Oral Photographs Research Articles

A Deep Learning System to Predict Epithelial Dysplasia in Oral Leukoplakia

Oral leukoplakia (OL) has an inherent disposition to develop oral cancer. OL with epithelial dysplasia (OED) is significantly likely to undergo malignant transformation; however, routine OED assessment is invasive and challenging. This study investigated whether a deep learning (DL) model can predict dysplasia probability among patients with leukoplakia using oral photographs. In addition, we assessed the performance of the DL model in comparison with clinicians’ ratings and in providing decision support on dysplasia assessment. Retrospective images of leukoplakia taken before biopsy/histopathology were obtained to construct the DL model (n = 2,073). OED status following histopathology was used as the gold standard for all images. We first developed, fine-tuned, and internally validated a DL architecture with an EfficientNet-B2 backbone that outputs the predicted probability of OED, OED status, and regions-of-interest heat maps. Then, we tested the performance of the DL model on a temporal cohort before geographical validation. We also assessed the model’s performance at external validation with opinions provided by human raters on OED status. Performance evaluation included discrimination, calibration, and potential net benefit. The DL model achieved good Brier scores, areas under the curve, and balanced accuracies of 0.124 (0.079–0.169), 0.882 (0.838–0.926), and 81.8% (76.5–87.1) at testing and 0.146 (0.112–0.18), 0.828 (0.792–0.864), and 76.4% (72.3–80.5) at external validation, respectively. In addition, the model had a higher potential net benefit in selecting patients with OL for biopsy/histopathology during OED assessment than when biopsies were performed for all patients. External validation also showed that the DL model had better accuracy than 92.3% (24/26) of human raters in classifying the OED status of leukoplakia from oral images (balanced accuracy: 54.8%–79.7%). Overall, the photograph-based intelligent model can predict OED probability and status in leukoplakia with good calibration and discrimination, which shows potential for decision support to select patients for biopsy/histopathology, obviate unnecessary biopsy, and assist in patient self-monitoring.

Standardization of Oral Photography in Dogs and Cats.

Oral photography is one of the essential methods of maintaining dental records. The primary purpose of photography is to record an image. Images of the oral cavity should have sufficient detail for discerning the features of the hard and soft tissues; in addition, they should accurately reproduce the real colors as they appear in the mouth. Photographs are of value during the repeated monitoring of patients and while making comparisons among historical cohort studies. This article describes the equipment and techniques for obtaining high-quality images of the oral cavities of dogs and cats in a standardized way.

Technological aspects and recommendations for applying teledentistry in oral medicine: a scoping review

BackgroundTeledentistry is applied in oral medicine to help dental practitioners and specialists manage complex oral conditions. This scoping review aims to synthesize the available evidence regarding the technical requirements and the provision of security services, as well as the recommendations on standard oral cavity photography methods for using teledentistry in oral medicine.MethodThe present scoping review was conducted in 2022 according to the Joanna Briggs Institute (JBI) manual. Four databases were systematically searched, including ISI Web of Science, PubMed, Scopus, and ProQuest, from January 1999 to December 2021. After finalizing the search strategy, all the original articles, reviews, editorials, letters, comments, and book chapters were included. All the dissertations uploaded in English with a full-text electronic file were also included to achieve a comprehensive picture of the available literature. Two types of analysis were applied: the quantitative descriptive analysis applying Excel 2016 and the qualitative thematic analysis using MAXQDA version 10.ResultsThe included papers were published during 1999–2021. Most of the included articles were from Brazil (n = 9, 21.43%), India (n = 6, 14.28%), the USA (n = 6, 14.28%), and the UK (n = 5, 11.90%). Social media usage as a teledentistry platform increased considerably between 2019 and 2021. Teledentistry platforms have been classified into four categories, including email, telephone, social network applications (WhatsApp, zoom, Google Meeting, Instagram, line Application, Viber, and Messenger), and teledentistry systems (teledentistry systems, which were designed specially to perform a teledentistry project). Teledentistry has been conducted to store and forward (n = 15, 30.71%), real-time (n = 9, 21.43%), or both (n = 12, 35.71%). According to the thematic results, basic technical requirements have been categorized into four, including teledentistry platform, type of interactions, internet connection, and photography tool. Other thematic results demonstrate seven categories of recommendations, including general factors, light, patient position, phone position, setting of a mobile phone camera, photo sending requirements, and photo visualization.ConclusionAccording to the results, over the past 7 years, especially during the COVID-19 pandemic, social network applications such as WhatsApp and Zoom have been used more frequently. Also, photography recommendations with a mobile phone camera depend on several factors such as general factors, light, patient position, phone position, setting of the mobile phone camera, and photo-sending requirements. Given the increased use of mobile cameras for oral photography, these recommendations can help improve the quality of images and subsequently improve the reliability of clinical decision-making in managing patients' conditions.

