Dissertation > Excellent graduate degree dissertation topics show

Research of Point Set Pattern Matching Based on Spectral Graph Method

Author: WangSongTao
Tutor: PanZuo
School: Wuhan University of Technology
Course: Computer Science and Technology
Keywords: Conversion and the corresponding Point pattern matching Spectral graph theory Neighboring matrix Laplacian matrix Image texture
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 85
Quote: 0
Read: Download Dissertation

Abstract


Point pattern matching computer vision and pattern recognition based on important issues , is currently a hot topic of concern in various fields . It is a widespread problem , not limited to the field of computer vision , Air -bit and posture estimation , target recognition , remote sensing image registration , medical image registration calculation has a wide range of biological and chemical applications . The spectral graph theory is applied to the study of the subject has high computational efficiency , better advantage . By Scott and Longuet-Higgins, Shapiro and Brady both point mode matching the lack of analysis of the spectrum , a new point pattern matching spectrum analysis method . A core part of the method is to construct a new neighboring matrix , the better the results obtained in the case of this method in more substantial image under the affine transformation and the point of jitter in the previous two methods , on the complexity of the algorithm with Spectrum diagram method has the advantage of efficiency than the iterative methods proposed in recent years , some high . The new methods have been more comprehensive testing , including tests on synthetic and real data , we have achieved good results . In this paper, based on previous literature spectral graph theory matching algorithm , make some improvements and refining the framework of an algorithm , the algorithm is divided into two steps , the first step is to choose the combination of distance function , the framework itself provides a group improved distance function , also depending on the application needs additional function . The final product of the distance function for the selected function , this product can be highlighted the role of the distance function for specific matching demand . The second step is to build distance function Gaussian weighted adjacent matrix or Laplacian matrix for the SVD decomposition , matching relation . The advantage of this framework can be generated in accordance with the needs of the specific application of a variety of algorithms have better flexibility and adaptability , implementation , every step of the generated code can be reused , easy programming .

Related Dissertations

  1. On the Ordering of Trees by the Laplacian Coefficients,O157.5
  2. Decomposition Design of Decentralized H_∞ Controllor for a Class of Multiagents System,TP13
  3. Research on Infrared Multi-camera Optical Tracking System for Surgical Robot,TP242
  4. Research on Point Pattern Matching Algorithm Based on Invariant Sequence,TP391.41
  5. Laplacian Spetral Radius and Signless Laplacian Spetral Radius of Graphs,O157.5
  6. Synchronization and Passivity Analysis of Complex Delayed Dynamical Networks,O157.5
  7. Mono-vision Based Aerocraft Pose Measurement Technology,TP391.41
  8. The Minimal Signless Laplacian Spectral Radius of Graphs with Grven Diameter,O157.5
  9. The Critical Group of Some Kinds of Corona,O157.5
  10. The Automatic Road Extraction Technology Based on Fuzzy Connectedness,TP751
  11. On Graphs Determined by Their Spectrum and Angle,O157.5
  12. On the Sandpile Group of Two Kinds of Graphs,O157.5
  13. The spectral radius of connected graphs and diagrams Las pull spectral estimation spectral radius,O157.5
  14. Based on micro-features fingerprint recognition algorithm,TP391.41
  15. The Molecular Dock of Combination Pharmacophore’s Assessment Method and Their Potential Use in the Development of New Drugs,TP391.9
  16. The Algebraic Connectivity of One Type Weighted Graph,O157.5
  17. Research on Algorithms of Point Pattern Matching Based on Graph Spectral Theory,TP391.41
  18. Largest and the second small figure Laplace eigenvalues,O157.5
  19. Research on the Image Matching Algorithm of LED Chip Check and Sorting,TP391.41
  20. Techniques on Interest Points Detection and Image Matching,TP391.4

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile