Symmetry and asymmetry analysis and its implications to computer-aided diagnosis: A review of the literature.

Abstract

Although clinicians have long sought to integrate computer-aided diagnostic (CAD) systems into routine clinical practice, it has proven to be extremely difficult to perform fully automated algorithmic analyses on lesions, based solely on the information contained in images. To increase the utility of computerized tools, it would be intuitive to incorporate anatomical and pathological knowledge and heuristics to help the system draw diagnostic inferences. In neuro-imaging applications, for example, one way to perform this knowledge integration is to uncover symmetry/asymmetry information from the corresponding regions of the head and to explore its implication to positive clinical findings. To correctly quantify asymmetric patterns in brain images, however, the symmetry axis, or the symmetry plane, needs to be appropriately oriented in space; i.e., the symmetry plane needs to be correctly identified either manually or using computerized methods. This review will provide an overview of the current state of knowledge of both symmetry axis/plane detection, and asymmetry quantification in neuro-images.