Segmentation of Volume Images Using Trilinear Interpolation of 2D Slices
|
|
- Alannah Perry
- 5 years ago
- Views:
Transcription
1 Segmentation of Volume Images Using Trilinear Interpolation of 2D Slices Pouyan TaghipourBibalan,Mehdi Mohammad Amini Australian National University Abstract Accurate segmentation of biomedical images can contribute to improved diagnosis, surgical planning and prognosis. A review of biomedical image segmentation methods using deformable models is presented. A 2D deformable model is applied to axial slices from a 3D MRI volume image of the human brain to determine the extent of the Occipital lobes. Trilinear interpolation is used to obtain slices inbetween or intersecting the available 2D slices. Using this technique the computational load is reduced compared to a direct 3D segmentation technique but it is shown that the applicability of this technique is bound to setting constraints between the plains. The difficulties of biomedical image segmentation are discussed and suggestions for addressing some of the issues are offered. 1 Introduction Recent advances in medical imaging allows the anatomic structure of internal organs to be viewed and gross pathology of organs and diseases to be recognized without the need to open the body. In addition to structural information, dynamic and functional information on biochemical and patho-physiologic processes may be imaged, all with unprecedented precision. For example fmri intensity measures the amount of oxyhemoglobin in the blood that has been converted to deoxyhemoglobin as a result of oxygen consumption by nearby cells. In 3D medical imaging applications conversion of the raw MRI data into an image useful to a radiologist involves the following processing: back projection to form 2D slices, noise reduction and image registration to align multiple 2D slices. The MRI data used in this project has already had this processing applied. We investigate further processing to establish the extent of objects of interest within these volume images. This is known as the image segmentation problem. 1.1 Image Segmentation The aim of image segmentation is to partition the image to meaningful regions, usually by establishing the boundaries of objects. It is an early step in many image processing applications. Difficulties arise in the face of image noise and inhomogeneities in the sensors. The nature of the object being segmented may introduce further difficulties, such as when the boundary is indistinct or not captured by the same image features over its entirety. Fully automated approaches are not generally applicable in the face of these problems and interaction with a human operator is usually required. Some systems allow the operator to add additional forces (e.g. a spring force) during deformation, to guide the model towards the desired boundary. The large body of medical image processing litterature offers a wide range of techniques applicable in different applications, with no single 1
2 technique being generally applicable. In this work we concentrate on segmentation using deformable models. 1.2 Deformable Models The goal of methods based on deformable models is to move an initial curve (in the 2D case) or surface (in the 3D case) towards the desired boundary, thus providing a model for the boundary. The boundary may be provided by edge detection methods based on image intensity or any other set of features that detect the desired boundary (e.g. texture, colour or other multi-spectral information such as provided by multiple types of MRI experiment on the same subject). A series of points lying on the curve/surface may be used to represent the current model during deformation. These are called parametric models. Parametric models vary in what forces are applied to the points to make the curve/surface move towards the boundary. Alternatively the current model may be implicitly represented by a level set, e.g. all the points which lie on a contour where some function has a constant value. These are called geometric models. In each iteration the level set is moved closer to the boundary under the influence of a speed function, without the need for the contour to be explicitly evaluated until sufficient convergence has been achieved. Geometric models vary in the nature of the speed function used. Parametric and geometric deformable models share common fundamentals and [17] discusses their inter-relationship. Simple segmentation techniques such as edge detection and thresholding are not generally sufficient for segmenting soft tissues, due to the difficulties mentioned in the introduction. Deformable models, also known as snakes, active contours or surfaces and balloons, can successfully deal with noise and indistinct edges. Figure 1: Active Contour from [10] 2 Parametric Models An active contour is defined by an ordered collection of points in the image plane (see figure 1). V = {v 1,... v n }; v i = (x i, y i ); i = 1... n (1) The points in the contour iteratively approach the boundary of an object. In this section the approaching process is demonstrated through the solution of an energy minimization problem. For each point in the neighborhood of V i, an energy term E i is computed, E i = αe int (v i ) + βe ext (v i ) (2) where E int (v i ) is an internal energy function dependent on the shape of the contour and E ext (v i ) is an external energy function dependent on the image properties, such as the gradient near the point v i. α and β are constants providing relative weighting of the energy terms. The value at the center of each matrix corresponds to the contour energy at point v i. If the energy functions are chosen correctly, the contour V approaches and stops at the object boundary. In this section an energy minimizing formulation for deriving parametric deformable models is presented. The solution to this equation 2
3 satisfies a minimum principle. Next finite differences is utilized to derive a numerical implementation. 2.1 Mathematical Representation The objective here is to iteratively find the curve X(s) = [x(s), y(s)], s [0, 1] that minimizes the energy functional E. It has been shown that the curve that minimizes E must satisfy the following Euler-Lagrange equation [9] [5] s (α X s ) 2 s 2 (β 2 X ) P (X) = 0 (3) s2 This equation has the same form as a force balance equation F int (X) + F pot (X) = 0 and can be interpreted as F int being responsible for stretching and bending and F pot pulling the contour toward the desired object boundary. Introduction of a new variable t helps in solving this equation, so that we can deal with X(s) as a function of time. The equation is thus presented in the following form. γ X t = s (α X s ) 2 s 2 (β 2 X ) P (X) s2 (4) The coefficient γ is introduced to make the units on the left side consistent with the right side. By utilizing dynamic force formulation, it can be shown that equation 4 can be changed to equation?? regardless of the type of the external energy; it can be a combination of forces either being potential forces or non potential forces. γ X t = F int(x) + F ext (X) (5) In the next section some forms of external forces are presented. 