Hybrid Method of Local Binary Pattern and Classification Tree for Early Breast Cancer Detection by Mammogram Classification

Abstract

- Breast cancer is an alarming disease due to a mutation in breast cells and it is one type of cancer among women which highly leads to their death. One of the most effective tools for early detection of breast cancer is mammography, which is a screening tool used to examine the human breast by using low-dose amplitude X-rays. Computer-Aided Diagnosis (CAD) is used as an important tool to help the medical professionals for classifying breast tissues into a different class. Computer-Aided Diagnosis (CAD) can be used to reduce human error in reading the mammograms and it shows effective results in the classification of benign and malignant abnormalities. The proposed method presents a new classification approach to detect the abnormalities in mammograms using Local Binary Pattern and Decision Tree Classification. A Uniform Local Binary Pattern(uLBP) is an extension of the original Local Binary Pattern in which only patterns that contain at most two transitions from 0 to 1 (or vice versa) are considered. In uniform Local Binary Pattern (LBP) mapping, there is a separate output label for each uniform pattern and all then on uniform patterns are assigned to a single label. These patterns are utilized to detect breast cancer by classification employing the Decision Tree Classification. Specificity and sensitivity are the two statistical measures used in this proposed method to verify and measure the significance of the test related to abnormalities in the breast tissues. Thus, it can be a measurement of performance tests for classifying the patients who do and do not suffer from cancer. The mini-MIAS mammography database is employed for testing the accuracy of the proposed method and the results are promising.