Optimization neural networks for the segmentation of magnetic resonance images

• Published in: IEEE transactions on medical imaging Vol. 11; no. 2; pp. 215 - 220
• Main Authors: Amartur, S.C., Piraino, D., Takefuji, Y.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 1992; Institute of Electrical and Electronics Engineers
• Subjects: Artificial neural networks; Biological and medical sciences; Clustering algorithms; Computer architecture; Humans; Image segmentation; Investigative techniques, diagnostic techniques (general aspects); Magnetic resonance; Magnetic resonance imaging; Medical sciences; Nervous system; Neural networks; Radiodiagnosis. Nmr imagery. Nmr spectrometry; Radiology; Visualization; Performance evaluation; Human; Segmentation; Central nervous system; Exploration; Nervous system; NMR spectrometry; Neural network; Nuclear magnetic resonance imaging; Optimization; Relaxation time; Spin spin relaxation; Mathematical analysis; Classification; Computerized processing; Proton; Soft tissue; Medical application; Computer aid; Hopfield model; Brain (vertebrata)
• ISSN: 0278-0062
• DOI: 10.1109/42.141645

The application of the Hopfield neural network for the multispectral unsupervised classification of MR images is reported. Winner-take-all neurons were used to obtain a crisp classification map using proton density-weighted and T/sub 2/-weighted images in the head. The preliminary studies indicate that the number of iterations needed to reach 'good' solutions was nearly constant with the number of clusters chosen for the problem.< >