Classification of Thermographic Images for Breast Cancer Detection

Abstract

Women are susceptible to developing breast cancer. The intensive care of the patient with breast cancer is expensive and, given the cost and value of the safety of the well-being of the citizen, the prevention of breast cancer has become a prime concern in public healthcare. In the past 20 years, several techniques have been proposed for this motive, such as mammography, which is commonly used for breast cancer prognosis. However, it is most like that a false positive of mammography can transpire in which the patient is diagnosed positive by another technique. In addition, the potential side effects of exercising mammography may inspire patients and physicians to look for other diagnostic methods & techniques. Thermography presents a basic thermal comparison between a normal breast and a cancerous breast which always shows an increase in thermal activity in the precancerous tissues and the areas circumjacent developing breast cancer. This paper primarily focuses on proposing an algorithm that helps in extracting significant and suitable features from thermal breast images for further processing. The features derived are then used for the classification of the subject's breast as healthy or sick by using Support Vector Machine (SVM). By using our proposed algorithm, a 100% of accuracy rate was obtained for classification in the dataset of 260 images.