Automated diagnosis of intestinal parasites: A new hybrid approach and its benefits.

Autor: Osaku D; Institute of Computing, University of Campinas, Brazil. Electronic address: danosaku@hotmail.com., Cuba CF; Institute of Computing, University of Campinas, Brazil. Electronic address: carolinacuba23@gmail.com., Suzuki CTN; Institute of Computing, University of Campinas, Brazil. Electronic address: celso.suzuki@gmail.com., Gomes JF; Institute of Computing, University of Campinas, Brazil. Electronic address: jgomes@ic.unicamp.br., Falcão AX; Institute of Computing, University of Campinas, Brazil. Electronic address: afalcao@ic.unicamp.br.
Jazyk: angličtina
Zdroj: Computers in biology and medicine [Comput Biol Med] 2020 Aug; Vol. 123, pp. 103917. Date of Electronic Publication: 2020 Jul 15.
DOI: 10.1016/j.compbiomed.2020.103917
Abstrakt: Intestinal parasites are responsible for several diseases in human beings. In order to eliminate the error-prone visual analysis of optical microscopy slides, we have investigated automated, fast, and low-cost systems for the diagnosis of human intestinal parasites. In this work, we present a hybrid approach that combines the opinion of two decision-making systems with complementary properties: (DS 1 ) a simpler system based on very fast handcrafted image feature extraction and support vector machine classification and (DS 2 ) a more complex system based on a deep neural network, Vgg-16, for image feature extraction and classification. DS 1 is much faster than DS 2 , but it is less accurate than DS 2 . Fortunately, the errors of DS 1 are not the same of DS 2 . During training, we use a validation set to learn the probabilities of misclassification by DS 1 on each class based on its confidence values. When DS 1 quickly classifies all images from a microscopy slide, the method selects a number of images with higher chances of misclassification for characterization and reclassification by DS 2 . Our hybrid system can improve the overall effectiveness without compromising efficiency, being suitable for the clinical routine - a strategy that might be suitable for other real applications. As demonstrated on large datasets, the proposed system can achieve, on average, 94.9%, 87.8%, and 92.5% of Cohen's Kappa on helminth eggs, helminth larvae, and protozoa cysts, respectively.
(Copyright © 2020 Elsevier Ltd. All rights reserved.)
Databáze: MEDLINE