Sex estimation: Anatomical references on panoramic radiographs using Machine Learning

Comparison of mandibular morphometric parameters in digital panoramic radiography in gender determination using machine learning

OBJECTIVE

This study aimed to evaluate the usability of morphometric features obtained from mandibular panoramic radiographs in gender determination using machine learning algorithms.

MATERIALS AND METHODS

High-resolution radiographs of 200 patients aged 20-77 (41.0 ± 12.7) were included in the study. Twelve different morphometric measurements were extracted from each digital panoramic radiography included in the study. These measurements were used as features in the machine learning phase in which six different machine learning algorithms were used (k-nearest neighbor, decision trees, support vector machines, naive Bayes, linear discrimination analysis, and neural networks). To evaluate the reliability, we have performed tenfold cross-validation and we repeated this 10 times for every classification process. This process enhances the reliability of the results for other datasets.

RESULTS

When all 12 features are used together, the accuracy rate is found to be 82.6 ± 0.5%. The classification accuracies are also compared using each feature alone. Three features that give the highest accuracy are coronoid height (80.9 ± 0.9%), condyle height (78.2 ± 0.5%), and ramus height (77.2 ± 0.4%), respectively. When compared to the classification algorithms, the highest accuracy was obtained with the naive Bayes algorithm with a rate of 84.0 ± 0.4%.

CONCLUSION

Machine learning techniques can accurately determine gender by analyzing mandibular morphometric structures from digital panoramic radiographs. The most precise results are achieved by evaluating the structures in combination, using attributes obtained from applying the MRMR algorithm to all features.

Sex Estimation From Measurements of the Mastoid Triangle and Volume of the Mastoid Air Cell System Using Classical and Machine Learning Methods: A Comparative Analysis

ABSTRACT

Previous studies on the sexual dimorphism of the mastoid triangle have typically focused on linear and area measurements. No studies in the literature have used mastoid air cell system volume measurements for direct anthropological or forensic sex determination. The aims of this study were to investigate the applicability of mastoid air cell system volume measurements and mastoid triangle measurements separately and combined for sex estimation, and to determine the accuracy of sex estimation rates using machine learning algorithms and discriminant function analysis of these data. On 200 computed tomography images, the distances constituting the edges of the mastoid triangle were measured, and the area was calculated using these measurements. A region-growing algorithm was used to determine the volume of the mastoid air cell system. The univariate sex determination accuracy was calculated for all parameters. Stepwise discriminant function analysis was performed for sex estimation. Multiple machine learning methods have also been used. All measurements of the mastoid triangle and volumes of the mastoid air cell system were higher in males than in females. The accurate sex estimation rate was determined to be 79.5% using stepwise discriminant function analysis and 88.5% using machine learning methods.

Size and shape of human mandible: Sex differences and influence of age on sex estimation accuracy

The degree of sexual dimorphism expressed by human bones is of primary importance for the development of accurate methods for sex estimation. The objective of the present study was to investigate sex differences in shape and size of the mandible using geometric morphometric methods. The study also aimed to examine the impact of age on the sex classification ability of the size and shape of the mandible. Computed tomography images of 190 Bulgarians (98 males and 92 females) were used in the study. Polygonal surface models of the skulls were generated and used for digitizing 45 landmarks located on the mandible. The raw three-dimensional coordinates of the landmarks were processed via generalized Procrustes superimposition. The sex differences in mandibular size and shape were evaluated for statistical significance. Multivariate regression was applied for correction of the allometric effect. Principal component analysis, discriminant analysis, and canonical variate analysis were also used in the study. Mandibular size differed significantly between males and females and achieved a sex classification accuracy of 87%. The significance of the sex differences in mandibular shape depended on the type of shape variables used in the analysis. The shape variables provided different classification accuracy: 78% using the Procrustes coordinates and 53% using the regression residuals. The male and female mandibles differed significantly in size and shape, including the allometric component. Mandibular size is a more effective sex indicator than shape. Age has an ambiguous effect on the classification accuracy of the size and shape variables of the mandible.

Contribution of Morphology of Frontal Sinuses (Linear and Volumetric Measurements) to Gender Identification Based on Cone Beam Computed Tomography Images (CBCT): A Systematic Review

Human identification is considered as an important step in the reconstruction of biological profiles, especially of unknown individuals. Frontal sinuses (FS) have been regarded as an ideal anatomical structure for individualisation because few pathological conditions can potentially alter their shape.

