Systems and methods for three-dimensional navigation of objects

Patent metadata

Field	Value
Publication number	US-12080003-B2
Application number	US-202017760963-A
Country	US
Kind code	B2
Filing date	Sep 24, 2020
Priority date	Sep 24, 2019
Publication date	Sep 3, 2024
Grant date	Sep 3, 2024

Abstract

Disclosed herein are systems and methods for registering a first three-dimensional medical image dataset taken with a first image capturing device with a second 3D dataset taken with a second image capturing device.

First claim

What is claimed is:

1. A method for registering a first three-dimensional (3D) medical image dataset taken with a first image capturing device with a second 3D dataset taken with a second image capturing device, the method comprising:
receiving first 3D medical image dataset of anatomical features of a subject with one or more markers, the first 3D medical image dataset acquired in a first 3D coordinate system using the first image capturing device;
receiving the second 3D dataset with the one or more markers, the second 3D medical image dataset in a second 3D coordinate system using the second image capturing device;
obtaining prior knowledge of the one or more markers;
finding a plurality of voxel blobs from the first 3D medical image dataset based on the prior knowledge of the one or more markers;
clustering the plurality of voxel blobs into a list of clusters of voxels, each of the clusters representing a candidate of the one or more markers;
for each cluster of voxels, finding a line passing at least a number of voxels in the cluster;
for each cluster of voxels, roughly fitting the cluster to one or more pre-determined marker types with one or more parameters, and fine-tune fitting the cluster to the one or more pre-determined marker types with more than two parameters, thereby
generating a corresponding fitting orientation; and
finding an optimal registration transformation among a plurality of combinations, wherein each combination includes at least three points in the first coordinate system and corresponding points in the second coordinate system determined by the second image capturing device.

2. The method of claim 1 comprising:
attaching one or more markers to the anatomical features of the subject;
acquiring the 3D medical image dataset of the anatomical features of the subject with the one or more markers in the first 3D coordinate system using the first image capturing device; and
acquiring the second 3D dataset with the one or more markers in the second 3D coordinate system using the second image capturing device.

3. The method of claim 1 , wherein finding the plurality of voxel blobs from the 3D medical image dataset based on prior knowledge of the one or more markers comprises:
using a threshold value to select a list of candidates, a binary mask is generated using the threshold value;
finding a list of connected candidates optionally using a connected components algorithm;
optionally applying one or more filters on the list of connected candidates thereby generating a plurality of voxel blobs.

4. The method of claim 3 , wherein finding the list of connected candidates uses a connected components algorithm; and
wherein finding the plurality of voxel blobs includes:
applying one or more filters on the list of connected candidates.

5. The method of claim 1 , wherein finding the plurality of voxel blobs from the 3D medical image dataset based on prior knowledge of the one or more markers comprises using a deep learning algorithm.

6. The method of claim 5 , wherein the deep learning algorithm comprises using a deep learning algorithm to segment a list of candidates; finding a list of connected candidates optionally using a connected components algorithm; and optionally applying one or more filters on the list of connected candidates thereby generating a plurality of voxel blobs.

7. The method of claim 1 , wherein the line is a main axis of a corresponding cluster.

8. The method of claim 1 , wherein the one or more parameters comprise an offset of the line and a rotation around the line.

9. The method claim 1 , wherein the more than two parameters include six parameters.

10. The method of claim 1 , wherein the fine-tune fitting includes using a cost function and a weighting.

11. The method of claim 1 , further comprising generating the plurality of combinations, each combination based on one or more candidates of the one or more markers and the corresponding fitting orientations.

12. The method of claim 1 , further comprising:
updating the list of clusters by merging two or more clusters of voxels of the list of clusters using the prior knowledge of the marker.

13. The method of claim 12 , wherein the prior knowledge includes a maximum distance between clusters on a main axis; and
wherein the prior knowledge includes a maximum radius.

14. The method of claim 12 , further comprising:
using random sample consensus to find a cluster.

15. The method of claim 12 , wherein the merging is based on weighting determined by distance between clusters.

16. The method of claim 1 , further comprising;
segmenting one or more anatomical features; and
ignoring one or more voxels based on the segmented anatomical features.

17. The method of claim 1 , further comprising:
coupling the one or more markers to one or more anatomical features of the subject;
generating the first 3D medical image dataset with the first image capturing device; and
generating the second 3D dataset with the second image capturing device.

18. The method of claim 1 , further comprising:
placing an implant after finding the optimal registration transformation.

Assignees

• Nuvasive Inc

Inventors

• Finley Eric
• Avitan Nissim
• Smith Justin
• Shilo Yehiel

Classifications

• G06T2219/2016

  Rotation, translation, scaling · CPC title

• G06T2219/2004

  Aligning objects, relative positioning of parts · CPC title

• G06T2207/30204

  Marker · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06T19/20

  Editing of three-dimensional [3D] images, e.g. changing shapes or colours, aligning objects or positioning parts · CPC title