Authors :
Cristian G. Vizgarra, Matias Roodschild, Jorge Gotay Sardiñas.
Volume/Issue :
Volume 4 - 2019, Issue 6 - June
Google Scholar :
https://goo.gl/DF9R4u
Scribd :
https://bit.ly/2NqxcTh
Abstract :
Several attempts have been made to automatically identify and classify pollen grains in microscopic images using computer algorithms. However, the success of pollen grain recognition depends completely on the determination of the most important features that can be used to describe it. The process of selecting the relevant characteristics is mostly done by the researcher who manually specifies the input characteristics given to the algorithm destined to solve the problem. In this article, three architectures of artificial neural networks have been selected to identify and classify pollen grains digital images without a priori establishment on the set of fundamental characteristics. For this study, eight different types of pollen grains belonging to the native flora of north-western Argentina have been utilized. The results show that the best neural classifier has an effectiveness of 95.03 % for the recognition of the eight pollen grains species. This percentage demonstrates that the methodology applied is satisfactory.
Keywords :
Pollen Grains; Classification; Neural Networks; Supervised Learning.
Several attempts have been made to automatically identify and classify pollen grains in microscopic images using computer algorithms. However, the success of pollen grain recognition depends completely on the determination of the most important features that can be used to describe it. The process of selecting the relevant characteristics is mostly done by the researcher who manually specifies the input characteristics given to the algorithm destined to solve the problem. In this article, three architectures of artificial neural networks have been selected to identify and classify pollen grains digital images without a priori establishment on the set of fundamental characteristics. For this study, eight different types of pollen grains belonging to the native flora of north-western Argentina have been utilized. The results show that the best neural classifier has an effectiveness of 95.03 % for the recognition of the eight pollen grains species. This percentage demonstrates that the methodology applied is satisfactory.
Keywords :
Pollen Grains; Classification; Neural Networks; Supervised Learning.
