Systems and methods for medical image processing

Patent metadata

Field	Value
Publication number	US-2024289924-A1
Application number	US-202418433370-A
Country	US
Kind code	A1
Filing date	Feb 5, 2024
Priority date	Aug 30, 2021
Publication date	Aug 29, 2024
Grant date	—

Abstract

Systems and methods for medical image processing. The method may include obtaining a contrast image of an object and at least one mask image of the object. The method may further include extracting a plurality of target structures from the at least one mask image by using one or more preset processing algorithms. The method may further include generating a subtracted image by subtracting the plurality of target structures from the contrast image.

First claim

1 . A method for medical image processing, comprising:
obtaining a contrast image of an object and at least one mask image of the object; extracting a plurality of target structures from the at least one mask image by using one or more preset processing algorithms, each of the plurality of target structures corresponding to a part of the contrast image; and generating a subtracted image by subtracting the plurality of target structures from the contrast image.

2 . The method of claim 1 , wherein the extracting a plurality of target structures from the at least one mask image by using one or more preset processing algorithms includes:
extracting a plurality of candidate target structures from the at least one mask image by processing the at least one mask image with the one or more preset processing algorithms; and determining the plurality of target structures based on the plurality of candidate target structures.

3 . The method of claim 2 , wherein the extracting a plurality of candidate target structures from the at least one mask image by processing the at least one mask image with the one or more preset processing algorithms includes:
processing each of the at least one mask image with a different preset processing algorithm from the one or more preset processing algorithms to extract the plurality of candidate target structures.

4 . The method of claim 2 , wherein the determining the plurality of target structures based on the plurality of candidate target structures includes:
determining, based on the contract image, a plurality of structure templates corresponding to the plurality of target structures by using the one or more preset processing algorithms; and determining, based on the plurality of candidate target structures and the plurality of structure templates, the plurality of target structures, each of the plurality of target structures corresponding to one of the plurality of structure templates.

5 . The method of claim 4 , wherein the determining, based on the plurality of candidate target structures and the plurality of structure templates, the plurality of target structures, each of the plurality of target structures corresponding to one of the plurality of structure templates, includes:
for each of the plurality of structure templates:
selecting, from the plurality of candidate target structures, at least two candidate target structures corresponding to a structure template;
generating a combined candidate target structure by combining the at least two candidate target structures; and
designating the combined candidate target structure as a target structure corresponding to the structure template.

6 . The method of claim 5 , wherein the generating a combined target structure by combining the at least two candidate target structures includes:
obtaining one or more high-frequency components and one or more low-frequency components of each of the at least two candidate target structures; and generating the combined candidate target structure by fusing a low-frequency component of one of the at least two candidate target structures and the high-frequency components of the at least two candidate target structures.

7 . The method of claim 6 , wherein the generating a combined candidate target structure by combining the at least two candidate target structures includes:
performing one or more rounds of decomposition to each of the at least two candidate target structures to generate a high-frequency component and a low-frequency component after each round of decomposition; obtaining a low-frequency component after the first round of decomposition of a candidate target structure of the at least two candidate target structures, a similarity between the candidate target structure and the structure template being a highest similarity among a plurality of similarities between the plurality of candidate target structures and the structure template; obtaining high-frequency components of the at least two candidate target structures after the one or more rounds of decomposition; fusing the high-frequency components of the at least two candidate target structures after the one or more rounds of decomposition to produce a fused high-frequency component; and generating the combined candidate target structure by fusing the low-frequency component of the candidate target structure and the fused high-frequency component.

8 . The method of claim 7 , wherein
the one or more rounds of the decomposition include four or five rounds; and the high-frequency components of the at least two candidate target structures used to generate the fused high-frequency component are generated in the last round of decomposition.

9 . The method of claim 5 , wherein the generating a combined candidate target structure by combining the at least two candidate target structures includes:
determining at least two similarities, each of the at least two similarities being a similarity between the structure template and one of the at least two candidate target structures; determining, based on the at least two similarities, at least two weights corresponding to the at least two candidate target structures; and determining the combined candidate target structure corresponding to the structure template by combining the plurality of corresponding candidate target structures based on the plurality of weights.

10 . The method of claim 1 , wherein
the plurality of target structures include at least one movement structure and at least one non-movement structure; or the plurality of target structures include a plurality of movement structures of different movement types.

11 . The method of claim 1 , further comprising:
determining a heartbeat status of the object corresponding to each of a plurality of mask images and a heartbeat status of the object corresponding to the contrast image; and selecting, from the plurality of mask images, the at least one mask image corresponding to heartbeat status that is the same as or substantially similar to the heartbeat status of the object corresponding to the contrast image.

12 . The method of claim 11 , wherein the selecting, from the plurality of mask images, the at least one mask image corresponding to heartbeat status that is the same as or substantially similar to the heartbeat status of the object corresponding to the contrast image includes:
determining image acquisition frequencies of the plurality of mask images and the contrast image and a cardiac cycle of the object; and selecting, from the plurality of mask images, based on the image acquisition frequencies of the plurality of mask images and the contrast image and a cardiac cycle of the object, the at least one mask image corresponding to heartbeat status that is the same as or substantially similar to the heartbeat status of the object corresponding to the contrast image.

13 . The method of claim 1 , further comprising:
determining a plurality of similarities, each of the plurality of similarities being a similarity between the contrast image and one of the plurality of mask images; and selecting, from the plurality of mask images, the at least one mask image based on the plurality of similarities.

14 . (canceled)

15 . (canceled)

16 . A method for medical image processing, comprising:
obtaining a contrast image of an object and a plurality of mask images of the object; selecting at least two mask images from the plurality of mask images based on the contrast image; obtaining one or more high-frequency components and one or more low-frequency components of each of the

Assignees

• Shanghai United Imaging Healthcare Co Ltd

Inventors

• Yue Liang
• Feng Juan
• Niu Jie
• Jiang Chunhua

Classifications

• G06T2207/30048

  Heart; Cardiac · CPC title

• G06T5/50 Primary

  using two or more images, e.g. averaging or subtraction · CPC title

• G06T2207/20224

  Image subtraction · CPC title

• G06V10/761

  Proximity, similarity or dissimilarity measures · CPC title

• G06T2207/20081

  Training; Learning · CPC title