Evaluation of transfer ensemble learning-based convolutional neural network models for the identification of chronic gingivitis from oral photographs

BackgroundTo evaluate the performances of several advanced deep convolutional neural network models (AlexNet, VGG, GoogLeNet, ResNet) based on ensemble learning for recognizing chronic gingivitis from screening oral images.MethodsA total of 683 intraoral clinical images acquired from 134 volunteers were used to construct the database and evaluate the models. Four deep ConvNet models were developed using ensemble learning and outperformed a single model. The performances of the different models were evaluated by comparing the accuracy and sensitivity for recognizing the existence of gingivitis from intraoral images.ResultsThe ResNet model achieved an area under the curve (AUC) value of 97%, while the AUC values for the GoogLeNet, AlexNet, and VGG models were 94%, 92%, and 89%, respectively. Although the ResNet and GoogLeNet models performed best in classifying gingivitis from images, the sensitivity outcomes were not significantly different among the ResNet, GoogLeNet, and Alexnet models (p>0.05). However, the sensitivity of the VGGNet model differed significantly from those of the other models (p < 0.001).ConclusionThe ResNet and GoogLeNet models show promise for identifying chronic gingivitis from images. These models can help doctors diagnose periodontal diseases efficiently or based on self-examination of the oral cavity by patients.

Using patients’ oral photographs for oral hygiene motivation promotes periodontal health: a prospective randomized controlled trial

ObjectivesThe aim of this prospective, randomized, controlled, single-centered, examiner-blinded clinical trial was to evaluate the effectiveness of a personalized and visual oral health education program in addition to conventional oral hygiene education.Materials and methodsFifty-six non-smoker, right-handed participants (aged 30.34 ± 11.46 years) without clinical signs of periodontitis were randomly grouped: the intervention group (n = 28) received a personalized visualized oral health education combined with conventional oral hygiene education, and the control group (n = 28) received conventional oral hygiene education only. All participants were assessed for improved periodontal parameters (PI, GI, BOP, and PPD) at baseline, first month, and third month.ResultsA significant reduction (p < 0.001) was observed in PI, GI, and BOP during two follow-up sessions compared to the baseline for the two groups. No differences were found for inter-group (p > 0.05) or intra-group (p = 1) comparison of PPD. PI (p = 0.012), GI (p = 0.032), and BOP (p = 0.024) scores were significantly reduced at the third-month follow-up assessment in the intervention group compared to the control group.ConclusionsClinical outcomes of periodontal health were significantly enhanced by the personalized and visual oral health education program applied in this study compared to the conventional oral hygiene education program.Clinical relevanceNumerous studies reported additional interventions to the oral hygiene education program. However, we did not find any published studies investigating the role of patients’ intra-oral photographs in oral care. This study’s results demonstrated that a visually aided education program for oral hygiene motivation may help improve oral health.Clinical trial registrationRegistration number is “NCT06316505” and date of registration is 18/03/2024.

