Rotation-invariant texture classification using feature distributions

A distribution-based classification approach and a set of recently developed texture measures are applied to rotation-invariant texture classification. The performance is compared to that obtained with the well-known circular-symmetric autoregressive random field (CSAR) model approach. A difficult classification problem of 15 different Brodatz textures and seven rotation angles is used in experiments. The results show much better performance for our approach than for the CSAR features. A detailed analysis of the confusion matrices and the rotation angles of misclassified samples produces several interesting observations about the classification problem and the features used in this study. Texture analysis Classification Feature distribution Rotation invariant Performance evaluation