In Figure 5, as the number of
iteration increases the volume area segmented decreases for both segmentation methods but ENGAC has 18.21% segmented area
volume decrement compare to GAC.

Table 2: Segmented volume area for GAC model and ENGAC model for caudate nucleus segmentation in trans-axial Human
Brain.


                                                   No of           (ENGAC)        (GAC)
                                                 iteration
                                                               Segmented area    Segmented
                                                                volume(cm3)         area
                                                                                volume(cm3)
                                                     30              5.69           5.67
                                                     60              4.55          4.71
                                                     90              4.03          4.24
                                                     120             3.12          3.85
                                                     150             3.01          3.68




                                                                       84
Computer Engineering and Intelligent Systems                                                                         www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.3, 2012



                                                                   Segmented area volume (ENGAC)
                                                                    Versus Segmented area volume
                                                                               (GAC)
                                                                   6



                                       Segmented volume area(cm3
                                                                   5                                   Segmented
                                                                   4                                   area
                                                                   3                                   volume(cm3
                                                                                                       ) ENGAC
                                                                   2
                                                                   1                                   Segmented
                                                                   0                                   area
                                                                                                       volume(cm3
                                                                       0         100       200
                                                                                                       ) GAC
                                                                           No of iteration

                               Figure 5: Segmented area volume (ENGAC Vs GAC)

Haudsorff distance

Haudsorff distance before and after segmentation was also calculated for the segmented features. Table 3(a) and (b) show the result
of statistics used to quantify the nearness of the points in the initial constraint to the feature extracted. The hausdorff distance is
considerably reduced in ENGAC model by 2.75mm compared to GAC model.

                  Table 3: Haudorff distance Before and After segmentation (a) GAC (b) ENGAC

                                    (a)
                                                                                                   (GAC)
                                                                                       Before segmentation   After
                                                                                                             segmentation
                                     Average                               Hausdorff   27.74mm               23.83mm
                                     distance


                                    (b)

                                                                                                     (ENGAC)
                                                                                       Before segmentation         After
                                                                                                               segmentation
                                      Average Hausdorff                                      27.81mm             21.08mm
                                          distance


Medical Expert’s assessment

The system for ENGAC model was evaluated using likert method based on two formulated criteria which are System Ease of Use
(SEU) and System Segmentation Quality (SSQ). The scale was from 1 to 5, with 5 being the highest. The response means of the
methods were calculated Table 4 depicts the Medical expert’s assessment for the features segmented. Segmentation quality of the
ENGAC segmentation model was rated higher due to the smoothness of the contour generated at the boundary of the feature
extracted follow by GAC model.

                                                                                          85
Computer Engineering and Intelligent Systems                                                      www.iiste.org
 ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
 Vol 3, No.3, 2012




     Table 4: Medical Expert’s assessment



                                            Ease of use     Segmentation Quality
                     Study        Model      response          response mean
                                               mean




                     Tran axial
                     CT image      GAC          3.40                  3.53


                                  ENGAC         3.42                  4.25




 5.    Conclusion

 Conclusively, GAC segmentation model is faster than ENGAC segmentation model in terms of segmentation time but ENGAC
 model has optimal Segmentation accuracy and quality because of its higher decrement rate in the final segmented area volume and
 hausdorff distance compared to Geometric active contour.


6.       References

Chunming Li, Rui Huang, Zhaohua Ding, J. Chris Gatenby, Dimitris N. (2011). “A Level Set Method for Image Segmentation in the Presence
of Intensity In homogeneities With Application to MRI” IEEE transactions on image processing, vol. 20, no. 7.

Chen Y, Thiruenkadam S., Tagare H., Huang F., Wilson D., and Geiser E. (2001) “On the incorporation of shape priors into geometric active
contours,” in Proc. IEEE Workshop Variational and Level Set Methods, pp. 145–152.

Wang Y. and Staib L. (1998) “Boundary finding with correspondence using statistical shape models,” IEEE Conf. Computer Vision and
Pattern Recognition, pp. 338–345.

Dambreville S., Rathi Y., and Tannenbaum A. (2008) .“A Framework for Image Segmentation Using Shape Models and Kernel Space Shape
Priors,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 30(8):1385–99,

Charpiat G., Faugeras O., and Keriven R. (2005) “Approximations of Shape Metrics and Application to Shape Warping and Empirical Shape
Statistics,” Foundations of Computational Mathematics, 5(1):1–58.

Leventon M., Grimson W., and Faugeras O. (2000) “Statistical Shape Influence in Geodesic Active Contours”. In Proc. IEEE CVPR.

Tsai A., Yezzi Jr A., Wells W., Tempany C., Tucker D., Fan A., Grimson W., and Willsky A. (2003) “A Shape-Based Approach to the
Segmentation of Medical Imagery Using Level Sets”. IEEE Trans. on Medical Imaging, 22(2):137–154.


                                                                    86
Computer Engineering and Intelligent Systems                                                      www.iiste.org
     ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
     Vol 3, No.3, 2012


Osher S. and Sethian J.A. (1988) “Fronts propagating with curvature dependent speed: algorithms based on Hamilton-Jacobi formulation,”
Journal of Computer Physics, vol. (79): 12-49.

Niethammer T. and Allen Tannenbaum (2004) “ Dynamic Geodesic Snakes for Visual Tracking”. Proceedings of the 2004 IEEE Computer
Society Conference on Computer Vision and Pattern Recognition CVPR (4) : 660-667

Xiao Han, Chenyang Xu, and Jerry L. Prince, (2003) “A Topology Preserving Level Set Method for Geometric Deformable Models” IEEE
Transactions on Pattern analysis and Machine intelligence, 25(6):56-70.




                                                                    87

Weitere ähnliche Inhalte

Was ist angesagt?