2.2 External Forces Gaussian Potential Force The Gaussian potential force is of the form P (x, y) = ω e [G σ (x, y) I(x, y)] 2 (6) where ω e is a positive weighting parameter, G σ (x, y) is a two dimensional Gaussian function with standard deviation σ and is the 2D image convolution operator. A larger σ implies a larger attraction range at the cost of losing the ability to track small features Pressure Force - Balloons The Gaussian potential force has only a short range. Adding a pressure force allows the curve/surface to move over areas of low gradient and pass through weak edges [5]. The pressure force F p is given by F p (X) = ω p N(X) (7) where N(X) is the inward unit normal of the model at the point X and ω p is a weighting parameter. The value of ω p determines the strength of the pressure force. It is selected in such a way that the pressure force is slightly smaller than the Gaussian potential force at significant edges but large enough to pass through weak edges. The pressure force keeps inflating or deflating the model until stopped by the Gaussian potential force. A disadvantage is that pressure forces may cause the curve/surface to cross itself and form loops [15] Distance Potential Force The distance potential force is designed to move a point towards the nearest edge, extending the attraction range. More details on this type of force can be found in [4] Gradient Vector Flow The Gradient Vector Flow (GVF) force is a diffusion of the gradient vectors [19] [20]. It has a wide attraction range and it deforms well in the presence of boundary concavities, unlike distance potential forces. In this method, first an edge map f(x, y) is derived from the image I(x, y) such that it is larger near the image edges. f(x, y) = E ext (x, y) (8) 3
4 wheree ext (x, y) can be P (x, y), the gaussian potential forece, or other external forces mentioned in literature. The gradient vector flow is defined to be the vector field v(x, y) = (u(x, y), v(x, y)), such that it minimizes the energy functional ε = µ(u 2 x+u 2 y+vx+v 2 y)+ f 2 2 v f 2 dxdy (9) It can be observed that when the field f is small the energy is dominated by the partial derivatives of the vector field, yielding a smooth field and when the field is large the second term dominates the integrand and is minimized by setting v = f. µ is a regularization parameter governing the tradeoff between the first term and the second term. µ is set according to the noise in the image. The more the noise in the image, the higher the value of µ. The GVF is found by solving the Euler equations µ 2 u (u f x )(f 2 x + f 2 y ) = 0 (10) µ 2 v (v f y )(f 2 x + f 2 y ) = 0 (11) where 2 is the Laplacian operator. The above equations are solved by treating u and v as functions of time and solving u t (x, y, t) = µ 2 u(x, y, t) (u(x, y, t) f x (x, y)).(f x (x, y) 2 + f y (x, y) 2 ) (12) approximating the derivatives in 5 with finite differences and converting to the vector notation [21] X n X n 1 = AX n + F ext (X n 1 ) (14) τ where τ = t γ and Xn, X n 1 and F ext (X n 1 ) are m 2 matrices, and A is an m m pentadiagonal banded matrix with m being the number of sample points. This is the formulation from which the MAT- LAB simulations are derived. By solving for X n the complete set of points which present the deformed curve X(s) in each iteration can be found. By choosing different types of F ext it is possible to overcome the issues concerned with the attraction range and boundary concavities. 2.4 Limitations of Parametric Models Although parametric models are successful in many applications, they have some serious limitations: the model must be explicitly evaluated at each iteration; as the model deforms points must be regularly redistributed over the curve/surface and more points are required as the model complexity increases. For complex 3D surfaces the computation can become prohibitive; v t (x, y, t) = µ 2 v(x, y, t) (v(x, y, t) f y (x, y)).(f x (x, y) 2 + f y (x, y) 2 ) (13) These equations are also known as generalized diffusion equations. The steady state solution as t of the above equations is the solution to the Euler equation. 2.3 Numerical Implementation Several numerical techniques have been reported in the literature, such as finite differencs [1], dynamic programming [1], greedy algorithm [18] and finite elements [18] [3] [16]. By as the model approaches the boundary the topology may change, e.g. a closed curve may pinch and split into two separate closed curves or merge back together. A parametric model can only handle such situations with computationally expensive reparameterization. 3 Geometric Models (Free Form) Geometric deformable models [11] address the limitations of parametric models described 4
5 4 Simulation Results 4.1 Tri-linear Interpolation within Volume Images Figure 2: Cuboid Plane Intersection from [8] above, in particular they naturally handle changes in topology and do not require redistribution of points along the contour. In this method the evolving model is represented implicitly as a level set, that is the contour for which some function has a constant value. The model is only explicitly evaluated after convergence to the boundary. We are not goint to go into details of geometric models since our simulations were based on parametric ones. An excellent introduction is available in [13]. InterpDemo computes the image on a plane of any orientation through a volume image. The world frame is taken to be that of the volume image. The plane is defined by z = 0 in the O frame, which is defined by a homogeneous transformation T of the world frame. The method described in [8] is used to compute the (3 to 6) vertices of the polygon of intersection of the plane with the volume image. The (x, y, z) coordinates of the vertices are computed in the world frame then transformed into (x, y, z ) coordinates in the O frame. The range of the x, y values of the vertices is computed in the O frame. A mesh of coordinates is constructed in the O frame over this range, with z = 0 fixed. This mesh is transformed into the world frame and the Matlab function interp3 is used to evaluate the intensity at each of these points by tri-linear interpolation. Figure 3 shows the InterpDemo controls used to generate the image shown in figure 4. Red circles in the image indicate the vertices of the polygon of intersection of the plane with the volume image. The parameters are: (θ x, θ x, θ x ) = (0, , 0) (x 0, y 0, z 0 ) = ( , , ) 3.1 Limitations of Geometric Models Because the topology is fixed in parametric models, they may be more appropriate in applications where the topology is known [21] [12] [2]. In cases where the volume image is too large to fit in main memory [7], out-of-core techniques may be applied [6] [14]. Figure 3: InterpDemo Settings 5
