Validity of Temporal Encoded Structured Light RGB-D Handheld Scanner for Digital Anthropometric Measurement

With the development of commercially affordable 3D optical imaging technologies, digital anthropometric measurement method which utilizes RGB-D handheld scanner emerges as a possible alternative to conventional anthropometric measurement method which utilizes analog instruments, offering various advantages such as quicker, cheaper, and non-invasive measurements in addition to yielding 3D body dimension data. While it promises many advantages over the conventional method, the validation of the digital method's performance is needed before its widespread use as an alternative to the conventional method. In this study, the digital method measurement result’s intra-measurer intra-subject inter-dimension accuracy and precision was analyzed and compared to that of the conventional method, to determine whether or not the digital method is acceptable as an alternative. Structured light RGB-D handheld scanner was used in the digital method. Absolute mean difference and coefficient of variation comparison were used to determine whether or not the digital method’s accuracy and precision is acceptable. It was found that the digital method’s accuracy was unacceptable based on the allowable observer error criterion for dimensions adapted from ANSUR II. Its precision was marginally acceptable based on the 1.5% and 5% coefficient of variation threshold adapted from various references. Overall, the digital method’s performance was lower than that of the conventional method, but may still be acceptable on various applications with larger room for errors and with proper compensation applied to deal with the systematic errors causing inaccuracies.