Publication: Understanding the learned behavior of customized convolutional neural networks toward malaria parasite detection in thin blood smear images
Issued Date
2018-07-01
Resource Type
ISSN
23294310
23294302
23294302
Other identifier(s)
2-s2.0-85050652206
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
Journal of Medical Imaging. Vol.5, No.3 (2018)
Suggested Citation
Sivaramakrishnan Rajaraman, Kamolrat Silamut, Md A. Hossain, I. Ersoy, Richard J. Maude, Stefan Jaeger, George R. Thoma, Sameer K. Antani Understanding the learned behavior of customized convolutional neural networks toward malaria parasite detection in thin blood smear images. Journal of Medical Imaging. Vol.5, No.3 (2018). doi:10.1117/1.JMI.5.3.034501 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/46570
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Understanding the learned behavior of customized convolutional neural networks toward malaria parasite detection in thin blood smear images
Abstract
© 2018 Society of Photo-Optical Instrumentation Engineers (SPIE). Convolutional neural networks (CNNs) have become the architecture of choice for visual recognition tasks. However, these models are perceived as black boxes since there is a lack of understanding of the learned behavior from the underlying task of interest. This lack of transparency is a serious drawback, particularly in applications involving medical screening and diagnosis since poorly understood model behavior could adversely impact subsequent clinical decision-making. Recently, researchers have begun working on this issue and several methods have been proposed to visualize and understand the behavior of these models. We highlight the advantages offered through visualizing and understanding the weights, saliencies, class activation maps, and region of interest localizations in customized CNNs applied to the challenge of classifying parasitized and uninfected cells to aid in malaria screening. We provide an explanation for the models' classification decisions. We characterize, evaluate, and statistically validate the performance of different customized CNNs keeping every training subject's data separate from the validation set.