Main Content

Load Images, Spatial Referencing Information, and Initial Transformation

You can load images into Registration Estimator from file or from the workspace. You can also provide optional spatial referencing information and an optional initial geometric transformation.

Load Images from File or Workspace

You can load images into Registration Estimator from file or from the workspace. Although you can load grayscale or color images, the app converts all RGB images to grayscale by using the rgb2gray function. Registration Estimator supports only 2-D images.

Loading images from file supports only BMP, JPG, JPEG, TIF, TIFF, PNG, and DCM file types. To work with a wider range of file formats, load images from the workspace. Registration Estimator supports any image read into the workspace by the imread function and DICOM image read into the workspace using the dicomread function.

Load images into the app by clicking the Load Images icon.

  • To load images from a file, select the Load from file option. In the dialog box, specify the file path of the moving and fixed images. Use Browse to navigate to a folder.

    "Load Images to Register" dialog box with fields to specify the file path of the moving and fixed images.

  • To load variables from the workspace, select the Load from workspace option. In the dialog box, select the name of the variable containing the moving image from the Moving Image menu and the variable containing the fixed image from the Fixed Image menu.

    "Load Images to Register" dialog box with dropdowns to select the moving image and fixed image variables from the workspace.

Provide Spatial Referencing Information

If you have 2-D spatial referencing objects in your workspace, or if you load DICOM images from a file, then you can provide optional spatial referencing information. Spatial referencing information is useful if you want to orient the images to a world coordinate system. For more information about spatial referencing objects, see imref2d.


If you load DICOM images into the workspace using dicomread, spatial referencing information in the metadata is no longer associated with the image data. To preserve spatial referencing information with DICOM images, either load the images from file or create an imref2d object from the image metadata. For more information about DICOM metadata, see Read Metadata from DICOM Files.

If you do not have spatial referencing information, then the Spatial Referencing Object and DICOM Metadata radio buttons are inactive.

Provide an Initial Geometric Transformation

You can provide an optional initial geometric transformation using affine2d and projective2d geometric transformation objects in your workspace. An initial geometric transformation is useful if you are processing a batch of images with similar initial misalignment. Once the first moving image has been registered, you can export the geometric transformation to the workspace and apply the transformation to other images in the series. See Export Results from Registration Estimator App.

If you do not have a geometric transformation object in your workspace, the Initial Transformation Object selection box is inactive.


If you have a 2-D geometric transformation object that uses the premultiply convention, such as an affinetform2d or projtform2d object, then you can convert the object to a projective2d object for use with Registration Estimator. First, get the A property of your geometric transformation object, then create a projective2d object using the transpose of the A property. For example, suppose you have a premultiply geometric transformation object called myTform:

A = myTform.A;
newTform = projective2d(A');

See Also



Related Topics