Texture analysis plays an important role in machine vision and pattern recognition. Along with the emergence of artificial intelligence (AI) comes an increase in applications requiring image texture analysis. Deep learning (DL)-based approaches are also gaining attention. A class of neural networks known as convolutional neural networks (CNNs) is a popular choice for DL-based machine vision applications.

This book presents various CNN-based approaches, in addition to covering the fundamentals and k-views-based image texture classification and analysis approaches. Its ten chapters are arranged in three parts.

Part 1’s four chapters present fundamentals, various texture features and models, and how those features and models are used for classification and dimensionality reduction. In all four chapters, the authors provide important basics such as spatial domain features, autocorrelation-based models, Markov random field-based models, Gabor and wavelet-based features and models, k-means approaches, and principal component analysis (PCA)-based approaches. These approaches are clearly presented with mathematical equations, illustrations, and experimental results.

Part 2 also has four chapters, all related to k-views-based approaches, from the basic k-views model to advanced models. All these models are neatly presented with step-by-step approaches, illustrations, and experimental results. The authors also stress the importance of rotation invariant approaches.

Part 3 (two chapters) is on CNN-based DL approaches. The first chapter of this part presents the basics of DL. The second chapter covers CNN-based approaches. Both chapters include illustrations and experimental results.

All ten chapters list useful references, and most chapters include exercises.

Overall, the authors attempt to provide a comprehensive look at texture analysis, from basic to advanced approaches. This book will be useful for advanced undergraduate and postgraduate computer science students, as well as researchers working in machine vision and pattern recognition.