Automatic segmentation of inferior Alveolar Nerve Canal in dental OPG images.

Accurate and automatic segmentation of Inferior Alveolar Nerve Canal (IAC) is most important for adequate pre diagnosis evaluation. The effective visualization of dental radiography is strongly recommended by dentists to evade IAC injury and complications. In this work, a shape and texture feature-based classification model is proposed for the segmentation of IAC in dental Orthopantomogram (OPG) images. Initially, the contrast levels of dental OPG images are improved by CLAHE and then sharpened. The boundary and texture features are extracted by shapelets, HOG, LFRT and GLCM feature descriptors. Then the features are classified as IAC and non-IAC by computational model classifiers. Finally, the assorted features are enabled in the OPG image to segment the IAC. The performances of different machine learning classifiers are analyzed and the Artificial Neural Network (ANN) model achieves an accuracy of 97.6%. The proposed method provides better IAC segmentation in panoramic dental images and avoids complexities during prognosis and at the time of dental implant surgery. [ABSTRACT FROM AUTHOR]