Objective: This study aimed to evaluate an automated detection system to detect and classify permanent teeth on orthopantomogram (OPG) images using convolutional neural networks (CNNs). Methods: In total, 591 digital OPGs were collected from patients older than 18 years. Three qualified dentists performed individual teeth labelling on images to generate the ground truth annotations. A three-step procedure, relying upon CNNs, was proposed for automated detection and classification of teeth. Firstly, U-Net, a type of CNN, performed preliminary segmentation of tooth regions or detecting regions of interest (ROIs) on panoramic images. Secondly, the Faster R-CNN, an advanced object detection architecture, identified each tooth within the ROI determined by the U-Net. Thirdly, VGG-16 architecture classified each tooth into 32 categories, and a tooth number was assigned. A total of 17,135 teeth cropped from 591 radiographs were used to train and validate the tooth detection and tooth numbering modules. 90% of OPG images were used for training, and the remaining 10% were used for validation. 10-folds cross-validation was performed for measuring the performance. The intersection over union (IoU), F1 score, precision, and recall (i.e. sensitivity) were used as metrics to evaluate the performance of resultant CNNs. Results: The ROI detection module had an IoU of 0.70. The tooth detection module achieved a recall of 0.99 and a precision of 0.99. The tooth numbering module had a recall, precision and F1 score of 0.98. Conclusion: The resultant automated method achieved high performance for automated tooth detection and numbering from OPG images. Deep learning can be helpful in the automatic filing of dental charts in general dentistry and forensic medicine.

Optical coherence tomography (OCT) is an imaging modality that allows for the noninvasive assessment and identification of the internal structures of the retina. The OCT image quality can often be degraded by speckle noise, which is inherent to this imaging technique. The properties of OCT backscattering signal and the associated speckle behavior have been studied in detail in Refs. 1 and 2. From a visual inspection perspective, the presence of speckle noise obscures subtle but important morphological details and thus is detrimental in clinical diagnosis. Speckle noise also negatively affects the automatic analysis methods intended for objective and accurate quantification of the images.

Objectives: There is a paucity of research evaluating the use of telehealth applications for preventive oral care, such as remote dental screening and oral health promotion. This study aimed to assess the efficacy of teledentistry in improving the oral health of school-aged children. Methods: In this cluster randomised controlled trial, a total of 175 children aged 4–15 years were enrolled from six schools across Western Australia. The schools were randomly assigned to either the teledentistry or the control group. The intervention consisted of dental screening and oral health promotion. Qualified oral health therapists (OHTs) performed in-person dental examinations on participants in the control group. Different OHTs conducted dental photography-based screenings for participants in the teledentistry group. Both groups received screening reports and educational leaflets. Nine months later, in-person examinations were conducted on all participants in both groups to assess their dental condition and adherence to the dental advice provided at baseline. The primary outcomes included decay experience (dft/DFT index) and the proportion of children converting from a ‘caries-free’ state to a ‘caries-active’ state at the follow-up. Results: A total of 164 children completed the follow-up (mean age, 7.7 ± 2.4 years). At baseline, the prevalence of dental caries in the control and teledentistry groups was 47% and 46%, respectively. The incidence of dental caries in the teledentistry (10%) and control (12%) groups at follow-up was not significantly different (p = 0.7). Conclusion: The findings suggest that teledentistry has comparable efficacy to traditional preventive oral care in maintaining oral health. Teledentistry may offer a viable solution for expanding access to preventive oral care, especially for disadvantaged communities.

OCT is a widely used clinical ophthalmic imaging technique, but the presence of speckle noise can obscure important pathological features and hinder accurate segmentation. This paper presents a novel method for denoising optical coherence tomography (OCT) images using a combination of texture loss and generative adversarial networks (GANs). Previous approaches have integrated deep learning techniques, starting with denoising Convolutional Neural Networks (CNNs) that employed pixel-wise losses. While effective in reducing noise, these methods often introduced a blurring effect in the denoised OCT images. To address this, perceptual losses were introduced, improving denoising performance and overall image quality. Building on these advancements, our research focuses on designing an image reconstruction GAN that generates OCT images with textural similarity to the gold standard, the averaged OCT image. We utilize the PatchGAN discriminator approach as a texture loss to enhance the quality of the reconstructed OCT images. We also compare the performance of UNet and ResNet as generators in the conditional GAN (cGAN) setting, as well as compare PatchGAN with the Wasserstein GAN. Using real clinical foveal-centered OCT retinal scans of children with normal vision, our experiments demonstrate that the combination of PatchGAN and UNet achieves superior performance (PSNR = 32.50) compared to recently proposed methods such as SiameseGAN (PSNR = 31.02). Qualitative experiments involving six masked clinical ophthalmologists also favor the reconstructed OCT images with PatchGAN texture loss. In summary, this paper introduces a novel method for denoising OCT images by incorporating texture loss within a GAN framework. The proposed approach outperforms existing methods and is well-received by clinical experts, offering promising advancements in OCT image reconstruction and facilitating accurate clinical interpretation.

D1ental caries remains the most common chronic disease in childhood, affecting almost half of all children globally. Dental care and examination of children living in remote and rural areas is an ongoing challenge that has been compounded by COVID. The development of a validated system with the capacity to screen large numbers of children with some degree of automation has the potential to facilitate remote dental screening at low costs. In this study, we aim to develop and validate a deep learning system for the assessment of dental caries using color dental photos. Three state-of-the-art deep learning networks namely VGG16, ResNet-50 and Inception-v3 were adopted in the context. A total of 1020 child dental photos were used to train and validate the system. We achieved an accuracy of 79% with precision and recall respectively 95% and 75% in classifying ‘caries’ versus ‘sound’ with inception-v3.

Optical Coherence Tomography (OCT) image reconstruction faces funda-mental challenges in balancing noise reduction with preservation of fine-scalestructures. Traditional pixel-based losses produce over-smoothed results,while global frequency constraints fail to target anatomically relevant spec-tral components. We propose Sinkhorn-based frequency band loss functionsthat combine anatomically guided band-partitioning with optimal transporttheory for principled spectral distribution matching. Comprehensive evalu-ation across 12 network architectures demonstrates statistically significantimprovements: +1.19 dB PSNR (33.49→34.68 dB) over global Fouriersupervision (p < 0.0001, 95% CI: [+1.01, +1.37] dB), +0.67 dB PSNR(34.01→34.68 dB) and +0.066 EPI (0.577→0.643, +11.4%) over pixel-wiseloss (p<0.001), and +0.81 dB PSNR (33.48→34.29 dB) over point-wise L1frequency loss (p<0.001). The best configuration achieves 34.68 dB PSNRand 0.9085 SSIM while maintaining superior retinal layer boundary preser-vation across all seven anatomical layers. The proposed optimal transportformulation consistently outperforms point-wise distance metrics by model-ing spectral distribution displacement rather than element-wise differences.We further introduce texture-overlay visualization that spatially localizesfrequency-specific artifacts by back-projecting band-wise power spectral den-sity deviations, establishing an objective pipeline for assessing spectral fidelityand structural accuracy in OCT reconstructions.