77 Views

OR46 - Second Mesio-Buccal Canal Segmentation with Yolov5 Architecture Using Cone Beam Computed Tomography Images

Thursday, May 4, 2023

1:15 PM – 1:30 PM CT

Location: E353B (Level 3)

Presenter(s)

Rohan Jagtap, D.D.S., M.S.

Director and Assistant Professor
University of Mississippi Medical Center
Flowood, Mississippi

Disclosure information not submitted.

Co-Author(s)

Roberta Pileggi, D.D.S., M.S.

Professor, Chair, Director, and Associate Dean
University of Florida
Gainesville, Florida

Disclosure information not submitted.
�D

Şuayip Burak Duman, D.D.S., Ph.D.

Associate Professor
Inonu University
Battalgazi/Malatya, Malatya, Turkey

Disclosure information not submitted.
DC

Duygu Celik Ozen, D.D.S.

Research Assistant
Inonu University
Malatya, Malatya, Turkey

Disclosure information not submitted.
IS

Ibrahim Sevki Bayrakdar, D.D.S., Ph.D.

Research Scholars
University of Mississippi
Jackson, Mississippi

Disclosure information not submitted.
OB

Oguzhan Baydar, D.D.S.

Research Assistant
Ege University, Izmir, Turkey
Bornova/İzmir, Izmir, Turkey

Disclosure information not submitted.
EA

Elham S Abu Alhaija, Ph.D.

Professor
QU Health, Qatar University
Doha, Ad Dawhah, Qatar

Disclosure information not submitted.
DH

Dilek Helvacioglu Yigit, D.D.S., Ph.D.

Associate Professor
QU Health, Qatar University
Doha, Ad Dawhah, Qatar

Disclosure information not submitted.
OC

Ozer Celik, Ph.D., M.Sc.

Assistant Professor
Eskişehir Osmangazi Üniversitesi
Odunpazarı, Eskisehir, Turkey

Disclosure information not submitted.
KO

Kaan Orhan, D.D.S., Ph.D., M.Sc., B.Sc.

Dean
Ankara University
Ankara, Ankara, Turkey

Disclosure information not submitted.

Introduction: The objective of this study is to use a deep learning model based on CNN architecture to detect the second mesiobuccal (MB2) canals, which are seen as a variation in maxillary molars root canals.

Methods: In the current study, 922 axial sections from 153 patients' cone beam computed tomography (CBCT) images were used. The segmentation method was employed to identify the MB2 canals in maxillary molars with no previous endodontic treatment. Labeled images were divided into training (80%), validation (10%), and testing (10%) groups. The artificial intelligence (AI) model was trained using the You Only Look Once v5 (YOLOv5x) architecture with 500 epochs and a learning rate of 0.01. Confusion matrix and Receiver Operating Characteristic (ROC) analysis were used in the statistical evaluation of the results.

Results: The accuracy of the MB2 canal segmentation model was 0.9253, the precision was 0.8322, and the F1 score value was 0.8763. The area under the curve (AUC) in the ROC graph of the model was 0.8436. The mAP value at 0.5 Inter-over Union (IoU) was found as 0.88323.

Conclusion: The deep learning algorithm showed high success in the detection of the MB2 canal. The success rate of the endodontic treatment can be increased, and clinicians' time can be preserved by using the newly created artificial intelligence-based models to identify variations in root canal anatomy before treatment.