Color morphology and segmentation of the breast thermography image

Breast cancer is one of the most feared cancers among women due its high incidence in females around the world. The screening tools that are used for breast cancer detection are mammogram, ultrasound, MRI, CT-scan, PET and thermogram. Although mammogram is considered as the gold standard screening tool, this procedure is invasive, expensive, time consuming and causes physical discomfort to the patient. Therefore, to overcome this problem, high interest has been given to thermogram, a non-ionizing and non-contact procedure. However, the absence of clear edges of breasts and low contrast may cause difficulty to analyze the thermogram. This study proposes a pre-processing method to detect breast and body boundaries to eliminate unwanted area and to extract colors in thermogram image. This pre-processing procedure is a technique done before the data is used for feature extraction and image classification process.