AIM

The aim of this review was to evaluate scientific evidence published since January 2010 and determine whether the dimensions and volume of FS might be useful parameters for gender determination and human identification, based only on cone beam computed tomography images (CBCT).

METHODS

This review was performed in accordance with the PRISMA statement. Four databases were searched for articles published between January 2010 and June 2022.

RESULTS

A total of 195 records were initially identified, and 90 remained after a manual duplicate check. Eight articles were selected for a full review according to the inclusion and exclusion criteria after title and abstract screening. A total of 718 participants (359 males and 359 females) were identified from the included studies. Frontal sinus volume (FSV) was significantly higher in male individuals. Frontal sinus height (FSH) and volume were the superior discriminating parameters for forensic identification.

CONCLUSIONS

This review demonstrates that assessment of FS based on CBCT images could be beneficial for gender identification in forensic science. According to the obtained studies, frontal sinus volume (FSV) and frontal sinus height (FSH) are significant greater in males than in females, providing an additional complementary method. Larger sample size and common measurement protocols are needed to verify their usefulness.

Forensic Gender Determination by Using Mandibular Morphometric Indices an Iranian Population: A Panoramic Radiographic Cross-Sectional Study

Gender determination is the first step in forensic identification, followed by age and height determination, which are both affected by gender. This study assessed the accuracy of gender estimation using mandibular morphometric indices on panoramic radiographs of an Iranian population. This retrospective study evaluated 290 panoramic radiographs (145 males and 145 females). The maximum and minimum ramus width, coronoid height, condylar height, antegonial angle, antegonial depth, gonial angle, and the superior border of mental foramen were bilaterally measured as well as bicondylar and bigonial breadths using Scanora Lite. Correlation of parameters with gender was analyzed by univariate, multiple, and best models. All indices except for gonial angle were significantly different between males and females and can be used for gender determination according to univariate model. Condylar height, coronoid height, and superior border of mental foramen and ramus were still significantly greater in males than in females after controlling for the effect of confounders (p < 0.05). Based on the best model, a formula including five indices of bicondylar breadth, condylar height, coronoid height, minimum ramus width, and superior border of mental foramen was used for gender determination. Values higher than 56% indicate male gender, while lower values indicate female gender, with 81.38% specificity for correct detection of females and 88.97% sensitivity for correct detection of males. Despite the satisfactory results, future research should focus on larger populations to verify the accuracy of the present findings.

Applications of contemporary artificial intelligence technology in forensic odontology as primary forensic identifier: A scoping review

Background

Forensic odontology may require a visual or clinical method during identification. Sometimes it may require forensic experts to refer to the existing technique to identify individuals, for example, by using the atlas to estimate the dental age. However, the existing technology can be a complicated procedure for a large-scale incident requiring a more significant number of forensic identifications, particularly during mass disasters. This has driven many experts to perform automation in their current practice to improve efficiency.

Objective

This article aims to evaluate current artificial intelligence applications and discuss their performance concerning the algorithm architecture used in forensic odontology.

Methods

This study summarizes the findings of 28 research papers published between 2010 and June 2022 using the Arksey and O'Malley framework, updated by the Joanna Briggs Institute Framework for Scoping Reviews methodology, highlighting the research trend of artificial intelligence technology in forensic odontology. In addition, a literature search was conducted on Web of Science (WoS), Scopus, Google Scholar, and PubMed, and the results were evaluated based on their content and significance.

Results

The potential application of artificial intelligence technology in forensic odontology can be categorized into four: (1) human bite marks, (2) sex determination, (3) age estimation, and (4) dental comparison. This powerful tool can solve humanity's problems by giving an adequate number of datasets, the appropriate implementation of algorithm architecture, and the proper assignment of hyperparameters that enable the model to perform the prediction at a very high level of performance.

Conclusion

The reviewed articles demonstrate that machine learning techniques are reliable for studies involving continuous features such as morphometric parameters. However, machine learning models do not strictly require large training datasets to produce promising results. In contrast, deep learning enables the processing of unstructured data, such as medical images, which require large volumes of data. Occasionally, transfer learning was used to overcome the limitation of data. In the meantime, this method's capacity to automatically learn task-specific feature representations has made it a significant success in forensic odontology.

