Application of deep learning and feature selection technique on external root resorption identification on CBCT images

BackgroundArtificial intelligence has been proven to improve the identification of various maxillofacial lesions. The aim of the current study is two-fold: to assess the performance of four deep learning models (DLM) in external root resorption (ERR) identification and to assess the effect of combin...

Full description

Bibliographic Details
Published in:BMC ORAL HEALTH
Main Authors: Reduwan, Nor Hidayah; Aziz, Azwatee Abdul Abdul; Razi, Roziana Mohd; Abdullah, Erma Rahayu Mohd Faizal; Nezhad, Seyed Matin Mazloom; Gohain, Meghna; Ibrahim, Norliza
Format: Article
Language:English
Published: BMC 2024
Subjects:
Online Access:https://www-webofscience-com.uitm.idm.oclc.org/wos/woscc/full-record/WOS:001166397200002
author Reduwan
Nor Hidayah; Aziz
Azwatee Abdul Abdul; Razi
Roziana Mohd; Abdullah
Erma Rahayu Mohd Faizal; Nezhad
Seyed Matin Mazloom; Gohain
Meghna; Ibrahim
Norliza
spellingShingle Reduwan
Nor Hidayah; Aziz
Azwatee Abdul Abdul; Razi
Roziana Mohd; Abdullah
Erma Rahayu Mohd Faizal; Nezhad
Seyed Matin Mazloom; Gohain
Meghna; Ibrahim
Norliza
Application of deep learning and feature selection technique on external root resorption identification on CBCT images
Dentistry, Oral Surgery & Medicine
author_facet Reduwan
Nor Hidayah; Aziz
Azwatee Abdul Abdul; Razi
Roziana Mohd; Abdullah
Erma Rahayu Mohd Faizal; Nezhad
Seyed Matin Mazloom; Gohain
Meghna; Ibrahim
Norliza
author_sort Reduwan
spelling Reduwan, Nor Hidayah; Aziz, Azwatee Abdul Abdul; Razi, Roziana Mohd; Abdullah, Erma Rahayu Mohd Faizal; Nezhad, Seyed Matin Mazloom; Gohain, Meghna; Ibrahim, Norliza
Application of deep learning and feature selection technique on external root resorption identification on CBCT images
BMC ORAL HEALTH
English
Article
BackgroundArtificial intelligence has been proven to improve the identification of various maxillofacial lesions. The aim of the current study is two-fold: to assess the performance of four deep learning models (DLM) in external root resorption (ERR) identification and to assess the effect of combining feature selection technique (FST) with DLM on their ability in ERR identification.MethodsExternal root resorption was simulated on 88 extracted premolar teeth using tungsten bur in different depths (0.5 mm, 1 mm, and 2 mm). All teeth were scanned using a Cone beam CT (Carestream Dental, Atlanta, GA). Afterward, a training (70%), validation (10%), and test (20%) dataset were established. The performance of four DLMs including Random Forest (RF) + Visual Geometry Group 16 (VGG), RF + EfficienNetB4 (EFNET), Support Vector Machine (SVM) + VGG, and SVM + EFNET) and four hybrid models (DLM + FST: (i) FS + RF + VGG, (ii) FS + RF + EFNET, (iii) FS + SVM + VGG and (iv) FS + SVM + EFNET) was compared. Five performance parameters were assessed: classification accuracy, F1-score, precision, specificity, and error rate. FST algorithms (Boruta and Recursive Feature Selection) were combined with the DLMs to assess their performance.ResultsRF + VGG exhibited the highest performance in identifying ERR, followed by the other tested models. Similarly, FST combined with RF + VGG outperformed other models with classification accuracy, F1-score, precision, and specificity of 81.9%, weighted accuracy of 83%, and area under the curve (AUC) of 96%. Kruskal Wallis test revealed a significant difference (p = 0.008) in the prediction accuracy among the eight DLMs.ConclusionIn general, all DLMs have similar performance on ERR identification. However, the performance can be improved by combining FST with DLMs.
BMC
1472-6831

2024
24
1
10.1186/s12903-024-03910-w
Dentistry, Oral Surgery & Medicine

WOS:001166397200002
https://www-webofscience-com.uitm.idm.oclc.org/wos/woscc/full-record/WOS:001166397200002
title Application of deep learning and feature selection technique on external root resorption identification on CBCT images
title_short Application of deep learning and feature selection technique on external root resorption identification on CBCT images
title_full Application of deep learning and feature selection technique on external root resorption identification on CBCT images
title_fullStr Application of deep learning and feature selection technique on external root resorption identification on CBCT images
title_full_unstemmed Application of deep learning and feature selection technique on external root resorption identification on CBCT images
title_sort Application of deep learning and feature selection technique on external root resorption identification on CBCT images
container_title BMC ORAL HEALTH
language English
format Article
description BackgroundArtificial intelligence has been proven to improve the identification of various maxillofacial lesions. The aim of the current study is two-fold: to assess the performance of four deep learning models (DLM) in external root resorption (ERR) identification and to assess the effect of combining feature selection technique (FST) with DLM on their ability in ERR identification.MethodsExternal root resorption was simulated on 88 extracted premolar teeth using tungsten bur in different depths (0.5 mm, 1 mm, and 2 mm). All teeth were scanned using a Cone beam CT (Carestream Dental, Atlanta, GA). Afterward, a training (70%), validation (10%), and test (20%) dataset were established. The performance of four DLMs including Random Forest (RF) + Visual Geometry Group 16 (VGG), RF + EfficienNetB4 (EFNET), Support Vector Machine (SVM) + VGG, and SVM + EFNET) and four hybrid models (DLM + FST: (i) FS + RF + VGG, (ii) FS + RF + EFNET, (iii) FS + SVM + VGG and (iv) FS + SVM + EFNET) was compared. Five performance parameters were assessed: classification accuracy, F1-score, precision, specificity, and error rate. FST algorithms (Boruta and Recursive Feature Selection) were combined with the DLMs to assess their performance.ResultsRF + VGG exhibited the highest performance in identifying ERR, followed by the other tested models. Similarly, FST combined with RF + VGG outperformed other models with classification accuracy, F1-score, precision, and specificity of 81.9%, weighted accuracy of 83%, and area under the curve (AUC) of 96%. Kruskal Wallis test revealed a significant difference (p = 0.008) in the prediction accuracy among the eight DLMs.ConclusionIn general, all DLMs have similar performance on ERR identification. However, the performance can be improved by combining FST with DLMs.
publisher BMC
issn 1472-6831

publishDate 2024
container_volume 24
container_issue 1
doi_str_mv 10.1186/s12903-024-03910-w
topic Dentistry, Oral Surgery & Medicine
topic_facet Dentistry, Oral Surgery & Medicine
accesstype
id WOS:001166397200002
url https://www-webofscience-com.uitm.idm.oclc.org/wos/woscc/full-record/WOS:001166397200002
record_format wos
collection Web of Science (WoS)
_version_ 1809678796445450240