6 4.2.3 External Force Figure 4: InterpDemo Interpolated Image 4.2 Segmentation of the Occipital lobes Overview Three axial slices through the Occipital lobes are segmented using the 2D parametric snake segmentation implementation provided in [22]. The initial contour for each slice is manually entered by the user. Interpolated slices midway between these slices are computed and also segmented. The resulting five curves delimiting part of the surface of the Occipital lobes are displayed in a 3D Matlab plot, allowing the user to explore the surface Image Preprocessing A parametric snake converges to a boundary under the action of an external force. In this case we applied a Gaussian blur with a standard deviation of 3 pixels to remove high frequencies and noise and used Canny edge detection with high and low thresholds set to 0.12 and 0.05 to define the boundaries. We experimented with a range of blur sizes and edge detection methods and threshold values before settling on these parameters. Clearly these steps are crucial to the result as they define the goal for subsequent processing. The resulting edge map was again blurred with a Gaussian (size 3 as before) and the Gradient Vector Flow (GVF) [20] external force field was computed from the result. Figure 5 shows an image slice, the corresponding edge map with the initial user defined contour and a small portion of the GVF field enlarged. We have also obtained similar results using the traditional gradient external force field. Computation of the field was much faster than with GVF, but it provided slower convergence to the boundary. From exploration of the volume image, it appears that the interface surface between lobes in the brain are often not clearly delimited by discontinuities in image intensity. Perhaps features other than intensity, such as texture, might detect these surfaces better. This is an area we have not further explored. If the snake exhibits a tendency to get caught on short discontinuous edges, the addition of an internal pressure force can overcome the external force towards these short edges, causing expansion to continue towards a more significant edge. We used a small pressure force, but obtained a similar results without it Segmentation Results Figure 6 shows the result of 2D snake segmentation of the five axial slices through the Occipital lobes. The segmentation is performed on the slices obtained from the interpolation of the available slices. Stacking up these segmented plains as in Figure 6, we obtain a 3D segmented representation. It was found that unless the different slices are perfectly alligned this technique does not give satisfactory results. 5 Conclusions Image preprocessing is a major part of medical image segmentation, with a large body of literature devoted to it and a range of methods required depending on the application. Image intensity may not be the best choice of feature 6
7 for detecting the interface surface between different lobes of the brain. Applying segmentation to each plane independently, without constraints between the planes, results in curves that are not smooth across the planes. Unsurprisingly, 3D techniques are required to generate good 3D surfaces. References Figure 5: Image, Edge map and GVF [1] A. A. Amini, T. E. Weymouth, and R. C. Jain. Using dynamic programming for solving variational problems in vision. IEEE Transactions. Pattern Analysis and Machine Inteligence, 12(9): , [2] D. L. Pham C. Xu and J. L. Prince. Finding the brain cortex using fuzzy segmentation, isosurfaces, and deformable surface models. Proc. Information Processing in Medical Imaging (IPMI97), page , [3] V. Caselles, F. Catt, T. Coll, and F. Dibos. A geometric model for active contours. Numerische Mathematik. [4] L. D. Cohen and I. Cohen. Finiteelement methods for active contour models and balloons for 2-d and 3-d images. IEEE Trans. Pattern. Analysis Machine Inteligence, 15(11): , [5] Cohen L D. On active contour models and balloons. CVGIP:Image Understanding, 53(2): , [6] R. Farias and C Silva. Out-of-core rendering of large unstructured grids. IEEE Comput. Graph. Appl, 21(4):4250, Figure 6: Occipital lobe segmentation [7] Leandro S. Marturelli Gilson A. Giraldi, Paulo S. Rodrigues and Rodrigo L. S. Silva. Improving the initialization, convergence, and memory utilization for deformable models. In Handbook of Biomedical Image Analysis, Segmentation Models, pages , New York, NY, USA, Kluwer Academic / Plenum. 7
8 [8] Gernot Hoffmann. Cube plane intersection, hoffmann/cubeplane pdf. [9] Terzopoulos D. Kass M, Witkin A. Snakes: Active contour models. International Journal of Computer Vision, 1(4): , [10] Blair Mackiewich. Intracranial boundary detection and radio frequency correction in magnetic resonance images. [20] C. Xu and J. L. Prince. Snakes, shapes, and gradient vector flow. IEEE Trans. Imag. Proc., 7(3):359369, stella/papers/blairthesis/main/node29.html. [21] Chenyang Xu, Dzung L. Pham, and [11] R. Malladi, J. Sethian, and B. Vemuri. Jerry L. Prince. Image segmentation using Shape modelling with front propagation. deformable models. pages The IEEE Trans. Pattern Anal. Machine Intell, 17: , John Hopkins University, [12] G. Sapiro P. C. Teo and B. A. Wandell. Creating connected representations of cortical gray matter for functional mri visualization. IEEE Trans. Med. Imag, 16:852863, [13] J. A. Sethian. Level set methods: An act of violence. American Scientist, sethian/2006/publications/popular/popular.html. [14] Jimenez W. Giraldi G. A. Silva R. Strauss, E. and A. F. Oliveira. A semi-automatic surface reconstruction framework based on t-surfaces and isosurface extraction methods. International Symposium on Computer Graphics, Image Processing and Vision (SIBGRAPI2002), [15] H. Tek and B. B. Kimia. Volumetric segmentation of medical images by threedimensional bubbles. Comp. Vis. Imag. Under., 65:246258, [16] D. Terzopoulos and D. Metaxas. Dynamic 3d models with local and global deformations:deformable superquadrics. IEEE Trans. Patt. Anal. Mach. Intell., 13: , [17] R. Kimmel V. Caselles and G. Sapiro. Geodesic active contours. Intl J. Comp. Vis, 22:6179, [18] D. J. Williams and M. Shah. A fast algorithm for active contours and curvature estimation. CVGIP:Image Understanding, 55(1):14 26, [19] C. Xu and J. L. Prince. Generalized gradient vector flow external forces for active contours. Signal Processing, 71(2):131139, [22] Chenyang Xu and Jerry L. Prince. Matlab implementation of gradient vector flow with a 2d parametric model, 8
Gradient Vector Flow: A New External Force for Snakes
66 IEEE Proc. Conf. on Comp. Vis. Patt. Recog. (CVPR'97) Gradient Vector Flow: A New External Force for Snakes Chenyang Xu and Jerry L. Prince Department of Electrical and Computer Engineering The Johns
More informationSnakes operating on Gradient Vector Flow
Snakes operating on Gradient Vector Flow Seminar: Image Segmentation SS 2007 Hui Sheng 1 Outline Introduction Snakes Gradient Vector Flow Implementation Conclusion 2 Introduction Snakes enable us to find
More informationImage Segmentation II Advanced Approaches
Image Segmentation II Advanced Approaches Jorge Jara W. 1,2 1 Department of Computer Science DCC, U. of Chile 2 SCIAN-Lab, BNI Outline 1. Segmentation I Digital image processing Segmentation basics 2.
