Iterative Estimation of 3D Transformations for Object Alignment
|
|
- Josephine Alexandrina Morris
- 5 years ago
- Views:
Transcription
1 Iterative Estimation of 3D Transformations for Object Alignment Tao Wang and Anup Basu Department of Computing Science, Univ. of Alberta, Edmonton, AB T6G 2E8, Canada Abstract. An Iterative Estimation Algorithm (IEA) of 3D transformations between two objects is presented in this paper. Skeletons of the 3D objects are extracted using a fully parallel thinning technique, feature point pairs (land markers) are extracted from skeletons automatically with a heuristic rule, and a least squares method and an iterative approach are applied to estimate the 3D transformation matrix. The algorithm has three advantages. First of all, no initial transformation matrix is needed. Secondly, user interaction is not required for identifying the land markers. Thirdly, the time complexity of this algorithm is polynomial. Experiments show that this method works quite well with high accuracy when the translations and rotation angles are small, even when noise exists in the data. 1 Introduction 3D alignment or registration algorithms have many applications in medical image processing. Consider two objects O1 and O2, such that O2=M*O1, where M is the 3D transformation matrix. In this context, the objective of 3D alignment or registration is to estimate the 3D transformation matrix M. Surveys [1-2] of 3D alignment methods are available in the literature. Since the previous decades, a lot of 3D alignment algorithms, including land marker based algorithm [3], have been proposed. This paper focuses on land marker based algorithms. The mean square distance () [3] is used as the metric. Point correspondence [4], which relates a land marker with its counterpart, is a crucial step for algorithms in this category. The advantage of this technique is that the set of land markers is relatively sparse compared to the original 3D object, so that the optimization process is relatively fast. However, this technique has some disadvantages. First of all, some algorithms, for instance, the ICP [3] algorithm, needs to know the initial transformation matrix. Secondly, user interaction is usually required for identifying the land markers. Last but not the least, the complexity of point correspondence [4] is non-polynomial since it is a combinatorial optimization problem. In this paper, an automatic Iterative Estimation Algorithm (IEA) is proposed. It has three advantages: (i) no initial transformation matrix is needed; (ii) user interaction is not required for identifying the land markers; (iii) the complexity of this algo-
2 rithm is polynomial. The disadvantage of this algorithm is that it does not guarantee achieving global optimization. We applied IEA in 3D medical image processing. We are interested in defining and tracking volume changes of airways caused by surgery, which increases volume. Doctors need to track the changes of airways for administering effective treatments. In our application, doctors take MRI scans of patients and save them in the DICOM [5] format. A commercial software ScanIP/FE [6] is used to segment the airways and remove noise semi-automatically. The skeletons are created to depict the segmented airway using a fully parallel 3D thinning algorithm developed by us. Then, we applied a heuristic algorithm to automatically extract feature point pairs (land markers) from skeletons. Since false point correspondence is inevitable, we use IEA to remove some point pairs and use the remainder point pairs to achieve smaller. The whole procedure has 5 main steps: segmentation, noise removal, thinning, finding feature point pairs and iterative transformation estimation. In the following sections, the algorithm is described in detail. In Section 2, we introduce the data collection and segmentation procedures. Section 3 will briefly discuss how to artificially create data for comparison. The fully parallel 3D thinning algorithm and the skeletonization process is summarized in Section 4. Section 5 focuses on the acquisition of feature point pairs (land markers), and the IEA. Results of experiments are presented in Section 6, before the work is concluded in Section 7. 2 Data collection, segmentation and noise removal A small portion of the upper airway is scanned for a child as an MRI image and saved in the DICOM format, which is the industry standard for MRI. ScanIP/FE is used to segment airways, remove noise, and semi-automatically create an airway model of 3D images. A 3D image is a mapping that assigns the value of 0 or 1 to each point in the 3D space. Points having the value of 1 are called black (object) points, while 0 s are called white (background) ones. Black points form objects of the image. Our test data has about 37,000 object points. Fig. 1 shows some image slices and the segmented airway (in red). The 3D models from different viewpoints are displayed in Fig. 2. Fig. 1. image segmented red) (Left) image slice center slice last slice. Some slices and airway (in The first (Middle) A (Right) The
3 Fig. 2. 3D airway model from 3 different points of view. 3 Data creation for estimation To verify our method, we artificially created some airway models for estimation. We applied 3D homogeneous transformations [7] to the 3D image. We use (alpha, beta, gama, dx, dy, dz) to represent the rotation angles and translations for x-, y- and z-axis of the Homogeneous Matrices. We also added some random noise to the data to create models for comparison. 4 Fully parallel 3D thinning algorithm and skeletonization 3D thinning for medial lines or surface approximation is a useful approach for many potential applications. The approach has been extensively researched in the last decade [8]. In the experiments, we applied our fully parallel 3D thinning algorithm to extract the skeletons from 3D images after a noise removal step. Our method is an improvement on Ma and Sonka s algorithm [8]. The algorithm in [8] has a number of advantages; however, we found that it cannot preserve connectivity of 3D objects. Chaturvedi et al. [9] also found this problem. We improved [8] by changing some masks in Class D to preserve connectivity. The skeletons are used to represent the airway models. The skeletons have about 500 object points, a large reduction from the 37,000 object points. Fig. 3 (Left) and (Middle) show the skeletons of the original model and the artificially created model with (alpha, beta, gama, dx, dy, dz) = (1, 1, 1, 0, 0, 0). We notice that: 1. The thinning algorithm is sensitive to rotations. Even when the rotation angles are very small, the two skeletons look quite different. 2. The thinning algorithm cannot provide unit-width structures. We can see that some regions on the skeletons are quite dense. However, we will show that these two drawbacks do little harm to the estimation of 3D transformations. Thus, we can still estimate the transformations precisely based on the thinning algorithm.
