Global optimization of weighted mutual information for multimodality image registration

Size: px
Start display at page:

Download "Global optimization of weighted mutual information for multimodality image registration"

Transcription

1 Global optimization of weighted mutual information for multimodality image registration C. E. RodrIguez-Carranza and M. H. Loew Department of Electrical Engineering and Computer Science Institute for Medical Imaging and Image Analysis The George Washington University, Washington DC ABSTRACT Failure to align images accurately often is due to the optimization algorithm being trapped in local maxima or spurious global maxima of the mutual information function. Strategies contemplated to improve registration involve modifying the optimization scheme or the registration measure itself. We recently found that normalized mutual information (for 2D image registration) provides a larger capture range and that is more robust, with respect to the optimization parameters, than the non-normalized measure. In this paper we assessed the utility of a stochastic global optimization technique for image registration using normalized and non-normalized mutual information. By conducting large-scale studies with patient data in 2D, we established a success rate baseline with the local optimizer only. Formal proof has not yet been found that incorporating the global optimizer does not impair performance. However, experiments to date indicate that its inclusion leads to better (i.e., higher probability of correct convergence) overall performance. More over, studies now underway show good effectiveness of our approach in a variety of 3D cases. Keywords: multimodality image registration, mutual information, topographical global optimization 1. INTRODUCTION Mutual information was first introduced as a measure for image alignment by Collignon, et al.' and Viola, et al.2 in Since then it has become a popular method for registration of mono- and multimodality images because of its robustness and accuracy. Reports of alignment studies using mutual information have come from a variety of sites.36 Yet there are still situations in which algorithms employing mutual information do not align images correctly. Failure to align images is due to the optimization algorithm being trapped in local maxima or spurious global maxima of the mutual information function. Strategies to improve registration can involve modifying the optimization scheme or the registration measure itself. In an earlier attempt to improve the behavior of mutual information, we introduced the use of a deterministic and weighted entropy measure.7 It was tested in 2D images (one modality) and results obtained indicate that normalized (or weighted) mutual information (using either Shannon's entropy or the deterministic entropy) provides a larger capture range and is more robust, with respect to the optimization parameters, than the non-normalized mutual information. For this paper, we look into the optimization part of the registration algorithm. A variety of optimization schemes have been applied to image registration using similarity measures: Powell's,"3 simplex,4'8 gradient descent, multiresolution exhaustive searches,9 stochastic optimization using derivative estimation,2 genetic algorithms,8 etc. In most cases the optimization strategy is local and requires a starting point (in other words, a starting transformation vector); the typical choice for it is the zero vector. Sometimes, the alternative is to select multiple starting points, and a decision has to be made of how to select such points. Some techniques, such as topographical global optimization, allow high-quality sampling of a large number of points. We wanted to determine whether it is possible that the incorporation of this global optimization scheme in image registration can refine the selection of the starting point and hence increase the probability of correct alignment. Therefore, in this paper we explore the use of topographical global optimization for 2D and 3D multimodality image registration using normalized mutual information (computed with either Shannon's entropy or the deterministic entropy). Other author information: (Send correspondence to C.ER.C.) M.H.L.: loew seas.gwu.edu C.E.R.C.: claudia seas.gwu.edu Part of the SPIE Conference on Image Processing San Diego, California February 1999 SPIE Vol X/99/$

2 2.1. Mutual information 2. BACKGROUND Mutual information has its roots in information theory. We refer the interested reader elsewhere' '11 for an in-depth cover of the subject. The basic equation for mutual information is the following: I(a,i3) = = H(a) H(aJ$) where H(a) refers to the marginal entropy or uncertainty of a, H(a, 3) refers to the joint entropy of a and j3, and H(a/13) is the conditional entropy. Entropy refers to the average information of an event. Mutual information indicates to what extent the realization of /3 lowers the uncertainty of a. The expression for the Shannon's entropy of event a with j outcomes A, each with probability of occurrence p(a), is defined as: H(a) = H(A1,..., A) = p(a)( log p(a)) Tn the context of image registration, mutual information has been used as a measure of alignment; the assumption is that mutual information between two images is maximized when both are aligned. The entropy of an image is computed based on the intensity distribution; this can be estimated either by a simple normalization of the histogram or by Parzen windowing. A totally homogeneous image will have zero entropy; an image with a different gray value at each pixel will have maximum entropy. Recently7 we proposed the use of a deterministic entropy measure to compute mutual information. The expression of the two-dimensional discrete case is: : 2f(x,y) 2x. I 1 L2f(x,y) H(f(,.); ) = (1) 2f(x,y) 2x For details in the derivation of this expression we refer the reader to the earlier paper Topographical global optimization The purpose of global optimization is to find the smallest of a number of local minima that a multimodal function I attains in the region of interest A C Stochastic algorithms for global optimization'2 consist of three distinct major steps: a sampling step, an optimization step, and a check of some stopping criterion. The sampling step consists of the generation of random points in the region of interest A, and the computation of the associated function value. As the number of sampled points approaches inffnity, the probability of finding the global minimum approaches one. The drawback, of course, is that the more trial points we use the longer the running time of the algorithm. This is something that should be avoided especially if the objective function is costly to evaluate. The optimization step consists of applying a local optimization routine to some of the sampled points. We can start a local minimization from each sampled point, but this simple strategy is highly inefficient. For example, a local optimizer will normally arrive at the same local solution several times. A more efficient strategy is to try to group (cluster) points belonging to basins of local minima and then start a local minimization from just one point in each identified basin. There are several stopping criteriathat can be used to stop the algorithm : (1) when there's sufficient evidence that the global optimum has been detected, (2) when the cost connected with searches for a better estimate has been exceeded, (3) when the number of prespecified number of steps or function evaluations has been reached. 90

3 In this paper we focus in the stochastic algorithm proposed by Torn and Vittanen'3 called topographical global optimization. In their algorithm, points are uniformly sampled (a good random number generator is vital to guarantee a uniform distribution of points in multiple dimensions). Ideally, sampling should be such that at least one of the sampled points lies in each basin of the global minima. The way TGO identifies clusters of points belonging to the same basin is through a directed graph of the topographical information of the objective function; this graph is called k-topoyraph. Each node in the graph represents a sampled point; the graph connects neighboring points with directed arcs pointing towards points of higher function values. The Euclidean distance is used to establish which are the closest neighbors. Figure 1 shows a very simple example. The function is one-dimensional and consecutive k-topographs are shown; thin arcs represent arcs from graphs with smaller k. Observe that the larger k is, the further away the neighbors are. For a k-topograph, nodes with no incoming arcs represent local minima. f(x) /R\\//1 ;1u' RE c3 1-topograph 2-topograph 3-topograph 0 = graph minimum = sampled point Figure 1. Example of a one-dimensional function and three k-topographs. When the sampling is terminated, the minima of the graph are presented and a local minimizer (the choice is not a central question) is started from a number of them. Minimization is started from, say, the three best minima in the graph and the corresponding minima found by the local minimization algorithm are presented; the smallest of the solutions corresponds to the global minima. The user has then the option to start additional local minimizations from other graph minima. The algorithm is simple but it has some disadvantages. The underlying assumption is that a lower function value implies a local minimum, which is not always the case. The algorithm will not perform too well if the problem function is very ragged within small areas. Large search areas exaggerate the effect of even small raggedness and renders the algorithm useless for problems of this nature unless the number of points, N, is chosen very large. Most other algorithms, however, will also have difficulties in finding the global solution for such problems Iterative topographical global optimization Torn and Viitanen'4 recently proposed an iterative version of topographical global optimization (I-TGO) which allows sampling of a large number of points in an efficient way. There are two versions of the algorithm, and we present one of them here. Both require constant storage complexity and linear computational complexity in the number of total points used (N). 91