Informative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical ImagesInformative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical ImagesMatt Moores
 
Iaetsd an enhanced circular detection technique rpsw using circular hough t...
Iaetsd an enhanced circular detection technique   rpsw using circular hough t...Iaetsd an enhanced circular detection technique   rpsw using circular hough t...
Iaetsd an enhanced circular detection technique rpsw using circular hough t...Iaetsd Iaetsd
 
D I G I T A L C O N T R O L S Y S T E M S J N T U M O D E L P A P E R{Www
D I G I T A L  C O N T R O L  S Y S T E M S  J N T U  M O D E L  P A P E R{WwwD I G I T A L  C O N T R O L  S Y S T E M S  J N T U  M O D E L  P A P E R{Www
D I G I T A L C O N T R O L S Y S T E M S J N T U M O D E L P A P E R{Wwwguest3f9c6b
 
CG OpenGL 3D object representations-course 8
CG OpenGL 3D object representations-course 8CG OpenGL 3D object representations-course 8
CG OpenGL 3D object representations-course 8fungfung Chen
 
Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)
Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)
Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)Joo-Haeng Lee
 
Robust Image Denoising in RKHS via Orthogonal Matching Pursuit
Robust Image Denoising in RKHS via Orthogonal Matching PursuitRobust Image Denoising in RKHS via Orthogonal Matching Pursuit
Robust Image Denoising in RKHS via Orthogonal Matching PursuitPantelis Bouboulis
 
Kccsi 2012 a real-time robust object tracking-v2
Kccsi 2012   a real-time robust object tracking-v2Kccsi 2012   a real-time robust object tracking-v2
Kccsi 2012 a real-time robust object tracking-v2Prarinya Siritanawan
 
Iaetsd a modified image fusion approach using guided filter
Iaetsd a modified image fusion approach using guided filterIaetsd a modified image fusion approach using guided filter
Iaetsd a modified image fusion approach using guided filterIaetsd Iaetsd
 
01 graphical models
01 graphical models01 graphical models
01 graphical modelszukun
 
Aj2418721874
Aj2418721874Aj2418721874
Aj2418721874IJMER
 
IRJET- An Image Cryptography using Henon Map and Arnold Cat Map
IRJET-  	  An Image Cryptography using Henon Map and Arnold Cat MapIRJET-  	  An Image Cryptography using Henon Map and Arnold Cat Map
IRJET- An Image Cryptography using Henon Map and Arnold Cat MapIRJET Journal
 
cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...
cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...
cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...zukun
 
Au2419291933
Au2419291933Au2419291933
Au2419291933IJMER
 
Gait using moment with gray and
Gait using moment with gray andGait using moment with gray and
Gait using moment with gray andijsc
 
Shearlet Frames and Optimally Sparse Approximations
Shearlet Frames and Optimally Sparse ApproximationsShearlet Frames and Optimally Sparse Approximations
Shearlet Frames and Optimally Sparse ApproximationsJakob Lemvig
 

Was ist angesagt? (17)

Informative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical ImagesInformative Priors for Segmentation of Medical Images
Informative Priors for Segmentation of Medical Images
 
Iaetsd an enhanced circular detection technique rpsw using circular hough t...
Iaetsd an enhanced circular detection technique   rpsw using circular hough t...Iaetsd an enhanced circular detection technique   rpsw using circular hough t...
Iaetsd an enhanced circular detection technique rpsw using circular hough t...
 
Image processing
Image processingImage processing
Image processing
 
D I G I T A L C O N T R O L S Y S T E M S J N T U M O D E L P A P E R{Www
D I G I T A L  C O N T R O L  S Y S T E M S  J N T U  M O D E L  P A P E R{WwwD I G I T A L  C O N T R O L  S Y S T E M S  J N T U  M O D E L  P A P E R{Www
D I G I T A L C O N T R O L S Y S T E M S J N T U M O D E L P A P E R{Www
 
CG OpenGL 3D object representations-course 8
CG OpenGL 3D object representations-course 8CG OpenGL 3D object representations-course 8
CG OpenGL 3D object representations-course 8
 
Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)
Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)
Note on Coupled Line Cameras for Rectangle Reconstruction (ACDDE 2012)
 
BMC 2012
BMC 2012BMC 2012
BMC 2012
 
Robust Image Denoising in RKHS via Orthogonal Matching Pursuit
Robust Image Denoising in RKHS via Orthogonal Matching PursuitRobust Image Denoising in RKHS via Orthogonal Matching Pursuit
Robust Image Denoising in RKHS via Orthogonal Matching Pursuit
 
Kccsi 2012 a real-time robust object tracking-v2
Kccsi 2012   a real-time robust object tracking-v2Kccsi 2012   a real-time robust object tracking-v2
Kccsi 2012 a real-time robust object tracking-v2
 
Iaetsd a modified image fusion approach using guided filter
Iaetsd a modified image fusion approach using guided filterIaetsd a modified image fusion approach using guided filter
Iaetsd a modified image fusion approach using guided filter
 
01 graphical models
01 graphical models01 graphical models
01 graphical models
 
Aj2418721874
Aj2418721874Aj2418721874
Aj2418721874
 
IRJET- An Image Cryptography using Henon Map and Arnold Cat Map
IRJET-  	  An Image Cryptography using Henon Map and Arnold Cat MapIRJET-  	  An Image Cryptography using Henon Map and Arnold Cat Map
IRJET- An Image Cryptography using Henon Map and Arnold Cat Map
 
cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...
cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...
cvpr2009 tutorial: kernel methods in computer vision: part II: Statistics and...
 
Au2419291933
Au2419291933Au2419291933
Au2419291933
 
Gait using moment with gray and
Gait using moment with gray andGait using moment with gray and
Gait using moment with gray and
 
Shearlet Frames and Optimally Sparse Approximations
Shearlet Frames and Optimally Sparse ApproximationsShearlet Frames and Optimally Sparse Approximations
Shearlet Frames and Optimally Sparse Approximations
 

Ähnlich wie Performance evaluation of geometric active contour (gac) and enhanced geometric active contour segmentation model (engac) for medical image segmentation

Image Texture Analysis
Image Texture AnalysisImage Texture Analysis
Image Texture Analysislalitxp
 
A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...
A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...
A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...IJERA Editor
 
New Mathematical Tools for the Financial Sector
New Mathematical Tools for the Financial SectorNew Mathematical Tools for the Financial Sector
New Mathematical Tools for the Financial SectorSSA KPI
 