More informationCHAPTER 3 Image Segmentation Using Deformable Models
CHAPTER Image Segmentation Using Deformable Models Chenyang Xu The Johns Hopkins University Dzung L. Pham National Institute of Aging Jerry L. Prince The Johns Hopkins University Contents.1 Introduction
More informationSnakes reparameterization for noisy images segmentation and targets tracking
Snakes reparameterization for noisy images segmentation and targets tracking Idrissi Sidi Yassine, Samir Belfkih. Lycée Tawfik Elhakim Zawiya de Noaceur, route de Marrakech, Casablanca, maroc. Laboratoire
More informationImplicit Active Contours Driven by Local Binary Fitting Energy
Implicit Active Contours Driven by Local Binary Fitting Energy Chunming Li 1, Chiu-Yen Kao 2, John C. Gore 1, and Zhaohua Ding 1 1 Institute of Imaging Science 2 Department of Mathematics Vanderbilt University
More informationAutomated Segmentation Using a Fast Implementation of the Chan-Vese Models
Automated Segmentation Using a Fast Implementation of the Chan-Vese Models Huan Xu, and Xiao-Feng Wang,,3 Intelligent Computation Lab, Hefei Institute of Intelligent Machines, Chinese Academy of Science,
More informationSNAKES [1], or active contours, are curves defined within
IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 7, NO. 3, MARCH 1998 359 Snakes, Shapes, and Gradient Vector Flow Chenyang Xu, Student Member, IEEE, and Jerry L. Prince, Senior Member, IEEE Abstract Snakes,
More informationActive contours in Brain tumor segmentation
Active contours in Brain tumor segmentation Ali Elyasi* 1, Mehdi Hosseini 2, Marzieh Esfanyari 2 1. Department of electronic Engineering, Young Researchers Club, Central Tehran Branch, Islamic Azad University,
More informationActive Geodesics: Region-based Active Contour Segmentation with a Global Edge-based Constraint
Active Geodesics: Region-based Active Contour Segmentation with a Global Edge-based Constraint Vikram Appia Anthony Yezzi Georgia Institute of Technology, Atlanta, GA, USA. Abstract We present an active
More informationChromosome Segmentation and Investigations using Generalized Gradient Vector Flow Active Contours
Published Quarterly Mangalore, South India ISSN 0972-5997 Volume 4, Issue 2; April-June 2005 Original Article Chromosome Segmentation and Investigations using Generalized Gradient Vector Flow Active Contours
More informationEdge Detection and Active Contours
Edge Detection and Active Contours Elsa Angelini, Florence Tupin Department TSI, Telecom ParisTech Name.surname@telecom-paristech.fr 2011 Outline Introduction Edge Detection Active Contours Introduction
More informationA REVIEW ON THE CURRENT SEGMENTATION ALGORITHMS FOR MEDICAL IMAGES
A REVIEW ON THE CURRENT SEGMENTATION ALGORITHMS FOR MEDICAL IMAGES Zhen Ma, João Manuel R. S. Tavares, R. M. Natal Jorge Faculty of Engineering, University of Porto, Porto, Portugal zhen.ma@fe.up.pt, tavares@fe.up.pt,
More informationDr. Ulas Bagci
Lecture 9: Deformable Models and Segmentation CAP-Computer Vision Lecture 9-Deformable Models and Segmentation Dr. Ulas Bagci bagci@ucf.edu Lecture 9: Deformable Models and Segmentation Motivation A limitation
More informationLevel Set Evolution without Reinitilization
Level Set Evolution without Reinitilization Outline Parametric active contour (snake) models. Concepts of Level set method and geometric active contours. A level set formulation without reinitialization.
More informationA MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING
Proceedings of the 1994 IEEE International Conference on Image Processing (ICIP-94), pp. 530-534. (Austin, Texas, 13-16 November 1994.) A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING
More informationExtract Object Boundaries in Noisy Images using Level Set. Literature Survey
Extract Object Boundaries in Noisy Images using Level Set by: Quming Zhou Literature Survey Submitted to Professor Brian Evans EE381K Multidimensional Digital Signal Processing March 15, 003 Abstract Finding
More informationGeometrical Modeling of the Heart
Geometrical Modeling of the Heart Olivier Rousseau University of Ottawa The Project Goal: Creation of a precise geometrical model of the heart Applications: Numerical calculations Dynamic of the blood
More informationAn Active Contour Model without Edges
An Active Contour Model without Edges Tony Chan and Luminita Vese Department of Mathematics, University of California, Los Angeles, 520 Portola Plaza, Los Angeles, CA 90095-1555 chan,lvese@math.ucla.edu
More informationBrain Structure Segmentation from MRI by Geometric Surface Flow
Brain Structure Segmentation from MRI by Geometric Surface Flow Greg Heckenberg Yongjian Xi Ye Duan Jing Hua University of Missouri at Columbia Wayne State University Abstract In this paper, we present
More informationA Geometric Contour Framework with Vector Field Support
A Geometric Contour Framework with Vector Field Support Zhenglong Li, Qingshan Liu, and Hanqing Lu National Laboratory of Pattern Recognition Automation of Institute, Chinese Academy of Sciences P.O. Box
More informationISSN: X International Journal of Advanced Research in Electronics and Communication Engineering (IJARECE) Volume 6, Issue 8, August 2017
ENTROPY BASED CONSTRAINT METHOD FOR IMAGE SEGMENTATION USING ACTIVE CONTOUR MODEL M.Nirmala Department of ECE JNTUA college of engineering, Anantapuramu Andhra Pradesh,India Abstract: Over the past existing
More informationObject Segmentation Using Graph Cuts Based Active Contours
Object Segmentation Using Graph Cuts Based Active Contours Ning Xu Beckman Institute & ECE Dept. University of Illinois Urbana, IL, USA ningxu@vision.ai.uiuc.edu Ravi Bansal Department of Psychiatry Columbia
More informationCerebral Artery Segmentation with Level Set Methods
H. Ho, P. Bier, G. Sands, P. Hunter, Cerebral Artery Segmentation with Level Set Methods, Proceedings of Image and Vision Computing New Zealand 2007, pp. 300 304, Hamilton, New Zealand, December 2007.
