Optimized multilayer perceptron using dynamic learning rate based microwave tomography breast cancer screening

Abstract

A performance of classification tool in a Computer Aided Diagnosis (CAD) software directly affects capacity of entire breast cancer screening. Most developed classification tools have mainly focused on standard techniques, for example, Magnetic Resonance Imaging (MRI), x-ray mammography, and ultrasound. With the advent of new technology, Microwave Tomography Imaging (MTI), it was inevitable to develop a desirable classification tool showing compromised performance. Yet, existing classification model using Artificial Neural Network (ANN) for handling data from the MTI shows non-negligible optimization scheme. In this paper, we present an improved model, Multilayer Perceptron (MLP) using Dynamic Learning Rate (DLR) in order to obtain better performance with optimized setting for binary classification that can be plugged into the CAD software platform. The proposed model has an optimized size of neural network so that it will not fall into indeterminate equation problem by having reasonable amount of weights between each perceptron Also, the proposed model will dynamically assign a learning rate onto each training points in the way that model earmarks a higher learning rate onto each training points belonging into minority class in order to escape from local minima which is a typical jeopardy of ANN. In experiment, we evaluated performance with following measures; precision, recall, specificity, accuracy, and Matthews Correlation Coefficient (MCC). Experimental result shows that MLP using DLR outperforms overall measures over existing ANN dealt with MTI.

DOI: 10.1145/2851613.2851825

Extracted Key Phrases

6 Figures and Tables

Cite this paper

@inproceedings{Pack2016OptimizedMP, title={Optimized multilayer perceptron using dynamic learning rate based microwave tomography breast cancer screening}, author={Chulwoo Pack and Sung Y. Shin and Hyung Do Choi and Soon-Ik Jeon and John Kim}, booktitle={SAC}, year={2016} }