In this paper, we present an approach to breast thermogram analysis that extracts features describing bilateral symmetries from an image and then utilizes a classifier ensemble for decision making.