Impact of Resolution and Texture of Laser Scanning Generated Three-Dimensional Models on Landmark Identification

The aim of the study was to determine the impact of the resolution and texture of three-dimensional (3D) models created through laser scanning on the measurement error (ME) of craniometric landmarks. Ten skulls were scanned at five different resolutions, and the generated 3D models were exported with and without texture. The 3D coordinates of 28 landmarks were derived. Each landmark was picked five times by one observer. The ME of a definite landmark was calculated as an average of distances between the repeated placements of the landmark by the observer and the landmark centroid. One-way analysis of variance was applied for detection of significant differences in the MEs between and within landmark types recorded at different resolutions. The MEs of landmark types in textured and nontextured models were compared by a paired test. Twelve linear measurements were calculated as interlandmark distances, and their values obtained on the models of different resolution were compared. The Frankfurt horizontal plane was constructed for each model and its deviation was calculated at different resolutions. Scan resolution impacted MEs of Type 1 and Type 2 landmarks but not the precision level of Type 3 landmarks. Texture most influenced the precise identification of Type 1 landmarks. The interlandmark distances between Type 2 landmarks were most consistent in their values, those between Type 1 landmarks showed deviations in low-resolution models, and distances between Type 3 landmarks demonstrated various patterns of transition of the values throughout the resolutions. Altogether, the use of textured high-resolution models would be preferable in morphometric studies. Anat Rec, 2019. (c) 2019 American Association for Anatomy