US 12,254,404 B2
Image registration in treatment planning
Niklas Gunnarsson, Stockholm (SE); and Jens Sjölund, Sundbyberg (SE)
Assigned to Elekta LTD., Crawley (GB)
Filed by Elekta Ltd, Crawley (GB)
Filed on Feb. 23, 2021, as Appl. No. 17/182,572.
Claims priority of provisional application 62/982,159, filed on Feb. 27, 2020.
Prior Publication US 2021/0268313 A1, Sep. 2, 2021
Int. Cl. G06N 3/08 (2023.01); G06T 7/246 (2017.01); G06T 7/73 (2017.01); A61N 5/10 (2006.01)
CPC G06N 3/08 (2013.01) [G06T 7/246 (2017.01); G06T 7/73 (2017.01); A61N 2005/1055 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method for image registration of images of a patient comprising:
receiving a first set of image data from the patient;
selecting one or more tracking regions in said first set of image data, each tracking region including multiple tracking targets;
receiving multiple sets of image data, each of the multiple sets of image data captured at different times from the patient;
identifying corresponding tracking targets in said multiple sets of image data captured at different times;
creating a sparse motion field describing motion vectors of the multiple tracking targets, the motion vectors specifying a position of a tracking target in reference to a previous position of the tracking target;
estimating a displacement field describing displacement vectors for a number of points in each respective set of said multiple sets of image data, the points located within said one or more tracking regions, and each displacement vector specifying a position of a point in the respective set of image data in reference to a previous position of the same point in another set of the image data captured at a previous time, wherein estimating the displacement field includes using sparse-to-dense interpolation of the sparse motion field to produce the displacement field; and
creating a representation of at least one image based on said displacement field.