三维与二维条件下人体面部测量点的转换与准确性

Conversion and accuracy of human facial measurement points in 3D and 2D images

目的在二维正面自然头位图像中通过其他可准确标记的面部测量点,标记足够接近真实眉间点、鼻根点、鼻尖点的替代测量点,为结合人工智能大数据分析的面部美学测量工作奠定基础。方法2020年6—9月,浙江省人民医院整形外科纳入无颌面畸形且无面部手术史的女性受试者50名,年龄20~29(24.4±2.7)岁。获取二维及三维图像,标记并测量真实眉间点、鼻根点、鼻尖点及其对应的备选测量点14个,对测量结果进行统计学分析。结果备选测量点G5与眉间点差值为(-0.04±1.81)mm,具有强相关性(r=0.756,P<0.01),N5与鼻根点差值(0.53±2.23)mm,相关性弱(r=0.260,P>0.05);P4与鼻尖点差值为(-0.01±0.83)mm,相关性极强(r=0.956,P<0.01)。推荐用G5、N5、P4点作为真实眉间点、鼻根点、鼻尖点的替代点。结论替代测量点能较准确地实现面部测量点在三维与二维条件下的转换。

Objective To use other accurately marked facial measurement points in the two-dimensional natural head position photo to mark alternative measurement points that are close enough to the real glabella,nasion,and pronasale for future analysis with artificial intelligence in order to lay the foundation for the assessment of facial aesthetics.Methods From June 2020 to September 2020,the study included 50 women aged 20-29 years(mean age,24.4±2.7 years)with no maxillofacial deformities and no history of facial surgery.Two-dimensional and three-dimensional images were obtained,and the real glabella,nasion,pronasale,and 14 corresponding alternative measurement points were marked and measured.Statistical analysis was performed on the measurement results.Results The difference between the alternative measurement point G5 and glabella was(-0.04±1.81)mm,with a strong correlation(r=0.756,P<0.01);the difference between N5 and nasion was 0.53±2.23 mm,the correlation was weak(r=0.260,P>0.05);the difference between P4 and pronasale was(-0.01±0.83)mm,with very strong correlation(r=0.956,P<0.01).Therefore,it is recommended to use the G5,N5,and P4 as the substitute points for the real glabella,nasion,pronasale.Conclusions The alternative measurement points selected in this study can accurately make the conversion of facial measurement points between three-dimensional and two-dimensional images.