Possibility of Human Gender Recognition Using Raman Spectra of Teeth

Gender determination of the human remains can be very challenging, especially in the case of incomplete ones. Herein, we report a proof-of-concept experiment where the possibility of gender recognition using Raman spectroscopy of teeth is investigated. Raman spectra were recorded from male and female molars and premolars on two distinct sites, tooth apex and anatomical neck. Recorded spectra were sorted into suitable datasets and initially analyzed with principal component analysis, which showed a distinction between spectra of male and female teeth. Then, reduced datasets with scores of the first 20 principal components were formed and two classification algorithms, support vector machine and artificial neural networks, were applied to form classification models for gender recognition. The obtained results showed that gender recognition with Raman spectra of teeth is possible but strongly depends both on the tooth type and spectrum recording site. The difference in classification accuracy between different tooth types and recording sites are discussed in terms of the molecular structure difference caused by the influence of masticatory loading or gender-dependent life events.

A Novel Ensemble Machine Learning Approach for Bioarchaeological Sex Prediction

I present a novel machine learning approach to predict sex in the bioarchaeological record. Eighteen cranial interlandmark distances and five maxillary dental metric distances were recorded from n = 420 human skeletons from the necropolises at Alfedena (600–400 BCE) and Campovalano (750–200 BCE and 9–11th Centuries CE) in central Italy. A generalized low rank model (GLRM) was used to impute missing data and Area under the Curve—Receiver Operating Characteristic (AUC-ROC) with 20-fold stratified cross-validation was used to evaluate predictive performance of eight machine learning algorithms on different subsets of the data. Additional perspectives such as this one show strong potential for sex prediction in bioarchaeological and forensic anthropological contexts. Furthermore, GLRMs have the potential to handle missing data in ways previously unexplored in the discipline. Although results of this study look promising (highest AUC-ROC = 0.9722 for predicting binary male/female sex), the main limitation is that the sexes of the individuals included were not known but were estimated using standard macroscopic bioarchaeological methods. However, future research should apply this machine learning approach to known-sex reference samples in order to better understand its value, along with the more general contributions that machine learning can make to the reconstruction of past human lifeways.

Findings from machine learning in clinical medical imaging applications - Lessons for translation to the forensic setting

Machine learning (ML) techniques are increasingly being used in clinical medical imaging to automate distinct processing tasks. In post-mortem forensic radiology, the use of these algorithms presents significant challenges due to variability in organ position, structural changes from decomposition, inconsistent body placement in the scanner, and the presence of foreign bodies. Existing ML approaches in clinical imaging can likely be transferred to the forensic setting with careful consideration to account for the increased variability and temporal factors that affect the data used to train these algorithms. Additional steps are required to deal with these issues, by incorporating the possible variability into the training data through data augmentation, or by using atlases as a pre-processing step to account for death-related factors. A key application of ML would be then to highlight anatomical and gross pathological features of interest, or present information to help optimally determine the cause of death. In this review, we highlight results and limitations of applications in clinical medical imaging that use ML to determine key implications for their application in the forensic setting.

Missing Value Imputation in Stature Estimation by Learning Algorithms Using Anthropometric Data: A Comparative Study

Estimating stature is essential in the process of personal identification. Because it is difficult to find human remains intact at crime scenes and disaster sites, for instance, methods are needed for estimating stature based on different body parts. For instance, the upper and lower limbs may vary depending on ancestry and sex, and it is of great importance to design adequate methodology for incorporating these in estimating stature. In addition, it is necessary to use machine learning rather than simple linear regression to improve the accuracy of stature estimation. In this study, the accuracy of statures estimated based on anthropometric data was compared using three imputation methods. In addition, by comparing the accuracy among linear and nonlinear classification methods, the best method was derived for estimating stature based on anthropometric data. For both sexes, multiple imputation was superior when the missing data ratio was low, and mean imputation performed well when the ratio was high. The support vector machine recorded the highest accuracy in all ratios of missing data. The findings of this study showed appropriate imputation methods for estimating stature with missing anthropometric data. In particular, the machine learning algorithms can be effectively used for estimating stature in humans.