Periodicity, Directionality, and Randomness: Wold Features for Image Modeling and Retrieval
IEEE Transactions on Pattern Analysis and Machine Intelligence
Texture Features for Browsing and Retrieval of Image Data
IEEE Transactions on Pattern Analysis and Machine Intelligence
NETRA: a toolbox for navigating large image databases
NETRA: a toolbox for navigating large image databases
Computer and Robot Vision
A region based representation for image and video retrieval
A region based representation for image and video retrieval
A study for comparative evaluation of the methods for image processing using texture characteristics
WSEAS Transactions on Information Science and Applications
Hi-index | 0.00 |
The paper presents a method for the classification and identification of texture image region. To compare regions, a data base with single texture images is used. The dimension of the reference images is greater than the analyzed region dimension. For the proper region recognition a decision theoretic method and two type of statistic texture feature are used. The first type features are the peak of grey level histogram and the texton contour pixel densities (edge densities) per unit of area. The second type feature derive from the medium co-occurrence matrices: contrast, energy, entropy, homogeneity, and variance. The algorithm are implemented in Visual C++2005 and Matlab and allows the simultaneously display of both the investigated regions pairs, and the euclidian distance between them. Our experimental results indicate the fact that the selected features which derive from medium co-occurrence matrices have a good discriminating power for texture classification. The results also confirm the fact that the distances between the similar regions are relatively small and the distances between regions from different textured images are relatively great.