Abstract: Automated dental age estimation is essential for clinical dentistry. However, traditional methods like Demirjian’s tooth development stage (TDS) require time-consuming manual annotation by trained experts. Moreover, population differences in dental development patterns can significantly affect the accuracy of age estimation, highlighting the need for an ethnicity-specific reference data set (RDS). This study aims to use deep neural networks (DNNs) for automated estimation of Demirjian’s TDS of molars on digital panoramic radiographs, integrating the most complete southern Chinese RDS for ethnicity-appropriate dental age assessment. Panoramic radiographs from individuals aged 2 to 25 years are annotated for molar TDS and used to train eight DNN architectures, including AlexNet, DenseNet-201, and ResNet-50. DenseNet-201 achieves the highest accuracy of 93% in classifying molar TDS. Most misclassifications involve adjacent stages. The integrated mean dental age (IMDA) is obtained by mapping the predicted molar TDS using the southern Chinese RDS. The mean difference and correlation coefficient between the estimated IMDA from the best-performing DNN (AlexNet) and chronological age are -0.063 years (-3.3 weeks) and r=0.898 (p<0.001), respectively. These findings demonstrate that combining DNN for TDS estimation with ethnic-specific RDS enables accurate and reliable dental age assessment.

基于深度学习和种族特异性数字化全景X光片参考数据集的可靠牙龄评估框架

[1]Abdulla MA, Ahmed I, Assawamakin A, et al., 2009. Mapping human genetic diversity in Asia. Science, 326(5959):1541-1545.

[2]Ahn J, Nguyen TP, Kim YJ, et al., 2022. Automated analysis of three-dimensional CBCT images taken in natural head position that combines facial profile processing and multiple deep-learning models. Comput Methods Programs Biomed, 226:107123.

[3]Aljameel SS, Althumairy L, Albassam B, et al., 2023. Predictive artificial intelligence model for detecting dental age using panoramic radiograph images. Big Data Cogn Comput, 7(1):8.

[4]AlQahtani SJ, Hector MP, Liversidge HM, 2010. Brief communication: the London atlas of human tooth development and eruption. Am J Phys Anthropol, 142(3):481-490.

[5]Banar N, Bertels J, Laurent F, et al., 2020. Towards fully automated third molar development staging in panoramic radiographs. Int J Leg Med, 134(5):1831-1841.

[6]Baydoğan MP, Baybars SC, Tuncer SA, 2023. Age-Net: an advanced hybrid deep learning model for age estimation using orthopantomograph images. Trait Signal, 40(4):1553-1563.

[7]Büyükçakır B, Bertels J, Claes P, et al., 2024. OPG-based dental age estimation using a data-technical exploration of deep learning techniques. J Forens Sci, 69(3):919-931.

[8]Chollet F, 2017. Xception: deep learning with depthwise separable convolutions. IEEE Conf on Computer Vision and Pattern Recognition, p.1800-1807.

[9]Cruz-Landeira A, Linares-Argote J, Martínez-Rodríguez M, et al., 2010. Dental age estimation in Spanish and Venezuelan children. Comparison of Demirjian and Chaillet’s scores. Int J Leg Med, 124(2):105-112.

[10]Demirjian A, Goldstein H, 1976. New systems for dental maturity based on seven and four teeth. Ann Human Biol, 3(5):411-421.

[11]Demirjian A, Goldstein H, Tanner JM, 1973. A new system of dental age assessment. Human Biol, 45(2):211-227.

[12]de Tobel J, Radesh P, Vandermeulen D, et al., 2017. An automated technique to stage lower third molar development on panoramic radiographs for age estimation: a pilot study. J Forens Odontostomatol, 35(2):42-54.

[13]Fan F, Ke WC, Dai XH, et al., 2023. Semi-supervised automatic dental age and sex estimation using a hybrid transformer model. Int J Leg Med, 137(3):721-731.

[14]Franco A, Murray J, Heng D, et al., 2024. Binary decisions of artificial intelligence to classify third molar development around the legal age thresholds of 14, 16 and 18 years. Sci Rep, 14(1):4668.

[15]Gu JX, Wang ZH, Kuen J, et al., 2018. Recent advances in convolutional neural networks. Patt Recogn, 77:354-377.

[16]Han MQ, Du SY, Ge YY, et al., 2022. With or without human interference for precise age estimation based on machine learning? Int J Leg Med, 136(3):821-831.

[17]He KM, Zhang XY, Ren SQ, et al., 2016. Deep residual learning for image recognition. IEEE Conf on Computer Vision and Pattern Recognition, p.770-778.

[18]Huang G, Liu Z, van der Maaten L, et al., 2017. Densely connected convolutional networks. IEEE Conf on Computer Vision and Pattern Recognition, p.2261-2269.

[19]Jayaraman J, Roberts G, 2016. Demirjian’s method is unsuitable for dental age estimation. Forens Sci Med Pathol, 12(4):532-533.

[20]Jayaraman J, King NM, Roberts GJ, et al., 2011. Dental age assessment: are Demirjian’s standards appropriate for southern Chinese children? J Forens Odontostomatol, 29(2):22-28.

