Classification of micro-calcification in mammograms using scalable linear Fisher discriminant analysis

Zobia Suhail,Reyer Zwiggelaar,Erika R E Denton

doi:10.1007/s11517-017-1774-z

Abstract

Breast cancer is one of the major causes of death in women. Computer Aided Diagnosis (CAD) systems are being developed to assist radiologists in early diagnosis. Micro-calcifications can be an early symptom of breast cancer. Besides detection, classification of micro-calcification as benign or malignant is essential in a complete CAD system. We have developed a novel method for the classification of benign and malignant micro-calcification using an improved Fisher Linear Discriminant Analysis (LDA) approach for the linear transformation of segmented micro-calcification data in combination with a Support Vector Machine (SVM) variant to classify between the two classes. The results indicate an average accuracy equal to 96% which is comparable to state-of-the art methods in the literature.Graphical Classification of Micro-calcification in Mammograms using Scalable Linear Fisher Discriminant Analysis

Highlights

Machine Learning is widely being used to solve problems involving high dimensional data
We propose a novel application of the scalable-Linear Discriminant Analysis (LDA) approach [14] for the classification of breast calcifications
We applied another version of LDA on our micro-calcification data for the benign and malignant classification that first transformed data to some low-dimensional space by using Principal component analysis (PCA) before applying linear transformation of data

Summary

Introduction

Machine Learning is widely being used to solve problems involving high dimensional data. In a large number of cases, the dimension of the data is much larger than the sample size, which is referred to as the undersampling problem [1]. High dimesionality and undersampling occurs in many applications [2, 3]. One of the solutions to deal with this undersampling problem is dimensionality reduction [3, 4]. Principal component analysis (PCA) is a procedure to convert a number of correlated variables into fewer variables called principal components [5], commonly used in fields of pattern recognition and computer vision [6, 7]. The purpose of PCA is to transform data to some low dimensional space and subsequently apply a classification method.

Objectives

Findings

Discussion

Conclusion