Fast automatic microstructural segmentation of ferrous alloy samples using optimum-path forest

Actas de congresos

Fecha

2010-05-21

Registro en:

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 6026 LNCS, p. 210-220.

0302-9743

1611-3349

http://hdl.handle.net/11449/71689

10.1007/978-3-642-12712-0_19

WOS:000279020400019

2-s2.0-77952364349

9039182932747194

Autor

Universidade Estadual Paulista (Unesp)

Technological Research Center

Universidade Estadual de Campinas (UNICAMP)

Faculty of Engineering

Institución

Universidade Estadual Paulista (Brasil)

Resumen

In this work we propose a novel automatic cast iron segmentation approach based on the Optimum-Path Forest classifier (OPF). Microscopic images from nodular, gray and malleable cast irons are segmented using OPF, and Support Vector Machines (SVM) with Radial Basis Function and SVM without kernel mapping. Results show accurate and fast segmented images, in which OPF outperformed SVMs. Our work is the first into applying OPF for automatic cast iron segmentation. © 2010 Springer-Verlag.

Mostrar el registro completo del ítem