4 Fig. 3. (Left) The skeleton of the original model. (Middle) The skeleton of the transformed model with (alpha, beta, gama, dx, dy, dz) = (1, 1, 1, 0, 0, 0). (Right) Line points (in red). 5 Feature point pair acquisition and the IEA Land marker based estimation algorithms have three main disadvantages as we discuss before. In this section, an Iterative Estimation Algorithm (IEA) is proposed to solve these problems. The disadvantage of this algorithm is that it does not guarantee global optimization. Definition 1: Connectivity number If P is an object point in a 3D image the connectivity number of P is the number of object points except itself in P s (3*3*3) neighborhood. In Fig. 4 (Left), the connectivity number of P is 4. Definition 2: Line point A line point is an object point with connectivity number equal to 2. In Fig. 3 (Right), the line points are displayed in red. Fig. 4. (Left) Connectivity number of point P is 4. A is an object point, A o is a background point. (Right) Point P and its 8 sub-neighborhood. Definition 3: Feature point (land marker) candidate A feature point (land marker) candidate is a line point with at least one non- line point in its (3 * 3 * 3) neighborhood. Definition 4: Feature point pair (land marker) I1 and I2 are two 3D images of the same object. I2=M*I1, where M is a 3D transformation matrix. A feature point pair (land marker) contains two object points P1 and P2, where P1 is on the skeleton of 3D image I1 and P2 is on the skeleton of 3D image I2, and P2=M*P The overdetermined linear system In general, M is a 4 * 4 Homogeneous Transformation Matrix. For point P1= (x1 y1 z1 1) T in 3D space, P2 = M * P1= (x2 y2 z2 1) T. The goal of our work is to estimate the 16 variables of matrix M in certain conditions. Since we need to calculate
5 16 variables, mathematically, we need 16 equations for this linear system. For each feature point pair, we can have 4 equations. So we need 4 feature point pairs in total to solve this problem. However, in most cases in our experiment, we can find more than 4 feature point pairs. This is a typical case of an over-determined linear system. An overdetermined linear system can be solved following well-established methods [10]. 5.2 The heuristic rule In our experiments, a heuristic rule is used to identify the feature point pairs. It requires two points in a feature point pair to have same "configurations" in its 8 subneighborhood. That is, the local topologies of two points in a feature point pair should be same. We used an adaptive method to search for the best neighborhood scale. Fig. 4 (Right) shows a point p and its 8 sub-neighborhood. Heuristic Rule: If two feature point candidates on different skeletons have the same number of object points and the same number of background points in all 8 subneighborhoods, these two points form a feature point pair. 5.3 The iterative estimation method Feature point correspondences is an open problem in Computer Vision. No general method can solve it in polynomial time. There are mainly two difficulties. Firstly, it is a combinatorial optimization problem. The time complexity is very high because of the large search space. Secondly, point correspondence is not one-one mapping. Some points, which are called outliers [4, 11], may not have counterparts because of occlusions, out of image transformations and errors in feature selection. A brief survey of previous work is available in the literature [4]. For several decades, researchers have proposed a number of point correspondence algorithms [4, 11]. Point correspondence algorithms have many applications [4, 12-13] including image alignment. The point correspondence algorithms are classified into two categories, globally optimal and non-globally optimal. A globally optimal algorithm [4] guarantees global optimality. However, the time complexity is non-polynomial. A non-globally optimal algorithm [11] is faster than a globally optimal algorithm. However, it does not guarantee global optimality. Another restriction is that some unwanted assumptions and constrains [11] are often required to detect the outliers. The IEA is a non-globally optimal algorithm. The complexity of this algorithm is polynomial. No user interaction, initial transformation knowledge and outlier detection are required. The motivation behind our work is that we realized that the characteristics of our application, as well as many other medical imaging applications, are: the number of feature point pairs is large, user interaction is unwanted, initial transformation knowledge is not easy to know and the outlier detection, which requires some unwanted assumptions, constrains, and is not very reliable, should be avoided. In these cases, we need a polynomial algorithm that requires no user interaction, no initial transformation information, implicitly removes outliers without any unwanted assumptions and constrains.
6 We notice that incorrect point correspondences and incorrect outlier detection negatively affect the estimation accuracy. However, we also notice that incorrect point correspondences and incorrect outlier detection are unavoidable. Therefore, in our algorithm, we use a very simple Heuristic Rule to do the point correspondences. There are some incorrect point correspondences and some outliers exist in the feature point pairs set. We do not know which point pair is correctly related and which point pair is incorrectly related. Since the incorrect point correspondences and incorrect outlier detection negatively affect the estimation, i.e., the is large, we iteratively remove one point pair from the whole set and test if the new is smaller. If the new is smaller than the old, we accept this removal. Otherwise, we undo the removal. The iteration stops when a preset threshold is achieved or only 4 pointpairs remain. This approach is very simple and easy to implement. It removes the incorrect point correspondences and removes outliers implicitly. The pseudo code of the Iterative Estimation Algorithm (IEA) is as follows: INPUT: 3D image I 1 and 3D image I 2, feature point candidate sets C 1 and C 2, and the threshold THRESHOLD OUTPUT: 3D transformation matrix M and the 1. Read I 1, I 2, C 1 and C 2 2. Select feature point pairs P = (P 1, P 2 ) with the Heuristic Rule, where P 1 C 1 and P 2 C 2. Denote by N FP the number of feature point pairs. 3. Initialize the mean square distance and the transformation matrix M with a predefined value. If N FP < 4, then go to Step Use least square method to solve the over-determined linear system with P, and get the 3D transformation matrix M. Create a new 3D image I 3, where I 3 =M *I 1 5. Calculate the between I 2 and I If < THRESHOLD, then go to Step 10. FOR (INDEX = 0; INDEX < N FP ; INDEX ++) 7. Remove a feature point pair with index = INDEX from P to create feature point pairs P, calculate the 3D transformation matrix M, create a new 3D image I 3, where I 3 = M * I 1 and calculate mean square distance between I 2 and I 3.
