Cite details and abstract

@CONFERENCE{Cox92b,
        AUTHOR             = {G. S Cox and F. J. Hoare and G. de Jager},
        BOOKTITLE          = {Third South African Workshop on Pattern Recognition},
        MONTH              = nov,
        PAGES              = {136-142},
        TITLE              = {Experiments in Lung Cancer Nodule Detection Using Texture Analysis and Neural Network Classifiers},
        YEAR               = {1992},
	ABSTRACT 	   = {

This paper discusses experiments in automatic lung cancer nodule detection in digitised chest x-rays. Texture features are extracted from subregions in an image. These features are used to train a neural network to classify the subregions as nodule tissue or otherwise. Two types of texture analysis, Haralick's co-occurrence matrices and Laws' texture measures, are used to generate texture feature vectors. Different neural network topologies, including back-propagation and learning vector quantisation networks, are used to evaluate the validity of the feature sets, and are investigated as the classification stage for nodule detection. } }