AU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSE
AU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSEAU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSE
AU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSEThiyagarajan G
 
Image Splicing Detection involving Moment-based Feature Extraction and Classi...
Image Splicing Detection involving Moment-based Feature Extraction and Classi...Image Splicing Detection involving Moment-based Feature Extraction and Classi...
Image Splicing Detection involving Moment-based Feature Extraction and Classi...IDES Editor
 
Complex and Social Network Analysis in Python
Complex and Social Network Analysis in PythonComplex and Social Network Analysis in Python
Complex and Social Network Analysis in Pythonrik0
 
Social Network Analysis
Social Network AnalysisSocial Network Analysis
Social Network Analysisrik0
 
TAO Fayan_X-Ray and MIP volume rendering
TAO Fayan_X-Ray and MIP volume renderingTAO Fayan_X-Ray and MIP volume rendering
TAO Fayan_X-Ray and MIP volume renderingFayan TAO
 
Vector-Based Back Propagation Algorithm of.pdf
Vector-Based Back Propagation Algorithm of.pdfVector-Based Back Propagation Algorithm of.pdf
Vector-Based Back Propagation Algorithm of.pdfNesrine Wagaa
 
Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}
Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}
Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}guest3f9c6b
 
Munich07 Foils
Munich07 FoilsMunich07 Foils
Munich07 FoilsAntonini
 
Optimal nonlocal means algorithm for denoising ultrasound image
Optimal nonlocal means algorithm for denoising ultrasound imageOptimal nonlocal means algorithm for denoising ultrasound image
Optimal nonlocal means algorithm for denoising ultrasound imageAlexander Decker
 
11.optimal nonlocal means algorithm for denoising ultrasound image
11.optimal nonlocal means algorithm for denoising ultrasound image11.optimal nonlocal means algorithm for denoising ultrasound image
11.optimal nonlocal means algorithm for denoising ultrasound imageAlexander Decker
 
An Efficient Convex Hull Algorithm for a Planer Set of Points
An Efficient Convex Hull Algorithm for a Planer Set of PointsAn Efficient Convex Hull Algorithm for a Planer Set of Points
An Efficient Convex Hull Algorithm for a Planer Set of PointsKasun Ranga Wijeweera
 
Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)IJERD Editor
 
Chapter-05c-Image-Restoration-(Reconstruction-from-Projections).ppt
Chapter-05c-Image-Restoration-(Reconstruction-from-Projections).pptChapter-05c-Image-Restoration-(Reconstruction-from-Projections).ppt
Chapter-05c-Image-Restoration-(Reconstruction-from-Projections).pptVSUDHEER4
 
Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...
Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...
Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...준식 최
 
MATHEON Center Days: Index determination and structural analysis using Algori...
MATHEON Center Days: Index determination and structural analysis using Algori...MATHEON Center Days: Index determination and structural analysis using Algori...
MATHEON Center Days: Index determination and structural analysis using Algori...Dagmar Monett
 

Ähnlich wie Performance evaluation of geometric active contour (gac) and enhanced geometric active contour segmentation model (engac) for medical image segmentation (20)

Image Texture Analysis
Image Texture AnalysisImage Texture Analysis
Image Texture Analysis
 
A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...
A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...
A Hybrid Technique for the Automated Segmentation of Corpus Callosum in Midsa...
 
New Mathematical Tools for the Financial Sector
New Mathematical Tools for the Financial SectorNew Mathematical Tools for the Financial Sector
New Mathematical Tools for the Financial Sector
 
AU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSE
AU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSEAU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSE
AU QP Answer key NOv/Dec 2015 Computer Graphics 5 sem CSE
 
Image Splicing Detection involving Moment-based Feature Extraction and Classi...
Image Splicing Detection involving Moment-based Feature Extraction and Classi...Image Splicing Detection involving Moment-based Feature Extraction and Classi...
Image Splicing Detection involving Moment-based Feature Extraction and Classi...
 
Complex and Social Network Analysis in Python
Complex and Social Network Analysis in PythonComplex and Social Network Analysis in Python
Complex and Social Network Analysis in Python
 
Social Network Analysis
Social Network AnalysisSocial Network Analysis
Social Network Analysis
 
TAO Fayan_X-Ray and MIP volume rendering
TAO Fayan_X-Ray and MIP volume renderingTAO Fayan_X-Ray and MIP volume rendering
TAO Fayan_X-Ray and MIP volume rendering
 
Vector-Based Back Propagation Algorithm of.pdf
Vector-Based Back Propagation Algorithm of.pdfVector-Based Back Propagation Algorithm of.pdf
Vector-Based Back Propagation Algorithm of.pdf
 
Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}
Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}
Computer Graphics Jntu Model Paper{Www.Studentyogi.Com}
 
Munich07 Foils
Munich07 FoilsMunich07 Foils
Munich07 Foils
 
Optimal nonlocal means algorithm for denoising ultrasound image
Optimal nonlocal means algorithm for denoising ultrasound imageOptimal nonlocal means algorithm for denoising ultrasound image
Optimal nonlocal means algorithm for denoising ultrasound image
 
11.optimal nonlocal means algorithm for denoising ultrasound image
11.optimal nonlocal means algorithm for denoising ultrasound image11.optimal nonlocal means algorithm for denoising ultrasound image
11.optimal nonlocal means algorithm for denoising ultrasound image
 
An Efficient Convex Hull Algorithm for a Planer Set of Points
An Efficient Convex Hull Algorithm for a Planer Set of PointsAn Efficient Convex Hull Algorithm for a Planer Set of Points
An Efficient Convex Hull Algorithm for a Planer Set of Points
 
Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)
 
Chapter-05c-Image-Restoration-(Reconstruction-from-Projections).ppt
Chapter-05c-Image-Restoration-(Reconstruction-from-Projections).pptChapter-05c-Image-Restoration-(Reconstruction-from-Projections).ppt
Chapter-05c-Image-Restoration-(Reconstruction-from-Projections).ppt
 
Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...
Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...
Paper Summary of Infogan-CR : Disentangling Generative Adversarial Networks w...
 
