Hot and Warm Region Segmentation of Breast Thermogram for Fractal Analysis based Cancer Detection

Abstract

One of the most chronic cancers among women is breast cancer. Accurate early detection of breast cancer may significantly lower the death rate of the condition. An infrared breast thermogram, a screening approach for a mammography, is applied in the latest study to determine the temperature distribution of breast tissue. Breast thermography has a significant advantage, including the fact that it is non-invasive, safe, and painless. To detect tumours, colour segmentation of infrared thermal images is essential. The temperature patterns of cancerous tissues differ from those of healthy tissues due to increased metabolic activity and angiogenesis. As a result of the fact that harmful breast tumours are hotter than benign or even healthy breast tumours. In this paper, K-means clustering is used for segmentation of the hot and warm regions of the suspected breast areas which provides an accurate temperature difference. Clusters are produced in MATLAB using this technique. Additionally, the influence of IR camera sensitivity on the number of segmentation clusters is examined. When using an ultrasensitive camera, the number of clusters evaluated may be enhanced. In this study, the prime objective is to analyze the segmented breast thermograms and by computing the fractal dimension for both hot and warm regions by using a unique technique as the Triangular Prism Surface Area method which helps in identifying malignancy and the significance of using thermal and fractal features in comparing thermograms of malignant and healthy female subjects.