Texture analysis using the trace transform

Texture analysis plays an important role in image analysis and pattern recognition. Also feature extraction is one of the most important tasks for efficient and accurate image retrieval purpose. This work concerns the analysis of textures and feature extraction. It deals with textures which are regularly shaped and sampled, irregularly sampled, and irregularly shaped. The Trace transform is used to extract thousands of features which can be investigated to identify those which are most relevant to the task. Also texture is widely used in content based image retrieval and there have been a number of studies over the years to establish which features are perceptually significant. However it is still difficult to retrieve reliably images that the human user would agree that are similar. In this work perceptual grouping and finding the most related features to human texture ranking are discussed. The results of using the Trace transform are compared with 10 other methods mainly based on Co-occurrence matrices and the Sum and difference histograms. Most image processing techniques assume that the image is represented by a rectangular grid of sampling points. This, however, need not be the case. The regularity of sampling is particularly important for texture analysis, where the relative spatial arrangement of the pixels is of paramount importance. In this work we investigate the way of using the Trace transform to recognise textures from irregularly sampled data. The Hough transform is used as an interface that allows us to identify tracing lines in the image and normalise convolution allows us to deal with the irregularly placed samples along the tracing lines in order to compute the trace functionals. In all cases we investigate and develop the use of the Trace transform and its functionals.