Automatic localization of inferior alveolar nerve canal in panoramic dental images

• Published in: Signal, image and video processing image and video processing, 2022-07, Vol.16 (5), p.1389-1397
• Main Authors: Pandyan, Uma Maheswari, Arumugam, Banumathi, Gurunathan, Ulaganathan, Kopuli Ashkar Ali, Shahul Hameed
• Format: Article
• Language: English
• Subjects: Computed tomography; Computer Imaging; Computer Science; Curve fitting; Dental implants; Equalization; Forecasting; Histograms; Image contrast; Image enhancement; Image processing; Image Processing and Computer Vision; Image reconstruction; Image segmentation; Localization; Medical imaging; Multimedia Information Systems; Nerves; Original Paper; Pattern Recognition and Graphics; Signal,Image and Speech Processing; Vision
• ISSN: 1863-1703; 1863-1711
• DOI: 10.1007/s11760-021-02091-1

Dental implants and maxillofacial procedures need forecast planning to circumvent inferior alveolar nerve canal (IAC) damage. Advancement in image processing techniques aids in the automation of IAC localization in maxillofacial regions of cone beam computed tomography-reconstructed panoramic X-ray images and reduces the complexities during the procedure. The proposed technique has been implemented on dental panoramic images and it involves intensity mapping and contrast limited adaptive histogram equalization technique to enhance the image and further the enhanced image is segregated as upper and lower jaws using B-spline technique. Then, the lower jaw is divided into two portions by vertical integral projection technique. Subsequently, the edge region of tooth and canal regions has been obtained by local phase congruency system. The predominant edge points extracted in the previous step are considered as Binary Robust Invariant Scalable Key points, which in turn are considered the feature descriptors for IAC segmentation. The feature points falling within the range of coordinates are connected using curve fitting approach to distinguish the IAC. Next, the performance of the work has been evaluated and finally it is compared with the recent techniques. The proposed method provides a higher accuracy of 92% and improved dice coefficient of 0.829 ± 0.13 compared to the deep learning method. This method provides the enhanced results and it has an ability to guide the dentist for the preoperative diagnostic process to locate the IAC successfully as well as to minimize the complexities associated with oral surgery and implantology.