4 In the I-TGO algorithm, points are iteratively sampled and minima are identified through the k-topograph. The parameters of the algorithm are (1) number of desired iterations, I, (2) number of points to sample per iteration, N1, and (3) the number of nearest neighbors k to determine the graph minima. A sketch of the I-TGO1 algorithm is shown next: 1. mo = 0, j = 1, and the k topograph is empty. 2. While the number of iterations j is less than I (a) sample N1 m11 points (b) compute Euclidean distances, form the k-topogragh, and identify the graph minima (m3) (c) eliminate all points except for the m3 graph minima 3. For each point in the graph (i.e. graph minima), run the local optimization algorithm. 4. Select the smallest of the solutions to be the global minima. As can be seen, the only information carried from one iteration to the next is the set of minima in the graph. The authors recommend that a relatively small k be used to avoid losing information. Only Nj(Nj 1)/2 distances need to be stored at any time. The number of distance calculations is I x N1(Nj 1)/2. Since the number of points sampled in the.7th iteration is N1 m_1, (where m3 is the number of minima in the th iteration), the total number of sampled points, N, can be written as: N = Nj m1_1 = IxNj m_1 = IxN1 Ixii where m: s the mean value of the m3's. Therefore, 1= N N1 ii Using this value of I in the expression for the number of distance calculations, we obtain implies linearity in N. x Ni( which 3. METHODS For our experiments, we assumed the following context. Given a pair of misaligned images, the idea is to present to a human observer the best set of starting points so that the images can be aligned with the smallest number of attempts. The observer picks one starting point, runs the registration algorithm, and then judges whether the registration is good enough or not; if it is not, then she/he repeats the sequence of selection and evaluation until satisfied with the alignment. We wanted to determine whether I-TGO could improve registration performance by providing a better selection of starting points; in other words, whether I-TGO could increase the probability of finding the global maxima of mutual information with fewer points. With this in mind, we compared the performance of registration with and without I-TGO by computing the probability that we need to sample j "bad" starting points (points that lead to a failed registration) before we 92

5 sample the first "good" starting point (points that lead to a successful registration). For registration without I-TGO, the baseline experiment consisted of running N registrations, each with a different starting point. The probability p of selecting a "good" starting point is the proportion of successful registrations out of the total N. With p we calculated the probability of selecting j "bad" sample points before the first "good" one as: P(F=j)=(1 _p)jp (2) For registration with I-TGO, we run M registration experiments. Each of them consisted of: (1) sampling points with I-TGO; the output of 1-TGO is the graph maxima, which constitute the list of starting points for registration, and (2) applying the local optimizer to each of the graph maxima. For each of this M lists of graph maxima, we observed the rank (or position in the list) of the first "good" maxima. With this information we computed the probability of j "bad" samples before the first "good" one as: P(F = j) = {number of "good" maxima with rank j + 1}/M (3) For example, the probability of no "bad" maxima corresponds to the proportion of "good" maxima with rank 1; the probability of one "bad" maxima corresponds to the proportion of "good" maxima with rank 2, etc. For all the experiments, the range of the starting points (ti, ti,, 0) was ([ 60mm, 60mm], [ 60mm, 6Omm}, [ 75, 75 ]). The number of registrations for the experiment without I-TGO was N= The parameters required for I-TGO are (a) number of iterations I, (b) number of samples per iteration N1, and (c) the extent of the topograph, k. For reasons explained next and based on results from the baseline experiment without T-TGO, we selected 1=4, N1=50, and k=3. The number of registrations with I-TGO was M = 100. The cost of incorporating I-TGO into image registration was measured in terms of number of function evaluations. The total cost of registration with I-TGO consists of two parts: (1) the cost of the sampling stage of I-TGO, and (2) the cost of running the local optimizer for each of the final graph maxima. For our experiments, we established that the cost of the sampling stage should not be more than the cost of running the local optimizer on one point. The latter was determined from the baseline experiment without I-TGO. The image dataset we used consisted of two slices of a registered pair of MR and CT images; they were reformatted to a size of 256x256. We applied a shift of 15mmin the y direction, +20mm in the x direction and +20 degrees to the MR image. This transformation vector becomes the reference transformation and we used it to establish registration accuracy of the experiments. Accuracy was calculated as the difference between the reference transformation vector and the computed transformation vector. A registration result was considered successful if the difference between the reference transformation and the computed transforniation was no more than 1mm for translations and 1 for the rotation. The local optimizer used in all experiments was Powell's direction set method. The parameters for this algorithm were the same for all experiments. The transformations were assumed to be rigid body. We used trilinear interpo lation, and only pixels that lie within the volume of overlap were retained; pixels that lied outside were assigned a value of zero. The number of histogram bins used to compute mutual information via Shannon's entropy was 256. For the case of normalized mutual information, we multiplied mutual information (using either Shannon's entropy measure or the deterministic entropy measure) by the ratio of the area of of overlap to the area of the smallest image. 4. RESULTS The results of the baseline experiment of registration without I-TGO are shown in Table 1. The mean number of function evaluations during optimization with Powell's algorithm is 243 and 301. As noted before, the cost of sampling in I-TGO should not exceed 301. Hence, by selecting the number of iterations to be 4 and the number of sampled points per iteration to be 50, we restricted the number of function evaluations in the sampling stage of I-TGO to be at most 4 x 50 =

6 Measure No. of successful starting points Function evaluations (mean stdfl from Powell's iterations Weighted Mldet Weighted Mlshan Table 1. Baseline results for registration without I-TGO Table 2. Registration with I-TGO In Table 2 we show the ranks of the "good" maxima obtained for the registration with I-TGO experiment. The average number of function evaluations during ITGO was 180 for mutual information using the deterministic entropy measure, and 179 for mutual information using Shannon's entropy measure. For registration without I-TGO, using Table 1 we calculate the probability of selecting a "good" starting point to be p = 0.35 for the deterministic entropy measure and p = for Shannon's entropy measure (the total number of registration experiments was 1000). With these values, we compute the probabilities of sampling j"bad" points before the first "good" point using Equation 2. The results are shown in rows 3 and 5 of Table 3. For the case of registration with I-TGO, we use Table 2 to compute the probabilities from Equation 3 (the total number of registration experiments was 100). The results are shown in rows 2 and 4 of Table 3. We look at up to a maximum of 5 samples; the last column corresponds to the probability of at least five consecutive failures. Measure Registration strategy p(f = 0) p(f = 1) p(f = 2) p(f = 3) p(f = 4) p(f > 4) weighted Midet with 1-TGO without I-TGO weighted Mlsha with I-TGO without I-TGO Table 3. Comparison of P(F = j) probabilities for registration with and without I-TGO. Table 3 indicates that, for mutual information with the deterministic entropy measure, the probability of a successful registration with one point is higher for the strategy with I-TGO than without I-TGO: 0.51 vs 0.35; the same is true for two points: 0.67 vs (adding the columns p(f = 0) and p(f = 1)). When we look at more than three points, the difference between the probabilities is minimal. The effect of I-TGO is more dramatic when using Shannon's entropy. The probability of selecting a "good" starting point at the first sample is 0.70 with I-TGO and without 1-TGO. The same holds if selecting two points (0.86 vs 0.374), three points (0.90 vs 0.505), etc. 5. DISCUSSION Topographical global optimization is a stochastic optimization scheme that provides a high-quality selection of points among which to search for a global optimum. The current results show that I-TGO increases the probability of finding the global maximum of mutual information with fewer points. This can be observed in Table 3, where normalized mutual information (using either entropy measure) with I-TGO shows higher probability of selecting a "good" starting point with less samples than when using the normalized mutual information without I-TGO. 94

7 The effect of I-TGO on the performance of registration is more important for mutual information based on Shannon's entropy measure than for mutual information based on the deterministic entropy measure. Taking as example the case of successful registration with only one sample, we observe that the improvement for the deterministicentropy-based measure with I-TGO is 1.45-fold; for the Shannon's entropy-based measure, it is 3.35-fold. Overall, Shannon's entropy based normalized mutual information with I-TGO shows the best performance of the four possibilities; it is followed by the deterministic-entropy-based measure with I-TGO and without I-TGO, and finally the Shannon's-entropy-based measure without I-TGO. For the case of sampling only one point, the probabilities are respectively, (0.7,0.51,0.35,0.209); when sampling two points the probabilities are (0.86,0.67,0.578,0.374). The cost of incorporating I-TGO into image registration was based on the number of function evaluations. We established that the maximum cost to pay for the sampling stage of I-ITGO was the cost of running the local optimizer on one point. For our experiments, the mean cost of sampling in I-TGO was 180 function evaluations, and the mean number of function evaluations for running the local optimizer were 243 and 301, respectively, for mutual information based on the deterministic entropy and Shannon's entropy. Hence, even though there is a (small) cost for using I-TGO, we believe that the observed improvement in the registration performance well justifies its use. Nevertheless, we need to perform further experiments to determine if the improvement in the performance is significant and if it can be generalized. We are currently testing normalized mutual information and the 1-TGO strategy for 3D image registration. Experiments now underway are showing good effectiveness of the I-TGO approach in a variety of 3D cases. ACKNOWLEDGMENTS C.E.R.C. thanks Josien Pluim for her kind help and fruitful discussions. REFERENCES 1. A. Collignon, F. Maes, D. Delaere, D. Vandermeulen, P. Suetens, and G. Marchal, "Automated multimodality image registration using information theory," Proceedings Information Processing in Medical Imaging (IPMI'95), (Dordrecht, The Netherlands), P. Viola and W. M. Wells, "Alignment by maximization of mutual information," Proceedings of the 5th International Conference on Computer Vision, pp , F. Maes, A. Collignon, D. Vandermeulen, G. Marchal, and P. Suetens, "Multimodality image registration by maximization of mutual information," IEEE Transactions on Medical Imaging 16(2), pp , C. R. Meyer, J. L. Boes, B. Kim, P. H. Bland, K. Zasadny, P. V. Kison, K. Koral, K. Prey, and R. L. Wahl, "Demonstration of accuracy and clinical versatility of mutual information for automatic multimodality image fusion using affine and thin plate spline warped geometric deformations," Medical Image Analysis 1(3), C. Studholme, D. L. G. Hill, and D. J. Hawkes, "Automated 3-d registration of mr and ct images of the head," Medical Image Analysis 1(2), J. West, M. Fitzpatrick, M. Wang, B. Dawant, C. Maurer, R. Kessler, R. Maciunas, C. Barillot, D. Lemoine, A. Collignon, F. Maes, P. Suetens, D. Vandermeulen, P. van den Elsen, S. Napel, T. Sumanaweera, B. Harkness, P. Hemler, D. Hill, D. Hawkes, C. Studholme, J. Maintz, M. Viergever, G. Malandain, X. Pennec, M. Noz, G. Maguire, M. Pollack, C. Pelizzari, R. Robb, D. Hanson, and It. Woods, "Comparison and evaluation of retrospective intermodality brain image registration techniques," Journal of Computer Assisted Tomography 21(4), pp , C. RodrIguez-Carranza and M. Loew, "Weighted and deterministic entropy measure for image registration using mutual information," in Image Processing, Proceedings SPIE 3338, pp , D. Hill, D. Hawkes, N. A. Harrison, and C. F. Ruff, "A strategy for automated multimodality image registration incorporating anatomical knowledge and imager characteristics," Proceedings Information Processing in Medical Imaging (IPMI'93), pp , Springer-Verlag, Berlin, P. A. van den Elsen, E. J. D. Pol, T. S. Sumanaweera, P. Hemler, S. Napel, and J. Adler, "Grey value correlation techniques used for automatic matching of Ct and mr brain and spine images," in Visualization in Biomedical Computing, Proceedings SPIE 2359, pp , T. M. Cover and J. A. Thomas, Elements of information theory, John Wiley and Sons, Inc., New York, A. M. Yaglom and I. M. Yaglom, Probability and Information, D. Reidel Publishing Company, Moscow,

8 F. Schoen, "Stochastic techniques for global optimization: a survey of recent advances," Journal of Global Optimization 1, pp , A. Torn and S. Viitanen, "Topographical global optimization," in Recent advances in global optimization, C. Floudas and P. Pardalos, eds., pp , Princeton University Press, A. TOrn and S. Viitanen, "Iterative topographical global optimization," in State of the art in global optimization, C. Floudas and P. Pardalos, eds., pp , Kiuwer Academic Publisher, 1996.

