Title :
Identification of the lung boundaries in MR images using neural networks
Author :
Middleton, I. ; Damper, R.I.
Author_Institution :
Dept. of Electron. & Comput. Sci., Southampton Univ., UK
fDate :
31 Oct-3 Nov 1996
Abstract :
Segmentation of medical images is an important first step in their interpretation. Thus an automatic segmentation technique would be of great benefit by reducing the manual effort currently required for this task. This paper reports on the use of neural networks in identifying the lung boundary in magnetic resonance images of the thorax. The networks are able to segment the slices used in training. Improved performance on unseen slices can be achieved using weight elimination and multi-slice training data
Keywords :
backpropagation; biomedical NMR; edge detection; generalisation (artificial intelligence); image segmentation; lung; medical image processing; multilayer perceptrons; MRI images; automatic segmentation technique; backpropagation; generalisation; image segmentation; lung boundaries identification; multi-slice training data; multilayer perceptron; neural networks use; thorax; weight elimination; Biomedical imaging; Image segmentation; Intelligent networks; Lungs; Magnetic resonance; Manuals; Neural networks; Speech; Thorax; Training data;
Conference_Titel :
Engineering in Medicine and Biology Society, 1996. Bridging Disciplines for Biomedicine. Proceedings of the 18th Annual International Conference of the IEEE
Conference_Location :
Amsterdam
Print_ISBN :
0-7803-3811-1
DOI :
10.1109/IEMBS.1996.652720
