RasterRegisterer

Transforms an image to minimize its difference with another.

Jump to Configuration

Typical Uses

  • Medical image processing
  • Image noise reduction
  • Computer Vision

How does it work?

The RasterRegisterer accepts two sets of raster features:

  • Fixed Images, through the Fixed port, which provide a reference onto which moving images will be transformed.
  • Moving Images, through the Moving port, which will be transformed to match the fixed one.

Fixed images and moving images will be grouped according to the Group By parameter. For each pair of matched features, a new feature will be output from the Registered output port. This feature will be the registered version of the moving image.

Non-raster features will be rejected.

The registration process

The registration process operates on two images. Once a moving image is paired with a fixed image, the registration process begins.

The moving image is transformed in the attempt to minimize the “difference” with the fixed image, which is determined by a metric. A metric is a function that takes to images and outputs a number that indicates how different they are. The metric value increases when the fixed and moving images are increasingly dissimilar. This transformer attempts to construct a transformation that, when applied to the moving image (creating the registered image), minimizes the metric value.

The construction of such a transformation is treated as an optimization problem. The RasterRegisterer starts from an initial transformation and iteratively modifies it in an attempt to minimize the metric value. This process can stop when either a certain tolerance is met, no further improvements can be made, or after a certain number of iterations.

Transformation modifications are performed according to the specified optimizer. The optimizer is a tool that evaluates whether the last change improved the metric value and takes action accordingly, to make a new change to the transformation. At every iteration the change made to the metric is called the step. The way the optimizer controls the transformation is by its transformation parameters. Transformation parameters are a set of free variables that uniquely determine a transformation and are exposed to the optimizer.

The RasterRegisterer offers a choice of Metric, Transformation, and Optimizer options.

Selecting Raster Bands and Palette

The RasterRegisterer operates on whole rasters and ignores any specific band or palette selection.

Alpha Bands and nodata values

Alpha bands can be used to determine whether a certain pixel of the image will be used for registration.

Both the moving and fixed image may have an alpha band. If, at a given point, either the moving or the fixed image’s alpha band have a zero value, that point will not be used for metric evaluation.

For all values other than zero, the metric will be evaluated at that point. What the value is makes no difference.

Rasters with more than one alpha band will be rejected.

Attributes

Attributes from the moving feature will be preserved when output via the Registered port. Rejected features will preserve attributes from both ports.

Coordinate system handling

Features must be in comparable coordinate systems to be registered.

If the Fixed Image of a group has specified a coordinate system then all the Moving Images must have a coordinate system as well. Moving features without a coordinate system would be rejected in this case.

Whereas, if the Fixed Image does not have coordinate system specified, moving features in the same group, that do specify a coordinate system, will be rejected.

If the coordinate systems of a Moving Image and Fixed Image differ, the RasterRegisterer will attempt to reproject the Moving Image onto the coordinate system of the Fixed Image before registering. Moving images that could not be reprojected will also be rejected.

The Registered feature will be in the coordinate system space of the Fixed image.

Usage Notes

Choosing a Raster Transformer

FME has an extensive selection of transformers for working with raster data. They can be generally categorized as working with whole rasters, bands, cells or palettes, and those designed for workflow control or combining raster with vector data.

For information on raster geometry and properties, see Rasters (IFMERaster).

Configuration

Input Ports

Output Ports

Parameters

Depending on the value of the Optimizer parameter, different parameters will be available in the Termination Conditions and Optimization Parameters sections. These parameters are described below.

Editing Transformer Parameters

Using a set of menu options, transformer parameters can be assigned by referencing other elements in the workspace. More advanced functions, such as an advanced editor and an arithmetic editor, are also available in some transformers. To access a menu of these options, click beside the applicable parameter. For more information, see Transformer Parameter Menu Options.

Defining Values

There are several ways to define a value for use in a Transformer. The simplest is to simply type in a value or string, which can include functions of various types such as attribute references, math and string functions, and workspace parameters. There are a number of tools and shortcuts that can assist in constructing values, generally available from the drop-down context menu adjacent to the value field.

Dialog Options - Tables

Transformers with table-style parameters have additional tools for populating and manipulating values.

Reference

Processing Behavior

Feature-Based

Feature Holding

No

Dependencies None
FME Licensing Level FME Professional Edition and above
Aliases  
History  

FME Community

The FME Community is the place for demos, how-tos, articles, FAQs, and more. Get answers to your questions, learn from other users, and suggest, vote, and comment on new features.

Search for all results about the RasterRegisterer on the FME Community.

 

Examples may contain information licensed under the Open Government Licence – Vancouver