Assessment of Nasal Aesthetic Parameters in Patients with Unilateral Cleft Lip and Palate – A Retrospective Study

Objectives To assess and compare the deviating nasal cephalometric parameters of patients with unilateral cleft lip and palate (UCLP) with a control group. The study also aims to correlate the deviating cephalometric parameters with two aesthetic scoring systems. Design A retrospective study. Setting Dental college and hospital. Participants Pre-treatment lateral cephalograms and extra oral photographs of 20 adult patients with repaired UCLP presenting for orthodontic treatment. 20 patients with age and type of malocclusion matched control were selected. Main Outcome The nasal cephalometric parameters of patients with UCLP and a control group were compared. The nasolabial region of patients with cleft lip and palate (CLP) was scored using Asher McDade Aesthetic Index (AMAI) and Cleft Aesthetic Rating Scale (CARS). The scoring was done by six different groups assessors to study their perception of the nasolabial region. Correlation between cephalometric parameters and the aesthetic scores was done. Results The study found significant differences in nasal length ( P = .003) and depth ( P &lt; .001) between UCLP and the non-cleft control group. In the aesthetic assessment, orthodontist gave the least scores, while layman group scored the highest. The CARS nose aesthetic scores showed statistically significant high, negative correlation with the nasal length ( P = .01). Conclusion The cephalometric parameters and the aesthetic indices can be aid the orthodontists in the assessment of nasolabial aesthetics and additionally refer for further definitive rhinoplasty to improve the patient's overall facial aesthetics.

Detecting white spot lesions on post-orthodontic oral photographs using deep learning based on the YOLOv5x algorithm: a pilot study

BackgroundDeep learning model trained on a large image dataset, can be used to detect and discriminate targets with similar but not identical appearances. The aim of this study is to evaluate the post-training performance of the CNN-based YOLOv5x algorithm in the detection of white spot lesions in post-orthodontic oral photographs using the limited data available and to make a preliminary study for fully automated models that can be clinically integrated in the future.MethodsA total of 435 images in JPG format were uploaded into the CranioCatch labeling software and labeled white spot lesions. The labeled images were resized to 640 × 320 while maintaining their aspect ratio before model training. The labeled images were randomly divided into three groups (Training:349 images (1589 labels), Validation:43 images (181 labels), Test:43 images (215 labels)). YOLOv5x algorithm was used to perform deep learning. The segmentation performance of the tested model was visualized and analyzed using ROC analysis and a confusion matrix. True Positive (TP), False Positive (FP), and False Negative (FN) values were determined.ResultsAmong the test group images, there were 133 TPs, 36 FPs, and 82 FNs. The model’s performance metrics include precision, recall, and F1 score values of detecting white spot lesions were 0.786, 0.618, and 0.692. The AUC value obtained from the ROC analysis was 0.712. The mAP value obtained from the Precision-Recall curve graph was 0.425.ConclusionsThe model’s accuracy and sensitivity in detecting white spot lesions remained lower than expected for practical application, but is a promising and acceptable detection rate compared to previous study. The current study provides a preliminary insight to further improved by increasing the dataset for training, and applying modifications to the deep learning algorithm.Clinical revelanceDeep learning systems can help clinicians to distinguish white spot lesions that may be missed during visual inspection.

Comparison of the effectiveness of two different pain assessment methods in different orthodontic procedures

