||Image Analysis > Classification > Similarity Map
The similarity map is a simple but efficient tool to detect spectroscopically similar regions of an image. The classification obtained with this tool is either based on the multidimensional distance to a reference spectrum ("end member"), or on some kind of correlation to it.
ImageLab supports the following similarity criteria:
- Euclidean distance
- Scaled Euclidean distance
- Mahalanobis distance
- Autoscaled correlation
- Significant correlation
- Q-normalized correlation
- Spectral angle mapper (SAM)
||Please follow these steps to create a similarity map:
- Select a list of descriptors which should form the basis of the similarity match. If you did not yet specify a set of descriptors, use the Spectral Descriptor Editor to create one.
- Select the similarity criterion
- Click the similarity map and hold the left mouse button pressed while moving the mouse. This moves the reference point and triggers the (re)calculation of the similarity map.
The resulting similarity map is either encoded by a red/blue color palette for correlations, or a red/black color palette for distances. In both cases the red areas are the areas of the highest similarity.