A Pyramid Approach For Multimodality Image Registration Based On Mutual Information

A Pyramid Approach For Multimodality Image Registration Based On Mutual Information Pyramid pproach For Multimodality Image Registration ased On Mutual Information Hua-mei Chen Department of Electrical Engineering and Computer Science 121 Link Hall Syracuse University Syracuse, NY 13244

More information

THE geometric alignment or registration of multimodality

THE geometric alignment or registration of multimodality IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 16, NO. 2, APRIL 1997 187 Multimodality Image Registration by Maximization of Mutual Information Frederik Maes,* André Collignon, Dirk Vandermeulen, Guy Marchal,

More information

Multi-modal Image Registration Using the Generalized Survival Exponential Entropy

Multi-modal Image Registration Using the Generalized Survival Exponential Entropy Multi-modal Image Registration Using the Generalized Survival Exponential Entropy Shu Liao and Albert C.S. Chung Lo Kwee-Seong Medical Image Analysis Laboratory, Department of Computer Science and Engineering,

More information

Image Registration. Prof. Dr. Lucas Ferrari de Oliveira UFPR Informatics Department

Image Registration. Prof. Dr. Lucas Ferrari de Oliveira UFPR Informatics Department Image Registration Prof. Dr. Lucas Ferrari de Oliveira UFPR Informatics Department Introduction Visualize objects inside the human body Advances in CS methods to diagnosis, treatment planning and medical

More information

Non-Rigid Registration of Medical Images: Theory, Methods and Applications

Non-Rigid Registration of Medical Images: Theory, Methods and Applications Non-Rigid Registration of Medical Images: Theory, Methods and Applications Daniel Rueckert Paul Aljabar Medical mage registration [1] plays an increasingly important role in many clinical applications

More information

Non-Rigid Registration of Medical Images: Theory, Methods and Applications

Non-Rigid Registration of Medical Images: Theory, Methods and Applications Non-Rigid Registration of Medical Images: Theory, Methods and Applications Daniel Rueckert Paul Aljabar Medical image registration [1] plays an increasingly important role in many clinical applications

More information

Fast Image Registration via Joint Gradient Maximization: Application to Multi-Modal Data

Fast Image Registration via Joint Gradient Maximization: Application to Multi-Modal Data MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.merl.com Fast Image Registration via Joint Gradient Maximization: Application to Multi-Modal Data Xue Mei, Fatih Porikli TR-19 September Abstract We

More information

2D-3D Registration using Gradient-based MI for Image Guided Surgery Systems

2D-3D Registration using Gradient-based MI for Image Guided Surgery Systems 2D-3D Registration using Gradient-based MI for Image Guided Surgery Systems Yeny Yim 1*, Xuanyi Chen 1, Mike Wakid 1, Steve Bielamowicz 2, James Hahn 1 1 Department of Computer Science, The George Washington

More information

Annales UMCS Informatica AI 1 (2003) UMCS. Registration of CT and MRI brain images. Karol Kuczyński, Paweł Mikołajczak

Annales UMCS Informatica AI 1 (2003) UMCS. Registration of CT and MRI brain images. Karol Kuczyński, Paweł Mikołajczak Annales Informatica AI 1 (2003) 149-156 Registration of CT and MRI brain images Karol Kuczyński, Paweł Mikołajczak Annales Informatica Lublin-Polonia Sectio AI http://www.annales.umcs.lublin.pl/ Laboratory

More information

The Correlation Ratio as a New Similarity Measure for Multimodal Image Registration

The Correlation Ratio as a New Similarity Measure for Multimodal Image Registration The Correlation Ratio as a New Similarity Measure for Multimodal Image Registration Alexis Roche, Gr~goire Malandain, Xavier Pennec, and Nicholas Ayache INRIA Sophia Antipolis, EPIDAURE project, France

More information

An Introduction to Statistical Methods of Medical Image Registration

An Introduction to Statistical Methods of Medical Image Registration his is page 1 Printer: Opaque this An Introduction to Statistical Methods of Medical Image Registration Lilla Zöllei, John Fisher, William Wells ABSRAC After defining the medical image registration problem,

More information

HST.582J / 6.555J / J Biomedical Signal and Image Processing Spring 2007

HST.582J / 6.555J / J Biomedical Signal and Image Processing Spring 2007 MIT OpenCourseWare http://ocw.mit.edu HST.582J / 6.555J / 16.456J Biomedical Signal and Image Processing Spring 2007 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

More information

IRE : an image registration environment for volumetric medical images

IRE : an image registration environment for volumetric medical images IRE : an image registration environment for volumetric medical images by David Kenneth Lyle A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Computer

More information

Medical Image Segmentation Based on Mutual Information Maximization

Medical Image Segmentation Based on Mutual Information Maximization Medical Image Segmentation Based on Mutual Information Maximization J.Rigau, M.Feixas, M.Sbert, A.Bardera, and I.Boada Institut d Informatica i Aplicacions, Universitat de Girona, Spain {jaume.rigau,miquel.feixas,mateu.sbert,anton.bardera,imma.boada}@udg.es

More information

Nonrigid Image Registration Using Free-Form Deformations with a Local Rigidity Constraint

Nonrigid Image Registration Using Free-Form Deformations with a Local Rigidity Constraint Nonrigid Image Registration Using Free-Form Deformations with a Local Rigidity Constraint Dirk Loeckx, Frederik Maes, Dirk Vandermeulen, and Paul Suetens Medical Image Computing (Radiology ESAT/PSI), Faculties

More information

Hit Percentages for Red/Blue Shapes

Hit Percentages for Red/Blue Shapes Mutual Information as a Stereo Correspondence Measure Georey Egnal GRASP Laboratory University of Pennsylvania Email: gegnal@grasp.cis.upenn.edu Abstract Traditional stereo systems often falter over changes

More information

Deformable Registration Using Scale Space Keypoints

Deformable Registration Using Scale Space Keypoints Deformable Registration Using Scale Space Keypoints Mehdi Moradi a, Purang Abolmaesoumi a,b and Parvin Mousavi a a School of Computing, Queen s University, Kingston, Ontario, Canada K7L 3N6; b Department

More information

Intensity gradient based registration and fusion of multi-modal images

Intensity gradient based registration and fusion of multi-modal images Intensity gradient based registration and fusion of multi-modal images Eldad Haber 1 and Jan Modersitzki 2 1 Mathematics and Computer Science, Emory University, Atlanta, GA, USA, haber@mathcs.emory,edu

More information

Mutual Information Based CT-MR Brain Image Registration. Using Generalized Partial Volume Joint Histogram Estimation

Mutual Information Based CT-MR Brain Image Registration. Using Generalized Partial Volume Joint Histogram Estimation 1 Mutual Information Based CT-MR Brain Image Registration Using Generalized Partial Volume Joint Histogram Estimation Hua-mei Chen and Pramod K. Varshney, Fellow, IEEE Electrical Engineering and Computer

More information

Using K-means Clustering and MI for Non-rigid Registration of MRI and CT

Using K-means Clustering and MI for Non-rigid Registration of MRI and CT Using K-means Clustering and MI for Non-rigid Registration of MRI and CT Yixun Liu 1,2 and Nikos Chrisochoides 2 1 Department of Computer Science, College of William and Mary, enjoywm@cs.wm.edu 2 Department

More information

Non linear Registration of Pre and Intraoperative Volume Data Based On Piecewise Linear Transformations

Non linear Registration of Pre and Intraoperative Volume Data Based On Piecewise Linear Transformations Non linear Registration of Pre and Intraoperative Volume Data Based On Piecewise Linear Transformations C. Rezk Salama, P. Hastreiter, G. Greiner, T. Ertl University of Erlangen, Computer Graphics Group

More information

Support Vector Machine Density Estimator as a Generalized Parzen Windows Estimator for Mutual Information Based Image Registration

Support Vector Machine Density Estimator as a Generalized Parzen Windows Estimator for Mutual Information Based Image Registration Support Vector Machine Density Estimator as a Generalized Parzen Windows Estimator for Mutual Information Based Image Registration Sudhakar Chelikani 1, Kailasnath Purushothaman 1, and James S. Duncan

More information

Distance Transforms in Multi Channel MR Image Registration

Distance Transforms in Multi Channel MR Image Registration Distance Transforms in Multi Channel MR Image Registration Min Chen 1, Aaron Carass 1, John Bogovic 1, Pierre-Louis Bazin 2 and Jerry L. Prince 1 1 Image Analysis and Communications Laboratory, 2 The Laboratory

More information

ECSE 626 Project Report Multimodality Image Registration by Maximization of Mutual Information

ECSE 626 Project Report Multimodality Image Registration by Maximization of Mutual Information ECSE 626 Project Report Multimodality Image Registration by Maximization of Mutual Information Emmanuel Piuze McGill University Montreal, Qc, Canada. epiuze@cim.mcgill.ca Abstract In 1997, Maes et al.

