A new method for wood recognition based on blocked HLAC

Abstract

This paper propose a new method for wood recognition based on texture analysis. At first, wood texture images are divided into several blocks in our method. Secondly, wood features are extracted from these blocked grey-scale images using different mask, which is knows as higher-order local autocorrelation (HLAC). Finally, Support Vector Machine (SVM) was used to verify the performance of the method. Experiments carried on the wood texture database demonstrate that our method outperforms the original HLAC method.