Skip to Main Content
Classification of real-world rough surface textures has been an active research topic during the past twenty years. The reason is that changing illumination directions can dramatically affect the appearances of images of these rough surfaces. In this paper, we introduce a simple texture classification method based on support vector machines (SVM) and wavelet packets. Our experiments show that the advantage of using SVM is that it can easily separate different textures using features after transformed in wavelet packet.