More information

Rigid and Deformable Vasculature-to-Image Registration : a Hierarchical Approach

Rigid 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 information

Multi-Modal Volume Registration Using Joint Intensity Distributions

Multi-Modal Volume Registration Using Joint Intensity Distributions Multi-Modal Volume Registration Using Joint Intensity Distributions Michael E. Leventon and W. Eric L. Grimson Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA leventon@ai.mit.edu

More information

A Novel Medical Image Registration Method Based on Mutual Information and Genetic Algorithm

A Novel Medical Image Registration Method Based on Mutual Information and Genetic Algorithm A Novel Medical Image Registration Method Based on Mutual Information and Genetic Algorithm Hongying Zhang, Xiaozhou Zhou, Jizhou Sun, Jiawan Zhang Visualization and Image Processing Group, SRDC, School

More information

Automatic Subthalamic Nucleus Targeting for Deep Brain Stimulation. A Validation Study

Automatic Subthalamic Nucleus Targeting for Deep Brain Stimulation. A Validation Study Automatic Subthalamic Nucleus Targeting for Deep Brain Stimulation. A Validation Study F. Javier Sánchez Castro a, Claudio Pollo a,b, Jean-Guy Villemure b, Jean-Philippe Thiran a a École Polytechnique

More information

Key words: automated registration, voxel similarity, multiresolution optimization, magnetic resonance, positron emission tomography

Key words: automated registration, voxel similarity, multiresolution optimization, magnetic resonance, positron emission tomography Automated three-dimensional registration of magnetic resonance and positron emission tomography brain images by multiresolution optimization of voxel similarity measures Colin Studholme, a) Derek L. G.

More information

William Yang Group 14 Mentor: Dr. Rogerio Richa Visual Tracking of Surgical Tools in Retinal Surgery using Particle Filtering

William Yang Group 14 Mentor: Dr. Rogerio Richa Visual Tracking of Surgical Tools in Retinal Surgery using Particle Filtering Mutual Information Computation and Maximization Using GPU Yuping Lin and Gérard Medioni Computer Vision and Pattern Recognition Workshops (CVPR) Anchorage, AK, pp. 1-6, June 2008 Project Summary and Paper

More information

OnDemand3D Fusion Technology

OnDemand3D Fusion Technology CYBERMED INC., ONDEMAND3D TECHNOLOGY INC. OnDemand3D Fusion Technology White Paper December 2009 USA Republic of Korea www.ondemand3d.com Introduction OnDemand3D TM Fusion is registration technology to

More information

1 Introduction Image fusion, in a clinical setting, is defined as the effective and meaningful integration of data from associated images to improve d

1 Introduction Image fusion, in a clinical setting, is defined as the effective and meaningful integration of data from associated images to improve d Likelihood Maximization Approach to Image Registration Yang-Ming Zhu Λ Member, IEEE and Steven M. Cochoff Nuclear Medicine Division, Marconi Medical Systems 595 Miner Road, Cleveland, Ohio 44143 Abstract

More information

Adaptive Local Multi-Atlas Segmentation: Application to Heart Segmentation in Chest CT Scans

Adaptive Local Multi-Atlas Segmentation: Application to Heart Segmentation in Chest CT Scans Adaptive Local Multi-Atlas Segmentation: Application to Heart Segmentation in Chest CT Scans Eva M. van Rikxoort, Ivana Isgum, Marius Staring, Stefan Klein and Bram van Ginneken Image Sciences Institute,

More information

A New & Robust Information Theoretic Measure and its Application to Image Alignment

A New & Robust Information Theoretic Measure and its Application to Image Alignment A New & Robust Information Theoretic Measure and its Application to Image Alignment F. Wang 1, B. C. Vemuri 1, M. Rao 2 and Y. Chen 2 1 Department of Computer & Information Sciences & Engr., 2 Department

More information

The Insight Toolkit. Image Registration Algorithms & Frameworks

The Insight Toolkit. Image Registration Algorithms & Frameworks The Insight Toolkit Image Registration Algorithms & Frameworks Registration in ITK Image Registration Framework Multi Resolution Registration Framework Components PDE Based Registration FEM Based Registration

More information

Operators-Based on Second Derivative double derivative Laplacian operator Laplacian Operator Laplacian Of Gaussian (LOG) Operator LOG

Operators-Based on Second Derivative double derivative Laplacian operator Laplacian Operator Laplacian Of Gaussian (LOG) Operator LOG Operators-Based on Second Derivative The principle of edge detection based on double derivative is to detect only those points as edge points which possess local maxima in the gradient values. Laplacian

More information

Image Registration Lecture 4: First Examples

Image Registration Lecture 4: First Examples Image Registration Lecture 4: First Examples Prof. Charlene Tsai Outline Example Intensity-based registration SSD error function Image mapping Function minimization: Gradient descent Derivative calculation

More information

Non-rigid Image Registration

Non-rigid Image Registration Overview Non-rigid Image Registration Introduction to image registration - he goal of image registration - Motivation for medical image registration - Classification of image registration - Nonrigid registration

More information

2D Rigid Registration of MR Scans using the 1d Binary Projections

2D Rigid Registration of MR Scans using the 1d Binary Projections 2D Rigid Registration of MR Scans using the 1d Binary Projections Panos D. Kotsas Abstract This paper presents the application of a signal intensity independent registration criterion for 2D rigid body

More information

Robust Linear Registration of CT images using Random Regression Forests

Robust Linear Registration of CT images using Random Regression Forests Robust Linear Registration of CT images using Random Regression Forests Ender Konukoglu a Antonio Criminisi a Sayan Pathak b Duncan Robertson a Steve White b David Haynor c Khan Siddiqui b a Microsoft

More information

Learning-based Neuroimage Registration

Learning-based Neuroimage Registration Learning-based Neuroimage Registration Leonid Teverovskiy and Yanxi Liu 1 October 2004 CMU-CALD-04-108, CMU-RI-TR-04-59 School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 Abstract

More information

Registration of 2D to 3D Joint Images Using Phase-Based Mutual Information

Registration of 2D to 3D Joint Images Using Phase-Based Mutual Information Registration of 2D to 3D Joint Images Using Phase-Based Mutual Information Rupin Dalvi a1, Rafeef Abugharbieh a, Mark Pickering b, Jennie Scarvell c, Paul Smith d a Biomedical Signal and Image Computing

More information

Utilizing Salient Region Features for 3D Multi-Modality Medical Image Registration

Utilizing Salient Region Features for 3D Multi-Modality Medical Image Registration Utilizing Salient Region Features for 3D Multi-Modality Medical Image Registration Dieter Hahn 1, Gabriele Wolz 2, Yiyong Sun 3, Frank Sauer 3, Joachim Hornegger 1, Torsten Kuwert 2 and Chenyang Xu 3 1

More information

Affine Image Registration Using A New Information Metric

Affine Image Registration Using A New Information Metric Affine Image Registration Using A New Information Metric Jie Zhang and Anand Rangarajan Dept. of Computer & Information Science & Engineering University of Florida, Gainesville, FL 36 {jiezhang,anand}@cise.ufl.edu

More information

ANALYSIS OF RELIABILITY AND IMPACT FACTORS OF MUTUAL INFORMATION SIMILARITY CRITERION FOR REMOTE SENSING IMAGERY TEMPLATE MATCHING

ANALYSIS OF RELIABILITY AND IMPACT FACTORS OF MUTUAL INFORMATION SIMILARITY CRITERION FOR REMOTE SENSING IMAGERY TEMPLATE MATCHING ANALYSIS OF RELIABILITY AND IMPACT FACTORS OF MUTUAL INFORMATION SIMILARITY CRITERION FOR REMOTE SENSING IMAGERY TEMPLATE MATCHING H.L. Wang a,, R. An b, * Q. Zhang b, C.Y. Chen b a College of Geography

More information

A Unifying Framework for Mutual Information methods for use in Non-linear Optimisation

A Unifying Framework for Mutual Information methods for use in Non-linear Optimisation A Unifying Framework for Mutual Information methods for use in Non-linear Optimisation Nicholas Dowson and Richard Bowden Centre for Vision Speed and Signal Processing University of Surrey, Guildford,

More information

A Radiometry Tolerant Method for Direct 3D/2D Registration of Computed Tomography Data to X-ray Images

A Radiometry Tolerant Method for Direct 3D/2D Registration of Computed Tomography Data to X-ray Images A Radiometry Tolerant Method for Direct 3D/2D Registration of Computed Tomography Data to X-ray Images Transfer Function Independent Registration Boris Peter Selby 1, Georgios Sakas 2, Stefan Walter 1,

More information

Ensemble registration: Combining groupwise registration and segmentation

Ensemble registration: Combining groupwise registration and segmentation PURWANI, COOTES, TWINING: ENSEMBLE REGISTRATION 1 Ensemble registration: Combining groupwise registration and segmentation Sri Purwani 1,2 sri.purwani@postgrad.manchester.ac.uk Tim Cootes 1 t.cootes@manchester.ac.uk

More information

Coupled Bayesian Framework for Dual Energy Image Registration

Coupled Bayesian Framework for Dual Energy Image Registration in Proc. IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), Vol. 2, pp 2475-2482, 2006 Coupled Bayesian Framework for Dual Energy Image Registration Hao Wu University of Maryland College Park,

More information

3D Voxel-Based Volumetric Image Registration with Volume-View Guidance

3D Voxel-Based Volumetric Image Registration with Volume-View Guidance 3D Voxel-Based Volumetric Image Registration with Volume-View Guidance Guang Li*, Huchen Xie, Holly Ning, Deborah Citrin, Jacek Copala, Barbara Arora, Norman Coleman, Kevin Camphausen, and Robert Miller

More information

A protocol for evaluation of similarity measures for non-rigid registration

A protocol for evaluation of similarity measures for non-rigid registration Available online at www.sciencedirect.com Medical Image Analysis 1 (8) 4 54 www.elsevier.com/locate/media A protocol for evaluation of similarity measures for non-rigid registration Darko Škerl, Boštjan

More information

Mutual Information Based Methods to Localize Image Registration

Mutual Information Based Methods to Localize Image Registration Mutual Information Based Methods to Localize Image Registration by Kathleen P. Wilkie A thesis presented to the University of Waterloo in fulfilment of the thesis requirement for the degree of Master of

More information

Multiresolution Registration of Remote-Sensing Images Using Stochastic Gradient

Multiresolution Registration of Remote-Sensing Images Using Stochastic Gradient Multiresolution Registration of Remote-Sensing Images Using Stochastic Gradient Arlene Cole-Rhodes 1, Kisha Johnson 1, Jacqueline Le Moigne 2 1. Dept of Electrical and Computer Engineering, Morgan State

More information

Elastic registration of medical images using finite element meshes

Elastic 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 information

Hierarchical Multi structure Segmentation Guided by Anatomical Correlations

Hierarchical Multi structure Segmentation Guided by Anatomical Correlations Hierarchical Multi structure Segmentation Guided by Anatomical Correlations Oscar Alfonso Jiménez del Toro oscar.jimenez@hevs.ch Henning Müller henningmueller@hevs.ch University of Applied Sciences Western

More information

Nonrigid Registration using Free-Form Deformations

Nonrigid Registration using Free-Form Deformations Nonrigid Registration using Free-Form Deformations Hongchang Peng April 20th Paper Presented: Rueckert et al., TMI 1999: Nonrigid registration using freeform deformations: Application to breast MR images

More information

Registration by continuous optimisation. Stefan Klein Erasmus MC, the Netherlands Biomedical Imaging Group Rotterdam (BIGR)

Registration by continuous optimisation. Stefan Klein Erasmus MC, the Netherlands Biomedical Imaging Group Rotterdam (BIGR) Registration by continuous optimisation Stefan Klein Erasmus MC, the Netherlands Biomedical Imaging Group Rotterdam (BIGR) Registration = optimisation C t x t y 1 Registration = optimisation C t x t y

More information

ABSTRACT 1. INTRODUCTION. we classified medical image registration methods according to nine criteria, viz.:

ABSTRACT 1. INTRODUCTION. we classified medical image registration methods according to nine criteria, viz.: General Multimodal Elastic Registration Based on Mutual Information J. B. Antoine Maintz, Erik H.W. Meijering, and Max A. Viergever Image Sciences Institute, Utrecht University, P.O.Box 80089, 3508 TB,

More information

Texture Classification by Combining Local Binary Pattern Features and a Self-Organizing Map

Texture Classification by Combining Local Binary Pattern Features and a Self-Organizing Map Texture Classification by Combining Local Binary Pattern Features and a Self-Organizing Map Markus Turtinen, Topi Mäenpää, and Matti Pietikäinen Machine Vision Group, P.O.Box 4500, FIN-90014 University

More information

Overview of Proposed TG-132 Recommendations

Overview of Proposed TG-132 Recommendations Overview of Proposed TG-132 Recommendations Kristy K Brock, Ph.D., DABR Associate Professor Department of Radiation Oncology, University of Michigan Chair, AAPM TG 132: Image Registration and Fusion Conflict

