Open Access. Powered by Scholars. Published by Universities.®

Engineering Commons

Open Access. Powered by Scholars. Published by Universities.®

Electrical and Computer Engineering

Series

1993

3-Layer Neural Network

Articles 1 - 1 of 1

Full-Text Articles in Engineering

Identification Of Variegated Coloring In Skin Tumors: Neural Network Vs. Rule-Based Induction Methods, Ajaya Durg, William V. Stoecker, J. P. Cookson, Scott E. Umbaugh, Randy Hays Moss Sep 1993

Identification Of Variegated Coloring In Skin Tumors: Neural Network Vs. Rule-Based Induction Methods, Ajaya Durg, William V. Stoecker, J. P. Cookson, Scott E. Umbaugh, Randy Hays Moss

Chemistry Faculty Research & Creative Works

The use of neural networks for automatic identification of variegated coloring, which is believed to be one of the most predictive features for malignant melanoma, is described. The Nestor development system (NDS) was chosen for neural network implementation. At the heart of NDS is a three-layer neural network called a restricted Coulomb energy (RCE) network. The learning scheme and the database for detection of variegated coloring are discussed. Results are reported