The proposed deep learning model, particularly the Vision Transformer, demonstrated high diagnostic accuracy and robust performance across gingival biotype classes, and showed improved recall and F1-scores, and better class separability in the confusion matrix, confirming its robustness and statistical reliability.
Abstract
Gingival biotype is a key factor influencing dental treatment outcomes. This study aimed to construct and validate an artificial intelligence (AI) model for objective and reproducible gingival biotype assessment based on intraoral photographs, thereby supporting clinical decision-making and personalized treatment planning.
A total of 1,600 participants (aged 24 ± 2 years; 720 males and 880 females) with healthy periodontal conditions were enrolled. Gingival biotype was clinically identified using the probe transparency method and categorized as thick, medium, or thin. The dataset (640 thick, 520 medium, 440 thin) was split into a training set (
n
= 1,500) and testing set (
n
= 100) with proportional distribution. Weighted cross-entropy was applied to account for class imbalance. The Vision Transformer (ViT) model was trained using AdamW with an initial learning rate of 0.001 and a batch size of 8 for 10 epochs, whereas Residual Network-18 (ResNet-18) was trained using Adam with a learning rate of 1 × 10
−4
and a batch size of 32 with early stopping. Offline data augmentation (rotation ±15°, horizontal/vertical flipping, contrast/gamma adjustment, and contrast-limited adaptive histogram equalization (CLAHE)) was applied in a 1:4 ratio with fixed random seeds (123) to expand the training set from 1,500 to 7,500 images, while the test set underwent fixed preprocessing only. Model performance was evaluated using F1-score and area under the receiver operating characteristic curve (AUC), with statistical significance and 95% confidence intervals estimated
via
bootstrap resampling and DeLong test.
The ViT model achieved superior diagnostic performance compared with ResNet-18. In the test set, the ViT model reached AUCs of 1.00 (thick), 0.98 (thin), and 0.97 (medium), significantly higher than those of ResNet-18 (0.82, 0.79, and 0.80;
p
< 0.001). The ViT model also showed improved recall and F1-scores, particularly for the “Thick” and “Thin” classes, and better class separability in the confusion matrix, confirming its robustness and statistical reliability.
The proposed deep learning model, particularly the Vision Transformer, demonstrated high diagnostic accuracy and robust performance across gingival biotype classes. The ViT model offers predictive capability and potential to support clinical decision-making in treatment planning.
This artificial intelligence (AI)-based approach enables non-invasive, objective and efficient gingival biotype assessment, facilitating early risk evaluation and personalized treatment planning. Integrated into digital diagnostic workflows, it can assist clinicians in selecting appropriate incision designs, restorative margin levels, and orthodontic force strategies according to biotype characteristics, thereby improving treatment predictability and patient outcomes.
Background: Periodontal diseases continue to be highly prevalent worldwide, and their early detection represents a clinical challenge, especially when based on non-standardized intraoral photographs. The present study develops an artificial intelligence-based prototype for the automatic classification of periodontal health, gingivitis, and periodontitis using Red, Green, Blue (RGB) images obtained in real conditions. Methods: A dataset comprising 1552 (306 healthy, 1019 with gingivitis, and 227 with periodontitis) was constructed by integrating proprietary clinical photographs with a public repository. A patient-level stratified split was enforced to prevent data leakage, ensuring that all images from the same patient remained within a single partition. This proposal uses EfficientNet-B2, which includes two-phase training, balanced focal loss, weighted sampling, CutMix/MixUp augmentation, and centered anatomical cropping to improve generalization across varied images. Results: The final model achieved an accuracy of 0.833, a macro F1-score (F1) of 0.832 [95% CI: 0.789–0.874], and a macro Area Under the Curve (AUC) of 0.962 [95% CI: 0.946–0.976] on an independent test set. A seven-configuration ablation study showed that each training component contributes to improved performance, and a baseline comparison with ResNet-50 demonstrated the superiority of EfficientNet-B2. Five-fold cross-validation with patient-level grouping yielded consistent results (F1 = 0.832 ± 0.016, AUC = 0.950 ± 0.005). Conclusions: These results demonstrate that EfficientNet-B2 is useful for assessing periodontal health using readily available RGB photographs, with potential for early detection, clinical triage, and remote assessment in modern dentistry.
