Segmentation of 3-D medical image data sets with a combination of region based initial segmentation and active surfaces
|
|
- Earl Edwards
- 5 years ago
- Views:
Transcription
1 Header for SPIE use Segmentation of 3-D medical image data sets with a combination of region based initial segmentation and active surfaces Regina Pohle, Thomas Behlau, Klaus D. Toennies Otto-von-Guericke University Magdeburg, Department of Simulation and Graphics ABSTRACT Segmentation is an essential step in the analysis of medical images. For segmentation of 3-D data sets in clinical practice segmentation methods are necessary which have a small user interaction time and which are highly flexible. For this purpose we propose a two-step segmentation approach. The first step results in a coarse segmentation using the Image Foresting Transformation. In the second step an active surface creates the final segmentation. Our segmentation method was tested for segmentation on real CT images. The performance was compared with the manual segmentation. We found our method to work reliable. Keywords: Segmentation, region based segmentation, active surface 1. INTRODUCTION The analysis of medical images for the purpose of computer-aided diagnosis and therapy planning includes segmentation as a preliminary stage for the visualization or quantification. For semi-automatic segmentation of structures in 3-D CT image interactive thresholding aided by morphological information 1, active surfaces and active shape models 3, the use of live wire technique 4, model-based adaptive region growing technique 5, or the watershed transformation 6 are among the many methods that were recently employed. Various models for including a-priori information about the expected content of the image are used in segmentation. The applied a-priori knowledge consists of a combination of anatomical / physiological information and of information about the image formation process. Accuracy and degree of completeness of model information greatly enhances the practicability of the outcome and degree of automatic of a segmentation process. However, model information in medical segmentation is often too complex or cannot be specified precisely, so that a complete automatic extraction is not feasible. Interaction is necessary, but it should be restricted to those aspects where information is entered that is readily available to the user and where it can be entered in a robust fashion. Further, the interaction should require little time. Some of the mentioned methods at the beginning do not fulfill these demands or they use to little a-priori knowledge so that they are usable only for simple segmentation tasks. Therefore, for the segmentation of complex and variable structure like it is typical in medical 3-D image the use of active shape models 3 and shape based adaptations of 3-D deformable models 7 gain importance because complex a- priori model information can be modeled by the two methods. Before a successful segmentation with these methods can be started, the user has to mark a large number of corresponding points in a training data set. From this data a statistic model of shape variation can be computed. This manual preprocessing step is very time consuming and it must repeated for every new segmentation task. For this reason we have developed an new two-level segmentation method in which the user interaction time is reduced. In a first coarse segmentation the user contributes complex a-priori knowledge with a few interaction steps in the algorithm. In the next step, the segmentation result of the first step is adjusted to the real object contour with the active surface approach. Details will be explained in the following section. Comparisons of the performance of our method with a manual segmentation of the liver in CT images will be presented. Results are shown in section 4.. TWO-LEVEL SEGMENTATION METHOD In our two-level segmentation process we use in the first step the region-based segmentation with the Image Foresting Transformation (IFT) 8 for the coarse segmentation. The result is a binary image in which position, shape and size of the searched structure is approximated. It is used to define a search corridor for an active surface, which is adjusted for finding the real object contour. In both steps the segmentation is carried out in 3-D.
2 .1. Coarse segmentation For reducing the computation time, we reduced the image resolution for the IFT algorithm 8. The original data set with a resolution in x- and y- direction of 51x51 voxel is mapped onto 64x64 voxel. This corresponds to a reduction of the image resolution by the factor 8. In z-direction the image resolution is smaller for the medical standard image processes (CT, MRT). For our CT data set the slice thickness was.5 mm in contrast to the pixel spacing of 0.74 mm x 0.74 mm so that the image resolution varies by the factor 4. Thu the image resolution in z-direction was reduced only by the factor to obtain the same relationships in all three directions. The 3-D data set with the reduced image resolution is called I. Furthermore an adjacent relation ρ in R 3 is defined by considering all pairs of voxel (p,q) I satisfying d(p,q) = 1, where d is the Euclidean distance between p und q. In the first step of the algorithm the user marks initial foreground and background regions by drawing lines interactively in few slices of this data set I. The interaction can occur alternatively on three section plane which are parallel to the three coordinate axes (see Figure 1 at the top). In our tests less than 10 lines were sufficient for the extraction of the area of interest. The marked lines are used for the definition of the initial path costs in the data set I, which is described by a weighted and non-oriented graph G during the region-oriented segmentation, based on the IFT. Each voxel in the data set is a node of the graph, and each pair (p,q) of ρ-adjacent voxel defines a non-oriented edge in G. For our computation we use the local neighborhood ρ = six voxel. The edge costs C(p,q) are defined as the absolute difference of the gray values of the adjacent voxel according to the equation 1. C ( p, q) I( p) I ( q) = (1) The neighborhood ρ and the costs C are used by IFT to reduce the region growing segmentation problem to an equivalent shortest-path forest problem. The initial path costs PF I for all manually marked voxels are the maximum costs from its incident edges for the defined neighborhood I i=1,..,6 where q i is a voxel in the direct neighborhood of p. ( C( p q )) PF = max, () i For all other voxel the initial path costs are infinite. Nodes are split into foreground and background according to the path costs. The optimal splitting of the data set is computed with dynamic programming. We use the following version of the algorithm according to 8 : 1) Set PF(p) to, set a label image lb(p) with the current label of each voxel to 0 for all voxel p I. ) Set PF(r) to PF I (r) according equation and lb(r) to its corresponding label for all marked voxel r 3) Put all voxel r in a priority queue Q 4) While Q is not empty do a) Remove the voxel p with the minimal PF(p) from Q and put p in a list L of voxel which have already been processed b) for each ρ-adjacent voxel q of p and q L do (i) compute the new path cost tmp based on PF(p) and C(p,q) (ii) if tmp < PF(q) then a) set PF(q) = tmp b) set lb(q) = lb(p) c) set the pointer with the current parents of q to p if q Q then insert q in Q else update position of q in Q After this first splitting of the image the user can introduce additional start regions as model information to improve the previous result. Outgoing from these new start regions a reorganization and re-classification of the regions starts until no node can be reached at lower cost by another path. The resulting binary image is then extended to its original resolution (see Figure 1 in the middle).
