Texture Classification

For adding features that could be used for the synapse classification, we will be looking into methods that are used for texture classification. Texture classification is one of the important components of texture analysis. The other components of texture analysis are texture segmentation, texture synthesis, shape identification from texture.

Texture classification is done by different types of Techniques. The major types are the following:

1. Statistical
1. Co-occurrence
2. Angular Second Moment
3. Contrast
4. Correlation
5. Entropy
   
2. Geometrical
   
3. Structural
4. Model-based
1. Multi-resolution autoregressive
   
5. Signal processing features
1. Heuristically designed filter banks
1. Laws Filter Masks
2. Ring and Wedge Filters
3. Dyadic Gabor Filter bank
4. Wavelet transfer, Packets and Frames
5. Discrete Cosine Transform
6. Quadrature Mirror Filters
7. Tree structured Gabor Filter bank
2. Optimized filter and filter banks
1. Eigenfilter
2. Prediction error filter
3. Optimized representation Gabor Filter bank
4. Optimized Two-Class Gabor Filter
5. Optimized Multi-Class Gabor Filter
6. Optimized Two-Texture FIR Filters
7. Optimized FIR Filter bank
8. Back Propagation Designed Mask
   

References:

1. Filtering for Texture Classification: A Comparative Study
   
2. http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/OJALA1/texclas.htm