Sample Selection in Textured Images

This paper proposes a texture learning method based on fractal compression and iterated function systems (IFS). This type of approach allows to extract self-similarities between blocks of a given image. The number of similarities for each element yields to a score of each blocks. The first blocks of this rating are considered as representative and are stored in a database in order to establish a learning process. Recognition is made by labeling blocks and pixels of the test image. The blocks of the new image are matched with the ones of the different texture databases. As an application, we used our method to recognize bridges and buildings on ground images.