YSC 2013
YSC 2013YSC 2013
YSC 2013
 
MATHEON Center Days: Index determination and structural analysis using Algori...
MATHEON Center Days: Index determination and structural analysis using Algori...MATHEON Center Days: Index determination and structural analysis using Algori...
MATHEON Center Days: Index determination and structural analysis using Algori...
 
D143136
D143136D143136
D143136
 

Mehr von Alexander Decker

Abnormalities of hormones and inflammatory cytokines in women affected with p...
Abnormalities of hormones and inflammatory cytokines in women affected with p...Abnormalities of hormones and inflammatory cytokines in women affected with p...
Abnormalities of hormones and inflammatory cytokines in women affected with p...Alexander Decker
 
A validation of the adverse childhood experiences scale in
A validation of the adverse childhood experiences scale inA validation of the adverse childhood experiences scale in
A validation of the adverse childhood experiences scale inAlexander Decker
 
A usability evaluation framework for b2 c e commerce websites
A usability evaluation framework for b2 c e commerce websitesA usability evaluation framework for b2 c e commerce websites
A usability evaluation framework for b2 c e commerce websitesAlexander Decker
 
A universal model for managing the marketing executives in nigerian banks
A universal model for managing the marketing executives in nigerian banksA universal model for managing the marketing executives in nigerian banks
A universal model for managing the marketing executives in nigerian banksAlexander Decker
 
A unique common fixed point theorems in generalized d
A unique common fixed point theorems in generalized dA unique common fixed point theorems in generalized d
A unique common fixed point theorems in generalized dAlexander Decker
 
A trends of salmonella and antibiotic resistance
A trends of salmonella and antibiotic resistanceA trends of salmonella and antibiotic resistance
A trends of salmonella and antibiotic resistanceAlexander Decker
 
A transformational generative approach towards understanding al-istifham
A transformational  generative approach towards understanding al-istifhamA transformational  generative approach towards understanding al-istifham
A transformational generative approach towards understanding al-istifhamAlexander Decker
 
A time series analysis of the determinants of savings in namibia
A time series analysis of the determinants of savings in namibiaA time series analysis of the determinants of savings in namibia
A time series analysis of the determinants of savings in namibiaAlexander Decker
 
A therapy for physical and mental fitness of school children
A therapy for physical and mental fitness of school childrenA therapy for physical and mental fitness of school children
A therapy for physical and mental fitness of school childrenAlexander Decker
 
A theory of efficiency for managing the marketing executives in nigerian banks
A theory of efficiency for managing the marketing executives in nigerian banksA theory of efficiency for managing the marketing executives in nigerian banks
A theory of efficiency for managing the marketing executives in nigerian banksAlexander Decker
 
A systematic evaluation of link budget for
A systematic evaluation of link budget forA systematic evaluation of link budget for
A systematic evaluation of link budget forAlexander Decker
 
A synthetic review of contraceptive supplies in punjab
A synthetic review of contraceptive supplies in punjabA synthetic review of contraceptive supplies in punjab
A synthetic review of contraceptive supplies in punjabAlexander Decker
 
A synthesis of taylor’s and fayol’s management approaches for managing market...
A synthesis of taylor’s and fayol’s management approaches for managing market...A synthesis of taylor’s and fayol’s management approaches for managing market...
A synthesis of taylor’s and fayol’s management approaches for managing market...Alexander Decker
 
A survey paper on sequence pattern mining with incremental
A survey paper on sequence pattern mining with incrementalA survey paper on sequence pattern mining with incremental
A survey paper on sequence pattern mining with incrementalAlexander Decker
 
A survey on live virtual machine migrations and its techniques
A survey on live virtual machine migrations and its techniquesA survey on live virtual machine migrations and its techniques
A survey on live virtual machine migrations and its techniquesAlexander Decker
 
A survey on data mining and analysis in hadoop and mongo db
A survey on data mining and analysis in hadoop and mongo dbA survey on data mining and analysis in hadoop and mongo db
A survey on data mining and analysis in hadoop and mongo dbAlexander Decker
 
A survey on challenges to the media cloud
A survey on challenges to the media cloudA survey on challenges to the media cloud
A survey on challenges to the media cloudAlexander Decker
 
A survey of provenance leveraged
A survey of provenance leveragedA survey of provenance leveraged
A survey of provenance leveragedAlexander Decker
 
A survey of private equity investments in kenya
A survey of private equity investments in kenyaA survey of private equity investments in kenya
A survey of private equity investments in kenyaAlexander Decker
 
A study to measures the financial health of
A study to measures the financial health ofA study to measures the financial health of
A study to measures the financial health ofAlexander Decker
 

Mehr von Alexander Decker (20)

Abnormalities of hormones and inflammatory cytokines in women affected with p...
Abnormalities of hormones and inflammatory cytokines in women affected with p...Abnormalities of hormones and inflammatory cytokines in women affected with p...
Abnormalities of hormones and inflammatory cytokines in women affected with p...
 
A validation of the adverse childhood experiences scale in
A validation of the adverse childhood experiences scale inA validation of the adverse childhood experiences scale in
A validation of the adverse childhood experiences scale in
 
A usability evaluation framework for b2 c e commerce websites
A usability evaluation framework for b2 c e commerce websitesA usability evaluation framework for b2 c e commerce websites
A usability evaluation framework for b2 c e commerce websites
 
A universal model for managing the marketing executives in nigerian banks
A universal model for managing the marketing executives in nigerian banksA universal model for managing the marketing executives in nigerian banks
A universal model for managing the marketing executives in nigerian banks
 
A unique common fixed point theorems in generalized d
A unique common fixed point theorems in generalized dA unique common fixed point theorems in generalized d
A unique common fixed point theorems in generalized d
 
A trends of salmonella and antibiotic resistance
A trends of salmonella and antibiotic resistanceA trends of salmonella and antibiotic resistance
A trends of salmonella and antibiotic resistance
 
A transformational generative approach towards understanding al-istifham
A transformational  generative approach towards understanding al-istifhamA transformational  generative approach towards understanding al-istifham
A transformational generative approach towards understanding al-istifham
 