3 .. Final segmentation The boundary between foreground and background voxel is smoothed with a boxcar filter. Subsequently, two threshold values T 1 > 0.5 and T < 0.5 are determined. It defines a search corridor for the active surface model. It consists of all voxel p with T p T 1. The initial position of the active surface may be either the internal boundary at T=T 1, the external boundary at T=T or the middle boundary at T = 0.5. For the example of the liver segmentation in 3-D CT images the derived search corridor from the coarse segmentation is illustrated in Figure 1 below. In the next step vertex points are determined for the initial location of the active surface (internal, external or middle boundary) by applying the marching cube algorithm 9. The number of the vertex points can be selected by the user. A small number of vertex points guarantee shorter computation time and a larger number of surface points guarantee a higher precision by the adaptation of the contour to the real contour. The surface υ is parametrically defined as υ ( s, r) = ( x( r), y( r), z( r)) (3) with x(r), y(r), and z(r) being the x, y, z co-ordinates on the surface and r [0, 1]. The energy function to be minimized can be defined by 1 1 υ υ υ υ υ E( υ ) = w10 + w01 + w11 + w0 + w01 + P( v( r)) dsdr (4) s r s r s r 0 0 where P is the potential associated with the external forces. The external forces cause the surface to be attracted to the real surface points of the object. They are dependent on the 3-D gradients in the image. The internal forces emanate from the shape of the surface. Their influence on the energy function depends on the selection of the coefficient w ij. The coefficients w 10 and w 01 describe the elasticity of the shape of the surface. The rigidity of the surface is determined by w 0 and w 0. The parameter w 11 characterized the resistance to twist. In our program the user has the choice to select the different values for these parameters. With the optimal parameter selection he/she can constrain the surface structure by adjusting boundary conditions..3. Defining of the external forces The power of the external forces is dependent on the gradient in the image at the location of the vertex points. These external forces should move the surface to positions where the real surface is located. Although it is often assumed that surfaces are to be found at locations with the highest gray level gradient, this is not true in CT images. Here, objects to be segmented may often have a smaller gradient value than other structures in the neighborhood, such as bone or air in the intestines. Therefore, we estimate an expected gradient range grad E from the result of the coarse segmentation. Then, we use these values for the normalization and computation of external forces according to the equation 5. P( v( r)) = grad ( I( x, y, z )) grad( I( x, y, z)) (5) E E E E In the first step we search corresponding vertex points between the internal and external surfaces of the result of the coarse segmentation. These corresponding points are in 3-d space connected with a line. Along these lines the gray values are selected and the gray value gradients are computed. In these gradient profiles we search for the two gradient maxima between the internal boundary and the middle boundary as well as between the middle boundary and the external boundary. Off these two maxima the one is selected as ideal gradient length for external force computation, which is located closer to the middle boundary. This procedure was chosen because it is supposed that the user has the coarse segmentation carried out in this way that the middle boundary is near to the real boundary of the searched structure. Our method is illustrated in Figure 3 and Figure 4 for the example line of Figure. The positions for the used estimated gradient values are shown in Figure 5. In this figure you can see that the most estimated gradient values are in the correct position on the boundary of the kidney. That way the computed external forces achieve an adaptation of the active surface to the real object boundary.
4 Figure : Part of the CT image with the kidney and internal, middle and external vertex point which are located in the slice. The line between the two corresponding point which are located, in this case in the same slice is marked black. Figure 3: Profile of the gray values for the marked line in Figure in direction from the internal vertex point to the external vertex point. The line shows the position of the middle boundary. Figure 4: Gradient profile along the marked line from Figure. The line shows again the position of the middle boundary. The marked gradient value was selected for the gradient estimation with our algorithm. Figure 5: Position for the estimated gradient values. The most estimated gradient values are in the correct position on the boundary between kidney tissue and surrounding structures. 3. PERFORMANCE EVALUATION OF THE ALGORITHM For evaluation of our segmentation algorithm we have used empirical discrepancy methods 10. These methods compare results with a gold standard. We selected discrepancy methods because they allow an objective and quantitative assessment of the segmentation algorithm with a close relationship to a concrete application. In our case we have used the results of the manual segmentation as the gold standard. For the tests we have had two manual segmentation results from different physicians for the three example data sets. So, we could measure a value for the intra-individual variability in the specification of the ground truth. In addition, we have had two results of the manual segmentation from one physician received on two different days for several slices of one data set. This delivered us a value for the intra-individual error.
5 Because the physician is interested on the volume of the segmented liver, we have used the value for the liver volume as error metric for assessment of the quality of the segmentation results. This metric has the advantage that it is independent of the region characteristics and of the segmentation method. In order to be independently of the user marking of the foreground and background regions all tests were carried out using three different coarse segmentation results. 4. RESULTS We have tested the described segmentation method for segmentation of the liver in CT data sets. The liver region exhibited only little PVE. The signal-to-noise-relation between liver tissue and surrounding structures is small. For our test data sets it was estimated to be 1:1.75. The required segmentation time was about three minutes for a data set of 51x51x90 voxel. About one minute of this time is necessary for the manual marking of the start regions and for the coarse segmentation step. The step of the active surface adaptation required about two minutes. In contrast to our two-step segmentation method the manual segmentation time of such data set was one hour. The average error for the volume determination compared with a manual segmentation that was carried out by a practicing surgeon was 8 percent. The main problem was the segmentation in the first and in the last slices. In these parts the segmented region was mostly too small because of the choice of the parameters for the internal force computation. The intra individual error for the volume determination on 5 by chance selected slices of two segmentations of one physician was 10 percent. A comparison between the inter individual error was 10 percent too. So our result was in the same range like the intra and inter individual variation. An example for the liver segmentation with our two level segmentation method is seen in Figure 6. In Figure 7 the result of this example is shown as a 3D visualization. Figure 6: Part of a slice of the CT data set with the liver region, left: result of the manual segmentation, right: result with our two-level segmentation approach
6 Figure 7: 3D visualization of the result of the liver segmentation 5. DISCUSSION Parametric deformable models have been applied for many different segmentation tasks. But, they have a limitation if the designed object boundary differ greatly in size and shape from the object boundary. In our approach, we avoid this problem using user interaction. In contrast to the other segmentation approaches with parametric deformable models we need only little user interaction for defining the search corridor. Because we work in the coarse segmentation step on an image with reduced image resolution the user marking has only an insignificant influence on the defined search corridor. This influence is further reduced through the possibility of the additional marking of object and background regions until the result of the coarse segmentation is satisfactory. By an optimal choice of the search corridor, we can guarantee that the surface is located near by the real boundary after the coarse segmentation. So, the shape of the initial surface is adapted on the real object shape too. This has the advantage that the structures are founded in a short time. According to the user marking of the object we can process different topological shape with our approach. 6. CONCLUSION In this paper we presented a new process of automatically segmentation based on a two level process. In the first step of the coarse segmentation we have extracted a initial surface and a search corridor using the region based segmentation approach based on the IFT algorithm. In the second step the initial surface was adapted on the real object surface using the active surface segmentation method. We tested the practicability of this segmentation process for segmentation of liver parenchyma in CT data sets. The results of our method were compared with a manual segmentation. We found our method to work reliable. The deviation of the measured volume value between our segmentation results and manual segmentation was in the same range like the deviation between two different manual segmentations results. The deviation resulted from the different decisions about the object affiliation in the first and last slices in which the object is included. In future work we will test our approach for other medical structures and for other medical image modalities. 7. REFERENCES 1. K.H. Höhne, W.A. Hanson, Interactive 3D segmentation of MRI and CT volumes using morphological operations, J. Comp. Assisted Tomogr, 16 (), pp , I. Cohen, L.D. Cohen, N. Ayache, Using deformable Surfaces to segment 3-D images and infer differential structures, CVGIP: Image Understanding, 56 (), pp. 4-63, T.F. Coote C.J. Taylo, Statistical models of appearance for medical image analysis and computer vision, Medical Imaging 001: Image Processing, Proc. SPIE, Vol. 43, pp , A. Schenk, G. Prause, H.O. Peitgen, Efficient Semiautomatic Segmentation of 3D Objects in Medical Images, Proc. of Medical Image Computing and Computer-assisted Intervention (MICCAI), Springer, LNCS, Vol. 1935, pp , R. Pohle, K.D. Toennie Self-learning model-based segmentation of medical images, Image Processing & Communication 7(3-4), pp , 001.
7 6. T. Schindewolf, H.O. Peitgen, Interaktive Bildsegmentierung von CT- und MR-Daten auf Basis einer modifizierten Wasserscheidentransformation, Bildverarbeitung für die Medizin, Springer, Series Informatik Aktuell, pp , V. Pekar, M.R. Krau C. Lorenz et. al., Shape model based adaption of 3-D deformable meshes for segmentation of medical images, Medical Imaging 001: Image Processing, Proc. SPIE, Vol. 43, pp , A.X. Falcao, R. de A. Lotufo, G. Araujo, The Image Foresting Transformation, Relatorio Tecnico IC-00-1, C. Lorenson: Marching Cube A High Resolution 3D Surface Construction Algorithm, Computer Graphics 1, pp , Zhang Y J, A Survey on Evaluation Methods for Image Segmentation, Pattern Recognition, 9(8), pp , 1996.
