An R Script to Estimate Which and How Many Landmarks Represent the Most Variable Parts of an Object
SUMMARY: The study of the shape variation in geometric morphometrics has an important limitation known as the Pinocchio effect. The Pinocchio effect produces artifactual variances of the landmarks and implies that it is not possible to know the morphological change structure of an object, other than by dividing the landmark sets and then comparing them. This, however, involves making prior assumptions about the pattern of variation of an object. In this study, we provide a code in R to iterate over a complete set of landmarks and test all possible combinations of landmarks until deliver those landmarks associated with the largest to the smallest morphological changes. We tested this on a sample of 28 landmarks in 143 3D models of human skulls. The results indicated that this process can result in a pooled variance of a subset of landmarks that is an order of magnitude larger than that of several other regions of the skull. This method makes it possible to describe the pattern of variation of any 2D or 3D object represented by fixed landmarks, to distinguish the shape features that have more morphological dispersion, and to avoid any aprioristic assumptions about how the morphological changes of an object behave.
| Main Authors: | , , , |
|---|---|
| Format: | Digital revista |
| Language: | English |
| Published: |
Sociedad Chilena de Anatomía
2024
|
| Online Access: | http://www.scielo.cl/scielo.php?script=sci_arttext&pid=S0717-95022024000200289 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | SUMMARY: The study of the shape variation in geometric morphometrics has an important limitation known as the Pinocchio effect. The Pinocchio effect produces artifactual variances of the landmarks and implies that it is not possible to know the morphological change structure of an object, other than by dividing the landmark sets and then comparing them. This, however, involves making prior assumptions about the pattern of variation of an object. In this study, we provide a code in R to iterate over a complete set of landmarks and test all possible combinations of landmarks until deliver those landmarks associated with the largest to the smallest morphological changes. We tested this on a sample of 28 landmarks in 143 3D models of human skulls. The results indicated that this process can result in a pooled variance of a subset of landmarks that is an order of magnitude larger than that of several other regions of the skull. This method makes it possible to describe the pattern of variation of any 2D or 3D object represented by fixed landmarks, to distinguish the shape features that have more morphological dispersion, and to avoid any aprioristic assumptions about how the morphological changes of an object behave. |
|---|