More informationTopologically Adaptable Snakes
Published in the Proc. of the Fifth Int. Conf. on Computer Vision (ICCV 95), Cambridge, MA, USA, June, 1995, 840 845.840 Topologically Adaptable Snakes Tim McInerney and Demetri Terzopoulos Department
More informationSegmentation Using Active Contour Model and Level Set Method Applied to Medical Images
Segmentation Using Active Contour Model and Level Set Method Applied to Medical Images Dr. K.Bikshalu R.Srikanth Assistant Professor, Dept. of ECE, KUCE&T, KU, Warangal, Telangana, India kalagaddaashu@gmail.com
More informationDeformable Contour Method: A Constrained Optimization Approach
Deformable Contour Method: A Constrained Optimization Approach Xun Wang, Lei He, Chia Y. Han, William G. Wee Electrical & Computer Engineering and Computer Science Department University of Cincinnati,
More informationDEFORMABLE contour models are commonly used in
640 IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 13, NO. 5, MAY 2004 RAGS: Region-Aided Geometric Snake Xianghua Xie and Majid Mirmehdi Abstract An enhanced, region-aided, geometric active contour that
More informationLevelset and B-spline deformable model techniques for image segmentation: a pragmatic comparative study.
Levelset and B-spline deformable model techniques for image segmentation: a pragmatic comparative study. Diane Lingrand, Johan Montagnat Rainbow Team I3S Laboratory UMR 6070 University of Nice - Sophia
More informationSCIENCE & TECHNOLOGY
Pertanika J. Sci. & Technol. 26 (1): 309-316 (2018) SCIENCE & TECHNOLOGY Journal homepage: http://www.pertanika.upm.edu.my/ Application of Active Contours Driven by Local Gaussian Distribution Fitting
More informationImage Segmentation. Ross Whitaker SCI Institute, School of Computing University of Utah
Image Segmentation Ross Whitaker SCI Institute, School of Computing University of Utah What is Segmentation? Partitioning images/volumes into meaningful pieces Partitioning problem Labels Isolating a specific
More informationB. Tech. Project Second Stage Report on
B. Tech. Project Second Stage Report on GPU Based Active Contours Submitted by Sumit Shekhar (05007028) Under the guidance of Prof Subhasis Chaudhuri Table of Contents 1. Introduction... 1 1.1 Graphic
More informationHierarchical Segmentation of Thin Structures in Volumetric Medical Images
Hierarchical Segmentation of Thin Structures in Volumetric Medical Images Michal Holtzman-Gazit 1, Dorith Goldsher 2, and Ron Kimmel 3 1 Electrical Engineering Department 2 Faculty of Medicine - Rambam
More informationImage Segmentation. Ross Whitaker SCI Institute, School of Computing University of Utah
Image Segmentation Ross Whitaker SCI Institute, School of Computing University of Utah What is Segmentation? Partitioning images/volumes into meaningful pieces Partitioning problem Labels Isolating a specific
More information1. Two double lectures about deformable contours. 4. The transparencies define the exam requirements. 1. Matlab demonstration
Practical information INF 5300 Deformable contours, I An introduction 1. Two double lectures about deformable contours. 2. The lectures are based on articles, references will be given during the course.
More informationBoundary Extraction Using Poincare Map Method
Boundary Extraction Using Poincare Map Method Ruchita V. Indulkar, Sanjay D. Jondhale ME Computer, Department of Computer Engineering,SVIT, Chincholi, Nashik, Maharastra,India. Associate Professor, Department
More informationPoint-Based Geometric Deformable Models for Medical Image Segmentation
Point-Based Geometric Deformable Models for Medical Image Segmentation Hon Pong Ho 1, Yunmei Chen 2, Huafeng Liu 1,3, and Pengcheng Shi 1 1 Dept. of EEE, Hong Kong University of Science & Technology, Hong
More informationCHAPTER-4 LOCALIZATION AND CONTOUR DETECTION OF OPTIC DISK
CHAPTER-4 LOCALIZATION AND CONTOUR DETECTION OF OPTIC DISK Ocular fundus images can provide information about ophthalmic, retinal and even systemic diseases such as hypertension, diabetes, macular degeneration
More informationEdge-Preserving Denoising for Segmentation in CT-Images
Edge-Preserving Denoising for Segmentation in CT-Images Eva Eibenberger, Anja Borsdorf, Andreas Wimmer, Joachim Hornegger Lehrstuhl für Mustererkennung, Friedrich-Alexander-Universität Erlangen-Nürnberg
More informationA Level-Set Based Volumetric CT Segmentation Technique: A Case Study with Pulmonary Air Bubbles
A Level-Set Based Volumetric CT Segmentation Technique: A Case Study with Pulmonary Air Bubbles José Silvestre Silva 1,2, Beatriz Sousa Santos 1,3, Augusto Silva 1,3, and Joaquim Madeira 1,3 1 Departamento
More informationBlood Vessel Diameter Estimation System Using Active Contours
Blood Vessel Diameter Estimation System Using Active Contours Ana Tizon, Jane Courtney School of Electronic & Communications Engineering Dublin Institute of Technology Dublin, Ireland atizon@yahoo.com
More informationMeshless Modeling, Animating, and Simulating Point-Based Geometry
Meshless Modeling, Animating, and Simulating Point-Based Geometry Xiaohu Guo SUNY @ Stony Brook Email: xguo@cs.sunysb.edu http://www.cs.sunysb.edu/~xguo Graphics Primitives - Points The emergence of points
More informationAcknowledgements. I also want to thank Mr. Bill Whyte for his suggestion on evaluation of system and project.
