Method and system for catheter tracking in fluoroscopic images using adaptive discriminant learning and measurement fusion

Patent metadata

Field	Value
Publication number	US-9014423-B2
Application number	US-201213412882-A
Country	US
Kind code	B2
Filing date	Mar 6, 2012
Priority date	Mar 14, 2011
Publication date	Apr 21, 2015
Grant date	Apr 21, 2015

Abstract

A method and system for adaptive discriminant learning and measurement fusion for image based catheter tracking is disclosed. An adaptive discriminant model is trained online based on a tracked object, such as a pigtail catheter tip, in at least one previous frame of a fluoroscopic image sequence. The object is tracked in the current frame of the fluoroscopic image sequence based at least on the adaptive discriminant model trained online. The object may be tracked in the current frame based on a fusion of three types of measurement models including the adaptive discriminant model trained online, an object detection model trained offline, and an online appearance model.

First claim

The invention claimed is:

1. A method for tracking an object in a sequence of fluoroscopic images, comprising:
training an adaptive discriminant model online based on a tracked object in at least one previous frame of the fluoroscopic image sequence; and
tracking the object in a current frame of the fluoroscopic image sequence based at least on the adaptive discriminant model trained online, wherein tracking the object in a current frame of the fluoroscopic image sequence based at least on the adaptive discriminant model trained online comprises:
tracking the object in the current frame using an overall measurement model that is a fusion of the adaptive discriminant model trained online, an object detection model trained offline, and an online appearance model.

2. The method of claim 1 , wherein the step of training an adaptive discriminant model online based on a tracked object in at least one previous frame of the fluoroscopic image sequence comprises:
extracting positive samples from the tracked object in the at least one previous frame and negative samples away from the tracked object in the at least one previous frame;
updating a linear discriminant vector based on the extracted positive samples and negative samples using a gradient descent method to reduce a Bayesian error; and
updating a probabilistic model based on the updated linear discriminant vector.

3. The method of claim 2 , wherein the step of training an adaptive discriminant model online based on a tracked object in at least one previous frame of the fluoroscopic image sequence further comprises:
repeating the steps of updating the linear discriminant vector and updating the probabilistic model until the linear discriminant vector converges.

4. The method of claim 2 , wherein an initial linear discriminant vector is trained offline based on annotated training data.

5. The method of claim 4 , wherein the initial linear discriminant vector is trained offline using principal component analysis (PCA) to reduce the dimensionality of the training data and Fisher discriminant analysis (FDA) applied to principal components result from the PCA to learn the initial discriminant vector.

6. The method of claim 1 , wherein the object detection model is trained offline based on annotated training data using a probabilistic boosting tree (PBT).

7. The method of claim 1 , wherein the online appearance model calculates a probabilities for image patches in the current frame based on differences between the image patches and an appearance template trained online from the tracked object in the at least one previous frame.

8. The method of claim 1 , wherein the step of tracking the object in the current frame using an overall measurement model that is a fusion of the adaptive discriminant model trained online, an object detection model trained offline, and an online appearance model comprises:
tracking the object in the current frame based on a motion prior probability which is propagated from the at least one previous frame and the overall measurement model.

9. The method of claim 1 , wherein the step of tracking the object in a current frame of the fluoroscopic image sequence based at least on the adaptive discriminant model trained online comprises:
detecting motion parameters m t in the object in the current frame t to maximize a posterior probability P(m t |Z 1:t ):

P
⁡

(

m
t

|

Z

1
:
t

)

∝

P
⁡

(

m
t

|

Z

1
:

t
-
1

)

⁢

∑

k
=
1

3

⁢

⁢

ω
k

⁢

P

(
k
)

⁡

(

Z
t

|

m
t

)

,

where Z i is observed image data of frame i, P(m t |Z 1:t-1 ) is a motion prior probability propagated from the at least one previous frame, P (t) (Z t |m t ) is the adaptive discriminant model trained online, P (2) (Z t |m t ) is a probabilistic measurement from an object detector trained offline, and P (3) (Z t |m t ) is an online appearance model.

10. The method of claim 1 , wherein the object is a pigtail catheter tip.

11. An apparatus for tracking an object in a sequence of fluoroscopic images, comprising:
means for training an adaptive discriminant model online based on a tracked object in at least one previous frame of the fluoroscopic image sequence; and
means for tracking the object in a current frame of the fluoroscopic image sequence based at least on the adaptive discriminant model trained online, comprising:
means for tracking the object in the current frame using an overall measurement model that is a fusion of the adaptive discriminant model trained online, an object detection model trained offline, and an online appearance model.

12. The apparatus of claim 11 , wherein the means for training an adaptive discriminant model online based on a tracked object in at least one previous frame of the fluoroscopic image sequence comprises:
means for extracting positive samples from the tracked object in the at least one previous frame and negative samples away from the tracked object in the at least one previous frame;
means for updating a linear discriminant vect

Assignees

• Wang Peng
• Zheng Yefeng
• John Matthias
• Boese Jan
• Funka-Lea Gareth
• Comaniciu Dorin
• Siemens Ag

Inventors

• Wang Peng
• Zheng Yefeng
• John Matthias
• Boese Jan
• Funka-Lea Gareth
• Comaniciu Dorin

Classifications

• A61B6/12

  Arrangements for detecting or locating foreign bodies · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06T2207/10064

  Fluorescence image · CPC title

• G06T2207/30021

  Catheter; Guide wire · CPC title

• A61B6/503

  for diagnosis of the heart · CPC title