Background: This study compares two pain ratings that patients use to indicate how uncomfortable they are during bonding andthe collection of orthodontic registration material.Methods: Two hundred people, ages eleven to twenty, participated in the study; 125 were female and 75 were male. Theparticipants’ discomfort levels were assessed during bonding operations and the initial registration material collection using thevisual analog scale (VAS) and the facial pain scale (FPS). During the first registration material collection in oral photography,dental impressions, X-rays, bonding operations, lip retractor insertion, polishing, acid and sealing, and bracket application,pain levels were recorded. The Mann-Whitney U test was used to compare the data between groups, while the Wilcoxon andFriedman tests were used to analyze the data within groups. The gathered data were statistically analyzed with a significancethreshold of 5%.Results: There was a statistically significant positive correlation between the VAS and FPS scales when evaluating pain duringintraoral photography, the dental impression procedure, X-rays, lip retractor insertion, polishing, acid and sealing, and bracketbonding (0.537&lt;r&lt;0.734; P=0.001). A statistically significant positive connection was also discovered when the genderdifference was included (0.261&lt;r&lt;0.42; P=0.001).Conclusion: Similar information was obtained during the orthodontic initial registration and bonding procedures using twodifferent pain assessment instruments. It is also believed to make it possible to compare research using different pain scales.

Artificial intelligence for classification and detection of oral mucosa lesions on photographs: a systematic review and meta-analysis.

This study aimed to review and synthesize studies using artificial intelligence (AI) for classifying, detecting, or segmenting oral mucosal lesions on photographs. Inclusion criteria were (1) studies employing AI to (2) classify, detect, or segment oral mucosa lesions, (3) on oral photographs of human subjects. Included studies were assessed for risk of bias using Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). A PubMed, Scopus, Embase, Web of Science, IEEE, arXiv, medRxiv, and grey literature (Google Scholar) search was conducted until June 2023, without language limitation. After initial searching, 36 eligible studies (from 8734 identified records) were included. Based on QUADAS-2, only 7% of studies were at low risk of bias for all domains. Studies employed different AI models and reported a wide range of outcomes and metrics. The accuracy of AI for detecting oral mucosal lesions ranged from 74 to 100%, while that for clinicians un-aided by AI ranged from 61 to 98%. Pooled diagnostic odds ratio for studies which evaluated AI for diagnosing or discriminating potentially malignant lesions was 155 (95% confidence interval 23-1019), while that for cancerous lesions was 114 (59-221). AI may assist in oral mucosa lesion screening while the expected accuracy gains or further health benefits remain unclear so far. Artificial intelligence assists oral mucosa lesion screening and may foster more targeted testing and referral in the hands of non-specialist providers, for example. So far, it remains unclear if accuracy gains compared with specialized can be realized.

Artificial intelligence for oral squamous cell carcinoma detection based on oral photographs: A comprehensive literature review.

Oral squamous cell carcinoma (OSCC) presents a significant global health challenge. The integration of artificial intelligence (AI) and computer vision holds promise for the early detection of OSCC through the analysis of digitized oral photographs. This literature review explores the landscape of AI-driven OSCC automatic detection, assessing both the performance and limitations of the current state of the art. An electronic search using several data base was conducted, and a systematic review performed in accordance with PRISMA guidelines (CRD42023441416). Several studies have demonstrated remarkable results for this task, consistently achieving sensitivity rates exceeding 85% and accuracy rates surpassing 90%, often encompassing around 1000 images. The review scrutinizes these studies, shedding light on their methodologies, including the use of recent machine learning and pattern recognition approaches coupled with different supervision strategies. However, comparing the results from different papers is challenging due to variations in the datasets used. Considering these findings, this review underscores the urgent need for more robust and reliable datasets in the field of OSCC detection. Furthermore, it highlights the potential of advanced techniques such as multi-task learning, attention mechanisms, and ensemble learning as crucial tools in enhancing the accuracy and sensitivity of OSCC detection through oral photographs. These insights collectively emphasize the transformative impact of AI-driven approaches on early OSCC diagnosis, with the potential to significantly improve patient outcomes and healthcare practices.

