Clinical oral investigations · 2025
Evaluation of an artificial intelligence-based model in diagnosing periodontal radiographic bone loss.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Cassiano Luanny de Brito Avelino, da Silva Jordão Paulino Cassiano, Martins Agnes Andrade, Barbosa Matheus Targino, Rodrigues Katryne Targino, Barbosa Ádylla Rominne Lima, da Silva Gomes Gabriela Ellen, Maia Paulo Raphael Leite, de Oliveira Patrícia Teixeira, de Sousa Lopes Maria Luiza Diniz, da Silva Ivanovitch Medeiros Dantas, de Aquino Martins Ana Rafaela Luz
Аннотация
OBJECTIVE: To develop an artificial intelligence model based on convolutional neural network for detecting and measuring periodontal radiographic bone loss (RBL).
MATERIALS AND METHODS: Keypoint annotations were carried out in 595 digital bitewing radiographic images using a Computer Vision Annotation Tool. The dataset was splitted: 416 of these images were trained using the You Only Look Once version 8 architecture with pose estimation (YOLO-v8-pose), 119 images were destined for the validation set, and 60 images were used for the test set, resulting in a model capable of detecting keypoints related to the cementoenamel junction (CEJ) and alveolar bone crest (ABC). In order to evaluate the performance of the obtained model, the following metrics were analyzed: F1-Score, precision, sensitivity and mean average precision (mAP). Then, an algorithm was implemented to measure the RBL by calculating the Euclidean distance between CEJ and ABC.
RESULTS: The model achieved an F1-Score of 66,89%, precision of 61,1%, a sensitivity of 73,9% and an mAP of 73.8%.
CONCLUSIONS: The developed model and its algorithm for identifying and measuring periodontal radiographic bone loss demonstrated promising performance, thereby presenting a potential tool for assisting in periodontal diagnosis. Further studies comparing the developed model with manual measurements performed by specialists are necessary for its validation.
CLINICAL RELEVANCE: Applying artificial intelligence in clinical dental practice can support diagnosis, streamline clinical workflows, and inform treatment planning, representing a significant advancement in dental automation.
MATERIALS AND METHODS: Keypoint annotations were carried out in 595 digital bitewing radiographic images using a Computer Vision Annotation Tool. The dataset was splitted: 416 of these images were trained using the You Only Look Once version 8 architecture with pose estimation (YOLO-v8-pose), 119 images were destined for the validation set, and 60 images were used for the test set, resulting in a model capable of detecting keypoints related to the cementoenamel junction (CEJ) and alveolar bone crest (ABC). In order to evaluate the performance of the obtained model, the following metrics were analyzed: F1-Score, precision, sensitivity and mean average precision (mAP). Then, an algorithm was implemented to measure the RBL by calculating the Euclidean distance between CEJ and ABC.
RESULTS: The model achieved an F1-Score of 66,89%, precision of 61,1%, a sensitivity of 73,9% and an mAP of 73.8%.
CONCLUSIONS: The developed model and its algorithm for identifying and measuring periodontal radiographic bone loss demonstrated promising performance, thereby presenting a potential tool for assisting in periodontal diagnosis. Further studies comparing the developed model with manual measurements performed by specialists are necessary for its validation.
CLINICAL RELEVANCE: Applying artificial intelligence in clinical dental practice can support diagnosis, streamline clinical workflows, and inform treatment planning, representing a significant advancement in dental automation.