Systems, methods, and computer-readable media for non-rigid registration of electromagnetic navigation space to CT volume

A system, comprising:

• a workstation operably coupled to a medical device, the workstation including a memory and a processor, the memory storing instructions thereon, which when executed, cause the processor to:
• generate a three-dimensional (3D) model of a luminal network of a patient, the 3D model defining an image-coordinate space;
• generate a navigation plan to an area of interest within the luminal network 3D model, the navigation plan including a plurality of reference points;
• define first fiducial points selected from the plurality of reference points;
• perform an initial registration of the image-coordinate space to a body-coordinate space, including:
• obtaining a plurality of survey points within the body-coordinate space, wherein the plurality of survey points corresponds to identified locations of the medical device within the luminal network;
• defining second fiducial points selected from the plurality of survey points;
• matching second fiducial points to first fiducial points to define matched pairs of fiducial points;
• and
• registering the image-coordinate space to the body-coordinate space using the matched pairs of fiducial points;
• and
• perform a secondary registration including:
• identifying non-fiducial survey points of the plurality of survey points;
• matching the non-fiducial survey points, the non-fiducial survey points including at least one survey point of the plurality of survey points other than the second fiducial points, to non-fiducial reference points, the non-fiducial reference points including at least one reference point of the plurality of reference points other than the first fiducial points, based on an inverse registration of the image-coordinate space to the body-coordinate space to define matched pairs of additional points, wherein the inverse registration includes a collection of optimized region registrations;
• deriving a transformation using the matched pairs of fiducial points and the matched pairs of additional points;
• and
• updating the mapping of the image-coordinate space to the body-coordinate space using the derived transformation.

A system, comprising:

• a workstation operably coupled to a medical device, the workstation including a memory and a processor, the memory storing instructions thereon, which when executed, cause the processor to:
• during a planning phase:
• generate a three-dimensional (3D) model of a luminal network of a patient, the 3D model defining an image-coordinate space;
• generate a navigation plan to an area of interest within the luminal network 3D model, the navigation plan including a plurality of reference points;
• and
• define first fiducial points, the first fiducial points selected from the plurality of reference points;
• and
• during a registration phase:
• obtain a plurality of survey points within the body-coordinate space, wherein the plurality of survey points corresponds to identified locations of the medical device within the luminal network;
• define second fiducial points, the second fiducial points selected from the plurality of survey points;
• match second fiducial points to first fiducial points;
• match non-fiducial survey points, the non-fiducial survey points including at least one survey point of the plurality of survey points other than the second fiducial points, to non-fiducial reference points, the non-fiducial reference points including at least one reference point of the plurality of reference points other than the first fiducial points, based on an inverse registration of the image-coordinate space to the body-coordinate space, wherein the inverse registration includes a collection of optimized region registrations;
• derive a transformation using the matched pairs of fiducial points and the matched pairs of non-fiducial points;
• and
• map the image-coordinate space to the body-coordinate space using the derived transformation.