Best Feature Selection for Texture Classification

Texture analysis techniques enable to determine the quality of product surfaces measured by image sensors. In previous works, wavelet texture analysis based on the conventional wavelet transform and wavelet packets have been recognized as the most successful technique for classifying steel quality. In this work, we propose a texture classification strategy based on a best feature selection method, which improves classification accuracy. Our proposed methodology has been applied and validated in the classification of the surface quality of rolled steel sheets.