Performance Evaluation for Three Classes of Textural Coarseness

Textural coarseness for textural feature are compared. The problem addressed is to determine which texture feature optimize retrieval rate. Many textural features have been proposed in different papers. No much focused on comparative textural coarseness study has appeared. The goal is compared and evaluating in a quantitative manner three types of textural coarseness, namely gray level co-occurrence textural coarseness, fractal dimension textural coarseness, Tamura textural model. Performance is assessed by the criterion of Human Vision System. Furthermore, a experiment of extraction textural coarseness with a standard Xinjiang Folk Art Patterns databases. The results show Tamura texture model performance of describing coarseness is the best followed fractal dimension. However, there is no universally best performance of textural. In the paper by comparison the performance of different textural feature and give the recommended models.