Summary This project has investigated methods for finding both the central axis and diameters of the blood vessels. Three kinds of methods maximum intensity project (MIP), shaded surface display (SSD)
More information3D Surface Reconstruction of the Brain based on Level Set Method
3D Surface Reconstruction of the Brain based on Level Set Method Shijun Tang, Bill P. Buckles, and Kamesh Namuduri Department of Computer Science & Engineering Department of Electrical Engineering University
More informationMEDICAL IMAGE NOISE REDUCTION AND REGION CONTRAST ENHANCEMENT USING PARTIAL DIFFERENTIAL EQUATIONS
MEDICAL IMAGE NOISE REDUCTION AND REGION CONTRAST ENHANCEMENT USING PARTIAL DIFFERENTIAL EQUATIONS Miguel Alemán-Flores, Luis Álvarez-León Departamento de Informática y Sistemas, Universidad de Las Palmas
More informationSubvoxel Segmentation and Representation of Brain Cortex Using Fuzzy Clustering and Gradient Vector Diffusion
Subvoxel Segmentation and Representation of Brain Cortex Using Fuzzy Clustering and Gradient Vector Diffusion Ming-Ching Chang Xiaodong Tao GE Global Research Center {changm, taox} @ research.ge.com SPIE
More informationNorbert Schuff VA Medical Center and UCSF
Norbert Schuff Medical Center and UCSF Norbert.schuff@ucsf.edu Medical Imaging Informatics N.Schuff Course # 170.03 Slide 1/67 Objective Learn the principle segmentation techniques Understand the role
More informationA Modified Image Segmentation Method Using Active Contour Model
nd International Conference on Electrical, Computer Engineering and Electronics (ICECEE 015) A Modified Image Segmentation Method Using Active Contour Model Shiping Zhu 1, a, Ruidong Gao 1, b 1 Department
More informationVariational Methods II
Mathematical Foundations of Computer Graphics and Vision Variational Methods II Luca Ballan Institute of Visual Computing Last Lecture If we have a topological vector space with an inner product and functionals
More informationSegmentation in Noisy Medical Images Using PCA Model Based Particle Filtering
Segmentation in Noisy Medical Images Using PCA Model Based Particle Filtering Wei Qu a, Xiaolei Huang b, and Yuanyuan Jia c a Siemens Medical Solutions USA Inc., AX Division, Hoffman Estates, IL 60192;
More informationMedical Image Segmentation by Active Contour Improvement
American Journal of Software Engineering and Applications 7; 6(): 3-7 http://www.sciencepublishinggroup.com//asea doi:.648/.asea.76. ISSN: 37-473 (Print); ISSN: 37-49X (Online) Medical Image Segmentation
More informationAutomatic Extraction of Femur Contours from Hip X-ray Images
Automatic Extraction of Femur Contours from Hip X-ray Images Ying Chen 1, Xianhe Ee 1, Wee Kheng Leow 1, and Tet Sen Howe 2 1 Dept of Computer Science, National University of Singapore, 3 Science Drive
More informationLocal Binary Signed Pressure Force Function Based Variation Segmentation Model.
Journal of Information & Communication Technology Vol. 9, No. 1, (Spring2015) 01-12 Local Binary Signed Pressure Force Function Based Variation Segmentation Model. Tariq Ali * Institute of Social Policy
More informationGlobal Minimization of the Active Contour Model with TV-Inpainting and Two-Phase Denoising
Global Minimization of the Active Contour Model with TV-Inpainting and Two-Phase Denoising Shingyu Leung and Stanley Osher Department of Mathematics, UCLA, Los Angeles, CA 90095, USA {syleung, sjo}@math.ucla.edu
More informationBiomedical Image Processing
Biomedical Image Processing Jason Thong Gabriel Grant 1 2 Motivation from the Medical Perspective MRI, CT and other biomedical imaging devices were designed to assist doctors in their diagnosis and treatment
More informationNIH Public Access Author Manuscript Proc Soc Photo Opt Instrum Eng. Author manuscript; available in PMC 2014 October 07.
NIH Public Access Author Manuscript Published in final edited form as: Proc Soc Photo Opt Instrum Eng. 2014 March 21; 9034: 903442. doi:10.1117/12.2042915. MRI Brain Tumor Segmentation and Necrosis Detection
More informationShape-based Diffeomorphic Registration on Hippocampal Surfaces Using Beltrami Holomorphic Flow
Shape-based Diffeomorphic Registration on Hippocampal Surfaces Using Beltrami Holomorphic Flow Abstract. Finding meaningful 1-1 correspondences between hippocampal (HP) surfaces is an important but difficult
More informationMultiple Contour Finding and Perceptual Grouping as a set of Energy Minimizing Paths
Multiple Contour Finding and Perceptual Grouping as a set of Energy Minimizing Paths Laurent D. COHEN and Thomas DESCHAMPS CEREMADE, UMR 7534, Université Paris-Dauphine 75775 Paris cedex 16, France cohen@ceremade.dauphine.fr
More informationLocal or Global Minima: Flexible Dual-Front Active Contours
Local or Global Minima: Flexible Dual-Front Active Contours Hua Li 1,2 and Anthony Yezzi 1 1 School of ECE, Georgia Institute of Technology, Atlanta, GA, USA 2 Dept. of Elect. & Info. Eng., Huazhong Univ.
