Three-Dimensional Local Energy-Based Shape Histogram ( 3 D-LESH )-Based Feature Extraction – A Novel Technique

In this paper, we present a novel technique (3D-LESH) for detecting breast cancer in volumetric medical images. This technique has been incorporated as part of an intelligent expert system that can help medical practitioners make decisions regarding the medical diagnosis of a patient. Analysis of these images, slice by slice, is cumbersome and inefficient. Hence, 3D-LESH is designed to compute a histogram-based feature set from a local energy map, calculated using a phase congruency (PC) measure of the volumetric MRI images in 3D space. 3D-LESH features are invariant to the variation in intensity of contrast within the different slices of the MRI image and hence are suitable for medical image analysis. The contribution of this article is manifold. First of all, we formulated a novel 3D-LESH feature extraction technique for 3D medical images to analyse volumetric images. Furthermore, the proposed 3D-LESH has for the first time been applied to breast MRI images. The final and foremost contribution is the design of an intelligent clinical decision support system (CDSS) as a hybrid approach, combining the novel 3D-LESH feature extraction technique with machine-learning classifiers to detect cancer from MRI images. The proposed system applies contrast-limited adaptive histogram equalisation (CLAHE) to the MRI images before extracting 3D-LESH features. Furthermore, a selected subset of these features is fed into a machine-learning classifier, namely, a support vector machine (SVM), an extreme learning machine (ELM) or an echo state network (ESN) classifier, to detect abnormalities and to distinguish between different stages of abnormality. We demonstrate the performance of this proposed technique by its application to breast cancer MRI images. The results indicate a high-performance accuracy of the proposed system (99.00±0.0050) with multiple classifiers. When compared with the wavelet-based feature extraction technique, statistical analysis provides conclusive evidence of the significance of our proposed algorithm. Keywords— clinical decision support system (CDSS), echo state network (ESN), extreme learning machine (ELM), local energybased shape histogram (LESH), magnetic resonance imaging (MRI), support vector machine (SVM). Corresponding author Email addresses: skw1@cs.stir.ac.uk (Summrina K. Wajid), ahu@cs.stir.ac.uk (Prof. Amir Hussain), kaizhu.huang@xjtlu.edu.cn (Prof. Kaizhu HUANG) Three-Dimensional Local Energy-Based Shape Histogram (3D-LESH)-Based Feature Extraction‒ A Novel Technique Summrina Kanwal Wajid a,* , Prof. Amir Hussain a,b , Prof. Kaizhu HUANG c