Sergio David Pintado-Brito, J. A. Izquierdo-Vega, Rocío Ortega-Palacios et al.· BioMedInformatics· 0 citations
BACKGROUND
Gingivitis is a common condition in individuals with inadequate oral hygiene. Although dental indices are widely used to assess periodontal status and educate patients, digital technologies have recently been incorporated into this field.
OBJECTIVES
The aim of the study was to provide an objective digital method to resolve the ambiguity between the first two scores of the gingival index (GI): absence of color alteration (score 0); and presence of color alteration in the gingiva (score 1), by designing a convolutional neural network (CNN) algorithm capable of detecting gingival color changes with high accuracy.
MATERIAL AND METHODS
In this cross-sectional study, 10 CNN models were developed and trained to distinguish between normal and abnormal gingival color. A total of 6,660 augmented, pre-classified frontalview images of the patients' gingiva were included. The dataset was divided into a training set (4,640 images; 70%) to train the CNN models and a test set (2,020 images; 30%) to test them. All images were previously classified by periodontists as representing either normal (0) or abnormal (1) gingival color. During training, the CNN models learned to classify gingival color based on these reference labels. Model performance was subsequently evaluated in the test phase by comparing the models' predictions with the reference diagnoses provided by periodontists.
RESULTS
To evaluate inter-rater reliability, model performance was examined using Cohen's kappa coefficient and classification accuracy relative to the reference diagnoses established by the periodontists. Model 4 demonstrated excellent inter-rater reliability (κ = 1.00) and achieved 99.9% accuracy, followed by Model 8 (99.7%) and Model 5 (99.3%). Models 7 and 3 showed lower accuracies of 81.1% and 88.1%, respectively. All models demonstrated statistically significant agreement with the reference diagnoses (p < 0.001).
CONCLUSIONS
Differentiation between GI scores of 0 and 1 using CNNs provides an objective method for detecting gingival color alterations. This approach has potential for improving patients' awareness of periodontal health, enhancing motivation to maintain oral hygiene, and encouraging dental visits.
A. R. AbdulAzeez, H. Hussein, Mahmood A. Jumaah et al.· Dental and Medical Problems· 0 citations
Artificial intelligence is increasingly explored in dentistry to improve workflow efficiency and support image-based analysis. This study benchmarks deep learning (DL) and machine learning (ML) approaches for classifying pediatric dental views using a publicly available dataset of 9,562 intraoral images from children aged 1-14 years, covering eight maxillary and mandibular view classes. Under 10-fold cross-validation, MobileNetV2 achieved the highest performance among DL models (accuracy 95.18%, F1-score 0.95, AUC 0.997), followed by InceptionV3 (93.76%) and Xception (93.07%). Among ML methods, Logistic Regression achieved 93.33% accuracy with an AUC of 0.996. A symmetry-aware architecture, DentSym, was further proposed, achieving an average accuracy of 98.92% with balanced precision, recall, and F1-score. Model interpretability was examined using Grad-CAM, indicating that predictions were based on relevant dental regions. The highest-performing model was integrated into a prototype iOS application for real-time classification as a proof of concept. However, as cross-validation was performed at the image level due to the absence of patient identifiers, the reported performance should be interpreted as an upper-bound estimate under the current experimental setting. The study provides baseline reference results for this dataset and highlights the potential of explainable, mobile-based AI systems for future dental applications.
E. Yasin, M. Koklu, Mohannad Alkanan et al.· BMC Oral Health· 0 citations
Dental image analysis can be affected by factors such as lighting, framing, image quality, and visual similarity among different oral conditions. This study aims to present the development and evaluation of OdontoAI, a minimum viable product designed for the automatic classification of dental images into six categories: calculus, caries, gingivitis, hypodontia, mouth ulcer, and tooth discoloration. The project used the public dataset Oral Diseases (Kaggle), submitted to a rigorous curation process including duplicate removal by MD5 hash, grouping by perceptual hash (pHash), and controlled partitioning at a 70/15/15% ratio. The final dataset contained 3,507 images. Classification was performed using a ResNet-50 architecture adapted through two-phase transfer learning. On the test set (524 images), the model achieved an accuracy of 96.18% and a macro F1-score of 90.78%. The best results were observed for gingivitis and hypodontia; caries presented a recall of 64.29%, reflecting difficulties associated with class imbalance and visual similarity between categories. Qualitative Grad-CAM analysis indicated activations in relevant dental regions, treated as a complementary resource rather than a diagnostic tool. The study demonstrates that rigorous curation protocols directly affect reported metrics, making direct comparisons with studies that do not control for data leakage across partitions invalid.
Ricardo Marciano dos Santos, Kayky Fernandes Gameiro, Vinícius Marques da Silva Ferreira et al.· Research, Society and Develo...· 0 citations
STATEMENT OF PROBLEM
Incomplete clinical records can be a significant hurdle in implant dentistry, transforming routine maintenance or a restorative task into a complex search. When the primary documentation-such as the implant passport or surgical report-is missing, the clinician is forced to rely on radiographic identification and trial-and-error, which increases the risk of component mismatch and patient dissatisfaction.
PURPOSE
The purpose of this study was to develop and validate an artificial intelligence (AI) algorithm capable of detecting dental implants on panoramic radiographs and classifying them by implant brand and prosthetic platform size.
MATERIAL AND METHODS
A dataset of 387 panoramic radiographs with 1004 dental implant images was randomly divided into training, validation, and test sets using an 80/10/10 stratified split across 25 independent partitions. Convolutional neural network (CNN) architectures were developed and trained using manually annotated images. Ground truth labels for implant brand and prosthetic platform size were obtained from patient records.
RESULTS
Among the evaluated architectures, a Faster region-based CNN (R-CNN) combined with an EfficientNet-B7 backbone demonstrated the highest diagnostic performance. Implant detection achieved a mean Intersection over Union (IoU) of 77.65% ±11.54% and an accuracy of 99.45% ±0.70% with an average of 0.60 ±0.76 false negatives per split (mean ±standard deviation across 25 partitions). Implant brand classification accuracy was 98.93% ±0.91% (Callus Pro), 97.81% ±1.43% (Denti Root Form), 98.85% ±1.19% (NobelReplace Conical Connection partially machined collar [PMC]), and 96.79% ±1.83% for implants of unknown brand, yielding an overall brand-classification accuracy of 95.66% ±2.02%. Platform-size classification accuracy was 87.62% ±2.94% for narrow, 87.78% ±3.15% for regular, and 96.79% ±1.83% for unknown sizes. Combined brand-and-platform classification achieved an overall accuracy of 85.60% ±3.27%.
CONCLUSIONS
The 2-stage CNN pipeline demonstrated clinically acceptable accuracy for automated dental implant detection and classification tasks on panoramic radiographs, supporting its potential integration into clinical workflows to improve diagnostic efficiency and standardization.
Marcell Mesterházi, Júlia Balogh, Anna Takács et al.· The Journal of prosthetic de...· 0 citations
Background: Artificial intelligence (AI) has made significant contributions to numerous technological advancements, particularly in radiological diagnostics, and continues to demonstrate progressive development within the domain of dentistry. This study aims to assess the efficacy of deep learning algorithms in the detection and numbering of primary dentition through panoramic radiographic analysis. Methods: This study incorporated a total of 701 panoramic radiographs obtained from pediatric individuals aged 5-9 years. The methodology employed a convolutional neural network (CNN)- based framework, specifically utilizing the mask region based (R)-CNN architecture, for automated dental detection and enumeration. The model’s performance metrics were systematically evaluated through confusion matrix analysis. Results: The model demonstrated superior performance metrics in the detection and numbering of deciduous dentition within panoramic radiographic analyses. The precision, recall, and F1-score values calculated using the confusion matrix were 0.973, 0.947, and 0.960 for the age group 5, 0.944, 0.904, and 0.923 for the age group 6, 0.959, 0.932, and 0.945 for the age group 7, 0.984, 0.913, and 0.947 for the age group 8, 0.920, 0.904, and 0.912 for the age group 9, respectively. Conclusion: Deep learning–based AI models are a promising approach for automatic detection and numbering of deciduous teeth in panoramic radiographs from children.
Şehnaz Vona, Huseyin Simsek, Yasin Yaşa et al.· Essentials of Dentistry· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
