Here are http://year2projectstextureanalysis.blogspot.co.uk/ , you will find the blog of year 2 student group project titled by 'Textures Analysis', by Kejia Wu, Wangting Zhang, Xinyi Li, Jieru Wang and Yang Zhao, which is supervised by Dr Dowsey Andrew.
This project aims at the classification a set of Brodatz texture images by extracting statistical features using MATLAB. Actually, it is a test of the machining learning ability of the program. Texture images could be divided into some parts, and they have their own features. Then the images have to be changed into Gray-level Co-occurrence and more advanced features will be collected. Half of the parts are used as the learning sample for the computer, and the other half parts could be put into the program to test if the program could estimate them as one image.
There are two prediction methods used in this program, which are K-Nearest Neighbor Analysis and Linear Discriminant Analysis. The performance of LDA is better tha the KNN method and the difference results will be shown in the blog. In the end, the program should calculate the percentage of the accuracy.