A time series analysis of the determinants of savings in namibia
A time series analysis of the determinants of savings in namibiaA time series analysis of the determinants of savings in namibia
A time series analysis of the determinants of savings in namibia
 
A therapy for physical and mental fitness of school children
A therapy for physical and mental fitness of school childrenA therapy for physical and mental fitness of school children
A therapy for physical and mental fitness of school children
 
A theory of efficiency for managing the marketing executives in nigerian banks
A theory of efficiency for managing the marketing executives in nigerian banksA theory of efficiency for managing the marketing executives in nigerian banks
A theory of efficiency for managing the marketing executives in nigerian banks
 
A systematic evaluation of link budget for
A systematic evaluation of link budget forA systematic evaluation of link budget for
A systematic evaluation of link budget for
 
A synthetic review of contraceptive supplies in punjab
A synthetic review of contraceptive supplies in punjabA synthetic review of contraceptive supplies in punjab
A synthetic review of contraceptive supplies in punjab
 
A synthesis of taylor’s and fayol’s management approaches for managing market...
A synthesis of taylor’s and fayol’s management approaches for managing market...A synthesis of taylor’s and fayol’s management approaches for managing market...
A synthesis of taylor’s and fayol’s management approaches for managing market...
 
A survey paper on sequence pattern mining with incremental
A survey paper on sequence pattern mining with incrementalA survey paper on sequence pattern mining with incremental
A survey paper on sequence pattern mining with incremental
 
A survey on live virtual machine migrations and its techniques
A survey on live virtual machine migrations and its techniquesA survey on live virtual machine migrations and its techniques
A survey on live virtual machine migrations and its techniques
 
A survey on data mining and analysis in hadoop and mongo db
A survey on data mining and analysis in hadoop and mongo dbA survey on data mining and analysis in hadoop and mongo db
A survey on data mining and analysis in hadoop and mongo db
 
A survey on challenges to the media cloud
A survey on challenges to the media cloudA survey on challenges to the media cloud
A survey on challenges to the media cloud
 
A survey of provenance leveraged
A survey of provenance leveragedA survey of provenance leveraged
A survey of provenance leveraged
 
A survey of private equity investments in kenya
A survey of private equity investments in kenyaA survey of private equity investments in kenya
A survey of private equity investments in kenya
 
A study to measures the financial health of
A study to measures the financial health ofA study to measures the financial health of
A study to measures the financial health of
 

Kürzlich hochgeladen

KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCostKubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCostMatt Ray
 
Introduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptxIntroduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptxMatsuo Lab
 
Computer 10: Lesson 10 - Online Crimes and Hazards
Computer 10: Lesson 10 - Online Crimes and HazardsComputer 10: Lesson 10 - Online Crimes and Hazards
Computer 10: Lesson 10 - Online Crimes and HazardsSeth Reyes
 
Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.YounusS2
 
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...DianaGray10
 
ADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDE
ADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDEADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDE
ADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDELiveplex
 
Anypoint Code Builder , Google Pub sub connector and MuleSoft RPA
Anypoint Code Builder , Google Pub sub connector and MuleSoft RPAAnypoint Code Builder , Google Pub sub connector and MuleSoft RPA
Anypoint Code Builder , Google Pub sub connector and MuleSoft RPAshyamraj55
 
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019IES VE
 
Machine Learning Model Validation (Aijun Zhang 2024).pdf
Machine Learning Model Validation (Aijun Zhang 2024).pdfMachine Learning Model Validation (Aijun Zhang 2024).pdf
Machine Learning Model Validation (Aijun Zhang 2024).pdfAijun Zhang
 
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve DecarbonizationUsing IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve DecarbonizationIES VE
 
AI Fame Rush Review – Virtual Influencer Creation In Just Minutes
AI Fame Rush Review – Virtual Influencer Creation In Just MinutesAI Fame Rush Review – Virtual Influencer Creation In Just Minutes
AI Fame Rush Review – Virtual Influencer Creation In Just MinutesMd Hossain Ali
 
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...Will Schroeder
 
UiPath Solutions Management Preview - Northern CA Chapter - March 22.pdf
UiPath Solutions Management Preview - Northern CA Chapter - March 22.pdfUiPath Solutions Management Preview - Northern CA Chapter - March 22.pdf
UiPath Solutions Management Preview - Northern CA Chapter - March 22.pdfDianaGray10
 
Bird eye's view on Camunda open source ecosystem
Bird eye's view on Camunda open source ecosystemBird eye's view on Camunda open source ecosystem
Bird eye's view on Camunda open source ecosystemAsko Soukka
 
UiPath Studio Web workshop series - Day 7
UiPath Studio Web workshop series - Day 7UiPath Studio Web workshop series - Day 7
UiPath Studio Web workshop series - Day 7DianaGray10
 
Igniting Next Level Productivity with AI-Infused Data Integration Workflows
Igniting Next Level Productivity with AI-Infused Data Integration WorkflowsIgniting Next Level Productivity with AI-Infused Data Integration Workflows
Igniting Next Level Productivity with AI-Infused Data Integration WorkflowsSafe Software
 
UiPath Studio Web workshop series - Day 6
UiPath Studio Web workshop series - Day 6UiPath Studio Web workshop series - Day 6
UiPath Studio Web workshop series - Day 6DianaGray10
 

Kürzlich hochgeladen (20)

KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCostKubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
 
Introduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptxIntroduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptx
 
Computer 10: Lesson 10 - Online Crimes and Hazards
Computer 10: Lesson 10 - Online Crimes and HazardsComputer 10: Lesson 10 - Online Crimes and Hazards
Computer 10: Lesson 10 - Online Crimes and Hazards
 
Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.
 
