Systems and methods for determining blood vessel conditions

Patent metadata

Field	Value
Publication number	US-10803583-B2
Application number	US-201816056535-A
Country	US
Kind code	B2
Filing date	Aug 7, 2018
Priority date	May 25, 2018
Publication date	Oct 13, 2020
Grant date	Oct 13, 2020

Abstract

The disclosure relates to systems and methods for determining blood vessel conditions. The method includes receiving a sequence of image patches along a blood vessel path acquired by an image acquisition device. The method also includes predicting a sequence of blood vessel condition parameters on the blood vessel path by applying a trained deep learning model to the acquired sequence of image patches on the blood vessel path. The deep learning model includes a data flow neural network, a recursive neural network and a conditional random field model connected in series. The method further includes determining the blood vessel condition based on the sequence of blood vessel condition parameters. The disclosed systems and methods improve the calculation of the sequence of blood vessel condition parameters through an end-to-end training model, including improving the calculation speed, reducing manual intervention for feature extraction, increasing accuracy, and the like.

First claim

What is claimed is:

1. A method for determining a blood vessel condition, comprising:
receiving a sequence of image patches on a blood vessel path acquired by an image acquisition device;
predicting, by a processor, a sequence of blood vessel condition parameters on the blood vessel path by applying a trained deep learning model to the acquired sequence of image patches on the blood vessel path, wherein the deep learning model includes a data flow neural network, a recursive neural network and a conditional random field model, connected in series, wherein the recursive neural network further comprises a forward processing layer and a backward processing layer that do not share a common edge; and
determining, by the processor, the blood vessel condition based on the sequence of blood vessel condition parameters.

2. The method of claim 1 , wherein, the blood vessel condition parameters include at least one of fractional flow reserve, blood flow, blood flow velocity, micro vascular resistance, blood flow pressure drop and vascular stenosis.

3. The method of claim 1 , wherein, the data flow neural network comprises multiple convolutional neural networks,
wherein the sequence of image patches includes multiple sequences of image patches on the blood vessel path,
wherein predicting the sequence of blood vessel condition parameters comprises:
inputting the acquired multiple sequences of image patches into the multiple convolutional neural networks respectively; and
providing the outputs of the multiple convolutional neural networks to the recursive neural network.

4. The method of claim 3 , wherein the image patches are 2D image patches or 3D image patches.

5. The method of claim 1 , wherein the recursive neural network comprises a bidirectional long short-term memory recursive neural network.

6. The method of claim 1 , further comprising:
receiving training data including a training sequence of image patches on the blood vessel path and a corresponding sequence of blood vessel condition parameters; and
training the deep learning model by using the training data.

7. The method of claim 1 , wherein the sequence of image patches on the blood vessel path is acquired by:
acquiring a medical image sequence of a blood vessel tree;
reconstructing a geometric model of the blood vessel tree on the basis of the acquired medical image sequence of the blood vessel tree;
extracting the blood vessel path and a center line of the blood vessel thereon from the geometric model of the blood vessel tree; and
intercepting the sequence of the image patches along the extracted center line of the blood vessel on the blood vessel path.

8. The method of claim 6 , wherein the corresponding sequence of blood vessel condition parameters are obtained by:
acquiring a medical image sequence of a blood vessel tree;
reconstructing a geometric model of the blood vessel tree on the basis of the acquired medical image sequence of the blood vessel tree; and
performing computational fluid dynamics (CFD) simulation on the reconstructed geometric model of the blood vessel tree, so as to obtain the corresponding sequence of blood vessel condition parameters.

9. The method of claim 6 , wherein the corresponding sequence of blood vessel condition parameters are obtained by measuring the blood vessel path.

10. The method of claim 7 , wherein predicting the sequence of the blood vessel condition parameters further comprises:
mapping the predicted sequence of the blood vessel condition parameters back to the blood vessel tree including the blood vessel path.

11. A method for determining a blood vessel condition, comprising:
receiving a sequence of image patches on a blood vessel path acquired by an image acquisition device, wherein the blood vessel path is a part of a blood vessel tree;
predicting, by a processor, a sequence of blood vessel condition parameters on the blood vessel path by applying a trained deep learning model to the acquired sequence of image patches on the blood vessel path, wherein the deep learning model includes a data flow neural network, a recursive neural network and a conditional random field model, connected in series, wherein predicting the sequence of the blood vessel condition parameters further comprises mapping the predicted sequence of the blood vessel condition parameters back to the blood vessel tree including the blood vessel path, wherein mapping the predicted sequence of the blood vessel condition parameters further comprises obtaining a sequence section of blood vessel condition parameters of an overlapping part in the blood vessel tree where individual blood vessel paths overlap with each other, on the basis of sequence sections in the overlapping part of the predicted sequences of the blood vessel condition parameters on the individual blood vessel paths overlapping with each other in the blood vessel tree; and
determining, by the processor, the blood vessel condition based on the sequence of blood vessel condition parameters.

12. A non-transitory computer readable storage medium storing computer executable instructions thereon, which, when executed by a processor, implement a method for determining a blood vessel condition, wherein the method comprises:
receiving a sequence of image patches on a blood vessel path acquired by an image acquisition device;
predicting a sequence of blood vessel condition parameters on the blood vessel path by applying a trained deep learning model to the acquired sequence of image patches, wherein the deep learning model includes a data flow neural network, a recursive neural network and a conditional random field model connected in series, wherein the recursive neural network further comprises a forward processing layer and a backward processing layer that do not share a common edge; and
determining the blood vessel condition based on the sequence of blood vessel condition parameters.

13. The computer readable storage medium of claim 12 , wherein the blood vessel condition parameters include at least one of fractional flow reserve, blood flow, blood flow velocity, micro vascular resistance, blood flow pressure drop and vascular stenosis.

14. The computer readable storage medium of claim 12 , wherein the recursive neural network comprises a bidirectional long short-term memory recursive neural network.

15. The computer readable storage medium of claim 12 , wherein the method further comprises:
receiving training data including a training sequence of image patches on the blood vessel path and the corresponding sequence of the blood vessel condition parameters; and
training the deep learning model by using the training data.

16. The computer readable storage medium of claim 12 , wherein the sequence of image patches on the blood vessel path is acquired by:
acquiring a medical image sequence of a blood vessel tree;
reconstructing a geometric model of the blood vessel tree on the basis of the acquired medical image sequence of the blood vessel tree;
extracting the blood vessel path and a center line of the blood vessel thereon from the geometric model of the blood vessel tree; and
intercepting the sequence of the image patches along the extracted center line of the blood vessel on the blood vessel path.

17. The computer readable storage medium of claim 15 , wherein the corresponding sequence of blood vessel condition parameters is obtained by:
acquiring a medical image sequence of a blood vessel tree;
reconstructing a geometric model of the blood vessel tree on the basis of the acquired medical image sequence of the blood

Assignees

• Shenzhen Keya Medical Tech Corporation

Inventors

• Wang Xin
• Yin Youbing
• Cao Kunlin
• Li Yuwei
• Bai Junjie
• Xu Xiaoyang

Classifications

• G06V10/85

  Markov-related models; Markov random fields · CPC title

• G06V40/14

  Vascular patterns · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/454

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title