8 Figure 1: At the top: selected region of the CT image of the abdomen with marks for the liver region and marks for the background region, in the middle: result of the coarse segmentation of the liver, below: extracted search corridor for the final segmentation
Segmentation of medical images using adaptive region growing
Header for SPIE use Segmentation of medical images using adaptive region growing Regina Pohle*, Klaus D. Toennies Otto-von-Guericke University Magdeburg, Department of Simulation and Graphics ABSTRACT
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 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 informationSimultaneous Model-based Segmentation of Multiple Objects
Simultaneous Model-based Segmentation of Multiple Objects Astrid Franz 1, Robin Wolz 1, Tobias Klinder 1,2, Cristian Lorenz 1, Hans Barschdorf 1, Thomas Blaffert 1, Sebastian P. M. Dries 1, Steffen Renisch
More information8/3/2017. Contour Assessment for Quality Assurance and Data Mining. Objective. Outline. Tom Purdie, PhD, MCCPM
Contour Assessment for Quality Assurance and Data Mining Tom Purdie, PhD, MCCPM Objective Understand the state-of-the-art in contour assessment for quality assurance including data mining-based techniques
More informationFully Automatic Model Creation for Object Localization utilizing the Generalized Hough Transform
Fully Automatic Model Creation for Object Localization utilizing the Generalized Hough Transform Heike Ruppertshofen 1,2,3, Cristian Lorenz 2, Peter Beyerlein 4, Zein Salah 3, Georg Rose 3, Hauke Schramm
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 informationOccluded Facial Expression Tracking
Occluded Facial Expression Tracking Hugo Mercier 1, Julien Peyras 2, and Patrice Dalle 1 1 Institut de Recherche en Informatique de Toulouse 118, route de Narbonne, F-31062 Toulouse Cedex 9 2 Dipartimento
More informationAutomated segmentation methods for liver analysis in oncology applications
University of Szeged Department of Image Processing and Computer Graphics Automated segmentation methods for liver analysis in oncology applications Ph. D. Thesis László Ruskó Thesis Advisor Dr. Antal
More informationFully Automatic Multi-organ Segmentation based on Multi-boost Learning and Statistical Shape Model Search
Fully Automatic Multi-organ Segmentation based on Multi-boost Learning and Statistical Shape Model Search Baochun He, Cheng Huang, Fucang Jia Shenzhen Institutes of Advanced Technology, Chinese Academy
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 informationComparison Study of Clinical 3D MRI Brain Segmentation Evaluation
Comparison Study of Clinical 3D MRI Brain Segmentation Evaluation Ting Song 1, Elsa D. Angelini 2, Brett D. Mensh 3, Andrew Laine 1 1 Heffner Biomedical Imaging Laboratory Department of Biomedical Engineering,
More informationA Multiple-Layer Flexible Mesh Template Matching Method for Nonrigid Registration between a Pelvis Model and CT Images
A Multiple-Layer Flexible Mesh Template Matching Method for Nonrigid Registration between a Pelvis Model and CT Images Jianhua Yao 1, Russell Taylor 2 1. Diagnostic Radiology Department, Clinical Center,
More informationGeneration of Triangle Meshes from Time-of-Flight Data for Surface Registration
Generation of Triangle Meshes from Time-of-Flight Data for Surface Registration Thomas Kilgus, Thiago R. dos Santos, Alexander Seitel, Kwong Yung, Alfred M. Franz, Anja Groch, Ivo Wolf, Hans-Peter Meinzer,
More information2 Michael E. Leventon and Sarah F. F. Gibson a b c d Fig. 1. (a, b) Two MR scans of a person's knee. Both images have high resolution in-plane, but ha
Model Generation from Multiple Volumes using Constrained Elastic SurfaceNets Michael E. Leventon and Sarah F. F. Gibson 1 MIT Artificial Intelligence Laboratory, Cambridge, MA 02139, USA leventon@ai.mit.edu
More informationExtracting consistent and manifold interfaces from multi-valued volume data sets
Extracting consistent and manifold interfaces from multi-valued volume data sets Stephan Bischoff, Leif Kobbelt Lehrstuhl für Informatik 8, RWTH Aachen, 52056 Aachen Email: {bischoff,kobbelt}@informatik.rwth-aachen.de
More informationActive double-contour for segmentation of vessels in digital subtraction angiography
Active double-contour for segmentation of vessels in digital subtraction angiography Manfred Hinz *, Klaus D. Toennies *, Markus Grohmann *, Regina Pohle * Computer Vision Group, Dept. Informatics, Otto-von-Guericke-Universität
More information2D Vessel Segmentation Using Local Adaptive Contrast Enhancement
2D Vessel Segmentation Using Local Adaptive Contrast Enhancement Dominik Schuldhaus 1,2, Martin Spiegel 1,2,3,4, Thomas Redel 3, Maria Polyanskaya 1,3, Tobias Struffert 2, Joachim Hornegger 1,4, Arnd Doerfler
More informationElastic registration of medical images using finite element meshes
Elastic registration of medical images using finite element meshes Hartwig Grabowski Institute of Real-Time Computer Systems & Robotics, University of Karlsruhe, D-76128 Karlsruhe, Germany. Email: grabow@ira.uka.de
More informationMatching 3D Lung Surfaces with the Shape Context Approach. 1)
Matching 3D Lung Surfaces with the Shape Context Approach. 1) Martin Urschler, Horst Bischof Institute for Computer Graphics and Vision, TU Graz Inffeldgasse 16, A-8010 Graz E-Mail: {urschler, bischof}@icg.tu-graz.ac.at
More informationLecture: Segmentation I FMAN30: Medical Image Analysis. Anders Heyden
Lecture: Segmentation I FMAN30: Medical Image Analysis Anders Heyden 2017-11-13 Content What is segmentation? Motivation Segmentation methods Contour-based Voxel/pixel-based Discussion What is segmentation?
More informationProstate Detection Using Principal Component Analysis
Prostate Detection Using Principal Component Analysis Aamir Virani (avirani@stanford.edu) CS 229 Machine Learning Stanford University 16 December 2005 Introduction During the past two decades, computed
More informationKnowledge-Based Organ Identification from CT Images. Masahara Kobashi and Linda Shapiro Best-Paper Prize in Pattern Recognition Vol. 28, No.
