Your browser doesn't support javascript.
loading
Recognition of Protozoa and Metazoa using image analysis tools, discriminant analysis, neural networks and decision trees.
Ginoris, Y P; Amaral, A L; Nicolau, A; Coelho, M A Z; Ferreira, E C.
Afiliación
  • Ginoris YP; Departamento de Engenharia Bioquímica, Escola de Química/UFRJ, Centro de Tecnologia, E-203, Rio de Janeiro, Brazil. yovankaperez@yahoo.com.br
Anal Chim Acta ; 595(1-2): 160-9, 2007 Jul 09.
Article en En | MEDLINE | ID: mdl-17605996
Protozoa and metazoa are considered good indicators of the treatment quality in activated sludge systems due to the fact that these organisms are fairly sensitive to physical, chemical and operational processes. Therefore, it is possible to establish close relationships between the predominance of certain species or groups of species and several operational parameters of the plant, such as the biotic indices, namely the Sludge Biotic Index (SBI). This procedure requires the identification, classification and enumeration of the different species, which is usually achieved manually implying both time and expertise availability. Digital image analysis combined with multivariate statistical techniques has proved to be a useful tool to classify and quantify organisms in an automatic and not subjective way. This work presents a semi-automatic image analysis procedure for protozoa and metazoa recognition developed in Matlab language. The obtained morphological descriptors were analyzed using discriminant analysis, neural network and decision trees multivariable statistical techniques to identify and classify each protozoan or metazoan. The obtained procedure was quite adequate for distinguishing between the non-sessile protozoa classes and also for the metazoa classes, with high values for the overall species recognition with the exception of sessile protozoa. In terms of the wastewater conditions assessment the obtained results were found to be suitable for the prediction of these conditions. Finally, the discriminant analysis and neural networks results were found to be quite similar whereas the decision trees technique was less appropriate.
Asunto(s)
Buscar en Google
Colección: 01-internacional Base de datos: MEDLINE Asunto principal: Procesamiento de Imagen Asistido por Computador / Árboles de Decisión / Análisis Discriminante / Redes Neurales de la Computación / Eucariontes Tipo de estudio: Health_economic_evaluation / Prognostic_studies Límite: Animals Idioma: En Revista: Anal Chim Acta Año: 2007 Tipo del documento: Article País de afiliación: Brasil Pais de publicación: Países Bajos
Buscar en Google
Colección: 01-internacional Base de datos: MEDLINE Asunto principal: Procesamiento de Imagen Asistido por Computador / Árboles de Decisión / Análisis Discriminante / Redes Neurales de la Computación / Eucariontes Tipo de estudio: Health_economic_evaluation / Prognostic_studies Límite: Animals Idioma: En Revista: Anal Chim Acta Año: 2007 Tipo del documento: Article País de afiliación: Brasil Pais de publicación: Países Bajos