Integration of Oral Health Screening, Intervention, and Referral Into the Pediatric Well-Child Visit

Parents of children usually understand the importance of getting primary care pediatric well-child checkups during certain stages of childhood, yet few know the importance of early routine dental visits to proactively incorporate good oral hygiene and relate the connection of oral care to overall systemic physical health. The purpose was to determine the impact of integrating oral health screening, intervention, and referral into the pediatric well-child visit. Children (aged 0-18 years) received oral health screening, oral photographs, fluoride varnish, oral health education, and referral during well-child visits. Forty-two percent of our population had never had a dental examination. Fifty-eight percent had no established dental home, and 73% drank sugar-sweetened beverages weekly. The overall impact of this model was providing comprehensive oral health care to children who have never been to the dentist and providing a smooth transition between medical and dental, increasing access to care.

Deep learning algorithms for classification and detection of recurrent aphthous ulcerations using oral clinical photographic images

Background/purposeThe application of artificial intelligence diagnosis based on deep learning in the medical field has been widely accepted. We aimed to evaluate convolutional neural networks (CNNs) for automated classification and detection of recurrent aphthous ulcerations (RAU), normal oral mucosa, and other common oral mucosal diseases in clinical oral photographs. Materials and methodsThe study included 785 clinical oral photographs, which was divided into 251 images of RAU, 271 images of the normal oral mucosa, and 263 images of other common oral mucosal diseases. Four and three CNN models were used for the classification and detection tasks, respectively. 628 images were randomly selected as training data. In addition, 78 and 79 images were assigned as validating and testing data. Main outcome measures included precision, recall, F1, specificity, sensitivity and area under the receiver operating characteristics curve (AUC). ResultsIn the classification task, the Pretrained ResNet50 model had the best performance with a precision of 92.86%, a recall of 91.84%, an F1 score of 92.24%, a specificity of 96.41%, a sensitivity of 91.84% and an AUC of 98.95%. In the detection task, the Pretrained YOLOV5 model had the best performance with a precision of 98.70%, a recall of 79.51%, an F1 score of 88.07% and an AUC of Precision-Recall curve 90.89%. ConclusionThe Pretrained ResNet50 and the Pretrained YOLOV5 algorithms were shown to have superior performance and acceptable potential in the classification and detection of RAU lesions based on non-invasive oral images, which may prove useful in clinical practice.

Employing CNN ensemble models in classifying dental caries using oral photographs

Dental caries is arguably the most persistent dental condition that affects most people over their lives. Carious lesions are commonly diagnosed by dentists using clinical and visual examination along with oral radiographs. In many circumstances, dental caries is challenging to detect with photography and might be mistaken as shadows for various reasons, including poor photo quality. However, with the introduction of Artificial Intelligence and robotic systems in dentistry, photographs can be a helpful tool in oral epidemiological research for the assessment of dental caries prevalence among the population. It can be used particularly to create a new automated approach to calculate DMF (Decay, Missing, Filled) index score. In this paper, an autonomous diagnostic approach for detecting dental cavities in photos is developed using deep learning algorithms and ensemble methods. The proposed technique employs a set of pretrained models including Xception, VGG16, VGG19, and DenseNet121 to extract essential characteristics from photographs and to classify images as either normal or caries. Then, two ensemble learning methods, E- majority and E-sum, are employed based on majority voting and sum rule to boost the performances of the individual pretrained model. Experiments are conducted on 50 images with data augmentation for normal and caries images, the employed E-majority and E-sum achieved an accuracy score of 96% and 97%, respectively. The obtained results demonstrate the superiority of the proposed ensemble framework in the detection of caries. Furthermore, this framework is a step toward constructing a fully automated, efficient decision support system to be used in the dentistry area.

