The Establishment Of A Method For Estimating Biological Age Based On Facial Morphology Of Children And Adolescents

Purposes:To explore the internal relationship between facial morphology and biological age(selecting the CHN 05 bone age standard as a specific evaluation index)for children and adolescents aged 8-14 and use mathematical statistics to establish a model equation for interpreting CHN 05 bone age.It provides a complete set of feasible methods for the use of facial morphological features to determine the CHN 05 bone age of children and adolescents conveniently,quickly and at low cost,and has a certain positive impact on the further popularization of bone age applications.Method:This study collected 398 males and females aged 8 to 14,and the subjects’high-definition facial photos and left wrist bone X-ray photos,used facial analysis software for data processing on each facial photo,analyzed and calculated 36 facial morphological feature values.At the same time,experts determine the bone age data of each subject in CHN 05.Firstly,find out some facial morphological feature indexes that are statistically correlated with CHN 05 bone age,then use these facial morphological indexes to perform multiple linear regression analysis with CHN 05 bone age to establish multiple linear regression equations.Results:A total of 14 facial morphological indicators were found to have a statistically significant correlation with the bone age of CHN 05:mouth width,nose length,lip thickness,face length,nose width,face width,mandibular length,Eyes distance,Nasal length,Nose tip-under nose,Inner canthals distance,Under NoseCorner of Mouth,Under nose-under chin,Under nose-under lip.Three sets of multiple linear regression equations are established:Full sample group:CHN 05 bone age=0.308+1.153*mandibular length+1.211*nose length+0.147*face length+0.368*inner canthals distance+0.349*eyes dist-ance,each variable has statistical significance(P<0.05).Male sample group:CHN 05 bone age=0.537+1.324*mandibular length+1.196*nose length+0.169*face length+1.121*Under Nose-Corner of Mouth+0.268*inner canthals distance+0.149*eyes distance,each variable has statistical significance(P<0.05).Female sample group:CHN 05 bone age=0.479+1.186*mandibular length+1.102*nose length+0.104*face length±1.063*Under Nose-Corner of Mouth+0.241*inner canthals distance+0.138*eyes distance,each variable has statistical significance(P<0.05).back-testing on the above three sets of equations:Full sample group:The correct rates within the three error intervals(±1.5,±1,±0.5)are:70.00%,63.33%,and 53.33%,respectively.Male sample group:The correct rates within the three error intervals(±1.5,±1,±0.5)are 73.33%,73.33%,and 53.33%,respectively.Female sample group:The correct rates within the three error intervals(±1.5,±1,±0.5)are:80.00%,73.33%,and 60.00%,respectively.Conclusion:This study designed a set of standard facial photo shooting process and compiled facial morphological feature data collection software,which can accurately,quickly and batch collect 36 facial feature index data in facial photos.A model equation for interpreting biological age based on facial morphological features was established.After back-testing,the accuracy rate is more than 70%in the±1.5 error interval,but the accuracy is only marginal in the smaller error interval(±0.5)Higher than 50%indicates that there are still many shortages in this study,and further improvement is needed.