Method and system for physiological image registration and fusion

Patent metadata

Field	Value
Publication number	US-9547902-B2
Application number	US-56248309-A
Country	US
Kind code	B2
Filing date	Sep 18, 2009
Priority date	Sep 18, 2008
Publication date	Jan 17, 2017
Grant date	Jan 17, 2017

Abstract

A method and system for physiological image registration and fusion is disclosed. A physiological model of a target anatomical structure in estimated each of a first image and a second image. The physiological model is estimated using database-guided discriminative machine learning-based estimation. A fused image is then generated by registering the first and second images based on correspondences between the physiological model estimated in each of the first and second images.

First claim

The invention claimed is:

1. A method for physiological registration and fusion of multiple images, comprising:
independently estimating a respective physiological model of a target anatomical structure in each of a first image and a second image; and
generating a fused image by registering the first and second images based on correspondences between a first plurality of control points representing the physiological model estimated in the first image and a second plurality of control points representing the physiological model estimated in the second image.

2. The method of claim 1 , wherein the physiological model estimated in each of the first image and the second image is generated based on a mathematical representation of the target anatomic structure in a set of annotated training data.

3. The method of claim 1 , wherein said step of independently estimating a respective physiological model of a target anatomical structure in each of a first image and a second image comprises:
fitting the respective physiological model to each of the first and second images using discriminative machine-learning based estimation.

4. The method of claim 3 , wherein said step of fitting the respective physiological model to each of the first and second image using discriminative machine-learning based estimation comprises:
estimating a position, orientation, and scale of the respective physiological model in each of the first and second images using marginal space learning (MSL) with a discriminative classifiers trained based on annotated training data.

5. The method of claim 4 , wherein said step of independently estimating a respective physiological model of a target anatomical structure in each of a first image and a second image further comprises:
refining the respective estimated physiological model in each of the first and second images using learning-based boundary detection.

6. The method of claim 1 , wherein said step of generating a fused image by registering the first and second images based on correspondences between a first plurality of control points representing the physiological model estimated in the first image and a second plurality of control points representing the physiological model estimated in the second image comprises:
estimating a transformation between corresponding ones of the first plurality of control points of the physiological model estimated in the first image and the second plurality of control points of the physiological model estimated in the second image.

7. The method of claim 6 , wherein said step of estimating a transformation between corresponding ones of the first plurality of control points of the physiological model estimated in the first image and the second plurality of control points of the physiological model estimated in the second image comprises:
estimating the transformation using a thin-plate-spline transformation.

8. The method of claim 1 , wherein said first and second images are generated using different imaging modalities.

9. The method of claim 1 , wherein the second image is a dynamic image sequence comprising a plurality of frames received in real-time during a medical procedure and said step of independently estimating a respective physiological model of a target anatomical structure in each of a first image and a second image comprises:
estimating the physiological model of the target anatomic structure in the first image in advance of the medical procedure to generate a patient specific physiological model of the target anatomic structure; and
estimating the patient specific physiological model of the target anatomic structure in each frame of said dynamic image sequence in real-time during the medical procedure as each frame of said dynamic image sequence is received.

10. The method of claim 9 , wherein said step of generating a fused image by registering the first and second images based on correspondences between a first plurality of control points representing the physiological model estimated in the first image and a second plurality of control points representing the physiological model estimated in the second image comprises:
registering the first image and each frame of said dynamic image sequence in real time during the medical procedure based on correspondences between the first plurality of control points of the patient specific physiological model of the target anatomic structure estimated in the first image and the second plurality of control points of the patient-specific physiological model estimated in each frame of said dynamic image sequence.

11. The method of claim 9 , wherein the first image is one of a computed tomography (CT) image and a magnetic resonance (MR) image of the target anatomic structure and the second image is an ultrasound image of the target anatomic structure.

12. The method of claim 1 , wherein the first image is a cardiac computed tomography (CT) image, the second image is a Doppler echocardiography image, and the physiological model is a physiological model one or more heart chambers.

13. The method of claim 1 , wherein the first and second images are images of a region of interest of a patient generated using the same imaging modality at different times during a particular medical treatment on a particular region of interest, and said step of generating a fused image by registering the first and second images based on correspondences between a first plurality of control points representing the physiological model estimated in the first image and a second plurality of control points representing the physiological model estimated in the second image comprises:
registering the first and second images based on correspondences between corresponding ones of the first plurality of control points and the second plurality of control points on a portion of the physiological model not affected by the treatment in each of the first and second images.

14. An apparatus for physiological registration and fusion of multiple images, comprising:
means for independently estimating a respective physiological model of a target anatomical structure in each of a first image and a second image; and
means for generating a fused image by registering the first and second images based on correspondences between a first plurality of control points representing the physiological model estimated in the first image and a second plurality of control points representing the physiological model estimated in the second image.

15. The apparatus of claim 14 , wherein the physiological model estimated in each of the first image and the second image is generated based on a mathematical representation of the target anatomic structure in a set of annotated training data.

16. The apparatus of claim 15 , wherein said means for independently estimating a respective physiological model of a target anatomical structure in each of a first image and a second image comprises:
means for fitting the respective physiological model to each of the first and second images using discriminative machine-learning based estimation.

17. The apparatus of claim 14 , wherein said means for generating a fused image by registering the first and second images based on correspondences between a first plurality of control points representing the physiological model estimated in the first image and a second plurality of control points representing the physiological model estimated in the second image comprises:
means for estimating a transformation between corresponding ones of the first plurality of control points of the physiological model esti

Assignees

• Ionasec Razvan
• Georgescu Bogdan
• Zheng Yefeng
• Comaniciu Dorin
• Siemens Healthcare Gmbh

Inventors

• Ionasec Razvan
• Georgescu Bogdan
• Zheng Yefeng
• Comaniciu Dorin

Classifications

• G06V10/7557

  based on appearance, e.g. active appearance models [AAM] · CPC title

• G06K2209/05

  Physics · mapped topic

• G06K9/621

  Physics · mapped topic

• G06T7/0034 Primary

  Physics · mapped topic

• G06T7/0032

  Physics · mapped topic