文摘
The second type of cancer that kills more women in the world is breast cancer. If the prognosis is done at an early stage of the disease, women can have a better chance of cure. However, the access to medical exams in poor countries is usually precarious. This work describes the study of a computer-assisted diagnostic system using thermal imaging. The images are generated by a thermographic camera that has a lower cost than the equipment used in conventional exams. We propose a system that classifies the thermographic breasts images in “normal” and “abnormal”. We have analyzed 8 statistical characteristics: mean, variance, standard deviation (SD), skewness, kurtosis, entropy, range and median. The classification used an Artificial Neural Network (ANN) and got a result of 87 % in sensitivity, 83 % in specificity and 85 % in accuracy.