An Extracting Method of Symmetry Plane from Head CT images for Surgery Based on OBB and Image Mutual Information

Abstract

In the field of oral and maxillofacial of surgery, in order to recover the facial symmetry of patients, head symmetry plane is calculated with points chosen manually by doctors. It is great clinical significance to find automatically the symmetry plane of the head before surgery. Aimed at this challenge, an extraction method for head symmetry plane extracting is proposed in this paper. First, region growing method is applied to extract brain tissues from head CT images. Then OBB bounding box method is used to surround the brain tissues to build an initial symmetry plane. Finally, the initial plane is shifted to the best position by mathematical translation and rotation operation with the mutual information. Furthermore, asymmetric index was used to evaluate the accuracy of the extracting symmetry plane. Experimental results showed that the method proposed can achieve the accuracy of manual extracted, but the stable level of this method significantly higher than the manual method.