Volume 38 -
J Dent Med-tums 2025, 38 - : 0-0 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Kalani H, Abbasi E. Application of support vector machine for detection of unilateral posterior crossbite in children based on surface electromyography signal. J Dent Med-tums 2025; 38 : 13
URL: http://jdm.tums.ac.ir/article-1-6338-en.html
URL: http://jdm.tums.ac.ir/article-1-6338-en.html
1- Assistant Professor, Department of Mechanical Engineering, Mechanics and Industries School, Sajjad University, Mashhad, Iran
2- Doctor, School of Medicine, Mashhad Azad University of Medical Sciences, Mashhad, Iran
2- Doctor, School of Medicine, Mashhad Azad University of Medical Sciences, Mashhad, Iran
Abstract: (810 Views)
Background and Aims: Posterior crossbite is a common malocclusion disorder in the primary dentition that affects masticatory function. Therefore, early detection and treatment of crossbite teeth is essential to prevent further dental complications and guarantee proper jaw development. This study investigated a reasonable and computationally efficient diagnostic system for detecting characteristics between children with and without unilateral posterior crossbite in the primary dentition using the surface electromyography (sEMG) activity of masticatory muscles.
Materials and Methods: The present study was an experimental in vitro study that used sEMG signals and support vector machine (SVM) to develop artificial intelligence systems capable of decoding muscle activity for diagnosing the crossbite. The core idea of SVM is to find the optimal separating hyperplane that maximizes the margin between two classes (presence or absence of crossbite disease) in the sEMG signal. In this study, 40 children (4 to 6 years old) were selected and divided into unilateral posterior crossbite (UPCB) (n=20) and normal occlusion (n=20) groups. The sEMG activity of the bilateral masticatory muscles was recorded during two 20-s gum-chewing sequences. Then, the time domain and frequency domain features had been obtained. In this study, eighteen time domain features and nine frequency domain features were employed. Finally, these features were used as inputs to the SVM method for data classification and crossbite disease diagnosis. In this paper, four kernel functions of SVM including linear, 2nd order polynomial, 3rd order polynomial and radial basis function were considered.
Results: Based on the obtained results, the crossbite disease had a significant effect on the EMG signals. The results demonstrated that this disease affected the amplitude of the signal more than the frequency. Therefore, using the time features of EMG signals, the SVM method was able to provide a more accurate prediction of crossbite disease. The findings indicated that the mean absolute value feature achieved a 95% accuracy in predicting posterior crossbite. Finally, the results revealed that the RBF method could exhibit superior performance.
Conclusion: The proposed method can be utilized in clinical applications for diagnoses of unilateral posterior crossbite. The findings of the study showed an influence of crossbite on the electrical activity of the temporal and masseter muscles. Therefore, the crossbite problem can be reasonably diagnosed by an appropriate learning strategy using EMG signals.
Materials and Methods: The present study was an experimental in vitro study that used sEMG signals and support vector machine (SVM) to develop artificial intelligence systems capable of decoding muscle activity for diagnosing the crossbite. The core idea of SVM is to find the optimal separating hyperplane that maximizes the margin between two classes (presence or absence of crossbite disease) in the sEMG signal. In this study, 40 children (4 to 6 years old) were selected and divided into unilateral posterior crossbite (UPCB) (n=20) and normal occlusion (n=20) groups. The sEMG activity of the bilateral masticatory muscles was recorded during two 20-s gum-chewing sequences. Then, the time domain and frequency domain features had been obtained. In this study, eighteen time domain features and nine frequency domain features were employed. Finally, these features were used as inputs to the SVM method for data classification and crossbite disease diagnosis. In this paper, four kernel functions of SVM including linear, 2nd order polynomial, 3rd order polynomial and radial basis function were considered.
Results: Based on the obtained results, the crossbite disease had a significant effect on the EMG signals. The results demonstrated that this disease affected the amplitude of the signal more than the frequency. Therefore, using the time features of EMG signals, the SVM method was able to provide a more accurate prediction of crossbite disease. The findings indicated that the mean absolute value feature achieved a 95% accuracy in predicting posterior crossbite. Finally, the results revealed that the RBF method could exhibit superior performance.
Conclusion: The proposed method can be utilized in clinical applications for diagnoses of unilateral posterior crossbite. The findings of the study showed an influence of crossbite on the electrical activity of the temporal and masseter muscles. Therefore, the crossbite problem can be reasonably diagnosed by an appropriate learning strategy using EMG signals.
Article number: 13
Type of Study: Research |
Subject:
Pediatric dentistry
Send email to the article author
| Rights and Permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
