MRI breast cancer analysis

Showing results in 'Publications'. Show all posts
Hassanien, A. E., and T. - H. Kim, "MRI Breast cancer diagnosis approach using support vector machine and pulse coupled neural networks", Journal of Applied Logic - Elsevier, 2012. Abstract

This article introduces a hybrid approach that combines the advantages of fuzzy sets, pulse coupled neural networks (PCNNs), and support vector machine, in conjunction with wavelet-based feature extraction. An application of MRI breast cancer imaging has been chosen and hybridization approach have been applied to see their ability and accuracy to classify the breast cancer images into two outcomes: normal or non-normal.
The introduced approach starts with an algorithm based on type-II fuzzy sets to enhance the contrast of the input images. This is followed by performing PCNN-based segmentation algorithm in order to identify the region of interest and to detect the boundary of the breast pattern. Then, wavelet-based features are extracted and normalized. Finally, a support vector machine classifier were employed to evaluate the ability of the lesion descriptors for discrimination of different regions of interest to determine whether they represent cancer or not. To evaluate the performance of presented approach, we present tests on different breast MRI images. The experimental results obtained, show that the overall accuracy offered by the employed machine learning techniques is high compared with other machine learning techniques including decision trees, rough sets, neural networks, and fuzzy ARTMAP.

M.Moftah, H., A. E. Hassanien, N. Ghali, and M. Shoman, Multi-objective optimization K-mean segmentation approach for MRI Breast Images, , 2012. Abstract

The objective of this paper is to evaluate a new approach intended for reliable MRI breast image segmentation. It is based on the concepts of multi-objective and adaptation to identify target objects through an optimization methodology which keeps the optimum result during its iterations. The proposed approach were used to improve and enhance the traditional k-means clustering algorithm to be more effective and efficient. The clustering and breast cancer segmentation are implemented in the proposed approach at the same time by using the concept of multiobjective, and adaptation continually, in each iteration and then maintaining the best results. To evaluate performance of the presented approach, we run tests over different MRI breast images. The experimental results show that the overall accuracy offered by the multiobjective proposed k-means is high compared with standard K-mean clustering technique.