Knowledge-Based Organ Identification from CT Images Masahara Kobashi and Linda Shapiro Best-Paper Prize in Pattern Recognition Vol. 28, No. 4 1995 1 Motivation The extraction of structure from CT volumes
More informationComputer-Aided Diagnosis in Abdominal and Cardiac Radiology Using Neural Networks
Computer-Aided Diagnosis in Abdominal and Cardiac Radiology Using Neural Networks Du-Yih Tsai, Masaru Sekiya and Yongbum Lee Department of Radiological Technology, School of Health Sciences, Faculty of
More information3D contour based local manual correction of tumor segmentations in CT scans
3D contour based local manual correction of tumor segmentations in CT scans Frank Heckel a, Jan Hendrik Moltz a, Lars Bornemann a, Volker Dicken a, Hans-Christian Bauknecht b,michaelfabel c, Markus Hittinger
More information3D Volume Mesh Generation of Human Organs Using Surface Geometries Created from the Visible Human Data Set
3D Volume Mesh Generation of Human Organs Using Surface Geometries Created from the Visible Human Data Set John M. Sullivan, Jr., Ziji Wu, and Anand Kulkarni Worcester Polytechnic Institute Worcester,
More informationMedical images, segmentation and analysis
Medical images, segmentation and analysis ImageLab group http://imagelab.ing.unimo.it Università degli Studi di Modena e Reggio Emilia Medical Images Macroscopic Dermoscopic ELM enhance the features of
More informationLIVER cancer has been among the 6 most common. Automatic Liver Segmentation based on Shape Constraints and Deformable Graph Cut in CT Images
IEEE TRANSACTIONS ON IMAGE PROCESSING 1 Automatic Liver Segmentation based on Shape Constraints and Deformable Graph Cut in CT Images Guodong Li #, Xinjian Chen #, Fei Shi, Weifang Zhu, Jie Tian*, Fellow,
More informationSegmentation of Bony Structures with Ligament Attachment Sites
Segmentation of Bony Structures with Ligament Attachment Sites Heiko Seim 1, Hans Lamecker 1, Markus Heller 2, Stefan Zachow 1 1 Visualisierung und Datenanalyse, Zuse-Institut Berlin (ZIB), 14195 Berlin
More informationSTIC AmSud Project. Graph cut based segmentation of cardiac ventricles in MRI: a shape-prior based approach
STIC AmSud Project Graph cut based segmentation of cardiac ventricles in MRI: a shape-prior based approach Caroline Petitjean A joint work with Damien Grosgeorge, Pr Su Ruan, Pr JN Dacher, MD October 22,
More informationSemi-Automatic Detection of Cervical Vertebrae in X-ray Images Using Generalized Hough Transform
Semi-Automatic Detection of Cervical Vertebrae in X-ray Images Using Generalized Hough Transform Mohamed Amine LARHMAM, Saïd MAHMOUDI and Mohammed BENJELLOUN Faculty of Engineering, University of Mons,
More informationSampling-Based Ensemble Segmentation against Inter-operator Variability
Sampling-Based Ensemble Segmentation against Inter-operator Variability Jing Huo 1, Kazunori Okada, Whitney Pope 1, Matthew Brown 1 1 Center for Computer vision and Imaging Biomarkers, Department of Radiological
More informationSemi-Automatic Segmentation of the Patellar Cartilage in MRI
Semi-Automatic Segmentation of the Patellar Cartilage in MRI Lorenz König 1, Martin Groher 1, Andreas Keil 1, Christian Glaser 2, Maximilian Reiser 2, Nassir Navab 1 1 Chair for Computer Aided Medical
More informationSegmentation of Neck Lymph Nodes in CT Datasets with Stable 3D Mass-Spring Models
Segmentation of Neck Lymph Nodes in CT Datasets with Stable 3D Mass-Spring Models Jana Dornheim 1, Heiko Seim 1, Bernhard Preim 1, Ilka Hertel 2, and Gero Strauss 2 1 Otto-von-Guericke-Universität Magdeburg,
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 informationAutomatic Generation of Shape Models Using Nonrigid Registration with a Single Segmented Template Mesh
Automatic Generation of Shape Models Using Nonrigid Registration with a Single Segmented Template Mesh Geremy Heitz, Torsten Rohlfing, and Calvin R. Maurer, Jr. Image Guidance Laboratories Department of
More informationImage Segmentation. Shengnan Wang
Image Segmentation Shengnan Wang shengnan@cs.wisc.edu Contents I. Introduction to Segmentation II. Mean Shift Theory 1. What is Mean Shift? 2. Density Estimation Methods 3. Deriving the Mean Shift 4. Mean
More informationVolume rendering for interactive 3-d segmentation
Volume rendering for interactive 3-d segmentation Klaus D. Toennies a, Claus Derz b a Dept. Neuroradiology, Inst. Diagn. Radiology, Inselspital Bern, CH-3010 Berne, Switzerland b FG Computer Graphics,
More informationBinary Shape Characterization using Morphological Boundary Class Distribution Functions
Binary Shape Characterization using Morphological Boundary Class Distribution Functions Marcin Iwanowski Institute of Control and Industrial Electronics, Warsaw University of Technology, ul.koszykowa 75,
More informationSemi-Automatic Topology Independent Contour- Based 2 ½ D Segmentation Using Live-Wire
Semi-Automatic Topology Independent Contour- Based 2 ½ D Segmentation Using Live-Wire Michael Knapp Vienna University of Technology Computer Graphics Group Favoritenstrasse 9-11/E186 1040 Wien, Austria
More informationObject Identification in Ultrasound Scans
Object Identification in Ultrasound Scans Wits University Dec 05, 2012 Roadmap Introduction to the problem Motivation Related Work Our approach Expected Results Introduction Nowadays, imaging devices like
More informationDecomposing and Sketching 3D Objects by Curve Skeleton Processing
Decomposing and Sketching 3D Objects by Curve Skeleton Processing Luca Serino, Carlo Arcelli, and Gabriella Sanniti di Baja Institute of Cybernetics E. Caianiello, CNR, Naples, Italy {l.serino,c.arcelli,g.sannitidibaja}@cib.na.cnr.it
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 informationHepaTux A Semiautomatic Liver Segmentation System
HepaTux A Semiautomatic Liver Segmentation System Andreas Beck and Volker Aurich Institut für Informatik, Heinrich-Heine-Universität Düsseldorf, D-40225 Düsseldorf becka-miccai@acs.uni-duesseldorf.de Abstract.
