Basic concepts for the Aurora Imaging Library Advanced Geometric Matcher module

The basic concepts and vocabulary conventions for the Aurora Imaging Library Advanced Geometric Matcher module are:

Composite-definition model. A model that can hold information about valid model appearances and background samples. This type of model requires training before it can be used to search for occurrences.
Coverage score. The percentage of the total length of the model's edges that are found in an occurrence.
Detection score. The likelihood that a found occurrence is a true occurrence.
Edge. A curve that delineates a boundary, which can be established from intensity transitions in an image.
Edgel. Elementary point (or edge element) within an edge.
Fit score. A measure of the correlation between model edgels and occurrence edgels.
Model. The information that defines the pattern of edges to find in the target.
Model occurrence. An instance of the model found in the target.
Model image. The model, when drawn, is a model image. For a single-definition model defined from an image, the model image is the entire source image or a copy of the region in the source image from which the model has been defined. For a single-definition model defined from an Edge Finder result buffer, the model image is the image source of the result buffer from which the model has been defined. For a single-definition model defined from a CAD DXF file, the model image is the image representation (edge map) of the model. For a composite-definition model, the model image is the entire source image; if a separate image has not been provided for the fit and coverage calculations, no model image exists.
Model origin. The point in the model's coordinate system that is considered to be (0, 0).
Model source. The immediate source of the model's edges. For a single-definition model, the model source is a single user-provided image, the buffer of the Edge Finder results, or a CAD DXF file. For a composite-definition model, the model source for detection is the train AGM result buffer; the model source for the fit and coverage can be the train AGM result buffer or a single user-provided image.
Negative model location. An area in a training image where a model is not found. Negative model locations are background elements and are delimited by red rectangles.
Positive model location. An area in a training image where a model is found. Positive model locations are model appearances and are delimited by blue rectangles.
Single-definition model. A model built from a single model appearance. This type of model can be used to search for occurrences without prior training.
Target. The image or Edge Finder result buffer in which to search for model occurrences.
Trained composite-definition model. A composite-definition model that has been successfully trained with model appearances and background samples, allowing you to use it to search for occurrences.
Training image. An image that contains appearances of the model and/or background.