View article

In recent years, there has been considerable developments in infrared imaging technology, and thermography is emerging as a promising modality for detecting breast cancer. Though thermography is a non-invasive, non-touch, and radiation-free imaging modality, it has not been fully used in clinical practice as manual interpretation of infrared breast thermograms is hard, subjective, and might be erroneous. A computer-aided diagnosis (CAD) engine with machine learning can help in improving the interpretation accuracy of thermal images. For this, automatic segmentation of the region of interest (ROI) from thermal images is an important preprocessing step into detecting malignant breast lesions, which is the topic of this chapter. Automated segmentation is also challenging, as thermal images can be taken at different views which are required for analysis of the entire breast region. This chapter discusses …