More informationAutomatic Segmentation of Parotids from CT Scans Using Multiple Atlases
Automatic Segmentation of Parotids from CT Scans Using Multiple Atlases Jinzhong Yang, Yongbin Zhang, Lifei Zhang, and Lei Dong Department of Radiation Physics, University of Texas MD Anderson Cancer Center
More informationAssessing the Local Credibility of a Medical Image Segmentation
Assessing the Local Credibility of a Medical Image Segmentation Joshua H. Levy, Robert E. Broadhurst, Surajit Ray, Edward L. Chaney, and Stephen M. Pizer Medical Image Display and Analysis Group (MIDAG),
More informationIterative CT Reconstruction Using Curvelet-Based Regularization
Iterative CT Reconstruction Using Curvelet-Based Regularization Haibo Wu 1,2, Andreas Maier 1, Joachim Hornegger 1,2 1 Pattern Recognition Lab (LME), Department of Computer Science, 2 Graduate School in
More informationAdaptive Fuzzy Connectedness-Based Medical Image Segmentation
Adaptive Fuzzy Connectedness-Based Medical Image Segmentation Amol Pednekar Ioannis A. Kakadiaris Uday Kurkure Visual Computing Lab, Dept. of Computer Science, Univ. of Houston, Houston, TX, USA apedneka@bayou.uh.edu
More informationMR IMAGE SEGMENTATION
MR IMAGE SEGMENTATION Prepared by : Monil Shah What is Segmentation? Partitioning a region or regions of interest in images such that each region corresponds to one or more anatomic structures Classification
More informationQuantitative Measurement of Kidney and Cyst Sizes in Patients with Autosomal Dominant Polycystic Kidney Disease (ADPKD)
Quantitative Measurement of Kidney and Cyst Sizes in Patients with Autosomal Dominant Polycystic Kidney Disease (ADPKD) V Daum 1, H Helbig 1, R Janka 3,K-U Eckardt 2 and R Zeltner 2 1 Friedrich Alexander
More informationDevelopment of 3D Model-based Morphometric Method for Assessment of Human Weight-bearing Joint. Taeho Kim
Development of 3D Model-based Morphometric Method for Assessment of Human Weight-bearing Joint Taeho Kim Introduction Clinical measurement in the foot pathology requires accurate and robust measurement
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 informationConstruction of Left Ventricle 3D Shape Atlas from Cardiac MRI
Construction of Left Ventricle 3D Shape Atlas from Cardiac MRI Shaoting Zhang 1, Mustafa Uzunbas 1, Zhennan Yan 1, Mingchen Gao 1, Junzhou Huang 1, Dimitris N. Metaxas 1, and Leon Axel 2 1 Rutgers, the
More informationSemantic Context Forests for Learning- Based Knee Cartilage Segmentation in 3D MR Images
Semantic Context Forests for Learning- Based Knee Cartilage Segmentation in 3D MR Images MICCAI 2013: Workshop on Medical Computer Vision Authors: Quan Wang, Dijia Wu, Le Lu, Meizhu Liu, Kim L. Boyer,
More informationMedicale Image Analysis
Medicale Image Analysis Registration Validation Prof. Dr. Philippe Cattin MIAC, University of Basel Prof. Dr. Philippe Cattin: Registration Validation Contents 1 Validation 1.1 Validation of Registration
More informationCalculating the Distance Map for Binary Sampled Data
MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.merl.com Calculating the Distance Map for Binary Sampled Data Sarah F. Frisken Gibson TR99-6 December 999 Abstract High quality rendering and physics-based
More informationInteractive segmentation of vascular structures in CT images for liver surgery planning
Interactive segmentation of vascular structures in CT images for liver surgery planning L. Wang¹, C. Hansen¹, S.Zidowitz¹, H. K. Hahn¹ ¹ Fraunhofer MEVIS, Institute for Medical Image Computing, Bremen,
More informationDepth-Layer-Based Patient Motion Compensation for the Overlay of 3D Volumes onto X-Ray Sequences
Depth-Layer-Based Patient Motion Compensation for the Overlay of 3D Volumes onto X-Ray Sequences Jian Wang 1,2, Anja Borsdorf 2, Joachim Hornegger 1,3 1 Pattern Recognition Lab, Friedrich-Alexander-Universität
More informationUsing Probability Maps for Multi organ Automatic Segmentation
Using Probability Maps for Multi organ Automatic Segmentation Ranveer Joyseeree 1,2, Óscar Jiménez del Toro1, and Henning Müller 1,3 1 University of Applied Sciences Western Switzerland (HES SO), Sierre,
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 informationHuman Heart Coronary Arteries Segmentation
Human Heart Coronary Arteries Segmentation Qian Huang Wright State University, Computer Science Department Abstract The volume information extracted from computed tomography angiogram (CTA) datasets makes
More informationMachine Learning for Medical Image Analysis. A. Criminisi
Machine Learning for Medical Image Analysis A. Criminisi Overview Introduction to machine learning Decision forests Applications in medical image analysis Anatomy localization in CT Scans Spine Detection
More informationAn Automated Image-based Method for Multi-Leaf Collimator Positioning Verification in Intensity Modulated Radiation Therapy
An Automated Image-based Method for Multi-Leaf Collimator Positioning Verification in Intensity Modulated Radiation Therapy Chenyang Xu 1, Siemens Corporate Research, Inc., Princeton, NJ, USA Xiaolei Huang,