More informationSnakes, level sets and graphcuts. (Deformable models)
INSTITUTE OF INFORMATION AND COMMUNICATION TECHNOLOGIES BULGARIAN ACADEMY OF SCIENCE Snakes, level sets and graphcuts (Deformable models) Centro de Visión por Computador, Departament de Matemàtica Aplicada
More informationNon-Rigid Image Registration III
Non-Rigid Image Registration III CS6240 Multimedia Analysis Leow Wee Kheng Department of Computer Science School of Computing National University of Singapore Leow Wee Kheng (CS6240) Non-Rigid Image Registration
More informationComputer Vision. Recap: Smoothing with a Gaussian. Recap: Effect of σ on derivatives. Computer Science Tripos Part II. Dr Christopher Town
Recap: Smoothing with a Gaussian Computer Vision Computer Science Tripos Part II Dr Christopher Town Recall: parameter σ is the scale / width / spread of the Gaussian kernel, and controls the amount of
More informationIntroduction to Medical Image Processing
Introduction to Medical Image Processing Δ Essential environments of a medical imaging system Subject Image Analysis Energy Imaging System Images Image Processing Feature Images Image processing may be
More informationLecture 12 Level Sets & Parametric Transforms. sec & ch. 11 of Machine Vision by Wesley E. Snyder & Hairong Qi
Lecture 12 Level Sets & Parametric Transforms sec. 8.5.2 & ch. 11 of Machine Vision by Wesley E. Snyder & Hairong Qi Spring 2017 16-725 (CMU RI) : BioE 2630 (Pitt) Dr. John Galeotti The content of these
More information1. Two double lectures about deformable contours. 4. The transparencies define the exam requirements. 1. Matlab demonstration
Practical information INF 5300 Deformable contours, II An introduction 1. Two double lectures about deformable contours. 2. The lectures are based on articles, references will be given during the course.
More informationA Geometric Flow Approach for Region-based Image Segmentation
A Geometric Flow Approach for Region-based Image Segmentation Juntao Ye Institute of Automation, Chinese Academy of Sciences, Beijing, China juntao.ye@ia.ac.cn Guoliang Xu Institute of Computational Mathematics
More informationAccurate Boundary Localization using Dynamic Programming on Snakes
Canadian Conference on Computer and Robot Vision Accurate Boundary Localization using Dynamic Programming on Snakes Akshaya Kumar Mishra, Paul Fieguth, D.A. Clausi VIP research Group, Systems Design Engineering,
More informationVolume visualization. Volume visualization. Volume visualization methods. Sources of volume visualization. Sources of volume visualization
Volume visualization Volume visualization Volumes are special cases of scalar data: regular 3D grids of scalars, typically interpreted as density values. Each data value is assumed to describe a cubic
More informationAbstract Constructing a mathematical representation of an object boundary (boundary mapping) from images is an important problem that is of importance
DEFORMABLE MODELS WITH APPLICATION TO HUMAN CEREBRAL CORTEX RECONSTRUCTION FROM MAGNETIC RESONANCE IMAGES by Chenyang Xu A dissertation submitted to the Johns Hopkins University in conformity with the
More informationMulti-Scale Free-Form Surface Description
Multi-Scale Free-Form Surface Description Farzin Mokhtarian, Nasser Khalili and Peter Yuen Centre for Vision Speech and Signal Processing Dept. of Electronic and Electrical Engineering University of Surrey,
More informationEE795: Computer Vision and Intelligent Systems
EE795: Computer Vision and Intelligent Systems Spring 2012 TTh 17:30-18:45 WRI C225 Lecture 02 130124 http://www.ee.unlv.edu/~b1morris/ecg795/ 2 Outline Basics Image Formation Image Processing 3 Intelligent
More informationA Multilayered Agent Society for Flexible Image Processing
The University of Hull Department of Computer Science A Multilayered Agent Society for Flexible Image Processing being a Thesis submited for the Degree of Doctor of Philosophy in the University of Hull
More information11/1/13. Visualization. Scientific Visualization. Types of Data. Height Field. Contour Curves. Meshes
CSCI 420 Computer Graphics Lecture 26 Visualization Height Fields and Contours Scalar Fields Volume Rendering Vector Fields [Angel Ch. 2.11] Jernej Barbic University of Southern California Scientific Visualization
More informationVisualization. CSCI 420 Computer Graphics Lecture 26
CSCI 420 Computer Graphics Lecture 26 Visualization Height Fields and Contours Scalar Fields Volume Rendering Vector Fields [Angel Ch. 11] Jernej Barbic University of Southern California 1 Scientific Visualization
More informationMultimodality Imaging for Tumor Volume Definition in Radiation Oncology
81 There are several commercial and academic software tools that support different segmentation algorithms. In general, commercial software packages have better implementation (with a user-friendly interface
More informationLaw, AKW; Zhu, H; Lam, FK; Chan, FHY; Chan, BCB; Lu, PP
Title Tumor boundary extraction in multislice MR brain images using region and contour deformation Author(s) Law, AKW; Zhu, H; Lam, FK; Chan, FHY; Chan, BCB; Lu, PP Citation International Workshop on Medical
More informationActive Contours Using a Constraint-Based Implicit Representation
To appear in Proceedings Computer Vision and Pattern Recognition, IEEE Computer Society Press, June 2005 Active Contours Using a Constraint-Based Implicit Representation Bryan S. Morse 1, Weiming Liu 1,
More informationFast 3D Brain Segmentation Using Dual-Front Active Contours with Optional User-Interaction
Fast 3D Brain Segmentation Using Dual-Front Active Contours with Optional User-Interaction Hua Li 1,2, Anthony Yezzi 1, and Laurent D. Cohen 3 1 School of ECE, Georgia Institute of Technology, Atlanta,
More informationNSCT BASED LOCAL ENHANCEMENT FOR ACTIVE CONTOUR BASED IMAGE SEGMENTATION APPLICATION
DOI: 10.1917/ijivp.010.0004 NSCT BASED LOCAL ENHANCEMENT FOR ACTIVE CONTOUR BASED IMAGE SEGMENTATION APPLICATION Hiren Mewada 1 and Suprava Patnaik Department of Electronics Engineering, Sardar Vallbhbhai
