Название: Modern Metaheuristics in Image Processing
Автор: Diego Oliva, Salvador Hinojosa, Noe Ortega-Sanchez
Издательство: CRC Press
Год: 2022
Страниц: 140
Язык: английский
Формат: pdf (true)
Размер: 10.3 MB

The use of metaheuristic algorithms (MA) has been increasing in recent years, and the image processing field is not the exempted of their application. In the last two years a big amount of MA has been introduced as alternatives for solving complex optimization problems. This book collects the most prominent metaheuristic algorithms and verifies its use in image processing tasks. In addition, literature review of both MA and digital image processing is presented as part of the introductory information. Each algorithm is detailed explained with special focus in the tuning parameters and the proper implementation for the image processing tasks. Besides several examples permits to the reader explore and confirm the use of this kind of intelligent methods.

Since image processing is widely used in different domains, this book considers different kinds of datasets that includes, magnetic resonance images, thermal images, agriculture images, among others. The reader then can have some ideas of implementation that complement the theory exposed of each optimization mechanism. Regarding the image processing problems this book consider the segmentation by using different metrics based on entropies or variances. In the same way, the identification of different shapes and the detection of objects are also covered in the corresponding chapters. Each chapter is complemented with a wide range of experiments and statistical analysis that permits the reader to judge about the performance of the MA. Finally, there is included a section that includes some discussion and conclusions. This section also provides some open questions and research opportunities for the audience.

The Chapter 7 is devoted to showing the code used in the algorithms, and it is analyzed and explained briefly for the enthusiastic reader. The structure given to the programs is organized in a series of files and follows the code’s aim which is to be able to use them in a general way. Keeping this in mind, it is possible to quickly adjust and implement any problem, knowing only the objective function and applying the changes in some conditionals. The codification was developed in MATLAB, and each file has the extension .m.