Systems and methods for image processing

1 . An image processing system, comprising:

at least one storage device including a set of instructions; and

at least one processor in communication with the at least one storage device, wherein when executing the set of instructions, the at least one processor is directed to cause the system to perform operations including:

obtaining an original segmented tubular structure of an image by automatically segmenting the image according to an automatic segmentation algorithm;

obtaining, on the original segmented tubular structure, at least one seed point input by a user via a user terminal;

identifying a target region of the image based on the at least on seed point, wherein the target region includes a tubular structure to be segmented;

constructing a distance field of the target region based on a trained model, wherein the distance field includes elements each of which is associated with a distance between the element and a predicted centerline of the tubular structure; and

segmenting the tubular structure to obtain a segmented tubular structure based on the distance field and an optimum path algorithm.

2 . The image processing system of claim 1 , wherein a location or a vicinity of each of the at least one seed point includes an unsegmented part of the tubular structure, and the unsegmented part of the tubular structure is not identified by the automatic segmentation algorithm.

3 . The image processing system of claim 2 , wherein the segmenting the tubular structure to obtain the segmented tubular structure based on the distance field and the optimum path algorithm includes:

determining a target centerline of the tubular structure based on the distance field and the optimum path algorithm; and

segmenting the tubular structure to obtain the segmented tubular structure based on the target centerline and the distance field.

4 . The image processing system of claim 3 , wherein the determining the target centerline of the tubular structure based on the distance field and the optimum path algorithm includes:

determining, based on the distance field and the optimum path algorithm, an optimum path including the at least one seed point; and

designating the optimum path as the target centerline of the tubular structure.

5 . The image processing system of claim 4 , wherein the determining, based on the distance field and the optimum path algorithm, the optimum path including the at least one seed point includes:

selecting one seed point from the at least one seed point as a starting point;

searching, from the starting point, the optimum path that passes through all of the at least one seed point based on the distance field according to the optimum path algorithm.

6 . The image processing system of claim 3 , wherein the segmenting the tubular structure to obtain the segmented tubular structure based on the target centerline and the distance field includes:

determining a growth condition based on the distance field; and

determining the segmented tubular structure by performing a region growing procedure on the target centerline based on the growth condition.

7 . The image processing system of claim 3 , wherein the segmenting the tubular structure to obtain the segmented tubular structure based on the target centerline and the distance field includes:

obtaining a distance between each element and the predicted centerline from the distance field;

identifying an element in the image that has a distance equal to the obtained distance from the distance field; and

segmenting the tubular structure from the identified element.

8 . The image processing system of claim 1 , wherein the constructing the distance field of the target region based on the trained model includes:

determining a plurality of sub-regions of the target region;

for each sub-region of the plurality of sub-regions, determining a sub-distance field by inputting the sub-region into the trained model; and

constructing the distance field of the target region based on the plurality of sub-distance fields.

9 . The image processing system of claim 1 , wherein the operations further include:

determining an optimum segmented tubular structure by processing the segmented tubular structure.

10 . The image processing system of claim 1 , wherein the trained model is obtained by a process including:

obtaining a plurality of training sample sets each of which includes a sample image and a sample distance field of the sample image; and

obtaining the trained model by training a preliminary model based on the plurality of training sample sets.

11 . The image processing system of claim 10 , wherein the obtaining the plurality of training sample sets includes:

obtaining a plurality of sample images;

for each sample image of the plurality of sample images, determining a sample distance filed.

12 . The image processing system of claim 11 , wherein for each sample image of the plurality of sample images, the determining the sample distance field includes:

for each sample image of the plurality of sample images,

determining a sample original segmented tubular structure;

determining a sample processed segmented tubular structure by processing the sample original segmented tubular structure; and

determining the sample distance field based on the sample processed segmented tubular structure.

13 . The image processing system of claim 1 , wherein the trained model includes a Vnet model.

14 . An image processing method, comprising:

identifying a target region of an image, wherein the target region includes a tubular structure to be segmented;

constructing a distance field of the target region based on a trained model, wherein the distance field includes elements each of which is associated with a distance between the element and a predicted centerline of the tubular structure; and

segmenting the tubular structure to obtain a segmented tubular structure based on the distance field and an optimum path algorithm.

15 . The image processing method of claim 14 , wherein the identifying the target region of the image includes:

obtaining at least one seed point; and

identifying the target region based on the at least one seed point.

16 . The image processing method of claim 15 , wherein the segmenting the tubular structure to obtain the segmented tubular structure based on the distance field and the optimum path algorithm includes:

determining a target centerline of the tubular structure based on the distance field and the optimum path algorithm; and

segmenting the tubular structure to obtain the segmented tubular structure based on the target centerline and the distance field.

17 . The image processing method of claim 16 , wherein the determining the target centerline of the tubular structure based on the distance field and the optimum path algorithm includes:

determining, based on the distance field and the optimum path algorithm, an optimum path including the at least one seed point; and

designating the optimum path as the target centerline of the tubular structure.

18 . The image processing method of claim 16 , wherein the segmenting the tubular structure to obtain the segmented tubular structure based on the target centerline and the distance field includes:

determining a growth condition based on the distance field; and

determining the segmented tubular structure by performing a region growing procedure on the target centerline based on the growth condition.

19 . A non-transitory computer readable medium, comprising executable instructions that, when executed by at least one processor, direct the at least one processor to perform a method, the method comprising:

obtaining an original segmented tubular structure of an image by automatically segmenting the image according to an automatic segmentation algorithm;

obtaining at least one seed point input by a user via a user terminal;

identifying a target region of the image, wherein the target region includes a tubular structure to be segmented;

constructing a distance field of the target region based on a trained model, wherein the distance field includes elements each of which is associated with a distance between the element and a predicted centerline of the tubular structure; and

segmenting the tubular structure to obtain a segmented tubular structure based on the distance field and an optimum path algorithm.