A direct way of combining texture and color for image segmentation

In this paper we present a new direct method for image segmentation using texture and color information combined through a multivariate linear discriminant algorithm. The color texture is computed in nine 3 x 3 masks obtained from each 3x3x3 spatio-spectral neighborhood in the image using the classical Haralick and Pressman texture features. Among these 9X 28 texture features the best set was extracted based on training set. The resulting set of 10 feature were used to segment an image into four different regions. The resulting segmentation was compared to classical color and texture segmentation methods using both box classifiers and maximum likelihood classification. It compared favourably on a test image from a Fastred-Lightgreen stained prostatic histological tissue section based on visual inspection. The classification accuracy of 97.5 % for the new method obtained on the training data was the best of the tested methods. If these results hold for a larger set of images this method should be a useful tool for segmenting images where both color and texture is relevant for the segmentation process. Such images are common in biomedicine as well as in remote sensing.