More information

Robust Multiresolution Alignment of MRI Brain Volumes

Robust Multiresolution Alignment of MRI Brain Volumes Robust Multiresolution Alignment of MRI Brain Volumes Oscar Nestares and David J. Heeger* Magnetic Resonance in Medicine 43:705 715 (2000) An algorithm for the automatic alignment of MRI volumes of the

More information

Image Registration I

Image Registration I Image Registration I Comp 254 Spring 2002 Guido Gerig Image Registration: Motivation Motivation for Image Registration Combine images from different modalities (multi-modality registration), e.g. CT&MRI,

More information

Lecture 13 Theory of Registration. ch. 10 of Insight into Images edited by Terry Yoo, et al. Spring (CMU RI) : BioE 2630 (Pitt)

Lecture 13 Theory of Registration. ch. 10 of Insight into Images edited by Terry Yoo, et al. Spring (CMU RI) : BioE 2630 (Pitt) Lecture 13 Theory of Registration ch. 10 of Insight into Images edited by Terry Yoo, et al. Spring 2018 16-725 (CMU RI) : BioE 2630 (Pitt) Dr. John Galeotti The content of these slides by John Galeotti,

More information

3D Multi-Modality Medical Image Registration Using Feature Space Clustering Andre Collignon, Dirk Vandermeulen, Paul Suetens, Guy Marchal Laboratory f

3D Multi-Modality Medical Image Registration Using Feature Space Clustering Andre Collignon, Dirk Vandermeulen, Paul Suetens, Guy Marchal Laboratory f 3D Multi-Modality Medical Image Registration Using Feature Space Clustering Andre Collignon, Dirk Vandermeulen, Paul Suetens, Guy Marchal Laboratory for Medical Imaging Research? Katholieke Universiteit

More information

Medical Image Registration by Maximization of Mutual Information

Medical Image Registration by Maximization of Mutual Information Medical Image Registration by Maximization of Mutual Information EE 591 Introduction to Information Theory Instructor Dr. Donald Adjeroh Submitted by Senthil.P.Ramamurthy Damodaraswamy, Umamaheswari Introduction

More information

Efficient 3D-3D Vascular Registration Based on Multiple Orthogonal 2D Projections

Efficient 3D-3D Vascular Registration Based on Multiple Orthogonal 2D Projections Efficient 3D-3D Vascular Registration Based on Multiple Orthogonal 2D Projections Ho-Ming Chan and Albert C. S. Chung Dept. of Computer Science, Hong Kong University of Science and Technology, HK {hmchan,achung}@cs.ust.hk

More information

Métodos de fusão e co-registo de imagem

Métodos de fusão e co-registo de imagem LICENCIATURA EM ENGENHARIA BIOMÉDICA Fundamentos de Imagem Diagnóstica e Terapeutica Métodos de fusão e co-registo de imagem Jorge Isidoro Sumário Introdução à imagem médica Registo de imagens Metodologia

More information

PET Image Reconstruction using Anatomical Information through Mutual Information Based Priors

PET Image Reconstruction using Anatomical Information through Mutual Information Based Priors 2005 IEEE Nuclear Science Symposium Conference Record M11-354 PET Image Reconstruction using Anatomical Information through Mutual Information Based Priors Sangeetha Somayajula, Evren Asma, and Richard

More information

Automatic Lung Surface Registration Using Selective Distance Measure in Temporal CT Scans

Automatic Lung Surface Registration Using Selective Distance Measure in Temporal CT Scans Automatic Lung Surface Registration Using Selective Distance Measure in Temporal CT Scans Helen Hong 1, Jeongjin Lee 2, Kyung Won Lee 3, and Yeong Gil Shin 2 1 School of Electrical Engineering and Computer

More information

Image Segmentation. Shengnan Wang

Image 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 information

Automated Medical Image Registration Using the Simulated Annealing Algorithm

Automated Medical Image Registration Using the Simulated Annealing Algorithm Automated Medical Image Registration Using the Simulated Annealing Algorithm Ilias Maglogiannis 1 and Elias Zafiropoulos 2 1 University of the Aegean, Dept. of Information and Communication Systems Engineering

More information

Image Registration for Dental X-Ray images using Hybrid Technique

Image Registration for Dental X-Ray images using Hybrid Technique Image Registration for Dental X-Ray images using Hybrid Technique Gurpreet Rathore 1, Dr. Vijay Dhir 2 Lecturer, Dept. of Computer Science, SBBSPGC, Jalandhar, India 1 Head of Department, Dept. of Computer

More information

Registration of Hyperspectral and Trichromatic Images via Cross Cumulative Residual Entropy Maximisation

Registration of Hyperspectral and Trichromatic Images via Cross Cumulative Residual Entropy Maximisation Registration of Hyperspectral and Trichromatic Images via Cross Cumulative Residual Entropy Maximisation Author Hasan, Mahmudul, Pickering, Mark, Robles-Kelly, Antonio, Zho Jun, Jia, Xiuping Published

More information

Retrospective evaluation of inter-subject brain registration

Retrospective evaluation of inter-subject brain registration Retrospective evaluation of inter-subject brain registration P. Hellier 1, C. Barillot 1, I. Corouge 1,B.Gibaud 2, G. Le Goualher 2,3, L. Collins 3, A. Evans 3, G. Malandain 4, N. Ayache 4 1 Projet Vista,

More information

RIGID IMAGE REGISTRATION

RIGID IMAGE REGISTRATION RIGID IMAGE REGISTRATION Duygu Tosun-Turgut, Ph.D. Center for Imaging of Neurodegenerative Diseases Department of Radiology and Biomedical Imaging duygu.tosun@ucsf.edu What is registration? Image registration

More information

Introduction to Medical Image Registration

Introduction to Medical Image Registration Introduction to Medical Image Registration Sailesh Conjeti Computer Aided Medical Procedures (CAMP), Technische Universität München, Germany sailesh.conjeti@tum.de Partially adapted from slides by: 1.

More information

Efficient population registration of 3D data

Efficient population registration of 3D data Efficient population registration of 3D data Lilla Zöllei 1, Erik Learned-Miller 2, Eric Grimson 1, William Wells 1,3 1 Computer Science and Artificial Intelligence Lab, MIT; 2 Dept. of Computer Science,

More information

Robust Realignment of fmri Time Series Data

Robust Realignment of fmri Time Series Data Robust Realignment of fmri Time Series Data Ben Dodson bjdodson@stanford.edu Olafur Gudmundsson olafurg@stanford.edu December 12, 2008 Abstract FMRI data has become an increasingly popular source for exploring

More information

CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION

CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION 6.1 INTRODUCTION Fuzzy logic based computational techniques are becoming increasingly important in the medical image analysis arena. The significant

More information

Image Segmentation Based on Watershed and Edge Detection Techniques

Image Segmentation Based on Watershed and Edge Detection Techniques 0 The International Arab Journal of Information Technology, Vol., No., April 00 Image Segmentation Based on Watershed and Edge Detection Techniques Nassir Salman Computer Science Department, Zarqa Private

More information

Efficient Computation of Histograms on the GPU

Efficient Computation of Histograms on the GPU Efficient Computation of Histograms on the GPU Alexander Kubias University of Koblenz-Landau Frank Deinzer Siemens Medical Solutions Dietrich Paulus University of Koblenz-Landau Matthias Kreiser Siemens

