BDBComp
Parceria:
SBC
Avaliação de Classificadores para o Reconhecimento Automático de Insetos

Vinícius M. A. de SouzaDiego F. SilvaPedro R. P. GarciaGustavo E. A. P. A. Batista

An important application in public health are the intelligent traps able to selectively capture insect species of interest, such as disease vectors, without affecting other beneficial species to the environment. The implementation on such trap requires the development of a sensor able to automatically detect the species of the insects that enter the trap. Recently, we proposed a sensor that uses lasers and machine learning techniques to automatically classify insects. Aiming to guide the choice of classifiers to be embedded in the intelligent trap, this work presents an experimental evaluation of several classifiers to automatically recognize insect species using the sensor data. Our results indicate that a simple kNN classifier outperforms the proposed baseline and is competitive with more sophisticated techniques such as SVM and GMM classifiers.

http://www.lbd.dcc.ufmg.br/colecoes/eniac/2013/005.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD