Image texture features based on gray-level co-occurence matrices (GLCMs) are useful in e.g. the analysis of MR images of tumours. However, the features can be quite sensitive to the number of grey-levels in the analysed image, in particular if the region of interest is small. In this work we propose a new method for computing the GLCM, based on Gaussian mixture models. The results show that the new method improves the estimation of the GLCM and at the same time eliminates the difficult task of selecting the number of grey-levels.
This abstract and the presentation materials are available to members only; a login is required.