7 8. If < THRESHOLD or N FP <4, then M=M and go to Step If <, then =, P = P. Otherwise, undo the remove operation in Step 6. END FOR 10. Output the M and. The time complexity of the IEA is polynomial. Step 1, 3, 6 and 10 of the IEA has constant complexity. Step 4 and 5 has polynomial complexity. The complexity of Step 2 is O(N C1 * N C2 ), where N C1 and N C2 are the number of feature point candidates in C 1 and C 2. The complexity of Step 7-9 is O(N FP ) * (T LS +T ), where T LS is the complexity of least square method and T is the complexity of calculating. Both of T LS and T are polynomial. So the total time complexity of IEA is polynomial. To test our approach, we integrate it into our previous framework of 3D transformation estimation. Experiments in Section 6 show that this approach with IEA can estimate 3D transformation with higher accuracy. The new framework is described as follows: INPUT: 3D image I 1 and I 2, the threshold THRESHOLD, and the range of neighborhood searching NB MIN and NB MAX. OUTPUT: 3D transformation matrix between I 1 and I 2 Read I 1 and I 2, create their skeletons S 1 and S 2, Init M EST and EST with a predefined value FOR (NB = NB MIN ; NB <= NB MAX ; NB += 2) 1. Calculate the feature point candidate sets C 1 from S 1, C 2 from S 2, within current neighborhood scale NB 2. Call the IEA to get M and 3. If < THRESHOLD, then M EST = M and go to Step If EST >, then EST = and M EST = M. END FOR 5. Output the estimated 3D transformation matrix M EST In our experiments, NB MIN = 3, NB MAX = 21 and THRESHOLD =1.0. The experiments and the results are described in Section 6.
8 6. Experimental results We first applied our new method with IEA to estimate the 3D translations along the x-, y- and z-axes. All the translations are positive integers in [0, 5]. Fig. 5 shows the for 3D translations. Result shows the is always 0 in both of the previous method and the new method. Both method can locate feature point pairs precisely and estimate the translations without error. Next, we use our new method with IEA to estimate the 3D rotations in x-, y- and z-axis. All the rotation angles are positive integers in [0, 5]. Fig. 6 shows the for 3D rotations. Result shows that the new method with the IEA has smaller. We also applied our new method with IEA to estimate the combinations of translations and rotations with some random noises (Figs. 7-10). All the translations and rotation angles are positive integers in [0, 5]. The random noise is in the range of [0%, 5%]. Results show that the new method with IEA has smaller. Translations dx + dy + dz Fig. 5. for 3D translations Rotations alpha + beta + gama Fig. 6. for 3D rotations
9 Trans. and Rots. in x, y and z axis without noise dx + dy + dz + alpha + beta + gama Fig. 7. for 3D translations and rotations without noise Trans. and Rots. in x, y and z axis with 1% random noise dx + dy + dz + alpha + beta + gama Fig. 8. for 3D translations and rotations with 1% random noise 3 Trans.and Rots. in x, y and z axis with 2% random noise dx + dy + dz + alpha + beta + gama Fig. 9. for 3D translations and rotations with 2% random noise
10 Trans. and Rots. in x, y and z axis with 5% random noise dx + dy + dz + alpha + beta + gama Fig. 10. for 3D translations and rotations with 5% random noise 7 Conclusions and future work In this paper, we demonstrated how to use an iterative algorithm to improve the estimation accuracy of 3D transformations. Result shows that this method works quite well when the differences in orientation are small. This work is a preliminary approach for defining and tracking the volume changes of airways with surgery. Because of constraints on time, we do not have airway model after surgery. Therefore, we had to create some models artificially to test our algorithm. In the near future, we will obtain real models after surgery and then compare our results using data from before and after surgery. Acknowledgements The support of ASRA and NSERC in making this work possible is gratefully acknowledged. References 1. J.B.A. Maintz and M.A. Viergever, A Survey of Medical Image Registration, Medical Image Analysis, vol.2, p.1-36, L. G. Brown, A survey of image registration techniques, ACM Computing Surveys (CSUR), 24(4), p , P. J. Besl and N. D. Mckay, A method for registration of 3D shapes, IEEE Trans. Pattern Anal. Machine Intell. 14(2), p , João Maciel, Global matching: optimal solution to correspondence problems, PhD Thesis, Universidade Técnica de Lisboa,
11 7. D. Hearn and M. Baker, Computer Graphics, Pearson Education. 8. C. M. Ma and M. Sonka, A fully parallel 3D thinning algorithm and its applications, CVIU, 64 (3), p , A. Chaturvedi, Z. Lee, Three-dimensional segmentation and skeletonization to build an airway tree data structure for small animals, 50(7), Physics in Medicine and Biology, p , J. J. Leader, Numerical analysis and scientific computation, Pearson Addison Wesley, Boston, P. Torr. Motion Segmentation and Outlier Detection. PhD thesis, U. Oxford, Justin Domke and Yiannis Aloimonos, A Probabilistic Notion of Correspondence and the Epipolar Constraint, Third International Symposium on 3D Data Processing, Visualization and Transmission, June Wei Zhang and Jana Kosecka, Image based Localization in Urban Environments, Third International Symposium on 3D Data Processing, Visualization and Transmission, June 2006.