[21]Jayaraman J, Roberts GJ, King NM, et al., 2012. Dental age assessment of southern Chinese using the United Kingdom Caucasian reference dataset. Forens Sci Int, 216(1-3):68-72.

[22]Jayaraman J, Wong HM, King NM, et al., 2013. The French–Canadian data set of Demirjian for dental age estimation: a systematic review and meta-analysis. J Forens Leg Med, 20(5):373-381.

[23]Jayaraman J, Roberts GJ, Wong HM, et al., 2016a. Ages of legal importance: implications in relation to birth registration and age assessment practices. Med Sci Law, 56(1):77-82.

[24]Jayaraman J, Wong HM, King NM, et al., 2016b. Development of a reference data set (RDS) for dental age estimation (DAE) and testing of this with a separate validation set (VS) in a southern Chinese population. J Forens Leg Med, 43:26-33.

[25]Jayaraman J, Roberts GJ, Wong HM, et al., 2018. Dental age estimation in southern Chinese population using panoramic radiographs: validation of three population specific reference datasets. BMC Med Imag, 18(1):5.

[26]Jayaraman J, Wong HM, Roberts GJ, et al., 2019. Age estimation in three distinct east Asian population groups using southern Han Chinese dental reference dataset. BMC Oral Health, 19(1):242.

[27]Kahaki SMM, Nordin MJ, Ahmad NS, et al., 2020. Deep convolutional neural network designed for age assessment based on orthopantomography data. Neur Comput Appl, 32(13):9357-9368.

[28]Kim M, Chung M, Shin YG, et al., 2023. Automatic registration of dental CT and 3D scanned model using deep split jaw and surface curvature. Comput Methods Programs Biomed, 233:107467.

[29]Kim S, Lee YH, Noh YK, et al., 2021. Age-group determination of living individuals using first molar images based on artificial intelligence. Sci Rep, 11(1):1073.

[30]Kim YR, Choi JH, Ko J, et al., 2023. Age group classification of dental radiography without precise age information using convolutional neural networks. Healthcare, 11(8):1068.

[31]Koch RM, Mentzel HJ, Heinrich A, 2025. Deep learning for forensic age estimation using orthopantomograms in children, adolescents, and young adults. Eur Radiol, 35(7):4191-4202.

[32]Krizhevsky A, Sutskever I, Hinton GE, 2017. ImageNet classification with deep convolutional neural networks. Commun ACM, 60(6):84-90.

[33]Lai MC, King NM, Wong HM, 2008. Dental development of Chinese children with cleft lip and palate. Cleft Palate Craniofac J, 45(3):289-296.

[34]Lee SE, Lee SH, Lee JY, et al., 2008. Age estimation of Korean children based on dental maturity. Forens Sci Int, 178(2-3):125-131.

[35]Li H, Cai XY, Winograd-Cort ER, et al., 2007. Mitochondrial DNA diversity and population differentiation in southern East Asia. Am J Phys Anthropol, 134(4):481-488.

[36]Li JZ, Absher DM, Tang H, et al., 2008. Worldwide human relationships inferred from genome-wide patterns of variation. Science, 319(5866):1100-1104.

[37]Maber M, Liversidge HM, Hector MP, 2006. Accuracy of age estimation of radiographic methods using developing teeth. Forens Sci Int, 159(Suppl 1):S68-S73.

[38]Mallineni SK, Jayaraman J, Wong HM, et al., 2019. Dental development in children with supernumerary teeth in the anterior region of maxilla. Clin Oral Investig, 23(7):2987-2994.

[39]Matthijs L, Delande L, de Tobel J, et al., 2024. Artificial intelligence and dental age estimation: development and validation of an automated stage allocation technique on all mandibular tooth types in panoramic radiographs. Int J Leg Med, 138(6):2469-2479.

[40]Merdietio Boedi R, Banar N, de Tobel J, et al., 2020. Effect of lower third molar segmentations on automated tooth development staging using a convolutional neural network. J Forens Sci, 65(2):481-486.

[41]Migliorelli L, Massini LS, Coccia M, et al., 2023. A deep learning-based telemonitoring application to automatically assess oral diadochokinesis in patients with bulbar amyotrophic lateral sclerosis. Comput Methods Programs Biomed, 242:107840.

[42]Milani OH, Atici SF, Allareddy V, et al., 2024. A fully automated classification of third molar development stages using deep learning. Sci Rep, 14(1):13082.

[43]Mohammad N, Ahmad R, Gaus MHA, et al., 2024. Accuracy of automated forensic dental age estimation lab (F-DentEst Lab) on large Malaysian dataset. Forens Sci Int, 361:112150.

[44]Mualla N, Houssein EH, Hassan MR, 2020. Dental age estimation based on X-ray images. Comput Mater Con, 62(2):591-605.

[45]Nolla CM, 1960. The development of the permanent teeth. J Dent Children, 27:254-266.

[46]Normile D, 2009. SNP study supports southern migration route to Asia. Science, 326(5959):1470.

[47]Oliveira W, Santos MA, Burgardt CAP, et al., 2024. Estimation of human age using machine learning on panoramic radiographs for Brazilian patients. Sci Rep, 14(1):19689.

