Contagem Semiautomatizada de Esporos de Fungos Micorrízicos Arbusculares Usando Redes Neurais Artificiais (Semi-Automated Counting of Arbuscular Mycorrhizal Fungi Spores Using Artificial Neural Network)

Clênia Andrade Oliveira Melo (clenia@uesb.edu.br)1, Juliane Gonçalves Lopes (enailuj.sepol@gmail.com)2, Alexsandra Oliveira Andrade (alexsandra@uesb.edu.br)2, Roque Mendes Prado Trindade (roquetrindade@uesb.edu.br)2, Robson Silva Magalhães (robsonmagalhaes@ufsb.edu.br)3


1Universidade Federal da Bahia
2Universidade Estadual do Sudoeste da Bahia
3Universidade Federal do Sul da Bahia

This paper appears in: Revista IEEE América Latina

Publication Date: Aug. 2017
Volume: 15,   Issue: 8 
ISSN: 1548-0992


Abstract:
Arbuscular Mycorrhizae (AM) are mutualistic associations between Arbuscular Mycorrhizal Fungi (AMF) and the roots of most plant species. The AMF spores originate filaments called hyphae that contribute to the absorption of water and nutrients, especially phosphorus. This paper presents a system that makes automatic counting of those fungi spores in slide images possible. The system facilitate and accelerate the task for specialists, since nowadays they still use manual counting. We built a database of spore images, and processed each of them through an artificial neural network to classify patterns automatically. We compared the results of our counting model based in Perceptron Artificial Neural Network with those of the Fuzzy Morphology method, using the same database. The agreement of the proposed counting model with the manual counting have been better than that of Fuzzy Morphology.

Index Terms:
Mycorrhizal Fungi Spores, Semi-Automated Counting, Artificial Neural Networks, Fuzzy Morphology, Image Process.   


Documents that cite this document
This function is not implemented yet.


[PDF Full-Text (592)]