Neural Networks for Medical Image Processing: A Study of Feature Identification
暂无分享,去创建一个
Abstract
Neural networks, a parallel computing architecture modelled on living nervous systems, are able to “learn” by example. The ability of a simulated neural network to distinguish among simulated microscopic amoebae nuclei images was studied. The neural network was successfully shown to organize feature detectors without the intermediate step of manual identification of salient features. The feature detectors were mapped onto the image format and the issue of redundancy was examined.