More informationMetaMorphs: Deformable Shape and Texture Models
MetaMorphs: Deformable Shape and Texture Models Xiaolei Huang, Dimitris Metaxas, Ting Chen Division of Computer and Information Sciences Rutgers University New Brunswick, NJ 8854, USA {xiaolei, dnm}@cs.rutgers.edu,
More informationBrain Structure Segmentation from MRI by Geometric Surface Flow
Hindawi Publishing Corporation International Journal of Biomedical Imaging Volume 2006, Article ID 86747, Pages 6 DOI 0.55/IJBI/2006/86747 Brain Structure Segmentation from MRI by Geometric Surface Flow
More informationThree-dimensional segmentation of bones from CT and MRI using fast level sets
Three-dimensional segmentation of bones from CT and MRI using fast level sets Jakub Krátký and Jan Kybic Center for Machine perception, Faculty of Electrical Engineering, Czech Technical University, Prague,
More informationImproved Geometric Constraints on Deformable Surface Model for Volumetric Segmentation
Improved Geometric Constraints on Deformable Surface Model for Volumetric Segmentation Jiuxiang Hu 1, Anshuman Razdan 1, Gregory M. Nielson 2, and Gerald E. Farin 2 1 Partnership for Research in Spatial
More informationA Survey of Image Segmentation Based On Multi Region Level Set Method
A Survey of Image Segmentation Based On Multi Region Level Set Method Suraj.R 1, Sudhakar.K 2 1 P.G Student, Computer Science and Engineering, Hindusthan College Of Engineering and Technology, Tamilnadu,
More informationSegmentation and Modeling of the Spinal Cord for Reality-based Surgical Simulator
Segmentation and Modeling of the Spinal Cord for Reality-based Surgical Simulator Li X.C.,, Chui C. K.,, and Ong S. H.,* Dept. of Electrical and Computer Engineering Dept. of Mechanical Engineering, National
More informationSkeleton Extraction via Anisotropic Heat Flow
DIREKOGLU ET AL.: SKELETON EXTRACTION VIA ANISOTROPIC HEAT FLOW 1 Skeleton Extraction via Anisotropic Heat Flow Cem Direkoglu cem.direkoglu@scss.tcd.ie Rozenn Dahyot rozenn.dahyot@scss.tcd.ie Michael Manzke
More informationAutomatic segmentation of the cortical grey and white matter in MRI using a Region Growing approach based on anatomical knowledge
Automatic segmentation of the cortical grey and white matter in MRI using a Region Growing approach based on anatomical knowledge Christian Wasserthal 1, Karin Engel 1, Karsten Rink 1 und André Brechmann
More informationTEMPLATE-BASED AUTOMATIC SEGMENTATION OF MASSETER USING PRIOR KNOWLEDGE
TEMPLATE-BASED AUTOMATIC SEGMENTATION OF MASSETER USING PRIOR KNOWLEDGE H.P. Ng 1,, S.H. Ong 3, P.S. Goh 4, K.W.C. Foong 1, 5, W.L. Nowinski 1 NUS Graduate School for Integrative Sciences and Engineering,
More informationEdge and local feature detection - 2. Importance of edge detection in computer vision
Edge and local feature detection Gradient based edge detection Edge detection by function fitting Second derivative edge detectors Edge linking and the construction of the chain graph Edge and local feature
More informationNon-Rigid Registration I
Non-Rigid Registration I CS6240 Multimedia Analysis Leow Wee Kheng Department of Computer Science School of Computing National University of Singapore Leow Wee Kheng (CS6240) Non-Rigid Registration I 1
More informationSUPPLEMENTARY FILE S1: 3D AIRWAY TUBE RECONSTRUCTION AND CELL-BASED MECHANICAL MODEL. RELATED TO FIGURE 1, FIGURE 7, AND STAR METHODS.
SUPPLEMENTARY FILE S1: 3D AIRWAY TUBE RECONSTRUCTION AND CELL-BASED MECHANICAL MODEL. RELATED TO FIGURE 1, FIGURE 7, AND STAR METHODS. 1. 3D AIRWAY TUBE RECONSTRUCTION. RELATED TO FIGURE 1 AND STAR METHODS
More informationMedical Image Segmentation using Level Sets
Medical Image Segmentation using Level Sets Technical Report #CS-8-1 Tenn Francis Chen Abstract Segmentation is a vital aspect of medical imaging. It aids in the visualization of medical data and diagnostics
More informationResearch on the Wood Cell Contour Extraction Method Based on Image Texture and Gray-scale Information.
, pp. 65-74 http://dx.doi.org/0.457/ijsip.04.7.6.06 esearch on the Wood Cell Contour Extraction Method Based on Image Texture and Gray-scale Information Zhao Lei, Wang Jianhua and Li Xiaofeng 3 Heilongjiang
More informationImage Segmentation and Registration
Image Segmentation and Registration Dr. Christine Tanner (tanner@vision.ee.ethz.ch) Computer Vision Laboratory, ETH Zürich Dr. Verena Kaynig, Machine Learning Laboratory, ETH Zürich Outline Segmentation
More informationKeywords segmentation, vector quantization
Volume 4, Issue 1, January 2014 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Colour Image
More informationsmooth coefficients H. Köstler, U. Rüde
A robust multigrid solver for the optical flow problem with non- smooth coefficients H. Köstler, U. Rüde Overview Optical Flow Problem Data term and various regularizers A Robust Multigrid Solver Galerkin
More informationNormalized cuts and image segmentation
Normalized cuts and image segmentation Department of EE University of Washington Yeping Su Xiaodan Song Normalized Cuts and Image Segmentation, IEEE Trans. PAMI, August 2000 5/20/2003 1 Outline 1. Image
More informationFeature Extraction and Image Processing, 2 nd Edition. Contents. Preface
, 2 nd Edition Preface ix 1 Introduction 1 1.1 Overview 1 1.2 Human and Computer Vision 1 1.3 The Human Vision System 3 1.3.1 The Eye 4 1.3.2 The Neural System 7 1.3.3 Processing 7 1.4 Computer Vision
More informationPROSTATE CANCER DETECTION USING LABEL IMAGE CONSTRAINED MULTIATLAS SELECTION
PROSTATE CANCER DETECTION USING LABEL IMAGE CONSTRAINED MULTIATLAS SELECTION Ms. Vaibhavi Nandkumar Jagtap 1, Mr. Santosh D. Kale 2 1 PG Scholar, 2 Assistant Professor, Department of Electronics and Telecommunication,
More informationColor Image Segmentation Editor Based on the Integration of Edge-Linking, Region Labeling and Deformable Model
This paper appears in: IEEE International Conference on Systems, Man and Cybernetics, 1999 Color Image Segmentation Editor Based on the Integration of Edge-Linking, Region Labeling and Deformable Model
More information