Breast tumor detection using empirical mode decomposition features

Abstract

Breast cancer is the second leading cause of cancer deaths among women worldwide. Microwave-based breast cancer detection has attracted increasing attention over the past two decades. Rather than recovering the image of the breast area and accurately determining the tumor location, machine-learningbased algorithms concentrate on detecting the existence of a tumor. Feature extraction is a key step in machine learning, and this step strongly impacts the final detection accuracy. Principal component analysis (PCA) is one of the most widely used feature extraction methods; however, PCA is negatively impacted by signal misalignment. This paper presents an empirical mode decomposition (EMD)-based feature extraction method that is more robust to signal misalignment. The statistical features are extracted from the decomposed subbands of the original signal. The experimental results obtained from clinical data indicate that the detection accuracy is improved by the combination of features from EMD and PCA.