Laws Energy Measure Based on Local Patterns for Texture Classification
This chapter presents the existing Local Binary Pattern (LBP) with its variants and Laws’ mask approach to present a new descriptor. Laws have recommended some labelled vectors that are combined to obtain two-dimensional convolution kernels. When the texture images are convolved with these masks, individual structural components of the image are extracted. The LBP operator that represents together structural and statistical approaches for texture analysis is an extremely versatile operator. Texture contains the vital part of information among all the characteristics present in an image. In texture analysis, while doing the classification, the most important part is to extract texture features by utilizing texture descriptors. A texture analysis using convolutions with different filter masks is named as Laws’ masks. Experimentally it has been proved that many of these are of suitable sizes and provide very useful information for individualizing different types of textures.