Deep learning: A primer for dentists and dental researchers

Journal of Dentistry, Journal Year: 2023, Volume and Issue: 130, P. 104430 - 104430

Published: Jan. 20, 2023

Language: Английский

---

AI in Orthodontics: Revolutionizing Diagnostics and Treatment Planning—A Comprehensive Review

Journal of Clinical Medicine, Journal Year: 2024, Volume and Issue: 13(2), P. 344 - 344

Published: Jan. 7, 2024

The advent of artificial intelligence (AI) in medicine has transformed various medical specialties, including orthodontics. AI shown promising results enhancing the accuracy diagnoses, treatment planning, and predicting outcomes. Its usage orthodontic practices worldwide increased with availability applications tools. This review explores principles AI, its orthodontics, implementation clinical practice. A comprehensive literature was conducted, focusing on dental diagnostics, cephalometric evaluation, skeletal age determination, temporomandibular joint (TMJ) decision making, patient telemonitoring. Due to study heterogeneity, no meta-analysis possible. demonstrated high efficacy all these areas, but variations performance need for manual supervision suggest caution settings. complexity unpredictability algorithms call cautious regular validation. Continuous learning, proper governance, addressing privacy ethical concerns are crucial successful integration into

Language: Английский

---

Artificial intelligence for radiographic imaging detection of caries lesions: a systematic review

BMC Oral Health, Journal Year: 2024, Volume and Issue: 24(1)

Published: Feb. 24, 2024

Abstract Background The aim of this systematic review is to evaluate the diagnostic performance Artificial Intelligence (AI) models designed for detection caries lesion (CL). Materials and methods An electronic literature search was conducted on PubMed, Web Science, SCOPUS, LILACS Embase databases retrospective, prospective cross-sectional studies published until January 2023, using following keywords: artificial intelligence (AI), machine learning (ML), deep (DL), neural networks (ANN), convolutional (CNN), (DCNN), radiology, detection, diagnosis dental (DC). quality assessment performed guidelines QUADAS-2. Results Twenty articles that met selection criteria were evaluated. Five periapical radiographs, nine bitewings, six orthopantomography. number imaging examinations included ranged from 15 2900. Four investigated ANN models, fifteen CNN two DCNN models. Twelve retrospective studies, prospective. achieved in detecting CL: sensitivity 0.44 0.86, specificity 0.85 0.98, precision 0.50 0.94, PPV (Positive Predictive Value) NPV (Negative 0.95, accuracy 0.73 area under curve (AUC) 0.84 intersection over union 0.3–0.4 0.78, Dice coefficient 0.66 0.88, F1-score 0.64 0.92. According QUADAS-2 evaluation, most exhibited a low risk bias. Conclusion AI-based have demonstrated good performance, potentially being an important aid CL detection. Some limitations these are related size heterogeneity datasets. Future need rely comparable, large, clinically meaningful Protocol PROSPERO identifier: CRD42023470708

Language: Английский

---

The relationships of personality traits on perceptions and attitudes of dentistry students towards AI

BMC Medical Education, Journal Year: 2025, Volume and Issue: 25(1)

Published: Jan. 6, 2025

Artificial intelligence (AI) has gained significant attention in dentistry due to its potential revolutionize practice and improve patient outcomes. However, dentists' views attitudes toward technology can affect the application of AI. This perception attitude be affected by personality traits individuals. study aims evaluate perceptions students cross-sectional was conducted on dental at Ordu University Faculty Dentistry, involving a sample 83 students. The utilized Big Five 50 Test 5-point Likert scale gather data 20 statements regarding AI dentistry. Data were analyzed using IBM SPSS Statistics software, chi-square test employed assess relationship between their towards artificial intelligence, as well gender intelligence. Statistical significance set P < 0.05. involved participants, with 29 male 54 female participants. most common Openness Agreeableness, whereas least Extraversion. Participants found useful believed it could help dentists radiographs. agreed statement that they would trust more than dentist evaluating radiograph results. A statistically difference personal expressions comparing Males familiar females. vary based traits. Developing educational strategies tailored these foster positive integration into practice.

Language: Английский

---

Advanced AI-driven detection of interproximal caries in bitewing radiographs using YOLOv8

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 7, 2025

Dental caries is a very common chronic disease that may lead to pain, infection, and tooth loss if its diagnosis at an early stage remains undetected. Traditional methods of tactile-visual examination bitewing radiography, are subject intrinsic variability due factors such as examiner experience image quality. This can result in inconsistent diagnoses. Thus, the present study aimed develop deep learning-based AI model using YOLOv8 algorithm for improving interproximal detection radiographs. In this retrospective on 552 radiographs, total 1,506 images annotated Tehran University Medical Science were processed. The was trained results evaluated terms precision, recall, F1 score, whereby it resulted precision 96.03% enamel 80.06% dentin caries, thus showing overall 84.83%, recall 79.77%, score 82.22%. proves reliability reducing false negatives diagnostic accuracy. enhances detection, offering reliable tool dental professionals improve accuracy clinical outcomes.

Language: Английский

---

Transforming Dental Caries Diagnosis Through Artificial Intelligence-Based Techniques

Cureus, Journal Year: 2023, Volume and Issue: unknown

Published: July 11, 2023

Diagnosing dental caries plays a pivotal role in preventing and treating tooth decay. However, traditional methods of diagnosing often fall short accuracy efficiency. Despite the endorsement radiography as diagnostic tool, identification through radiographic images can be influenced by individual interpretation. Incorporating artificial intelligence (AI) into holds significant promise, potentially enhancing precision efficiency diagnoses. This review introduces fundamental concepts AI, including machine learning deep algorithms, emphasizes their relevance potential contributions to diagnosis caries. It further explains process gathering pre-processing data for AI examination. Additionally, techniques are explored, focusing on image processing, analysis, classification models predicting risk severity. Deep applications using convolutional neural networks presented. Furthermore, integration systems practice is discussed, challenges considerations implementation well ethical legal aspects. The breadth technologies prospective utility clinical scenarios from radiographs outlines advancements its revolutionizing diagnosis, encouraging research development this rapidly evolving field.

