Project Journal

介紹截取特徵的方法和目的. (for OCT)
目前只看到 chapter 1
因為後面的演算法用了很多 "fourier transformation" 和 "線代"
要去複習才看得懂QQ...

《FEATURE EXTRACTION METHODS FOR CHARACTER RECOGNITION--A SURVEY》 (1995)

目錄

1. Introduction
1.1. lnvariants
1.2. Reconstructability

2. Features Extracted from Gray-scale Images
2.1. Template matching
2.2. Deformable templates
2.3. Unitary image transforms
2.4. Zoning
2.5. Geometric moment invariants
2.6. Zernike moments

3. Features Extracted from Binary Images
3.1. Template matching
3.2. Unitary image transforms
3.3. Projection histograms
3.4. Zoning
3.5. Geometric moment invariants

4. Features Extracted from the Binary Contour
4.1. Contour profiles
4.2. Zoning
4.3. Spline curve approximation
4.4. Elliptic Fourier descriptors
4.5. Other Fourier descriptors
4.6. Evaluation studies

5. Features Extracted from the Vector Presentation
5.1. Template matching
5.2. Deformable templates
5.3. Graph description
5.4. Discrete features
5.5. Zoning
5.6. Fourier descriptors

6. Neural Network Classifiers
7. Discussion
8. Summary