Objectives

This study aims to develop a tooth detection model for panoramic X-ray images by fine-tuning a pre-trained instance segmentation model using a limited dataset.

Methods

Three hundred panoramic X-ray images were selected from a pool of 41,990. These images were annotated with polygonal shapes and labeled for 28 teeth based on the FDI system. The annotated data were used to fine-tune the Mask R-CNN model from Detectron2. Three groups（100, 200, and 300 cases）were evaluated using 5-fold cross-validation. The model's performance was assessed through average precision（AP）, detection rate, and accuracy. Statistical analysis was conducted using one-way ANOVA and Tukey's multiple comparison test.

Results

The model's AP improved with increasing dataset size, showing values of 42.90, 51.42, and 56.05 for groups of 100, 200, and 300 cases, respectively. Tooth detection rates were consistent across all groups（0.964, 0.970, and 0.968）, while accuracy in tooth number identification increased significantly with larger datasets, from 0.635 to 0.756. Despite these improvements, misclassifications of tooth types remained.

Conclusions

Fine-tuning a pre-trained instance segmentation model with a small dataset can achieve high accuracy in tooth detection for panoramic X-ray images. This approach may reduce the need for large, annotated datasets, streamlining the development process.

The objective of this study was to investigate the corrosion resistance of titanium alloys in acidic saline solution containing 0.33% fluoride. The effects of fluoride on orthodontic multi-brackets and on dental implants and prosthetic devices were evaluated and the discoloration and ion release behavior of titanium alloys were compared. Commercially pure titanium and two titanium alloys（Ti-6Al-4V and Ti-7Nb-6Al）were used and immersed in acidic saline solution（NAF, pH 5.0）containing 1500 ppm fluoride for 7 days for surface analysis. The color changes of titanium alloys were observed, and the color differences of titanium alloy had larger than commercially pure titanium. No significant difference was observed in the amount of titanium ion release in NAF among the alloys. The titanium alloy surface immersed in NAF was observed to contain fluoride precipitates, while a rough surface was observed on commercially pure titanium as a result of the dissolution of titanium. Therefore, titanium and titanium alloys could exhibit dissolution with corrosion in acid saline solution containing 1500 ppm fluoride, and the precipitation on titanium alloys were occurred. In conclusion, when titanium alloys are used as dental materials in orthodontic brackets and prosthetic devices set in the oral cavity, the possible risk of corrosion due to the application of acidic fluoride may need to know.

Objective: We produced and verified the effectiveness of video materials for training disaster responders dealing with bereaved families.

Methods: The video of “disaster players” interacting with family members of missing people at the 2022 Iwate Prefecture Comprehensive Disaster Prevention Training was recorded. The recordings were taken from the perspective of the bereaving families, and edited and processed for production of training materials. These materials reflected the opinions expressed at a review meeting held with the training participants. DVD of the video were distributed to municipalities, the police, and coast guard departments in Iwate Prefecture, whose members were asked to watch the videos in order to answer a questionnaire survey. Results: Of the participants that answered（response rate: 54%）, all stated that the contents were comprehensible and useful for future training, drills, or contingencies. Forty-three percent of the respondents were unaware of any response protocol before watching the video, 79% felt they could empathize with the bereaved families, and 68% desired participation in similar drills.

Conclusions: The video materials effectively enabled the disaster preparedness-related desk training of responders through discussions（i.e., face-to-face meetings or online）without time constraints.