Sum and Difference Histograms for Texture Classification
IEEE Transactions on Pattern Analysis and Machine Intelligence
Texture descriptors based on co-occurrence matrices
Computer Vision, Graphics, and Image Processing
| Hi-index | 0.00 |
This article describes the algorithms and the mechatronic system developed for the clustering and classification of marble slabs on production line according to their texture. The method used for the recognition of textures is based on the Sum and Difference Histograms, a faster version of the Co-occurrence Matrices, and the classifier has been implemented by using an LVQ neural network. For each pattern (a marble slab color image), a set of statistical, texture-dependant, parameters is extracted. The input of the classifier (the LVQ network) is the set of parameters calculated before, normalized in the range [0,1], which forms a vector that characterize the pattern shown to the net; the desired output of the network is the class where the pattern belongs to (supervised learning). In our tests, seven different color spaces were used, each one with three different neighbourhoods of pixels. The selected samples chosen for testing the algorithms have been marble slabs of "Crema Marfil Sierra de la Puerta" type. The neural network has been implemented by using MATLAB.