DEEP LEARNING-DRIVEN SEGMENTATION OF DENTAL IMPLANTS AND PERI-IMPLANTITIS DETECTION IN ORTHOPANTOMOGRAPHS: A NOVEL DIAGNOSTIC TOOL
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
INTRODUCTION AND OBJECTIVE: Dental implants are well-established for restoring partial or complete tooth loss, with osseointegration being essential for their long-term success. Peri-implantitis, marked by inflammation and bone loss, compromises implant longevity. Current diagnostic methods for peri-implantitis face challenges such as subjective interpretation and time consumption. Our deep learning-based approach aims to address these limitations by providing a more accurate and efficient solution. This study aims to develop a deep learning-based approach for segmenting dental implants and detecting peri-implantitis in orthopantomographs (OPGs), enhancing diagnostic accuracy and efficiency. MATERIALS AND METHODS: After applying exclusion criteria, 7696 OPGs were used in the study, which was ethically authorized by the Near East University Ethics Review Board. Using the Python-implemented U-Net architecture, the DICOM-formatted images were segmented and converted into PNG files. The classification model used a convolutional neural network (CNN) for distinguishing between healthy implants and those affected by peri-implantitis, leveraging features extracted from the segmented regions to enhance diagnostic accuracy. The model was trained for 500 epochs using the Adam optimizer, with the dataset split into training (70%), validation (15%), and test (15%) sets. Dice similarity coefficient (DSC) and accuracy were used to assess segmentation performance. Three medical professionals used precision, recall, and F1-score to assess the classification model after segmentation, which determined whether implants were showing signs of peri-implantitis. RESULTS: The segmentation model achieved a test accuracy of 0.999, Dice Similarity Coefficient (DSC) of 0.986, and Intersection over Union (IoU) of 0.974. For classification, out of 3693 implants, 638 were clinically identified as having peri-implantitis. The model correctly identified 576 of these, with 165 false positives. Performance metrics included a precision of 0.777, recall of 0.903, and F1-score of 0.835. CONCLUSION: The deep learning-based approach for segmentation and classification of dental implants and peri-implantitis in OPGs is highly effective, providing reliable tools for enhancing clinical diagnosis and treatment planning.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.005 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it