Chen, Y.Q., Nixon, M.S. and Thomas, D.W.
Satistical Geometrical Texture Description.
Full text not available from this repository.
Texture plays an important role in image analysis and understanding. Its potential applications include remote sensing, quality control, and medical diagnosis etc. As a front end in a typical classification system, texture feature extraction is of key significance to the overall system performance. There have been many papers, proposing various approaches to this challenging problem. Structural approaches are based on the theory of formal languages: a texture image is regarded as generated from a set of texture primitives using a set of placement rules. These approaches work well on ``deterministic'' textures but most natural textures, unfortunately, are not of this type. From a statistical point of view, texture images are complicated pictorial patterns on which sets of statistics can be defined to characterise these patterns. Aside from the most popularly used Spatial Grey Level Dependence Matrix (SGLDM), there are also other statistics such as the recently proposed Statistical Feature Matrix (SFM). These statistics, however, are largely heuristic, resulting in limited discrimination ability. Fourier transform based methods usually perform well on textures showing strong periodicity. Their performance significantly deteriorates, though, when the periodicity weakens. Stochastic models such as two-dimensional ARMA, Markov random fields etc. can also be used for texture feature extraction via parameter estimation. These approaches consider textures as realisations of a random process. We have developed a novel set of texture features - Statistical Geometrical Features (SGF) - based on the statistics of geometrical properties of connected regions in a stack of binary images obtained from a texture image.
Actions (login required)