Dataset Information

Ensemble landmarking of 3D facial surface scans.

ABSTRACT: Landmarking of 3D facial surface scans is an important analysis step in medical and biological applications, such as genome-wide association studies (GWAS). Manual landmarking is often employed with considerable cost and rater dependent variability. Landmarking automatically with minimal training is therefore desirable. We apply statistical ensemble methods to improve automated landmarking of 3D facial surface scans. Base landmarking algorithms using features derived from 3D surface scans are combined using either bagging or stacking. A focus is on low training complexity of maximal 40 training samples with template based landmarking algorithms that have proved successful in such applications. Additionally, we use correlations between landmark coordinates by introducing a search strategy guided by principal components (PCs) of training landmarks. We found that bagging has no useful impact, while stacking strongly improves accuracy to an average error of 1.7 mm across all 21 landmarks in this study, a 22% improvement as compared to a previous, comparable algorithm. Heritability estimates in twin pairs also show improvements when using facial distances from landmarks. Ensemble methods allow improvement of automatic, accurate landmarking of 3D facial images with minimal training which is advantageous in large cohort studies for GWAS and when landmarking needs change or data quality varies.

SUBMITTER: de Jong MA

PROVIDER: S-EPMC5758814 | biostudies-literature | 2018 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Ensemble landmarking of 3D facial surface scans.

de Jong Markus A MA Hysi Pirro P Spector Tim T Niessen Wiro W Koudstaal Maarten J MJ Wolvius Eppo B EB Kayser Manfred M Böhringer Stefan S

Scientific reports 20180108 1

Landmarking of 3D facial surface scans is an important analysis step in medical and biological applications, such as genome-wide association studies (GWAS). Manual landmarking is often employed with considerable cost and rater dependent variability. Landmarking automatically with minimal training is therefore desirable. We apply statistical ensemble methods to improve automated landmarking of 3D facial surface scans. Base landmarking algorithms using features derived from 3D surface scans are co ...[more]

PMID: 29311563

Similar Datasets

Project description:Virtual anthropology (VA) is based on applying anthropological methods currently used to analyse bones to 3D models of human remains. While great advances have been made in this endeavour in the past decade, several interrogations concerning how reliable these models are and what their proper use should be remain unanswered. In this research, a fundamental assumption of VA has been investigated: if the way we perceive and apply an anthropological method is truly similar when looking at bones macroscopically and through various 3D media. In order to answer, 10 skulls of known age and sex were scanned using a computed tomography (CT) scanner and a 3D surface scanner. Two observers separately applied a defined staging method to eight suture sites on these skulls, first looking at the bone macroscopically, then at the 3D surface scan, and finally on the CT scan. Two rounds of observation were carried out by each observer. Intra- and inter-observer error were evaluated, and two sample t-tests used to evaluate if the different types of medium used yielded significantly different observations. The results show a high degree of inter-observer error, and that data obtained from 3D surface scans differ from macroscopic observation (confidence level 95%, P ≤ 0.05). CT scans, in these settings, yielded results comparable to those obtained through macroscopic observations. These results offer many possibilities for future research, including indications on the kind of anthropological methods and anatomical landmarks that might be reliably transferable to the virtual environment. All current methods used in traditional anthropology should be tested, and if they prove unreliable, new techniques to analyse bones from virtual models should be developed.Key pointsLarge discrepancies between observation on dry bones and computer-generated 3D models (surface scans or CT scans) could lead to the re-evaluation of the suitability of traditional anthropological methods for application on 3D models.This preliminary study evaluates whether macroscopic, 3D surface scans, and CT scans viewings generate different observations.The results indicate that the data are not always coherent across all three media of observation.Explanations include the aspect given to the bone by the 3D software, differences between handling bones in real life versus on a computer, and level of expertise of the observers.

Dataset Information

Ensemble landmarking of 3D facial surface scans.

Publications

Ensemble landmarking of 3D facial surface scans.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets