Recognition and registration methods

A part of the CVonline computer vision resource summarizing different approaches to structure registration and classification as commonly used in computer vision and image processing.

1. Statistical classification methods

   1. Bayesian classifier

   2. Contextual image classification

   3. Decision trees

   4. Learning classifier systems

   5. Feature-based

   6. Fuzzy classification

   7. Hough forest

   8. k-nearest neighbor classification

   9. Linear and higher order discriminant functions

   10. Markov random field based classification

   11. Minimum distance estimation

   12. Multi-classifier fusion

   13. Neural networks

   14. Random Forest

   15. Sparse coding

   16. Support vector machines and two-sided SVMs

   17. Vector quantization based classification

2. General reasoning methods used in vision

• 1. Ambiguous images

3. Geometric model matching, Feature correspondence, Shape correspondence

• 1. General matching control and search algorithms

     1. Hypothesize and test

     2. Heuristic search

     3. Recognition by components

     4. Interpretation trees and other search tree variations

     5. Rule-based systems

     6. Theorem proving

     7. RAST algorithms

  2. General recognition methods

• • 1. Appearance-based, Iconic, View-based recognition

    2. Boltzmann machine, Hopfield net, Simulated annealing

    3. Constraint-based matching

    4. Context-based matching

    5. Template matching

       1. Deformable/elastic template matching

       2. Pictorial structures

       3. Sub-template matching

    6. Geometric hashing

    7. Image stitching

    8. Image registration

    9. Inverse compositional method

    10. Iterated closest point/Iterative closest point (ICP)

• • • 1. Non-Rigid ICP

• • 11. MAPSAC

    12. MLESAC

    13. Model based recognition

    14. Multi-scale contour segmentation

    15. Non-rigid alignment (See also Introduction)

    16. Object categorization representation

• • • 1. Cognitive neuroscience of visual object recognition

      2. Bag of words

      3. Fisher Vector

      4. Structured models

      5. Subcategory recognition

      6. Vocabulary trees

• • 17. Object Size/Scale/distance estimation

    18. Polygon matching

    19. Pose clustering

    20. RANSAC

    21. Relaxation labeling

    22. Softassign algorithm

    23. Structural description

    24. Template and cross-correlation matching

    25. View-based object recognition

4. Identity verification/alignment

• 1. Geometric feature proximity

  2. Pose consistency

5. Model based indexing, invocation

6. Special feature matching

• 1. 2D to 2D point feature matching

     1. Proximity matrices

  2. 2D to 3D point feature matching

  3. 3D to 3D point feature matching

  4. 2D, 3D point to structure matching

  5. Aspect graph matching

  6. Boundary/contour/curve/edge matching

• • 1. TERCOM

• 7. Line matching

  8. Phase matching

  9. Property-based matching

  10. Optical/Appearance flow matching

  11. Polygon matching

  12. Needle-map matching

  13. Region matching

  14. Spatial pyramid matching

  15. Surface matching

  16. Texture classification

  17. Texture matching

  18. Volume matching

7. Syntactic pattern matching

• 1. 1D/String matching

  2. 2D/Pattern grammars