Some studies on analysis of breast thermograms with pattern classification

Abstract

Breast cancer has increased the fatality rate among women and it has become a global concern to curb the incident rates. The overall statistics around the world, project India as one among the developing countries to be prone to such non-communicable disease. The treatment cost is high and the survival is uncertain if the disease advances in its stages. The key to curing this disease is by detecting the abnormality in the breast at the earliest stage. Early stages of detecting cancer require periodical screening and selfexamination of the breast. Gold standard imaging tool such as mammography gives high false positive rates when they are implemented on the denser breast. As most the Indian women who are diagnosed with breast cancer fall under the average age of 35 to 50, their denser breast makes it inconvenient to be detected in the mammograms. Breast thermography being non-invasive, non-contact and non-radiative, makes the screening possible for any age group of women, restricts the fear of harmful radiation on pregnant women, and can detect a tumor in the developing stages itself. Thermography has an underlying principle which works based on the surface mapping of the temperature. The metabolic rate underneath the skin is reflected as thermal variations and it is captured by the infrared (IR) thermal camera. These thermal variations are indicated in the spectrum of colour and a thermal profile known as thermogram is produced using the IR camera. These thermograms do not depend on any skin tone or any ethnicity. But it is necessary to study the nature of breast constitution for women from different regions of the world. In this research work, a database has been created which consists of breast thermograms captured from the women of the different age group in the rural places of South India. Compared to the early years of using low newline