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
 
20150722 - AGV
20150722 - AGV20150722 - AGV
20150722 - AGV
 
ADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDE
ADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDEADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDE
ADOPTING WEB 3 FOR YOUR BUSINESS: A STEP-BY-STEP GUIDE
 
201610817 - edge part1
201610817 - edge part1201610817 - edge part1
201610817 - edge part1
 
Anypoint Code Builder , Google Pub sub connector and MuleSoft RPA
Anypoint Code Builder , Google Pub sub connector and MuleSoft RPAAnypoint Code Builder , Google Pub sub connector and MuleSoft RPA
Anypoint Code Builder , Google Pub sub connector and MuleSoft RPA
 
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
 
Machine Learning Model Validation (Aijun Zhang 2024).pdf
Machine Learning Model Validation (Aijun Zhang 2024).pdfMachine Learning Model Validation (Aijun Zhang 2024).pdf
Machine Learning Model Validation (Aijun Zhang 2024).pdf
 
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve DecarbonizationUsing IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
 
AI Fame Rush Review – Virtual Influencer Creation In Just Minutes
AI Fame Rush Review – Virtual Influencer Creation In Just MinutesAI Fame Rush Review – Virtual Influencer Creation In Just Minutes
AI Fame Rush Review – Virtual Influencer Creation In Just Minutes
 
20230104 - machine vision
20230104 - machine vision20230104 - machine vision
20230104 - machine vision
 
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
 
UiPath Solutions Management Preview - Northern CA Chapter - March 22.pdf
UiPath Solutions Management Preview - Northern CA Chapter - March 22.pdfUiPath Solutions Management Preview - Northern CA Chapter - March 22.pdf
UiPath Solutions Management Preview - Northern CA Chapter - March 22.pdf
 
Bird eye's view on Camunda open source ecosystem
Bird eye's view on Camunda open source ecosystemBird eye's view on Camunda open source ecosystem
Bird eye's view on Camunda open source ecosystem
 
UiPath Studio Web workshop series - Day 7
UiPath Studio Web workshop series - Day 7UiPath Studio Web workshop series - Day 7
UiPath Studio Web workshop series - Day 7
 
Igniting Next Level Productivity with AI-Infused Data Integration Workflows
Igniting Next Level Productivity with AI-Infused Data Integration WorkflowsIgniting Next Level Productivity with AI-Infused Data Integration Workflows
Igniting Next Level Productivity with AI-Infused Data Integration Workflows
 
UiPath Studio Web workshop series - Day 6
UiPath Studio Web workshop series - Day 6UiPath Studio Web workshop series - Day 6
UiPath Studio Web workshop series - Day 6
 

Performance evaluation of geometric active contour (gac) and enhanced geometric active contour segmentation model (engac) for medical image segmentation

  • 1. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 Performance Evaluation of Geometric Active Contour (GAC) and Enhanced Geometric Active Contour Segmentation Model (ENGAC) for Medical Image Segmentation Ajala Funmilola Alaba*, Emuoyibofarhe Justice O Department of Computer Science and Engineering, LAUTECH Ogbomoso, Oyo state, Nigeria. *E-mail of the corresponding author: funfaith2003@yahoo.co.uk. Abstract Segmentation is an aspect of computer vision that deals with partitioning of an image into homogeneneous region. Medical image segmentation is an indispensable tool for medical image diagnoses. Geometric active contour (GAC) segmentation is one of the outstanding model used in machine learning community to solve the problem of medical image segmentation. However, It has problem of deviation from the true outline of the target feature and it generates spurious edge caused by noise that normally stop the evolution of the surface to be extracted. In this paper, enhanced Geometric active contour was formulated by using Kernel Principal Component Analysis(KPCA) with the existing Geometric active contour segmentation model and performance evaluation of the formulated model was carried out. Keyword: Geometric active contour, Segmentation, Medical image, Kernel Principal Component Analysis. 1. Introduction Image segmentation is an aspect of computer vision that deals with the partitioning an image into regions having homogeneous meaning. It is used to separate an object from its background and extracts meaningful objects lying in images either by dividing images into contiguous semantic regions, or by extracting one or more specific objects in images such as medical structures. However, we have different types of image segmentation methods which are not suitable for medical images since medical images are diverse, complex and vary from natural images. Some of the methods used for medical image segmentation fall under deformable models. The use of deformable model was popularized by Kass et al. (1990), snake model was used to develop active contour model that minimizes energy functional under the influence of forces which are internal force, image force and external constraint force. Deformable models can be classified into three categories which are Free-form, Parametric and Geometric active contour model. Geometric active contour model This make use of level set method that represent contour or surfaces as Zero level set of a higher dimensional function usually called a level set function Chunming et al, (2011). Level set can handle topological changes in curve evolution which is not possible with the classical active contour model. Osher and Sethian (1998) offer a natural and numerical implementation of curve evolution equation using level set. Readers are refer to Neithammer and Tannenbum (2004) for some previous work on Geometric active contour model. Xiao et al., (2003) gave the numeric scheme used for GAC model, which is describe in the following equation. Let x= (x, y) denote a point in a curve C which is a member in image domain Ω and let the curve be represented by Lipchitz- continuous function using signed distance function A parameterized differentiable curve C is a differential map . C : a, b      x( p), y( p)  x( p) T (1) 78
  • 2. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 The rule that determine how each point C moves starting from initial curve C o is dC P, t   vC P, t , t  dt (2) Where V = vector field that refers to as speed, Ignoring C and V in equation 2 gives dC v dt (3) Equation (2) gives a partial differential equation that determines how curve C evolves and often called flow. Minimized the energy below   E image  , T  H  G (T ) 2      2    H x  G T , I x (4) Where Eimage is the energy that represents distance between the shape of the region defined by the interior of the curve C and the shape G extracted from the image. Effective minimization of the energy in (4) amount to realizing a trade-off between segmented object and the background. Enhanced Geometric Active contour model The enhanced model (ENGAC) was formulated using Kernel Principal Component Analysis (KPCA) and Geometric active contour (GAC) segmentation model. KPCA was used to get shape variability within the Geometric active contour segmentation model. The training set of MRI and CT scan of tran-axial medical images were gotten from University College Hospital, Ibadan, Oyo state. The medical images were registered using appropriate registration method to pre-process the medical images. Figure 1 gives the overview of the enhanced geometric active contour segmentation model. Figure 1: Overview of enhanced Geometric Active Contour model The enhancement process started for the conventional Geometric active contour segmentation model by building Kernel Principal Component Analysis (KPCA) feature space for medical shapes to be extracted and segmented. The steps involves were; 79
  • 3. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 (i) Get a training set of registered medical images (ii) Represent the contour of the registered medical images with high dimensional function using signed distance function, such that a point was represented by (Xi ,Xj) where i = 1……N and j = 1…..N (iii) Calculation of mean offset map represented by φ for the points using  1 N  X i , X j N i 1   (5) (iv) The centered kernel was formed for the points this equation K (i,j)=  X i    ,  X j       X i ,  X j   K X i X j  (6) X i   X i    where (v) The centered kernel was decomposed by getting (a) Diagonal matrix that represent Eigen values represented by (α1……αn) (b) Orthonormal matrix that contain Eigen vectors represented by (M i1……MiN) (c) Covariance matrix calculated using C  1 N    Xi , X j  Xi , X j N i 1 T (vi) The Eigenvector of Covariance matrix was computed for the decomposed kernel denoted by   N M ki Vk    Xi i 1  k (7) (vii) Any subsequent projection of element X of the input space I on the KPCA space was denoted by P' X  and the co- ordinate of P'  X  on kth component Vk , was calculated by 1 N Bk  U ki K  X , X i  k i 1 (8) (viii) The projection of   X  onto the subspace spanned by the first l eigenvectors was calculated by l P '   Bk Vk   k 1 (9) The square distance d between   X  and its projection on the KPCA space was denoted by 2 F (ix) d F   X , P'  X     X   P'  X  2 2 (10) To hybridize GAC and Kernel PCA, these are the steps; (i) Perform nonlinear PCA on the constraint SDF using Polynomial kernel function. K o i, j   i ,  j c (11) From equation (11), when c = 2 and the square bracket is opened, c 2 2 i ,  j  i   j   i   j (12) where c= variance parameter. 80
  • 4. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 2 i   j = square distance between two SDFs Φi and Φj o = nonlinear mapping corresponding to Polynomial kernel (2) Minimize the energy functional equation below; Eenergy ( x)   xi , Z ' xi  F (13) where F Eenergy denotes the facts that the shape knowledge is expressed as a distance in feature space. x = test shape represented by Signed Distance Function (SDF). using  dx   Eenergy   x Eenergy. F F t d (14) where   g i  ko i ,  j  N  E F c energy i 1 2. Methods of Performance Evaluation The performance evaluations of the models were carried out using the following metrics by Lauren (2001). These metrics were used to evaluate the performance of enhanced Phasewire image segmentation on Livewire image segmentation. (a) Segmentation/Running time This is the time taken to compute the segmentation including the running time of the algorithm or the time spent to perform interactive segmentation. The research was implemented on Compaq Presario CQ60 notebook PC with AMD Athlon Dual-core QL-62 2.00 GHz processor. (b) Segmentation Accuracy The accuracy of the segmentation algorithms is generally based on Volume measurement. This is the volume of segmented feature in the image domain. It is calculated by multiplying number of voxel by the volume of each voxel. (c) Hausdsorff distance This evaluating metric measures the degree of mismatch between the initial constraint and the boundary of the extracted feature. It identifies the point in one set that is the farthest from the other and outputs the distance to the nearest point in the other set. For two sets of points A  a1 ,....., a m and B  b1 ,....., bn the hausdorff distance H(A,B) is defined as; H ( A, B) = max( h( A, B), h(B, A)) Where h(A, B) = max min || a- b || aε A bεB (d) Assessment by the Medical expert. This is the evaluation based on Expert comment on the output of the system. The questionnaire and the system were made available for the Medical expert on the ease of usage and segmentation quality of the output 3 Discussion of Results 81
