Introduction

Abstract

Pattern recognition techniques have been consistently applied in various domains, ranging from remote sensing and medicine to engineering, among others. The literature is vast and dominated mainly by Neural Networks in the past, followed by Support Vector Machines, and recently by the so-called Deep Learning approaches. However, there is always room for improvements and novel techniques, for there is no approach that can lead to the best results in all situations. This chapter introduces this book, which concerns the Optimum-Path Forest, a framework for designing graph-based classifiers based on optimum connectivity among samples. We highlight new ideas and applications throughout the book, and future trends will foster the related literature in the following years. © 2022 Copyright