More information

Good Morning! Thank you for joining us

Good Morning! Thank you for joining us Good Morning! Thank you for joining us Deformable Registration, Contour Propagation and Dose Mapping: 101 and 201 Marc Kessler, PhD, FAAPM The University of Michigan Conflict of Interest I receive direct

More information

Image Registration in Hough Space Using Gradient of Images

Image Registration in Hough Space Using Gradient of Images Image Registration in Hough Space Using Gradient of Images Ramtin Shams Research School of Information Sciences and Engineering (RSISE) The Australian National University (ANU) Canberra, ACT 2 ramtin.shams@anu.edu.au

More information

Cluster of Workstation based Nonrigid Image Registration Using Free-Form Deformation

Cluster of Workstation based Nonrigid Image Registration Using Free-Form Deformation Cluster of Workstation based Nonrigid Image Registration Using Free-Form Deformation Xiaofen Zheng, Jayaram K. Udupa, and Xinjian Chen Medical Image Processing Group, Department of Radiology 423 Guardian

More information

Digital subtraction CT angiography based on efficient 3D registration and refinement

Digital subtraction CT angiography based on efficient 3D registration and refinement Computerized Medical Imaging and Graphics 28 (2004) 391 400 www.elsevier.com/locate/compmedimag Digital subtraction CT angiography based on efficient 3D registration and refinement Sung Min Kwon a, Yong

More information

Nonrigid Registration with Adaptive, Content-Based Filtering of the Deformation Field

Nonrigid Registration with Adaptive, Content-Based Filtering of the Deformation Field Nonrigid Registration with Adaptive, Content-Based Filtering of the Deformation Field Marius Staring*, Stefan Klein and Josien P.W. Pluim Image Sciences Institute, University Medical Center Utrecht, P.O.

More information

2 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

2 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 information

Image Co-Registration II: TG132 Quality Assurance for Image Registration. Image Co-Registration II: TG132 Quality Assurance for Image Registration

Image Co-Registration II: TG132 Quality Assurance for Image Registration. Image Co-Registration II: TG132 Quality Assurance for Image Registration Image Co-Registration II: TG132 Quality Assurance for Image Registration Preliminary Recommendations from TG 132* Kristy Brock, Sasa Mutic, Todd McNutt, Hua Li, and Marc Kessler *Recommendations are NOT

More information

Reproducibility of interactive registration of 3D CT and MR pediatric treatment planning head images

Reproducibility of interactive registration of 3D CT and MR pediatric treatment planning head images JOURNAL OF APPLIED CLINICAL MEDICAL PHYSICS, VOLUME 2, NUMBER 3, SUMMER 2001 Reproducibility of interactive registration of 3D CT and MR pediatric treatment planning head images Jaap Vaarkamp* Department

More information

A Joint Histogram - 2D Correlation Measure for Incomplete Image Similarity

A Joint Histogram - 2D Correlation Measure for Incomplete Image Similarity A Joint Histogram - 2D Correlation for Incomplete Image Similarity Nisreen Ryadh Hamza MSc Candidate, Faculty of Computer Science & Mathematics, University of Kufa, Iraq. ORCID: 0000-0003-0607-3529 Hind

More information

Copyright 2009 Society of Photo Optical Instrumentation Engineers. This paper was published in Proceedings of SPIE, vol. 7260, Medical Imaging 2009:

Copyright 2009 Society of Photo Optical Instrumentation Engineers. This paper was published in Proceedings of SPIE, vol. 7260, Medical Imaging 2009: Copyright 2009 Society of Photo Optical Instrumentation Engineers. This paper was published in Proceedings of SPIE, vol. 7260, Medical Imaging 2009: Computer Aided Diagnosis and is made available as an

More information

Improvement of SURF Feature Image Registration Algorithm Based on Cluster Analysis

Improvement of SURF Feature Image Registration Algorithm Based on Cluster Analysis Sensors & Transducers 2014 by IFSA Publishing, S. L. http://www.sensorsportal.com Improvement of SURF Feature Image Registration Algorithm Based on Cluster Analysis 1 Xulin LONG, 1,* Qiang CHEN, 2 Xiaoya

More information

Department of ECE, SCSVMV University, Kanchipuram

Department of ECE, SCSVMV University, Kanchipuram Medical Image Registering: A Matlab Based Approach [1] Dr.K.Umapathy, [2] D.Vedasri, [3] H.Vaishnavi [1] Associate Professor, [2][3] UG Student, Department of ECE, SCSVMV University, Kanchipuram Abstract:-

More information

3D Registration based on Normalized Mutual Information

3D Registration based on Normalized Mutual Information 3D Registration based on Normalized Mutual Information Performance of CPU vs. GPU Implementation Florian Jung, Stefan Wesarg Interactive Graphics Systems Group (GRIS), TU Darmstadt, Germany stefan.wesarg@gris.tu-darmstadt.de

More information

ABSTRACT 1. INTRODUCTION 2. METHODS

ABSTRACT 1. INTRODUCTION 2. METHODS Finding Seeds for Segmentation Using Statistical Fusion Fangxu Xing *a, Andrew J. Asman b, Jerry L. Prince a,c, Bennett A. Landman b,c,d a Department of Electrical and Computer Engineering, Johns Hopkins

More information

The organization of the human cerebral cortex estimated by intrinsic functional connectivity

The organization of the human cerebral cortex estimated by intrinsic functional connectivity 1 The organization of the human cerebral cortex estimated by intrinsic functional connectivity Journal: Journal of Neurophysiology Author: B. T. Thomas Yeo, et al Link: https://www.ncbi.nlm.nih.gov/pubmed/21653723

More information

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm Group 1: Mina A. Makar Stanford University mamakar@stanford.edu Abstract In this report, we investigate the application of the Scale-Invariant

More information

Fast 3D Mean Shift Filter for CT Images

Fast 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 information

Align3_TP Manual. J. Anthony Parker, MD PhD Beth Israel Deaconess Medical Center Boston, MA Revised: 26 November 2004

Align3_TP Manual. J. Anthony Parker, MD PhD Beth Israel Deaconess Medical Center Boston, MA Revised: 26 November 2004 Align3_TP Manual J. Anthony Parker, MD PhD Beth Israel Deaconess Medical Center Boston, MA J.A.Parker@IEEE.org Revised: 26 November 2004 General ImageJ is a highly versatile image processing program written

More information

AUTOMATIC REGISTRATION OF OPTICAL IMAGERY WITH 3D LIDAR DATA USING LOCAL COMBINED MUTUAL INFORMATION

AUTOMATIC REGISTRATION OF OPTICAL IMAGERY WITH 3D LIDAR DATA USING LOCAL COMBINED MUTUAL INFORMATION AUTOMATIC REGISTRATION OF OPTICAL IMAGERY WITH 3D LIDAR DATA USING LOCAL COMBINED MUTUAL INFORMATION Ebadat G. Parmehr a,b, *, Clive S. Fraser a,b, Chunsun Zhang a,c, Joseph Leach b a Cooperative Research

More information