Feed-forward neural network optimized by hybridization of PSO and ABC for abnormal brain detection

ABSTRACT Automated and accurate classification of MR brain images is of crucially importance for medical analysis and interpretation. We proposed a novel automatic classification system based on particle swarm optimization (PSO) and artificial bee colony (ABC), with the aim of distinguishing abnorma...

Full description

Saved in:
Bibliographic Details
Published in:International journal of imaging systems and technology 2015-06, Vol.25 (2), p.153-164
Main Authors: Wang, Shuihua, Zhang, Yudong, Dong, Zhengchao, Du, Sidan, Ji, Genlin, Yan, Jie, Yang, Jiquan, Wang, Qiong, Feng, Chunmei, Phillips, Preetha
Format: Article
Language:English
Subjects:
artificial bee colony
Brain
Classification
Datasets
feed-forward neural network
hybridization
magnetic resonance imaging
Methods
Neural networks
particle swarm optimization
pattern recognition
principle component analysis
stationary wavelet transform
Wavelet transforms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:ABSTRACT Automated and accurate classification of MR brain images is of crucially importance for medical analysis and interpretation. We proposed a novel automatic classification system based on particle swarm optimization (PSO) and artificial bee colony (ABC), with the aim of distinguishing abnormal brains from normal brains in MRI scanning. The proposed method used stationary wavelet transform (SWT) to extract features from MR brain images. SWT is translation‐invariant and performed well even the image suffered from slight translation. Next, principal component analysis (PCA) was harnessed to reduce the SWT coefficients. Based on three different hybridization methods of PSO and ABC, we proposed three new variants of feed‐forward neural network (FNN), consisting of IABAP‐FNN, ABC‐SPSO‐FNN, and HPA‐FNN. The 10 runs of K‐fold cross validation result showed the proposed HPA‐FNN was superior to not only other two proposed classifiers but also existing state‐of‐the‐art methods in terms of classification accuracy. In addition, the method achieved perfect classification on Dataset‐66 and Dataset‐160. For Dataset‐255, the 10 repetition achieved average sensitivity of 99.37%, average specificity of 100.00%, average precision of 100.00%, and average accuracy of 99.45%. The offline learning cost 219.077 s for Dataset‐255, and merely 0.016 s for online prediction. Thus, the proposed SWT + PCA + HPA‐FNN method excelled existing methods. It can be applied to practical use.
ISSN:0899-9457
1098-1098
DOI:10.1002/ima.22132