More informationFast 3D Mean Shift Filter for CT Images
Fast 3D Mean Shift Filter for CT Images Gustavo Fernández Domínguez, Horst Bischof, and Reinhard Beichel Institute for Computer Graphics and Vision, Graz University of Technology Inffeldgasse 16/2, A-8010,
More informationAN essential part of any computer-aided surgery is planning
1 A Model Based Validation Scheme for Organ Segmentation in CT Scan Volumes Hossein Badakhshannoory, Student Member, IEEE, and Parvaneh Saeedi, Member, IEEE Abstract In this work, we propose a novel approach
More informationAutomatic Parameter Optimization for De-noising MR Data
Automatic Parameter Optimization for De-noising MR Data Joaquín Castellanos 1, Karl Rohr 2, Thomas Tolxdorff 3, and Gudrun Wagenknecht 1 1 Central Institute for Electronics, Research Center Jülich, Germany
More informationMedical Image Segmentation
Medical Image Segmentation Xin Yang, HUST *Collaborated with UCLA Medical School and UCSB Segmentation to Contouring ROI Aorta & Kidney 3D Brain MR Image 3D Abdominal CT Image Liver & Spleen Caudate Nucleus
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 informationInteractive Differential Segmentation of the Prostate using Graph-Cuts with a Feature Detector-based Boundary Term
MOSCHIDIS, GRAHAM: GRAPH-CUTS WITH FEATURE DETECTORS 1 Interactive Differential Segmentation of the Prostate using Graph-Cuts with a Feature Detector-based Boundary Term Emmanouil Moschidis emmanouil.moschidis@postgrad.manchester.ac.uk
More informationKidney Segmentation in Ultrasound Images Using Curvelet Transform and Shape Prior
013 International Conference on Communication Systems and Network Technologies Kidney Segmentation in Ultrasound Images Using Curvelet Transform and Shape Prior Ehsan Jokar 1, Hossein Pourghassem Department
More informationAutomatic Detection and Segmentation of Kidneys in Magnetic Resonance Images Using Image Processing Techniques
Biomedical Statistics and Informatics 2017; 2(1): 22-26 http://www.sciencepublishinggroup.com/j/bsi doi: 10.11648/j.bsi.20170201.15 Automatic Detection and Segmentation of Kidneys in Magnetic Resonance
More informationShape-Based Kidney Detection and Segmentation in Three-Dimensional Abdominal Ultrasound Images
University of Toronto Shape-Based Kidney Detection and Segmentation in Three-Dimensional Abdominal Ultrasound Images Authors: M. Marsousi, K. N. Plataniotis, S. Stergiopoulos Presenter: M. Marsousi, M.
More informationTopic 4 Image Segmentation
Topic 4 Image Segmentation What is Segmentation? Why? Segmentation important contributing factor to the success of an automated image analysis process What is Image Analysis: Processing images to derive
More informationInteractive 3D Heart Chamber Partitioning with a New Marker-Controlled Watershed Algorithm
Interactive 3D Heart Chamber Partitioning with a New Marker-Controlled Watershed Algorithm Xinwei Xue School of Computing, University of Utah xwxue@cs.utah.edu Abstract. Watershed transform has been widely
More information3D VISUALIZATION OF SEGMENTED CRUCIATE LIGAMENTS 1. INTRODUCTION
JOURNAL OF MEDICAL INFORMATICS & TECHNOLOGIES Vol. 10/006, ISSN 164-6037 Paweł BADURA * cruciate ligament, segmentation, fuzzy connectedness,3d visualization 3D VISUALIZATION OF SEGMENTED CRUCIATE LIGAMENTS
More informationIntroduction to Medical Imaging (5XSA0) Module 5
Introduction to Medical Imaging (5XSA0) Module 5 Segmentation Jungong Han, Dirk Farin, Sveta Zinger ( s.zinger@tue.nl ) 1 Outline Introduction Color Segmentation region-growing region-merging watershed
More informationHybrid Spline-based Multimodal Registration using a Local Measure for Mutual Information
Hybrid Spline-based Multimodal Registration using a Local Measure for Mutual Information Andreas Biesdorf 1, Stefan Wörz 1, Hans-Jürgen Kaiser 2, Karl Rohr 1 1 University of Heidelberg, BIOQUANT, IPMB,
More informationTexture-Based Detection of Myositis in Ultrasonographies
Texture-Based Detection of Myositis in Ultrasonographies Tim König 1, Marko Rak 1, Johannes Steffen 1, Grit Neumann 2, Ludwig von Rohden 2, Klaus D. Tönnies 1 1 Institut für Simulation & Graphik, Otto-von-Guericke-Universität
More informationUser-Defined B-Spline Template-Snakes
User-Defined B-Spline Template-Snakes Tim McInerney 1,2 and Hoda Dehmeshki 1 1 Dept. of Math, Physics and Computer Science, Ryerson Univ., Toronto, ON M5B 2K3, Canada 2 Dept. of Computer Science, Univ.
More informationSkull Segmentation of MR images based on texture features for attenuation correction in PET/MR
Skull Segmentation of MR images based on texture features for attenuation correction in PET/MR CHAIBI HASSEN, NOURINE RACHID ITIO Laboratory, Oran University Algeriachaibih@yahoo.fr, nourine@yahoo.com
More informationContours & Implicit Modelling 4
Brief Recap Contouring & Implicit Modelling Contouring Implicit Functions Visualisation Lecture 8 lecture 6 Marching Cubes lecture 3 visualisation of a Quadric toby.breckon@ed.ac.uk Computer Vision Lab.
More informationComparison of Vessel Segmentations using STAPLE
Comparison of Vessel Segmentations using STAPLE Julien Jomier, Vincent LeDigarcher, and Stephen R. Aylward Computer-Aided Diagnosis and Display Lab The University of North Carolina at Chapel Hill, Department
More informationVOLUMETRIC HARMONIC MAP
COMMUNICATIONS IN INFORMATION AND SYSTEMS c 2004 International Press Vol. 3, No. 3, pp. 191-202, March 2004 004 VOLUMETRIC HARMONIC MAP YALIN WANG, XIANFENG GU, AND SHING-TUNG YAU Abstract. We develop
More informationA Method of Automated Landmark Generation for Automated 3D PDM Construction
A Method of Automated Landmark Generation for Automated 3D PDM Construction A. D. Brett and C. J. Taylor Department of Medical Biophysics University of Manchester Manchester M13 9PT, Uk adb@sv1.smb.man.ac.uk
More informationAutomatic Model-Based Segmentation of Medical Images
Automatic Model-Based Segmentation of Medical Images Cristian Lorenz Jochen Peters, Fabian Wenzel, Jürgen Weese May 26 th 2014 Need Medical imaging systems produce a huge amount of patient images with
More informationHistogram and watershed based segmentation of color images
Histogram and watershed based segmentation of color images O. Lezoray H. Cardot LUSAC EA 2607 IUT Saint-Lô, 120 rue de l'exode, 50000 Saint-Lô, FRANCE Abstract A novel method for color image segmentation
More informationFast Interactive Region of Interest Selection for Volume Visualization
Fast Interactive Region of Interest Selection for Volume Visualization Dominik Sibbing and Leif Kobbelt Lehrstuhl für Informatik 8, RWTH Aachen, 20 Aachen Email: {sibbing,kobbelt}@informatik.rwth-aachen.de
More informationOpen Topology: A Toolkit for Brain Isosurface Correction
Open Topology: A Toolkit for Brain Isosurface Correction Sylvain Jaume 1, Patrice Rondao 2, and Benoît Macq 2 1 National Institute of Research in Computer Science and Control, INRIA, France, sylvain@mit.edu,
More informationVertebrae Segmentation in 3D CT Images based on a Variational Framework
Vertebrae Segmentation in 3D CT Images based on a Variational Framework Kerstin Hammernik, Thomas Ebner, Darko Stern, Martin Urschler, and Thomas Pock Abstract Automatic segmentation of 3D vertebrae is
More informationarxiv: v1 [cs.cv] 6 Jun 2017
Volume Calculation of CT lung Lesions based on Halton Low-discrepancy Sequences Liansheng Wang a, Shusheng Li a, and Shuo Li b a Department of Computer Science, Xiamen University, Xiamen, China b Dept.
More informationExtraction and recognition of the thoracic organs based on 3D CT images and its application
1 Extraction and recognition of the thoracic organs based on 3D CT images and its application Xiangrong Zhou, PhD a, Takeshi Hara, PhD b, Hiroshi Fujita, PhD b, Yoshihiro Ida, RT c, Kazuhiro Katada, MD
More informationAvailable Online through
Available Online through www.ijptonline.com ISSN: 0975-766X CODEN: IJPTFI Research Article ANALYSIS OF CT LIVER IMAGES FOR TUMOUR DIAGNOSIS BASED ON CLUSTERING TECHNIQUE AND TEXTURE FEATURES M.Krithika
More informationRigid and Deformable Vasculature-to-Image Registration : a Hierarchical Approach
Rigid and Deformable Vasculature-to-Image Registration : a Hierarchical Approach Julien Jomier and Stephen R. Aylward Computer-Aided Diagnosis and Display Lab The University of North Carolina at Chapel
More informationMorphological track 1
Morphological track 1 Shapes Painting of living beings on cave walls at Lascaux [about 1500 th BC] L homme qui marche by Alberto Giacometti, 1948, NOUVELLES IMAGES Editor (1976) Les lutteurs by Honoré
More informationA Non-Linear Image Registration Scheme for Real-Time Liver Ultrasound Tracking using Normalized Gradient Fields
A Non-Linear Image Registration Scheme for Real-Time Liver Ultrasound Tracking using Normalized Gradient Fields Lars König, Till Kipshagen and Jan Rühaak Fraunhofer MEVIS Project Group Image Registration,
More informationAutomated Model-Based Rib Cage Segmentation and Labeling in CT Images
Automated Model-Based Rib Cage Segmentation and Labeling in CT Images Tobias Klinder 1,2,CristianLorenz 2,JensvonBerg 2, Sebastian P.M. Dries 2, Thomas Bülow 2,andJörn Ostermann 1 1 Institut für Informationsverarbeitung,
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 informationMethodological progress in image registration for ventilation estimation, segmentation propagation and multi-modal fusion
Methodological progress in image registration for ventilation estimation, segmentation propagation and multi-modal fusion Mattias P. Heinrich Julia A. Schnabel, Mark Jenkinson, Sir Michael Brady 2 Clinical
More informationdoi: /
Yiting Xie ; Anthony P. Reeves; Single 3D cell segmentation from optical CT microscope images. Proc. SPIE 934, Medical Imaging 214: Image Processing, 9343B (March 21, 214); doi:1.1117/12.243852. (214)
More informationMedical Image Segmentation Using Descriptive Image Features
YANG, YUAN, LI, YAN: MEDICAL IMAGE SEGMENTATION 1 Medical Image Segmentation Using Descriptive Image Features Meijuan Yang meijuan.yang@opt.ac.cn Yuan Yuan yuany@opt.ac.cn Xuelong Li xuelong_li@opt.ac.cn
More information