Detection of Cavities from Dental Panoramic X-ray Images Using Nested U-Net Models

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(23), P. 12771 - 12771

Published: Nov. 28, 2023

Dental caries is one of the most prevalent and chronic diseases worldwide. X-ray radiography considered a standard tool valuable resource for radiologists to identify dental problems that are hard recognize by visual inspection alone. However, available panoramic image datasets extremely limited only include small number images. U-Net deep learning networks showing promising performance in medical segmentation. In this work, different models applied images detect lesions. The Detection, Numbering, Segmentation Panoramic Images (DNS) dataset, which includes 1500 obtained from Ivisionlab, used experiment. major objective work extend DNS dataset detecting cavities generating binary ground truth use as evaluation models. These truths revised experts ensure their robustness correctness. Firstly, we expand images’ truth. Secondly, apply U-Net, U-Net++ U-Net3+ expanded learn hierarchical features enhance cavity boundary. results show outperforms other versions with 95% testing accuracy.

Language: Английский

---

Teeth and Technology: The Responsibility of Artificial Intelligence Techniques in the Dental Field- A Literature Review

Wasit Journal of Computer and Mathematics Science, Journal Year: 2024, Volume and Issue: 3(1), P. 1 - 17

Published: March 30, 2024

With the significant growth of modern technology and its integration into many different industries, especially in healthcare sector, artificial intelligence is one critical methods contributing to development medical fields, including dentistry. It possesses important influential techniques that contribute improving results patient care, diagnosis, treatment planning, tracking spread diseases. These play a major role assisting dentists diagnosing patients with high efficiency accuracy. In this review, developing field dentistry will be reviewed by highlighting most literature which these are involved. A search was conducted Web Science, Scopus, PubMed databases from 2018 2023, where articles were found (n=432), did not meet selection criteria excluded, resulting thirty included. involve six areas: periodontal, dental implantology, forensic dentistry, oral medicine pathology, orthodontics, diagnostics/dentistry. addition, review presents matters related data security, ethical concerns, dentists' skills. This article finds deep learning widely utilized as show accuracy obtained, equivalent professionals, it contributes reducing human errors revolutionizing improvement outcomes.

Language: Английский

---

Deep Learning for Predicting the Difficulty Level of Removing the Impacted Mandibular Third Molar

International Dental Journal, Journal Year: 2024, Volume and Issue: unknown

Published: July 1, 2024

Preoperative assessment of the impacted mandibular third molar (LM3) in a panoramic radiograph is important surgical planning. The aim this study was to develop and evaluate computer-aided visualisation–based deep learning (DL) system using predict difficulty level removal an LM3. included 1367 LM3 images from 784 patients who presented 2021–2023 University Dental Hospital; were collected retrospectively. surgically removing LM3s assessed via our newly developed DL system, which seamlessly integrated 3 distinct models. ResNet101V2 handled binary classification for identifying radiographs, RetinaNet detected precise location LM3, Vision Transformer performed multiclass image levels model achieved accuracy 0.8671. demonstrated exceptional detection performance, with mean average precision 0.9928. Additionally, delivered 0.7899 predicting levels. development 3-phase has yielded very good performance radiographs

Language: Английский

---

DFECF-DET: All-Weather Detector Based on Differential Feature Enhancement and Cross-Modal Fusion With Visible and Infrared Sensors

IEEE Sensors Journal, Journal Year: 2023, Volume and Issue: 23(23), P. 29200 - 29210

Published: Oct. 19, 2023

In complex environments such as night, fog, and battlefield camouflage, a single camera sensor is not sufficient to reflect scene information, multisource can raise environmental awareness. A visible light with an infrared efficient combination. However, there are huge differences in the inputs from different sensors, how fuse information two sensors apply it specific task problem that needs be solved. Hence, input detection algorithm combination of cameras proposed this article. Its purpose solve low accuracy changing environments. First, differential feature enhancement (DFE) module enhance features constantly degraded during network transmission designed Second, cross-modal fusion (CF) multiple sources designed. Finally, modules embedded two-stream network. Experiments on publicly available FLIR LLVIP datasets show article improves mAP75 by 8.3/4.3 compared single-source detector. some special environments, uses 0.1 MB storage at cost 1.7 mAP boost! Extensive ablation experiments demonstrate lightweight, efficient, plug-and-play.

Language: Английский

---

An efficient but effective writer: Diffusion-based semi-autoregressive transformer for automated radiology report generation

Biomedical Signal Processing and Control, Journal Year: 2023, Volume and Issue: 88, P. 105651 - 105651

Published: Nov. 2, 2023

Language: Английский

---