Theoretical background and related works

• Autor(a) principal: Afonso, Luis C.S. [UNESP]
• Data de Publicação: 2022
• Outros Autores: Falcão, Alexandre Xavier, Papa, João Paulo [UNESP]
• Tipo de documento: Conjunto de dados
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP (dados de pesquisa)
• Texto Completo: http://dx.doi.org/10.1016/B978-0-12-822688-9.00010-4; http://hdl.handle.net/11449/242083
• topic: Image-forest transform; Machine learning; Optimum-path forest; Pattern recognition

The Optimum-Path Forest (OPF) is a framework for the design of graph-based classifiers, which covers supervised, semisupervised, and unsupervised applications. The OPF is mainly characterized by its low training and classification times as well as competitive results against well-established machine learning techniques, such as Support Vector Machine and Artificial Neural Networks. Besides, the framework allows the design of different approaches based on the problem itself, which means a specific OPF-based classifier can be built for a given particular task. This paper surveyed several works published in the past years concerning OPF-based classifiers and sheds light on future trends concerning such a framework in the context of the deep learning era. © 2022 Copyright