[48]Olze A, Bilang D, Schmidt S, et al., 2005. Validation of common classification systems for assessing the mineralization of third molars. Int J Leg Med, 119(1):22-26.

[49]Ong SH, Kim H, Song JS, et al., 2024. Fully automated deep learning approach to dental development assessment in panoramic radiographs. BMC Oral Health, 24(1):426.

[50]Ozlu Ucan G, Gwassi OAH, Apaydin BK, et al., 2025. Automated age estimation from OPG images and patient records using deep feature extraction and modified genetic-random forest. Diagnostics, 15(3):314.

[51]Pintana P, Upalananda W, Saekho S, et al., 2022. Fully automated method for dental age estimation using the ACF detector and deep learning. Egypt J Forens Sci, 12(1):54.

[52]Roberts G, Lucas VS, McDonald F, et al., 2017. In our opinion. Br Dent J, 222(12):918-921.

[53]Şahin TN, Kölüş T, 2024. Age and sex estimation in children and young adults using panoramic radiographs with convolutional neural networks. Appl Sci, 14(16):7014.

[54]Sathyavathi S, Baskaran KR, 2024. Human age estimation using deep convolutional neural network based on dental images (orthopantomogram). IETE J Res, 70(2):1585-1592.

[55]Shang SN, Shi YQ, Zhang YJ, et al., 2024. Artificial intelligence for brain disease diagnosis using electroencephalogram signals. J Zhejiang Univ-Sci B, 25(10):914-940.

[56]Sharifonnasabi F, Jhanjhi NZ, John J, et al., 2022. Hybrid HCNN-KNN model enhances age estimation accuracy in orthopantomography. Front Publ Health, 10:879418.

[57]Shi YC, Ye ZL, Guo JX, et al., 2024. Deep learning methods for fully automated dental age estimation on orthopantomograms. Clin Oral Invest, 28(3):198.

[58]Simonyan K, Zisserman A, 2015. Very deep convolutional networks for large-scale image recognition.

[59]Sivri MB, Taheri S, Kırzıoğlu Ercan RG, et al., 2024. Dental age estimation: a comparative study of convolutional neural network and Demirjian’s method. J Forens Leg Med, 103:102679.

[60]Sokolova M, Lapalme G, 2009. A systematic analysis of performance measures for classification tasks. Inform Process Manag, 45(4):427-437.

[61]Stoneking M, Delfin F, 2010. The human genetic history of East Asia: weaving a complex tapestry. Curr Biol, 20(4):R188-R193.

[62]Szegedy C, Liu W, Jia YQ, et al., 2015. Going deeper with convolutions. IEEE Conf on Computer Vision and Pattern Recognition, p.1-9.

[63]Szegedy C, Vanhoucke V, Ioffe S, et al., 2016. Rethinking the inception architecture for computer vision. IEEE Conf on Computer Vision and Pattern Recognition, p.2818-2826.

[64]Vila-Blanco N, Carreira MJ, Varas-Quintana P, et al., 2020. Deep neural networks for chronological age estimation from OPG images. IEEE Trans Med Imag, 39(7):2374-2384.

[65]Wang J, Dou JW, Han JX, et al., 2023. A population-based study to assess two convolutional neural networks for dental age estimation. BMC Oral Health, 23(1):109.

[66]Wang P, Bai YL, Xiao Y, et al., 2025. Aberrant network topological structure of sensorimotor superficial white-matter system in major depressive disorder. J Zhejiang Univ-Sci B, 26(1):39-51.

[67]Wang XD, Liu YL, Miao XY, et al., 2022. DENSEN: a convolutional neural network for estimating chronological ages from panoramic radiographs. BMC Bioinform, 23(3):426.

[68]Wodzinski M, Daniol M, Socha M, et al., 2022. Deep learning-based framework for automatic cranial defect reconstruction and implant modeling. Comput Methods Programs Biomed, 226:107173.

[69]Wong HM, Peng SM, Wen YF, et al., 2016a. A population survey of early factors associated with permanent tooth emergence: findings from a prospective cohort study. Commun Dent Oral Epidemiol, 44(5):467-475.

[70]Wong HM, Wen YF, Jayaraman J, et al., 2016b. Northern Chinese dental ages estimated from southern Chinese reference datasets closely correlate with chronological age. Heliyon, 2(12):e00216.

[71]Wong HM, Peng SM, Yang YQ, et al., 2017. Tooth eruption and obesity in 12-year-old children. J Dent Sci, 12(2):126-132.

[72]Wong HM, Peng SM, McGrath CPJ, 2019. Association of infant growth with emergence of permanent dentition among 12 year-aged southern Chinese school children. BMC Oral Health, 19(1):47.

[73]Yavuz BS, Ekmekcioglu O, Ankarali H, 2025. Comparison of different dental age estimation methods with deep learning: Willems, Cameriere-European, London Atlas. Int J Leg Med, 139(4):1661-1672.

[74]Yeom HG, Lee BD, Lee W, et al., 2023. Estimating chronological age through learning local and global features of panoramic radiographs in the Korean population. Sci Rep, 13(1):21857.

Similar articles