Point rendering

This page describes how landmark points are rendered as "spots" in the BigWarp windows.

Transform

The BigWarp currentTransform is an InvertibleRealTransform whose forward direction maps points from target to moving space*, and whose inverse direction maps points from moving to target space.

*This may seem counter-intuitive, but is because transforming moving images to target space requires transforming points from target to moving space

Landmarks

Moving and target landmarks are stored in the LandmarkTableModel.

Landmarks are in the physical (not pixel) space of the moving and physical images, and there are three distinct spaces in which points can live:

1. Moving image space
2. Transformed moving image space
3. Target (fixed) image space

Landmarks in moving image space are stored in movingPts, landmarks in transformed moving image space in warpedPoints, and landmarks in target image space in targetPts.

Landmark placement

Moving

When a landmark p is placed in moving image space:

1. p is added to the movingPts list directly (in physical units).
2. The warped point is set if its available, or flagged to be updated when needed.

Transformed moving

When a landmark p is placed in transformed moving space:

1. p is added to the warpedPoints list directly.
2. p is transformed through the forward currentTransform and the result is added to movingPts

Target

When a landmark p is placed in target space:

1. p is added directly to targetPts