Abstract
In this work, we present an ensemble of descriptors for the classification of virus images acquired using transmission electron microscopy. We trained multiple support vector machines on different sets of features extracted from the data. We used both handcrafted algorithms and a pretrained deep neural network as feature extractors. The proposed fusion strongly boosts the performance obtained by each stand-alone approach, obtaining state of the art performance.
Highlights
Recognizing and classifying viruses is fundamental to the medical field for both diagnosis and research. Since this task requires highly qualified medical-staff, there is a growing interest in making this process automatic. Such images can be acquired using electronic microscopy, which are currently not used in clinical practice, and that could be an innovative diagnostic tool
A support vector machines (SVM) is a trainable classifier that finds a separating hyperplane among the samples in a high dimensional space by maximizing the distance between data points and the hyperplane. They are used in two label classification problems, they can be generalized to n class problems by training a set of SVMs, each of which is trained to detect whether a samples belongs to a given class
1500 transmission electron microscopy (TEM) images of size 41 × 41 of viruses belonging to 15 different species (: Adenovirus, Astrovirus, CCHF, Cowpox, Dengue, Ebola, Influenza, Lassa, Marburg, Norovirus, Orf, Papilloma, Rift Valley, Rotavirus, Westnile)
Summary
Recognizing and classifying viruses is fundamental to the medical field for both diagnosis and research. Since this task requires highly qualified medical-staff, there is a growing interest in making this process automatic. Such images can be acquired using electronic microscopy, which are currently not used in clinical practice, and that could be an innovative diagnostic tool. The main difficulty in classifying viruses is their large number, due to the introduction of the DNA sequencing technique that made the number of classified viruses grow exponentially. There are other factors that make the creation of an accurate virus taxonomy very complex: their replication and genetic heritage
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
