Statistical Comparison of 3D Scanning and Manual Measurement Methods in Anthropometry

Abstract

3D scanner and Martin ruler are two common measurement methods to perform anthropometric measurements. This paper compare the two measurement methods for chest width and heel width by statistical method. Firstly, necessary descriptive statistical analysis are performed to obtain general knowledge of the data. The results show that using the manual measurement method, the means of chest width and heel width both are relatively small, the data are more concentrated, and there is no abnormal point. Additionally, we conduct normality test for the chest width and the heel width then select appropriate method for detecting the outliers. If the data follows the normality, we use the Grubbs test to detect the outliers; if not, boxplots are used to determine whether there are outliers. The results show that there is no abnormal point in the chest width and the heel width under the manual measurement method, and there are outliers in the chest width and the heel width under the 3D scanning method. Finally, a two-sample T-test are performed on chest width and heel width. It is found that there is a significant difference between manual measurement and 3D scanning, regardless of chest width or heel width. The result provides a basis for the correct use of the 3D scanner.