Training a learning vector quantization network for biomedical classification

Anagnostopoulos, C; Anagnostopoulos, J; Vergados, DD; Kayafas, E; Loumos, V; Theodoropoulos, G

dc.contributor.author	Anagnostopoulos, C	en
dc.contributor.author	Anagnostopoulos, J	en
dc.contributor.author	Vergados, DD	en
dc.contributor.author	Kayafas, E	en
dc.contributor.author	Loumos, V	en
dc.contributor.author	Theodoropoulos, G	en
dc.date.accessioned	2014-03-01T02:42:02Z
dc.date.available	2014-03-01T02:42:02Z
dc.date.issued	2001	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/30738
dc.subject	Image processing	en
dc.subject	Learning vector quantization	en
dc.subject	Neural networks	en
dc.subject	Parasite identification	en
dc.subject.other	Biomedical engineering	en
dc.subject.other	Feature extraction	en
dc.subject.other	Image analysis	en
dc.subject.other	Learning systems	en
dc.subject.other	Medical imaging	en
dc.subject.other	Object recognition	en
dc.subject.other	Relational database systems	en
dc.subject.other	Vector quantization	en
dc.subject.other	Biomedical classification	en
dc.subject.other	Learning vector quantization	en
dc.subject.other	Parasite identification	en
dc.subject.other	Neural networks	en
dc.title	Training a learning vector quantization network for biomedical classification	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/IJCNN.2001.938761	en
heal.identifier.secondary	http://dx.doi.org/10.1109/IJCNN.2001.938761	en
heal.publicationDate	2001	en
heal.abstract	A competitive Learning Vector Quantization (LVQ) Artificial Neural Network (ANN) was trained to identify third stage parasitic strongyle larvae from domestic animals on the basis of quantitative data obtained from processed digital images of larvae. For this reason various novel quantitative features obtained from processed digital images of larvae were tested whether they are variant or invariant to the shape taken by the motile larvae during image recording. A total of 255 images of 57 individual larvae in various shapes belonging to 5 genera were recorded. Following image processing 16 novel features were measured of which 7 were selected as invariant to larva shape. By trial and error two of those novel features ""area"" and ""perimeter"" along with the quantitative features used in conventional identification, ""overall body length"", ""width"" and ""tail of sheath"" were used as an effective training data set for the ANN. This ANN coupled with an image analysis facility and a knowledge relational database became the basis for developing a computer-based larva identification system whose overall identification performance was 91.9%. The advantages of this system are its speed and objectivity. The objectivity of the system is based on the fact that it is not subject to inter- and intra-observer variability arising from the user's profile of competency in interpreting subjective and non-quantifiable descriptions. The limitations of the system are that it can not handle raw images but only data extracted from images, its performance depends on the reliability of the input vectors used as training data for the ANN and its use is restricted only in well equipped laboratories due to its requirement for expensive instrumentation.	en
heal.journalName	Proceedings of the International Joint Conference on Neural Networks	en
dc.identifier.doi	10.1109/IJCNN.2001.938761	en
dc.identifier.volume	4	en
dc.identifier.spage	2506	en
dc.identifier.epage	2511	en