Estudo do grau de risco de câncer de mama utilizando a dimensão fractal em imagens infravermelhas

Abstract

Breast cancer is malignant tumor that increases as a result of genetic changes in breast cells that start to divide with no control. It is a disease with high degree of cure if it be diagnosed in early stages. In this study was developed a method for analysis of the level of risk of breast cancer using dimensionless temperature and calculating by the fractal dimension in infrared images. These images can be analyzed in order to bring information about possible areas that can be diagnosed as a possible breast cancer (malignant or benign). To perform such analyzes, it can consider a standard temperature of a region of the image captured by the camera, and use the temperature difference of symmetric regions of the body as a reference. Moreover, the calculus of fractal dimension can be used to analyze the irregularity of the edge of a region of interest (ROI), such as a breast mass, where it may exhibit some characteristics of a malignant or benign cancer. It was considered temperature information of the zone of interest selected by the user, treated with some image processing techniques such as segmentation by thresholding and edge detection. The results earned were classified into five grades of risk and also malignant tumor or benign tumor. It was analyzed five sets of images and ten separate images. In these analyzes were carried out 30 test cases taking into consideration the average adimensional temperature and 17 test cases taking into consideration the fractal dimension. In the case of average adimensional temperature the rating remained between 3 degrees lower risk and fractal dimension all cases were classified as benign tumor.