A method for quantitative measurement of gas volume changes in upper airway
A method for quantitative measurement of gas volume changes in upper airway TAO WANG AND ANUP BASU DEPARTMENT OF COMPUTING SCIENCE, UNIVERSITY OF ALBERTA Abstract A method for quantitative measurement
More informationA Study of Medical Image Analysis System
Indian Journal of Science and Technology, Vol 8(25), DOI: 10.17485/ijst/2015/v8i25/80492, October 2015 ISSN (Print) : 0974-6846 ISSN (Online) : 0974-5645 A Study of Medical Image Analysis System Kim Tae-Eun
More informationAlgorithm research of 3D point cloud registration based on iterative closest point 1
Acta Technica 62, No. 3B/2017, 189 196 c 2017 Institute of Thermomechanics CAS, v.v.i. Algorithm research of 3D point cloud registration based on iterative closest point 1 Qian Gao 2, Yujian Wang 2,3,
More informationCOMPARATIVE STUDY OF DIFFERENT APPROACHES FOR EFFICIENT RECTIFICATION UNDER GENERAL MOTION
COMPARATIVE STUDY OF DIFFERENT APPROACHES FOR EFFICIENT RECTIFICATION UNDER GENERAL MOTION Mr.V.SRINIVASA RAO 1 Prof.A.SATYA KALYAN 2 DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING PRASAD V POTLURI SIDDHARTHA
More informationCHAPTER 3 DISPARITY AND DEPTH MAP COMPUTATION
CHAPTER 3 DISPARITY AND DEPTH MAP COMPUTATION In this chapter we will discuss the process of disparity computation. It plays an important role in our caricature system because all 3D coordinates of nodes
More informationShape Modeling of A String And Recognition Using Distance Sensor
Proceedings of the 24th IEEE International Symposium on Robot and Human Interactive Communication Kobe, Japan, Aug 31 - Sept 4, 2015 Shape Modeling of A String And Recognition Using Distance Sensor Keisuke
More informationSurface Registration. Gianpaolo Palma
Surface Registration Gianpaolo Palma The problem 3D scanning generates multiple range images Each contain 3D points for different parts of the model in the local coordinates of the scanner Find a rigid
More informationOne Dim~nsional Representation Of Two Dimensional Information For HMM Based Handwritten Recognition
One Dim~nsional Representation Of Two Dimensional Information For HMM Based Handwritten Recognition Nafiz Arica Dept. of Computer Engineering, Middle East Technical University, Ankara,Turkey nafiz@ceng.metu.edu.
More informationRobust Point Matching for Two-Dimensional Nonrigid Shapes
Robust Point Matching for Two-Dimensional Nonrigid Shapes Yefeng Zheng and David Doermann Language and Media Processing Laboratory Institute for Advanced Computer Studies University of Maryland, College
More informationA MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING
Proceedings of the 1994 IEEE International Conference on Image Processing (ICIP-94), pp. 530-534. (Austin, Texas, 13-16 November 1994.) A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING
More informationRotation Invariant Image Registration using Robust Shape Matching
International Journal of Electronic and Electrical Engineering. ISSN 0974-2174 Volume 7, Number 2 (2014), pp. 125-132 International Research Publication House http://www.irphouse.com Rotation Invariant
More informationInverse Kinematics II and Motion Capture
Mathematical Foundations of Computer Graphics and Vision Inverse Kinematics II and Motion Capture Luca Ballan Institute of Visual Computing Comparison 0 1 A B 2 C 3 Fake exponential map Real exponential
More informationSegmentation and Tracking of Partial Planar Templates
Segmentation and Tracking of Partial Planar Templates Abdelsalam Masoud William Hoff Colorado School of Mines Colorado School of Mines Golden, CO 800 Golden, CO 800 amasoud@mines.edu whoff@mines.edu Abstract
More informationStructured Light II. Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov
Structured Light II Johannes Köhler Johannes.koehler@dfki.de Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov Introduction Previous lecture: Structured Light I Active Scanning Camera/emitter
More informationA 3-D Scanner Capturing Range and Color for the Robotics Applications
J.Haverinen & J.Röning, A 3-D Scanner Capturing Range and Color for the Robotics Applications, 24th Workshop of the AAPR - Applications of 3D-Imaging and Graph-based Modeling, May 25-26, Villach, Carinthia,
More informationTracking of Human Body using Multiple Predictors
Tracking of Human Body using Multiple Predictors Rui M Jesus 1, Arnaldo J Abrantes 1, and Jorge S Marques 2 1 Instituto Superior de Engenharia de Lisboa, Postfach 351-218317001, Rua Conselheiro Emído Navarro,
More informationImage 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 informationAn explicit feature control approach in structural topology optimization
th World Congress on Structural and Multidisciplinary Optimisation 07 th -2 th, June 205, Sydney Australia An explicit feature control approach in structural topology optimization Weisheng Zhang, Xu Guo
More informationADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N.
ADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N. Dartmouth, MA USA Abstract: The significant progress in ultrasonic NDE systems has now
More informationMoving Object Segmentation Method Based on Motion Information Classification by X-means and Spatial Region Segmentation
IJCSNS International Journal of Computer Science and Network Security, VOL.13 No.11, November 2013 1 Moving Object Segmentation Method Based on Motion Information Classification by X-means and Spatial
More informationClassification of Subject Motion for Improved Reconstruction of Dynamic Magnetic Resonance Imaging
1 CS 9 Final Project Classification of Subject Motion for Improved Reconstruction of Dynamic Magnetic Resonance Imaging Feiyu Chen Department of Electrical Engineering ABSTRACT Subject motion is a significant
More informationLandmark Detection on 3D Face Scans by Facial Model Registration
Landmark Detection on 3D Face Scans by Facial Model Registration Tristan Whitmarsh 1, Remco C. Veltkamp 2, Michela Spagnuolo 1 Simone Marini 1, Frank ter Haar 2 1 IMATI-CNR, Genoa, Italy 2 Dept. Computer
More information2 ATTILA FAZEKAS The tracking model of the robot car The schematic picture of the robot car can be seen on Fig.1. Figure 1. The main controlling task
NEW OPTICAL TRACKING METHODS FOR ROBOT CARS Attila Fazekas Debrecen Abstract. In this paper new methods are proposed for intelligent optical tracking of robot cars the important tools of CIM (Computer
More informationDetection of Edges Using Mathematical Morphological Operators
OPEN TRANSACTIONS ON INFORMATION PROCESSING Volume 1, Number 1, MAY 2014 OPEN TRANSACTIONS ON INFORMATION PROCESSING Detection of Edges Using Mathematical Morphological Operators Suman Rani*, Deepti Bansal,
More informationarxiv: v1 [cs.cv] 28 Sep 2018
Camera Pose Estimation from Sequence of Calibrated Images arxiv:1809.11066v1 [cs.cv] 28 Sep 2018 Jacek Komorowski 1 and Przemyslaw Rokita 2 1 Maria Curie-Sklodowska University, Institute of Computer Science,
More informationDetermination of a Vessel Tree Topology by Different Skeletonizing Algorithms
Determination of a Vessel Tree Topology by Different Skeletonizing Algorithms Andre Siegfried Prochiner 1, Heinrich Martin Overhoff 2 1 Carinthia University of Applied Sciences, Klagenfurt, Austria 2 University
More information3D Computer Vision. Structured Light II. Prof. Didier Stricker. Kaiserlautern University.
3D Computer Vision Structured Light II Prof. Didier Stricker Kaiserlautern University http://ags.cs.uni-kl.de/ DFKI Deutsches Forschungszentrum für Künstliche Intelligenz http://av.dfki.de 1 Introduction
More informationAugmented reality with the ARToolKit FMA175 version 1.3 Supervisor Petter Strandmark By Olle Landin
Augmented reality with the ARToolKit FMA75 version.3 Supervisor Petter Strandmark By Olle Landin Ic7ol3@student.lth.se Introduction Agumented Reality (AR) is the overlay of virtual computer graphics images
More informationStereo and Epipolar geometry
Previously Image Primitives (feature points, lines, contours) Today: Stereo and Epipolar geometry How to match primitives between two (multiple) views) Goals: 3D reconstruction, recognition Jana Kosecka
More informationRegistration of Dynamic Range Images
Registration of Dynamic Range Images Tan-Chi Ho 1,2 Jung-Hong Chuang 1 Wen-Wei Lin 2 Song-Sun Lin 2 1 Department of Computer Science National Chiao-Tung University 2 Department of Applied Mathematics National
More informationNavigation System for ACL Reconstruction Using Registration between Multi-Viewpoint X-ray Images and CT Images
Navigation System for ACL Reconstruction Using Registration between Multi-Viewpoint X-ray Images and CT Images Mamoru Kuga a*, Kazunori Yasuda b, Nobuhiko Hata a, Takeyoshi Dohi a a Graduate School of
More informationHigh-speed Three-dimensional Mapping by Direct Estimation of a Small Motion Using Range Images
MECATRONICS - REM 2016 June 15-17, 2016 High-speed Three-dimensional Mapping by Direct Estimation of a Small Motion Using Range Images Shinta Nozaki and Masashi Kimura School of Science and Engineering
More informationIterative Closest Point Algorithm in the Presence of Anisotropic Noise
Iterative Closest Point Algorithm in the Presence of Anisotropic Noise L. Maier-Hein, T. R. dos Santos, A. M. Franz, H.-P. Meinzer German Cancer Research Center, Div. of Medical and Biological Informatics
More informationof Iterative Closest Point ICP algorithm to
Iterative Closest Point Algorithm. L Error Approach via Linear Programming Yaroslav O. Halchenko PhD Student, CS@NJIT yh42@njit.edu Abstract The original goal of this project was a straightforward implementation
More information3D Shape Registration using Regularized Medial Scaffolds
3D Shape Registration using Regularized Medial Scaffolds 3DPVT 2004 Thessaloniki, Greece Sep. 6-9, 2004 Ming-Ching Chang Frederic F. Leymarie Benjamin B. Kimia LEMS, Division of Engineering, Brown University
More informationStructured Light II. Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov
Structured Light II Johannes Köhler Johannes.koehler@dfki.de Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov Introduction Previous lecture: Structured Light I Active Scanning Camera/emitter
More informationA Robust and Efficient Motion Segmentation Based on Orthogonal Projection Matrix of Shape Space
A Robust and Efficient Motion Segmentation Based on Orthogonal Projection Matrix of Shape Space Naoyuki ICHIMURA Electrotechnical Laboratory 1-1-4, Umezono, Tsukuba Ibaraki, 35-8568 Japan ichimura@etl.go.jp
More informationADAPTIVE GRAPH CUTS WITH TISSUE PRIORS FOR BRAIN MRI SEGMENTATION
ADAPTIVE GRAPH CUTS WITH TISSUE PRIORS FOR BRAIN MRI SEGMENTATION Abstract: MIP Project Report Spring 2013 Gaurav Mittal 201232644 This is a detailed report about the course project, which was to implement
More informationSubpixel Corner Detection Using Spatial Moment 1)
Vol.31, No.5 ACTA AUTOMATICA SINICA September, 25 Subpixel Corner Detection Using Spatial Moment 1) WANG She-Yang SONG Shen-Min QIANG Wen-Yi CHEN Xing-Lin (Department of Control Engineering, Harbin Institute
