Breast cancer is the most commonly diagnosed form of cancer in women, accounting for about 30% of all cases. Despite earlier, less encouraging studies, which were based on low-capability and poorly calibrated equipment, thermal infrared imaging has been shown to be well suited for the task of detecting breast cancer, in particular when the tumor is in its early stages or in dense tissue. Early detection is important as it provides significantly higher chances of survival, and in this respect, infrared imaging outperforms the standard method of mammography. While mammography can detect tumors only once they exceed a certain size, even small tumors can be identified using thermography due to the high metabolic activity of cancer cells that leads to an increase in local temperature that can be picked up in the infrared.
In this chapter, we derive a number of image features from breast thermograms. These features are designed to describe the bilateral differences between regions of interest of the left and right breast. We then use these features in a pattern classification process to discriminate malignant cases from benign ones. Our pattern classification systems are based on rules, and we employ two different approaches to generate rule bases. The first utilizes fuzzy if-then rules and applies a genetic algorithm to optimize the rule base, while the second uses an ant colony optimization classification algorithm. Both approaches are shown to provide good classification accuracy.
Online access to SPIE eBooks is limited to subscribing institutions.