Análise de fotos do sorriso postadas em redes sociais como alternativa para a identificação humana

ABSTRACT Photos of smiles posted on social media contain information on individuals’ anatomical and oral morphological characteristics, which can be important for ante and post-mortem human identification, during confrontation for forensic purposes. Objective: The study aimed to evaluate the feasibility of using smile photographs on social networks as a source of information for human forensic identification. Methods: The study sample consisted of forty individuals, randomly divided into four equal groups, two groups in Instagram (IG®) and two groups in Facebook (FB®). Standardized oral photographs of the anterior teeth of the participants, from three different angles (post-mortem photos) were taken using a Nikon® EOS 550D camera. Photos of smiles posted in FB and IG by the participants were also collected (ante-mortem photos). The analysis were carried out by 4 forensic experts, 18 dental professors, and 21 dental students. They compared simulated ante and post-mortem photos, to identify the alleged victims. Results: The correct identification score ranged from 28.6% (students) to 100% (forensic experts). The most frequently reported dental characteristics used for the identification were morphology of the anterior teeth, zenith, and gingival recessions. There was no statistically significant association between the rate of correct identification and the degree of difficulty reported during the analysis (p= 0,068), whereas there was also no association between this index and or the duration of the analysis (p=0,884). Conclusions: Therefore, the photographs of the smile posted on social media proved to be a database of dental information, and with potential to assist in identification with dental forensic purposes.

Dental hygiene students' matching accuracy when comparing antemortem dental radiographs and oral photographs to simulated postmortem WinID3® odontograms.

Matching dental antemortem (AM) and postmortem (PM) data for human identification is especially challenging when the workforce is limited. Dental hygienists have served mass fatality incidents (MFIs) due to dental-related expertise. However, forensics within dental hygiene education and research on transferable skills is limited. This qualitative balance design study assessed senior dental hygiene students' match accuracy of simulated cases varying in dental identifiers based on AM full mouth series (FMS) radiographs and oral photographs to PM WinID3® odontograms to demonstrate possible disaster victim identification (DVI) transferable skills gained during formal education. A convenience sample of senior dental hygiene students (n=31) was presented information on WinID3® interpretation, then presented with 5 mismatched cases and asked to visually interpret each to make 10 total matches; five based on AM FMS with simulated PM WinID3® odontograms and five based on AM photographs with PM WinID3® odontograms. Match accuracy scores ranged from 41.9% to 58.1% for cases with 1-10 identifiers, and 77.4% to 93.5% for cases with 11-40 identifiers. Accuracy when matching AM radiographs to PM odontograms versus AM photographs to PM odontograms was compared and revealed no statistical differences in match accuracy depending on image type (p=0.388 to 1.000). Results of this pilot study suggests transferable match accuracy skills resulted from the participants' dental hygiene formal education. These baseline skills with additional specialized training support the rationale for dental hygienists serving on DVI teams. More research is needed in education and practice when preparing dental hygienists for forensic-based service.

The Retention Effect of Resin-Based Desensitizing Agents on Hypersensitivity-A Randomized Controlled Trial.

Recently, the development of dental materials has increased the availability of various hyperesthesia desensitizers. However, there are no studies on the duration of retreatment in terms of adherence rates. Thus, the adhesion rates of resin-based desensitizers were investigated. We used a conventional desensitizer and a recently developed desensitizer containing calcium salt of 4-methacryloxyethyl trimellitic acid (C-MET) and 10-methacryloyloxydecyl dihydrogen calcium phosphate (MDCP). These colored agents were applied to the surfaces of premolars and molars, and the area was measured from weekly oral photographs. Areas were statistically analyzed and mean values were calculated using 95% confidence intervals. A p-value of <0.05 was considered statistically significant. These rates were significantly higher on the buccal side of the maxilla and lower on the lingual side of the maxilla. In addition, the desensitizer containing C-MET and MDCP displayed significantly higher adhesion rates. It is suggested that this will require monthly follow-ups and reevaluation because both agents cause less than 10% adherence and there is almost no sealing effect after 4 weeks. In addition, the significantly higher adhesion rate of the desensitizer containing C-MET and MDCP indicated that the novel monomer contributed to the improvement in the adhesion ability.