More informationShweta Gandhi, Dr.D.M.Yadav JSPM S Bhivarabai sawant Institute of technology & research Electronics and telecom.dept, Wagholi, Pune
Face sketch photo synthesis Shweta Gandhi, Dr.D.M.Yadav JSPM S Bhivarabai sawant Institute of technology & research Electronics and telecom.dept, Wagholi, Pune Abstract Face sketch to photo synthesis has
More informationTensor Sparse PCA and Face Recognition: A Novel Approach
Tensor Sparse PCA and Face Recognition: A Novel Approach Loc Tran Laboratoire CHArt EA4004 EPHE-PSL University, France tran0398@umn.edu Linh Tran Ho Chi Minh University of Technology, Vietnam linhtran.ut@gmail.com
More information3D face recognition based on a modified ICP method
University of Wollongong Research Online Faculty of Informatics - Papers (Archive) Faculty of Engineering and Information Sciences 2011 3D face recognition based on a modified ICP method Kankan Zhao University
More informationMethod of Background Subtraction for Medical Image Segmentation
Method of Background Subtraction for Medical Image Segmentation Seongjai Kim Department of Mathematics and Statistics, Mississippi State University Mississippi State, MS 39762, USA and Hyeona Lim Department
More informationA 3D Point Cloud Registration Algorithm based on Feature Points
International Conference on Information Sciences, Machinery, Materials and Energy (ICISMME 2015) A 3D Point Cloud Registration Algorithm based on Feature Points Yi Ren 1, 2, a, Fucai Zhou 1, b 1 School
More informationImprovement 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 informationAutomated Lesion Detection Methods for 2D and 3D Chest X-Ray Images
Automated Lesion Detection Methods for 2D and 3D Chest X-Ray Images Takeshi Hara, Hiroshi Fujita,Yongbum Lee, Hitoshi Yoshimura* and Shoji Kido** Department of Information Science, Gifu University Yanagido
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 informationSilhouette-based Multiple-View Camera Calibration
Silhouette-based Multiple-View Camera Calibration Prashant Ramanathan, Eckehard Steinbach, and Bernd Girod Information Systems Laboratory, Electrical Engineering Department, Stanford University Stanford,
More informationEE795: Computer Vision and Intelligent Systems
EE795: Computer Vision and Intelligent Systems Spring 2012 TTh 17:30-18:45 FDH 204 Lecture 14 130307 http://www.ee.unlv.edu/~b1morris/ecg795/ 2 Outline Review Stereo Dense Motion Estimation Translational
More informationMotion. 1 Introduction. 2 Optical Flow. Sohaib A Khan. 2.1 Brightness Constancy Equation
Motion Sohaib A Khan 1 Introduction So far, we have dealing with single images of a static scene taken by a fixed camera. Here we will deal with sequence of images taken at different time intervals. Motion
More informationGeneric Face Alignment Using an Improved Active Shape Model
Generic Face Alignment Using an Improved Active Shape Model Liting Wang, Xiaoqing Ding, Chi Fang Electronic Engineering Department, Tsinghua University, Beijing, China {wanglt, dxq, fangchi} @ocrserv.ee.tsinghua.edu.cn
More informationSkew Detection and Correction of Document Image using Hough Transform Method
Skew Detection and Correction of Document Image using Hough Transform Method [1] Neerugatti Varipally Vishwanath, [2] Dr.T. Pearson, [3] K.Chaitanya, [4] MG JaswanthSagar, [5] M.Rupesh [1] Asst.Professor,
More informationRenyan Ge and David A. Clausi
MORPHOLOGICAL SKELETON ALGORITHM FOR PDP PRODUCTION LINE INSPECTION Renyan Ge and David A. Clausi Systems Design Engineering University of Waterloo, 200 University Avenue West Waterloo, Ontario, Canada
More informationIRIS SEGMENTATION OF NON-IDEAL IMAGES
IRIS SEGMENTATION OF NON-IDEAL IMAGES William S. Weld St. Lawrence University Computer Science Department Canton, NY 13617 Xiaojun Qi, Ph.D Utah State University Computer Science Department Logan, UT 84322
More informationLow Cost Motion Capture
Low Cost Motion Capture R. Budiman M. Bennamoun D.Q. Huynh School of Computer Science and Software Engineering The University of Western Australia Crawley WA 6009 AUSTRALIA Email: budimr01@tartarus.uwa.edu.au,
More informationIdentifying and Reading Visual Code Markers
O. Feinstein, EE368 Digital Image Processing Final Report 1 Identifying and Reading Visual Code Markers Oren Feinstein, Electrical Engineering Department, Stanford University Abstract A visual code marker
More informationPROSTATE CANCER DETECTION USING LABEL IMAGE CONSTRAINED MULTIATLAS SELECTION
PROSTATE CANCER DETECTION USING LABEL IMAGE CONSTRAINED MULTIATLAS SELECTION Ms. Vaibhavi Nandkumar Jagtap 1, Mr. Santosh D. Kale 2 1 PG Scholar, 2 Assistant Professor, Department of Electronics and Telecommunication,
More informationAn automatic correction of Ma s thinning algorithm based on P-simple points
Author manuscript, published in "Journal of Mathematical Imaging and Vision 36, 1 (2010) 54-62" DOI : 10.1007/s10851-009-0170-1 An automatic correction of Ma s thinning algorithm based on P-simple points
More informationDEFORMABLE MATCHING OF HAND SHAPES FOR USER VERIFICATION. Ani1 K. Jain and Nicolae Duta
DEFORMABLE MATCHING OF HAND SHAPES FOR USER VERIFICATION Ani1 K. Jain and Nicolae Duta Department of Computer Science and Engineering Michigan State University, East Lansing, MI 48824-1026, USA E-mail:
More informationHybrid Approach for MRI Human Head Scans Classification using HTT based SFTA Texture Feature Extraction Technique
Volume 118 No. 17 2018, 691-701 ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version) url: http://www.ijpam.eu ijpam.eu Hybrid Approach for MRI Human Head Scans Classification using HTT
More informationResearch and application of volleyball target tracking algorithm based on surf corner detection
Acta Technica 62 No. 3A/217, 187 196 c 217 Institute of Thermomechanics CAS, v.v.i. Research and application of volleyball target tracking algorithm based on surf corner detection Guowei Yuan 1 Abstract.
More informationResearch on QR Code Image Pre-processing Algorithm under Complex Background
Scientific Journal of Information Engineering May 207, Volume 7, Issue, PP.-7 Research on QR Code Image Pre-processing Algorithm under Complex Background Lei Liu, Lin-li Zhou, Huifang Bao. Institute of
More informationOn Skeletons Attached to Grey Scale Images. Institute for Studies in Theoretical Physics and Mathematics Tehran, Iran ABSTRACT
On Skeletons Attached to Grey Scale Images M. Karimi Behbahani, Arash Rafiey, 2 Mehrdad Shahshahani 3 Institute for Studies in Theoretical Physics and Mathematics Tehran, Iran ABSTRACT In [2], [3] and
More informationRecognition-based Segmentation of Nom Characters from Body Text Regions of Stele Images Using Area Voronoi Diagram
Author manuscript, published in "International Conference on Computer Analysis of Images and Patterns - CAIP'2009 5702 (2009) 205-212" DOI : 10.1007/978-3-642-03767-2 Recognition-based Segmentation of
More informationShape Similarity Measurement for Boundary Based Features
Shape Similarity Measurement for Boundary Based Features Nafiz Arica 1 and Fatos T. Yarman Vural 2 1 Department of Computer Engineering, Turkish Naval Academy 34942, Tuzla, Istanbul, Turkey narica@dho.edu.tr
More informationGraph Matching Iris Image Blocks with Local Binary Pattern
Graph Matching Iris Image Blocs with Local Binary Pattern Zhenan Sun, Tieniu Tan, and Xianchao Qiu Center for Biometrics and Security Research, National Laboratory of Pattern Recognition, Institute of
More informationDetecting Multiple Symmetries with Extended SIFT
1 Detecting Multiple Symmetries with Extended SIFT 2 3 Anonymous ACCV submission Paper ID 388 4 5 6 7 8 9 10 11 12 13 14 15 16 Abstract. This paper describes an effective method for detecting multiple
More informationMeasurement of 3D Foot Shape Deformation in Motion
Measurement of 3D Foot Shape Deformation in Motion Makoto Kimura Masaaki Mochimaru Takeo Kanade Digital Human Research Center National Institute of Advanced Industrial Science and Technology, Japan The
More informationTask analysis based on observing hands and objects by vision
Task analysis based on observing hands and objects by vision Yoshihiro SATO Keni Bernardin Hiroshi KIMURA Katsushi IKEUCHI Univ. of Electro-Communications Univ. of Karlsruhe Univ. of Tokyo Abstract In
More informationCombining Appearance and Topology for Wide
Combining Appearance and Topology for Wide Baseline Matching Dennis Tell and Stefan Carlsson Presented by: Josh Wills Image Point Correspondences Critical foundation for many vision applications 3-D reconstruction,
More information3D Environment Reconstruction
3D Environment Reconstruction Using Modified Color ICP Algorithm by Fusion of a Camera and a 3D Laser Range Finder The 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems October 11-15,
More informationComputer Science 426 Midterm 3/11/04, 1:30PM-2:50PM
NAME: Login name: Computer Science 46 Midterm 3//4, :3PM-:5PM This test is 5 questions, of equal weight. Do all of your work on these pages (use the back for scratch space), giving the answer in the space
More informationModels and The Viewing Pipeline. Jian Huang CS456
Models and The Viewing Pipeline Jian Huang CS456 Vertex coordinates list, polygon table and (maybe) edge table Auxiliary: Per vertex normal Neighborhood information, arranged with regard to vertices and
More informationChapter 3 Image Registration. Chapter 3 Image Registration
Chapter 3 Image Registration Distributed Algorithms for Introduction (1) Definition: Image Registration Input: 2 images of the same scene but taken from different perspectives Goal: Identify transformation
More informationSample Based Texture extraction for Model based coding
DEPARTMENT OF APPLIED PHYSICS AND ELECTRONICS UMEÅ UNIVERISTY, SWEDEN DIGITAL MEDIA LAB Sample Based Texture extraction for Model based coding Zhengrong Yao 1 Dept. Applied Physics and Electronics Umeå
More informationAC : EDGE DETECTORS IN IMAGE PROCESSING
AC 11-79: EDGE DETECTORS IN IMAGE PROCESSING John Schmeelk, Virginia Commonwealth University/Qatar Dr. John Schmeelk is a Professor of mathematics at Virginia Commonwealth University teaching mathematics
More informationHand-Eye Calibration from Image Derivatives
Hand-Eye Calibration from Image Derivatives Abstract In this paper it is shown how to perform hand-eye calibration using only the normal flow field and knowledge about the motion of the hand. The proposed
More informationFace Recognition Based on LDA and Improved Pairwise-Constrained Multiple Metric Learning Method
Journal of Information Hiding and Multimedia Signal Processing c 2016 ISSN 2073-4212 Ubiquitous International Volume 7, Number 5, September 2016 Face Recognition ased on LDA and Improved Pairwise-Constrained
More informationTEXTURE OVERLAY ONTO NON-RIGID SURFACE USING COMMODITY DEPTH CAMERA
TEXTURE OVERLAY ONTO NON-RIGID SURFACE USING COMMODITY DEPTH CAMERA Tomoki Hayashi 1, Francois de Sorbier 1 and Hideo Saito 1 1 Graduate School of Science and Technology, Keio University, 3-14-1 Hiyoshi,
More informationGrasp Recognition using a 3D Articulated Model and Infrared Images
Grasp Recognition using a 3D Articulated Model and Infrared Images Koichi Ogawara Institute of Industrial Science, Univ. of Tokyo, Tokyo, Japan Jun Takamatsu Institute of Industrial Science, Univ. of Tokyo,
More informationQuaternion-based color difference measure for removing impulse noise in color images
2014 International Conference on Informative and Cybernetics for Computational Social Systems (ICCSS) Quaternion-based color difference measure for removing impulse noise in color images Lunbo Chen, Yicong
More informationImage Thickness Correction for Navigation with 3D Intra-cardiac Ultrasound Catheter
Image Thickness Correction for Navigation with 3D Intra-cardiac Ultrasound Catheter Hua Zhong 1, Takeo Kanade 1,andDavidSchwartzman 2 1 Computer Science Department, Carnegie Mellon University, USA 2 University
More informationDeformable 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 informationFiltering of impulse noise in digital signals using logical transform
Filtering of impulse noise in digital signals using logical transform Ethan E. Danahy* a, Sos S. Agaian** b, Karen A. Panetta*** a a Dept. of Electrical and Computer Eng., Tufts Univ., 6 College Ave.,
More informationAn Efficient Character Segmentation Based on VNP Algorithm
Research Journal of Applied Sciences, Engineering and Technology 4(24): 5438-5442, 2012 ISSN: 2040-7467 Maxwell Scientific organization, 2012 Submitted: March 18, 2012 Accepted: April 14, 2012 Published:
More informationDigital Image Processing
Digital Image Processing Third Edition Rafael C. Gonzalez University of Tennessee Richard E. Woods MedData Interactive PEARSON Prentice Hall Pearson Education International Contents Preface xv Acknowledgments
More informationAutomated Digital Conversion of Hand-Drawn Plots
Automated Digital Conversion of Hand-Drawn Plots Ruo Yu Gu Department of Electrical Engineering Stanford University Palo Alto, U.S.A. ruoyugu@stanford.edu Abstract An algorithm has been developed using
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 informationWhole Body MRI Intensity Standardization
Whole Body MRI Intensity Standardization Florian Jäger 1, László Nyúl 1, Bernd Frericks 2, Frank Wacker 2 and Joachim Hornegger 1 1 Institute of Pattern Recognition, University of Erlangen, {jaeger,nyul,hornegger}@informatik.uni-erlangen.de
More informationEstimation of common groundplane based on co-motion statistics
Estimation of common groundplane based on co-motion statistics Zoltan Szlavik, Laszlo Havasi 2, Tamas Sziranyi Analogical and Neural Computing Laboratory, Computer and Automation Research Institute of
More informationBased on Regression Diagnostics
Automatic Detection of Region-Mura Defects in TFT-LCD Based on Regression Diagnostics Yu-Chiang Chuang 1 and Shu-Kai S. Fan 2 Department of Industrial Engineering and Management, Yuan Ze University, Tao
More informationCamera Calibration. Schedule. Jesus J Caban. Note: You have until next Monday to let me know. ! Today:! Camera calibration
Camera Calibration Jesus J Caban Schedule! Today:! Camera calibration! Wednesday:! Lecture: Motion & Optical Flow! Monday:! Lecture: Medical Imaging! Final presentations:! Nov 29 th : W. Griffin! Dec 1
More informationStereo Vision. MAN-522 Computer Vision
Stereo Vision MAN-522 Computer Vision What is the goal of stereo vision? The recovery of the 3D structure of a scene using two or more images of the 3D scene, each acquired from a different viewpoint in
More informationPHOTOGRAMMETRIC TECHNIQUE FOR TEETH OCCLUSION ANALYSIS IN DENTISTRY
PHOTOGRAMMETRIC TECHNIQUE FOR TEETH OCCLUSION ANALYSIS IN DENTISTRY V. A. Knyaz a, *, S. Yu. Zheltov a, a State Research Institute of Aviation System (GosNIIAS), 539 Moscow, Russia (knyaz,zhl)@gosniias.ru
More informationInternational Journal of Signal Processing, Image Processing and Pattern Recognition Vol.9, No.2 (2016) Figure 1. General Concept of Skeletonization
Vol.9, No.2 (216), pp.4-58 http://dx.doi.org/1.1425/ijsip.216.9.2.5 Skeleton Generation for Digital Images Based on Performance Evaluation Parameters Prof. Gulshan Goyal 1 and Ritika Luthra 2 1 Associate
More informationEfficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest.
Efficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest. D.A. Karras, S.A. Karkanis and D. E. Maroulis University of Piraeus, Dept.
More information2. TARGET PHOTOS FOR ANALYSIS
Proceedings of the IIEEJ Image Electronics and Visual Computing Workshop 2012 Kuching, Malaysia, November 21-24, 2012 QUANTITATIVE SHAPE ESTIMATION OF HIROSHIMA A-BOMB MUSHROOM CLOUD FROM PHOTOS Masashi
More informationCHAPTER 1 INTRODUCTION
CHAPTER 1 INTRODUCTION 1.1 Introduction Pattern recognition is a set of mathematical, statistical and heuristic techniques used in executing `man-like' tasks on computers. Pattern recognition plays an
More informationReconstruction of complete 3D object model from multi-view range images.
Header for SPIE use Reconstruction of complete 3D object model from multi-view range images. Yi-Ping Hung *, Chu-Song Chen, Ing-Bor Hsieh, Chiou-Shann Fuh Institute of Information Science, Academia Sinica,
More informationA Connection between Network Coding and. Convolutional Codes
A Connection between Network Coding and 1 Convolutional Codes Christina Fragouli, Emina Soljanin christina.fragouli@epfl.ch, emina@lucent.com Abstract The min-cut, max-flow theorem states that a source
More information