Facial soft tissue thicknesses in craniofacial identification: Data collection protocols and associated measurement errors

Abstract

Facial soft tissue thicknesses (FSTT) form a key component of craniofacial identification methods, but as for any data, embedded measurement errors are highly pertinent. These in part dictate the effective resolution of the measurements. As herein reviewed, measurement methods are highly varied in FSTT studies and associated measurement errors have generally not been paid much attention. Less than half (44%) of 95 FSTT studies comment on measurement error and not all of these provide specific quantification. Where informative error measurement protocols are employed (5% of studies), the mean error magnitudes range from 3% to 45% rTEM and are typically in the order of 10-20%. These values demonstrate that FSTT measurement errors are similar in size to (and likely larger than) the magnitudes of many biological effects being chased. As a result, the attribution of small millimeter or submillimeter differences in FSTT to biological variables must be undertaken with caution, especially where they have not been repeated across different studies/samples. To improve the integrity of FSTT studies and the reporting of FSTT measurement errors, we propose the following standard:(1) calculate the technical error of measurement (TEM or rTEM) in any FSTT research work; (2) assess the error embedded in the full data collection procedure; and (3) conduct validation testing of FSTT means proposed for point estimation prior to publication to ensure newly calculated FSTT means provide improvements. In order to facilitate the latter, a freely available R tool TDValidator that uses the C-Table data for validation testing is herein provided.

B‐mode Ultrasound Measurement of Facial Soft Tissue Thickness for Craniofacial Identification: A Standardized Approach

“…A‐mode ultrasound was first used for FSTT measurement in craniofacial identification by both Russian and German teams; however, Helmer was the first to publish German FSTT tables using the method . In 2000, Manhein et al were the first to publish B‐mode recorded FSTT.…”

“…This applies to B‐mode ultrasound data as much as it applies to any other FSTT data. Ideally, these validation tests should report the standard error of the estimate and can be undertaken using hold‐out groups or testing against other known samples, such as previously acquired data retained in the freely available C‐Table . Free computer scripts are available for conducting these tests .…”

“…Ideally, these validation tests should report the standard error of the estimate and can be undertaken using hold‐out groups or testing against other known samples, such as previously acquired data retained in the freely available C‐Table . Free computer scripts are available for conducting these tests . Similarly, for all FSTT investigations, test‐retest error metrics should be reported .…”

“…Free computer scripts are available for conducting these tests . Similarly, for all FSTT investigations, test‐retest error metrics should be reported . In these instances, the entire process should be replicated, from image acquisition to physical measurement of the FSTTs .…”

A Synthetic Feature Skull Descriptor for 3D Skull Similarity Measurement

“…e 3D skull is an intrinsic creature of the human face, which is a well preserved bone under the effect of fire, humidity, and temperature changes. It can be widely applied in the fields of craniofacial identification [1][2][3], craniofacial reconstruction [4][5][6], and so on. e similarity measurement of the skull is one of the key techniques in the abovementioned craniofacial research.…”