Nesta seção, fornecemos uma lista selecionada de conteúdo de vídeo semanal do nosso canal do YouTube, focando em publicações recentes, conceitos-chave e tendências em Inteligência Artificial aplicada à Odontologia. Cada entrada inclui metadados essenciais para ajudar os usuários a navegar rapidamente por nossa biblioteca de conteúdo educacional. Todos os vídeos são gratuitos e estão disponíveis publicamente.
Detecting early gingivitis with Faster R-CNN
Authors: Dr.Nielsen Santos Pereira
Year: 2025
Description: This video reviews a study applying Faster R-CNN to detect early signs of gingivitis in orthodontic patients using intraoral photographs. With a dataset of 134 images and a patient-level split (107 train / 27 test), the model achieved 90% accuracy and 87% mAP, showing strong potential for clinical screening support. DOI: https://doi.org/10.3390/ijerph17228447
Keywords:
Predicting oral cancer risk using WSI and DL
Authors: Dr.Nielsen Santos Pereira
Year: 2025
Description: This study presents a deep learning–based prognostic model (OMRS) using whole slide histopathology images to predict malignant transformation risk in oral leukoplakia patients. Slides scanned at 40× were analyzed using a CNN trained on nondysplastic oral mucosa and oral squamous cell carcinoma. Patch-level predictions were aggregated into slide-level risk scores, demonstrating superior prognostic performance compared to conventional dysplasia grading systems.
Keywords:
AI predicts Fishman skeletal maturity from hand-wrist X-rays
Authors: Dr.Nielsen Santos Pereira
Year: 2025
Description: This study evaluates an artificial intelligence–based automated system for predicting Fishman’s Skeletal Maturity Indicators (SMI) using handwrist radiographs. The proposed hybrid system integrates GreulichPyle, TannerWhitehouse 3, and Fishman’s SMI methods. The workflow includes automated ROI detection, regionwise classification, and SMI stage mapping. The model achieved clinically reliable performance, with an overall accuracy of 0.772 and MAE of 0.27 SMI stages.
Keywords:
AI segments zygomatic bones on CBCT scans
Authors: Dr.Nielsen Santos Pereira
Year: 2026
Description: This proof-of-concept study proposes a deep learning–based framework for automatic segmentation of zygomatic bones from cone-beam computed tomography (CBCT) images. The system combines a VGG-16 slice classification network with a 3D U-Net segmentation backbone enhanced by edge supervision, and includes explainability via Grad-CAM/Guided Grad-CAM. Trained and evaluated on 130 CBCT scans (6:2:2 split), the model achieved high segmentation accuracy and reduced segmentation time from ~49 minutes (dentists) to ~17 seconds per scan. Clinical applications include digital planning for zygoma reconstruction, orbital surgery, zygomatic implant surgery, and orthodontics.
Keywords:
AI detects pulpal calcifications on BW radiographs
Authors: Dr.Nielsen Santos Pereira
Year: 2026
Description: This research article investigates the use of the YOLOv4 deep learning algorithm to automate the identification of pulpal calcifications within dental bite-wing radiographs. By training the model on 2,000 images labeled by expert radiologists, the study sought to create a reliable clinical decision support system for dentists. The results demonstrated that the artificial intelligence achieved high accuracy and precision in both locating pulp chambers and detecting the presence of calcified masses. This technology is particularly significant because pulp stones can complicate root canal treatments and lead to procedural errors if not identified beforehand. Ultimately, the authors conclude that deep learning can effectively assist practitioners by providing a diagnostic performance that rivals human expertise
Keywords: