Blood vessels extraction using Fuzzy Mathematical Morphology

Daniel Oliveira Dantas; Diego de Souza Oliveira; Helton Danilo Passos Leal

ICASSP 2017accepted0 citations

Blood vessels extraction using Fuzzy Mathematical Morphology

Daniel Oliveira Dantas, Diego de Souza Oliveira, Helton Danilo Passos Leal

Abstract

Blood vessel extraction from retinography images is useful for detection of many retinopathies. In this paper we propose a way to improve blood vessel detection by use of Fuzzy Mathematical Morphology (FMM) operators. The proposed pipeline, although simple, was found to have the highest accuracy on the STARE dataset, and second highest on the DRIVE dataset. We also present a parallel implementation of the FMM operators, in OpenCL, up to about 500 times faster than their counterpart in C++.

BibTeX

@inproceedings{icassp2017_bloodvesselsextr,
  title = {Blood vessels extraction using Fuzzy Mathematical Morphology},
  author = {Daniel Oliveira Dantas and Diego de Souza Oliveira and Helton Danilo Passos Leal},
  booktitle = {ICASSP 2017},
  year = {2017}
}