Actually extracts features from an image.
Actually extracts features from an image. The image and featurePoint arguments are arguably necessary for all feature extractor implementations. The additional two parameters (mesh and point id) are passed in because the feature extractor might need that information. For instance, the NormalDirectionFeatureExtractor needs that information to determine the "spatial orientation" of the line on which the image points to be examined have to lie.
the image to extract features in
the point in space to extract features at
a mesh instance in correspondence with the image
a point id on the mesh, corresponding to a profiled point id.
Return the points at which feature components are extracted for a given mesh and point.
Return the points at which feature components are extracted for a given mesh and point.
This is mainly intended for visualization, and only really makes sense if there exists a one-to-one correspondence
between such points and the actual feature components returned by the apply()
.
In other words: if a particular implementation uses a method for determining features where there is no correspondence between the feature components and particular points, the method should return None. Otherwise, it should return Some(IndexedSeq(...)), where the length of the sequence matches the length of the feature vector.
For the difference between profilePointId and featurePoint, see the documentation of the apply method.
the mesh to determine the feature points for
the profile point id to determine the feature points for
the actual point in space where the features are to be extracted
a sequence of points, or None if there is no sensible correspondence between feature and points, as outlined above.