ModelMetrics
Implements utility functions for evaluating the quality of a StatisticalMeshModel
Attributes
- Graph
- Supertypes
- class Objecttrait Matchableclass Any
- Self type
- ModelMetrics.type
Members list
Value members
Concrete methods
Returns the generalization metric of the Statistical Mesh Model, that is how well the model can represent unseen data
Returns the generalization metric of the Statistical Mesh Model, that is how well the model can represent unseen data
Attributes
- dc
test data collection that is in correspondence with the model reference The implementation of this metric is inspired from : Styner, Martin A., et al. "Evaluation of 3D correspondence methods for model building." Information processing in medical imaging. Springer Berlin Heidelberg, 2003. For every mesh in the test data, we project the mesh into the model (that is find the closest shape in the model space to the given mesh) and compute the average mesh distance (see scalismo.mesh.MeshMetrics) between the mesh and the projection. To be able to perform the projection, it is important that the data collection is in correspondence with the model. The returned value is a scala.util.Try containing the average over all test data in case of success, or an Exception otherwise
- pcaModel
Statistical Mesh Model to be evaluated
Returns the specificity metric of the Statistical Mesh Model, that is how close the model remains to the category of shapes it is supposed to represent
Returns the specificity metric of the Statistical Mesh Model, that is how close the model remains to the category of shapes it is supposed to represent
Attributes
- data
test data to verify specificity against
- nbSamples
number of samples drawn to compute the average The implementation of this metric is inspired from : Styner, Martin A., et al. "Evaluation of 3D correspondence methods for model building." Information processing in medical imaging. Springer Berlin Heidelberg, 2003. The general idea is as follows : 1 - sample a shape from the mesh model 2- compute the average mesh distance (see scalismo.mesh.MeshMetrics) of the sample to all elements of the given sequence of meshes and select the minimum distance These steps are then repeated nbSamples times and the average value is returned.
- pcaModel
model to be evaluated