Effects of Individualized Tooth-brushing Instruction Using Oral Photographs on Elementary School Students

Effects of Individualized Tooth-brushing Instruction Using Oral Photographs on Elementary School Students

Detection of dental caries in oral photographs taken by mobile phones based on the YOLOv3 algorithm.

BackgroundThe aim of this study was to explore the value of using the YOLOv3 algorithm for detection and diagnosis of dental caries in oral photographs taken with mobile phones.MethodsOral photographs taken with the mobile phones of 570 patients were used as 3 datasets: the augmented images (n=3,990), the enhanced images (n=3,990), and the combined augmented and enhanced images (n=7,980). Oral photographs taken by mobile phones from another 70 patients were used as an independent test set. We used the YOLOv3 network for migration learning for modelling. Diagnostic precision, recall, F1-score, and mean average precision (mAP) were calculated to obtain the detection and diagnostic performance of the YOLOv3 algorithm.ResultsAfter 3 independent training, the mAP value of the original group YOLOv3 algorithm was 56.20%, in which the precision for primary caries recognition was 76.92%, recall was 49.59%, and F1-score was 0.60; the precision for secondary caries recognition was 91.67%, recall was 52.38%, and F1-score was 0.67. The mAP value of the enhance group algorithm was 66.69%, in which the precision for primary caries identification was 81.82%, recall was 52.07%, and F1-score was 0.64, and the precision for secondary caries identification was 100%, recall was 33.33%, and F1-score was 0.50. The mAP value of the comprehensive group algorithm was 85.48%, in which the precision for primary caries identification was 93.33%, recall was 69.42%, F1-score was 0.80, and the F1-score for secondary caries identification was 0.50; precision was 100%, recall was 52.38%, and F1-score was 0.69.ConclusionsThe caries detection capability based on the YOLOv3 algorithm highlights the potential utility of deep learning in caries detection and diagnosis. Comparing the 3 experiments, the detection of the model trained after using image augmented and enhancement techniques was significantly improved.

Performance of deep convolutional neural network for classification and detection of oral potentially malignant disorders in photographic images

Oral potentially malignant disorders (OPMDs) are a group of conditions that can transform into oral cancer. The purpose of this study was to evaluate convolutional neural network (CNN) algorithms to classify and detect OPMDs in oral photographs. In this study, 600 oral photograph images were collected retrospectively and grouped into 300 images of OPMDs and 300 images of normal oral mucosa. CNN-based classification models were created using DenseNet-121 and ResNet-50. The detection models were created using Faster R-CNN and YOLOv4. The image data were randomly selected and assigned as training, validating, and testing data. The testing data were evaluated to compare the performance of the CNN models with the diagnosis results produced by oral and maxillofacial surgeons. DenseNet-121 and ResNet-50 were found to produce high efficiency in diagnosis of OPMDs, with an area under the receiver operating characteristic curve (AUC) of 95%. Faster R-CNN yielded the highest detection performance, with an AUC of 74.34%. For the CNN-based classification model, the sensitivity and specificity were 100% and 90%, respectively. For the oral and maxillofacial surgeons, these values were 91.73% and 92.27%, respectively. In conclusion, the DenseNet-121, ResNet-50 and Faster R-CNN models have potential for the classification and detection of OPMDs in oral photographs.

Respectability – Armour Against Inferiority

Respectability has been a significant feature of status in the Cape, shaped by racial dynamics that have their foundation in slavery and colonialism and later formalised in apartheid legislation. The issue of representation is central to the paper and draws on the work of Stuart Hall who suggests that the concept of representation plays a more active and creative role in how we think about the world and our place in it. Drawing on oral history and personal family photographs taken during apartheid, I postulate that, through acts of performance such as dressing up and sitting for photographs, people who were classified coloured, attempted to take control of the way they were represented. In so doing, they actively resisted their dehumanisation and racial subjugation. In this regard I argue that the photographs defy and resist the memories that we have of apartheid and testify to a will to freedom and humanity.