Filtering for Texture Classification: A Comparative Study
IEEE Transactions on Pattern Analysis and Machine Intelligence
Pattern Recognition with Fuzzy Objective Function Algorithms
Pattern Recognition with Fuzzy Objective Function Algorithms
Multiresolution Signal Decomposition: Transforms, Subbands, and Wavelets
Multiresolution Signal Decomposition: Transforms, Subbands, and Wavelets
Texture representation based on pattern map
Signal Processing
Practical Genetic Algorithms with CD-ROM
Practical Genetic Algorithms with CD-ROM
Embedded image coding using zerotrees of wavelet coefficients
IEEE Transactions on Signal Processing
A random set view of texture classification
IEEE Transactions on Image Processing
Design-based texture feature fusion using Gabor filters and co-occurrence probabilities
IEEE Transactions on Image Processing
Texture Analysis and Classification With Linear Regression Model Based on Wavelet Transform
IEEE Transactions on Image Processing
Texture analysis and classification with tree-structured wavelet transform
IEEE Transactions on Image Processing
Hi-index | 0.00 |
The wavelet transform is an important analysis used in the field of texture classification. It decomposes an image into subbands. Some of the subbands contain more significant coefficients than others. Based on this property, we propose a texture analysis and classification approach using a combination of the fuzzy C-means clustering method (FCM) and the wavelet transform. By taking the energy coefficients of two pairs of frequency channels resulting from 2D wavelet transform, and grouping the data into a specific number of clusters, we were able to build a feature list for each texture. The feature list is obtained by applying the FCM on each frequency channel pair. The centers obtained are used as the features for every combination of frequency channel pair; the partition matrix generated from the FCM is used as a method for determining the k-nearest neighbors of an unknown texture. The subband effect of the wavelet FCM features is studied by varying the number of decomposition levels of the wavelet tree. Optimal number of features was obtained by varying the number of clusters and the k-nearest neighbors of the FCM. Experiments show that this method outperformed other methods (linear regression model, Gabor transform).