Language: Английский

---

Evaluation of Attitudes and Perceptions in Students about the Use of Artificial Intelligence in Dentistry

Dentistry Journal, Journal Year: 2023, Volume and Issue: 11(5), P. 125 - 125

Published: May 5, 2023

The implementation of artificial intelligence brings with it a great change in health care, however, there is discrepancy about the perceptions and attitudes that dental students present towards these new technologies.The study design was observational, descriptive, cross-sectional. A total 200 who met inclusion criteria were surveyed online. For qualitative variables, descriptive statistical measures obtained, such as absolute relative frequencies. comparison main variables type educational institution, sex level education, chi-square test or Fisher's exact used according to established assumptions significance p < 0.05 confidence 95%.The results indicated 86% agreed will lead advances dentistry. However, 45% participants disagreed would replace dentists future. In addition, respondents use should be part undergraduate postgraduate studies 67% 72% agreement rates respectively.The indicate This suggests bright future for relationship between intelligence.

Language: Английский

---

Evaluation of accuracy of deep learning and conventional neural network algorithms in detection of dental implant type using intraoral radiographic images: A systematic review and meta-analysis

Journal of Prosthetic Dentistry, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 1, 2024

Statement of problemWith the growing importance implant brand detection in clinical practice, accuracy machine learning algorithms has become a subject research interest. Recent studies have shown promising results for use detection. However, despite these findings, comprehensive evaluation is needed.PurposeThe purpose this systematic review and meta-analysis was to assess accuracy, sensitivity, specificity deep using 2-dimensional images such as from periapical or panoramic radiographs.Material methodsElectronic searches were conducted PubMed, Embase, Scopus, Scopus Secondary, Web Science databases. Studies that met inclusion criteria assessed quality Quality Assessment Diagnostic Accuracy Studies-2 (QUADAS-2) tool. Meta-analyses performed random-effects model estimate pooled performance measures 95% confidence intervals (CIs) STATA v.17.ResultsThirteen selected review, 3 used meta-analysis. The found overall CNN detecting dental implants radiographic 95.63%, with sensitivity 94.55% 97.91%. highest reported 99.08% Multitask ResNet152 algorithm, 100.00% 98.70% respectively (Neuro-T version 2.0.1) algorithm Straumann SLActive BLT brand. All had low risk bias.ConclusionsThe algorithms.

Language: Английский

---

Diagnostic performance of artificial intelligence-aided caries detection on bitewing radiographs: a systematic review and meta-analysis

Japanese Dental Science Review, Journal Year: 2024, Volume and Issue: 60, P. 128 - 136

Published: Feb. 29, 2024

The accuracy of artificial intelligence-aided (AI) caries diagnosis can vary considerably depending on numerous factors. This review aimed to assess the diagnostic AI models for detection and classification bitewing radiographs. Publications after 2010 were screened in five databases. A customized risk bias (RoB) assessment tool was developed applied 14 articles that met inclusion criteria out 935 references. Dataset sizes ranged from 112 3686 While 86 % studies reported a model with an ≥80 %, most exhibited unclear or high bias. Three compared model's performance dentists, which consistently showed higher average sensitivity. Five included bivariate random-effects meta-analysis overall detection. odds ratio 55.8 (95 CI= 28.8 – 108.3), summary sensitivity specificity 0.87 (0.76 0.94) 0.89 (0.75 0.960), respectively. Independent meta-analyses dentin enamel conducted sensitivities 0.84 (0.80 0.87) 0.71 (0.66 0.75), Despite promising models, lack high-quality, adequately reported, externally validated highlight current challenges future research needs.

Language: Английский

---

Visual Diagnostics of Dental Caries through Deep Learning of Non-Standardised Photographs Using a Hybrid YOLO Ensemble and Transfer Learning Model

International Journal of Environmental Research and Public Health, Journal Year: 2023, Volume and Issue: 20(7), P. 5351 - 5351

Published: March 31, 2023

Background: Access to oral healthcare is not uniform globally, particularly in rural areas with limited resources, which limits the potential of automated diagnostics and advanced tele-dentistry applications. The use digital caries detection progression monitoring through photographic communication, influenced by multiple variables that are difficult standardize such settings. objective this study was develop a novel cost-effective virtual computer vision AI system predict dental cavitations from non-standardised photographs reasonable clinical accuracy. Methods: A set 1703 augmented images obtained 233 de-identified teeth specimens. Images were acquired using consumer smartphone, without any standardised apparatus applied. utilised state-of-the-art ensemble modeling, test-time augmentation, transfer learning processes. “you only look once” algorithm (YOLO) derivatives, v5s, v5m, v5l, v5x, independently evaluated, an best results augmented, learned ResNet50, ResNet101, VGG16, AlexNet, DenseNet. outcomes evaluated precision, recall, mean average precision (mAP). Results: YOLO model achieved (mAP) 0.732, accuracy 0.789, recall 0.701. When transferred final demonstrated diagnostic 86.96%, 0.89, 0.88. This surpassed all other base methods object free-hand smartphone photographs. Conclusion: system, blending ensemble, deep processes, developed can improve access has aid

Language: Английский

---