AUTOMATIC LOCALIZATION OF LANDMARKS IN CEPHALOMETRIC IMAGES Via MODIFIED U-Net

Cephalometric analysis is basic assessment aid for orthodontics, oral & maxillofacial surgery and treatment planning. The identification of landmark locations on lateral cephalograms plays a critical role in the clinical analysis, automating this process could reduce possible human errors and ti...

Full description

Saved in:

Bibliographic Details
Published in:	2019 10th International Conference on Computing, Communication and Networking Technologies (ICCCNT) pp. 1 - 6
Main Authors:	Goutham, E.N.D., Vasamsetti, Srikanth, Kishore, P.V.V., Sardana, H.K.
Format:	Conference Proceeding
Language:	English
Published:	IEEE 01-07-2019
Subjects:	cephalometric landmarks convolution neural networks deep neural networks lateral cephalograms X-ray images
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Cephalometric analysis is basic assessment aid for orthodontics, oral & maxillofacial surgery and treatment planning. The identification of landmark locations on lateral cephalograms plays a critical role in the clinical analysis, automating this process could reduce possible human errors and time consumption. This work explores the automation of landmark detection process using deep learning framework. The modified U-Net model has been trained on seven landmarks and implemented to locate the same in test images. The proposed model has been evaluated with dice metrics which was observed to be approximately 88 percent for each landmark. The successful detection rate has been calculated and compared with two other prominent methods, comparable results were achieved which comes under clinically acceptable range. This work is in the early stages as not many deep learning methods explored in this particular domain, which demonstrates assurance for improvement in computer-aided treatment and surgery planning.
DOI:	10.1109/ICCCNT45670.2019.8944411