  • 5. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 The training sets used for the evaluation were obtained from University College Hospital Ibadan (UCH). Trans-axial Human Brain medical images for both Computed tomography (CT) and Magnetic Resonance Imaging (MRI) were used. Figure 2(a) is a MRI scan of a human brain that serves as the test image for the registration, while Figure 2(b) is a CT scan of a human brain that was used as the refrence image for the registration. Figure 2(c) is a registered image for both CT and MRI (a) (b) (c) (d) Figure 2: Medical image Registration (a) MRI scan of trans-axial Human Brain (b) CT scan of trans-axial Human Brain (c) Registered image (d) Region of interest for segmentation (a) (b) (c) (d) (e) (f) (g) (h) (i) Figure 3: Segmentation Results for segmenting Caudate nucleus from trans-axial Human Brain (GAC) (a) contour at iteration = 100 (b) contour at iteration = 250 (c) Final segmentation contour at iteration =350 (d) magnitude of Eigenvalue (e) 3D view of segmented area. 82
  • 6. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 (ENGAC) (f) contour at iteration = 100 (g) contour at iteration = 250 (h) Final segmentation contour at iteration =350 (i) 3D view of segmented area. Figure 2(d) shows a region of interest or an initial constraint for the region to be segmented. The constraint then evolved according to specified time steps (0.005s). During the process, embedding functions continues to change the topology little by little according to the given time steps. The embedding function then reconstructed after each set of the time-steps using incremental solver in GAC. The evolving energy stops when the initial constraint minimized the energy functional to reach the shape in kernel space in ENGAC. Figure 3(a) and (b) depict intermediate results of ENGAC Segmentation model at iteration equal 100 and 250 respectively. Figure 3(c) gives the final segmentation result at iteration equal 350. The caudate nucleus in trans-axial human brain was fully extracted using the model. Figure 3(d) gives the magnitude of Eigen value and isovalues for ENGAC image segmentation that results in a set of shape parameter vectors. From this set, the mean parameter vector and the covariance matrix as well as the Eigen vector and Eigenvalues were interpreted as mode of shape variation. At different points in the feature space on the co-ordinates x and y, x represents 0.002 and y represents 0.039. The nine (9) vectors found were represented as shown in the graph and the non-linearity of these Eigen vectors was what kernel identifies by keeping the Eigen vectors higher than tolerance of 1e -18 and by normalizing these Eigen vectors to get the exact shape. Figure 3(e) gives the 3D view of the final segmentation of the specified region of interest. This could help the physician to locate the position of caudate nucleus in the entire image domain. Figure 3(f) and (g) depict intermediate results of GAC Segmentation model at iteration equal 100 and 250 respectively. Figure 3(h) gives the final segmentation result at iteration equal 350. The caudate nucleus in trans-axial human brain was fully extracted using the model. Figure 3(e) is the 3D view of the final segmentation of the specified region of interest 4 Performance Evaluation The results of the performance evaluation were given below; Segmentation time The segmentation time was evaluated for Geometric active contour (GAC) and enhanced Geometric active contour (ENGAC). Table 1 gives the result for the evaluation. Figure 4 shows that as the number of iteration increases the segmentation time increases for both segmentation methods but ENGAC model’segmentation time is 16.03% slower to segmentation time in Geometric Active Contour (GAC) model. Table 1: Segmentation time for GAC model and ENGAC model for caudate nucleus segmentation in trans-axial Human Brain No of (ENGAC) (GAC) iteration Segmentation Segmentation time(sec) time(sec) 30 58.65 45.42 60 67.29 60.17 90 79.31 78.49 120 91.18 80.35 150 100.49 84.38 83
  • 7. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 Segmentation time (ENGAC) Versus Segmentation time (GAC) 120 100 Time (sec) 80 Segmentation 60 time(sec) ENGAC 40 Segmentation 20 time(sec) GAC 0 0 100 200 No of iteration Figure 4: Segmentation Time (ENGAC Vs GAC) Segmentation accuracy The segmented area volume that was used to measure the accuracy was calculated by multiplying number of voxel by the volume of each voxel. Table 2 gives the results obtained for GAC and ENGAC segmentation model. In Figure 5, as the number of iteration increases the volume area segmented decreases for both segmentation methods but ENGAC has 18.21% segmented area volume decrement compare to GAC. Table 2: Segmented volume area for GAC model and ENGAC model for caudate nucleus segmentation in trans-axial Human Brain. No of (ENGAC) (GAC) iteration Segmented area Segmented volume(cm3) area volume(cm3) 30 5.69 5.67 60 4.55 4.71 90 4.03 4.24 120 3.12 3.85 150 3.01 3.68 84
  • 8. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 Segmented area volume (ENGAC) Versus Segmented area volume (GAC) 6 Segmented volume area(cm3 5 Segmented 4 area 3 volume(cm3 ) ENGAC 2 1 Segmented 0 area volume(cm3 0 100 200 ) GAC No of iteration Figure 5: Segmented area volume (ENGAC Vs GAC) Haudsorff distance Haudsorff distance before and after segmentation was also calculated for the segmented features. Table 3(a) and (b) show the result of statistics used to quantify the nearness of the points in the initial constraint to the feature extracted. The hausdorff distance is considerably reduced in ENGAC model by 2.75mm compared to GAC model. Table 3: Haudorff distance Before and After segmentation (a) GAC (b) ENGAC (a) (GAC) Before segmentation After segmentation Average Hausdorff 27.74mm 23.83mm distance (b) (ENGAC) Before segmentation After segmentation Average Hausdorff 27.81mm 21.08mm distance Medical Expert’s assessment The system for ENGAC model was evaluated using likert method based on two formulated criteria which are System Ease of Use (SEU) and System Segmentation Quality (SSQ). The scale was from 1 to 5, with 5 being the highest. The response means of the methods were calculated Table 4 depicts the Medical expert’s assessment for the features segmented. Segmentation quality of the ENGAC segmentation model was rated higher due to the smoothness of the contour generated at the boundary of the feature extracted follow by GAC model. 85
  • 9. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 Table 4: Medical Expert’s assessment Ease of use Segmentation Quality Study Model response response mean mean Tran axial CT image GAC 3.40 3.53 ENGAC 3.42 4.25 5. Conclusion Conclusively, GAC segmentation model is faster than ENGAC segmentation model in terms of segmentation time but ENGAC model has optimal Segmentation accuracy and quality because of its higher decrement rate in the final segmented area volume and hausdorff distance compared to Geometric active contour. 6. References Chunming Li, Rui Huang, Zhaohua Ding, J. Chris Gatenby, Dimitris N. (2011). “A Level Set Method for Image Segmentation in the Presence of Intensity In homogeneities With Application to MRI” IEEE transactions on image processing, vol. 20, no. 7. Chen Y, Thiruenkadam S., Tagare H., Huang F., Wilson D., and Geiser E. (2001) “On the incorporation of shape priors into geometric active contours,” in Proc. IEEE Workshop Variational and Level Set Methods, pp. 145–152. Wang Y. and Staib L. (1998) “Boundary finding with correspondence using statistical shape models,” IEEE Conf. Computer Vision and Pattern Recognition, pp. 338–345. Dambreville S., Rathi Y., and Tannenbaum A. (2008) .“A Framework for Image Segmentation Using Shape Models and Kernel Space Shape Priors,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 30(8):1385–99, Charpiat G., Faugeras O., and Keriven R. (2005) “Approximations of Shape Metrics and Application to Shape Warping and Empirical Shape Statistics,” Foundations of Computational Mathematics, 5(1):1–58. Leventon M., Grimson W., and Faugeras O. (2000) “Statistical Shape Influence in Geodesic Active Contours”. In Proc. IEEE CVPR. Tsai A., Yezzi Jr A., Wells W., Tempany C., Tucker D., Fan A., Grimson W., and Willsky A. (2003) “A Shape-Based Approach to the Segmentation of Medical Imagery Using Level Sets”. IEEE Trans. on Medical Imaging, 22(2):137–154. 86
  • 10. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012 Osher S. and Sethian J.A. (1988) “Fronts propagating with curvature dependent speed: algorithms based on Hamilton-Jacobi formulation,” Journal of Computer Physics, vol. (79): 12-49. Niethammer T. and Allen Tannenbaum (2004) “ Dynamic Geodesic Snakes for Visual Tracking”. Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition CVPR (4) : 660-667 Xiao Han, Chenyang Xu, and Jerry L. Prince, (2003) “A Topology Preserving Level Set Method for Geometric Deformable Models” IEEE Transactions on Pattern analysis and Machine intelligence, 25(6):56-70. 87