Classification of maxillofacial deformities using SIFT algorithm in MSCT images

There are many maxillofacial abnormalities in the world, but only a few of these maxillofacial deformities are common. As One of these deformities, we focus on the Hyper/Hypo-plasia or dysplasia of maxilla or mandibular. Cranio-maxillofacial (CMF) surgeries are the only known solution for them, which the outcome depends on surgeon’s experience and requires extensive pre-surgical planning. Classification of these abnormalities can aid the surgeon at all stages to plan and operate to achieve the desired shape. Previous works had the anatomical landmarks extracted manually. We will find relevant local features automatically using SIFT algorithm which provides many descriptor key points. These feature vectors are feed to the classifier and the results for two classifiers are compared

Keywords:

classification , local feature extraction , keypoint descriptor vector , maxillofacial dysplasia , PCA-SIFT