Disclosure of Invention
The invention aims to provide a liver image reconstruction method and a liver image reconstruction system, which can automatically and accurately identify the relative position of hepatic vessels and liver tumors and facilitate doctors to make an operation scheme according to the relative position of the hepatic vessels and the liver tumors of each patient.
In a first aspect, the present invention provides a liver image reconstruction method, including:
acquiring a CT image of a patient in a multi-scale receptive field, wherein the CT image is a liver tumor image or a liver blood vessel image of the patient;
respectively extracting a global semantic feature map and a local visual feature map based on the CT image;
performing feature fusion on the global semantic feature map and the local visual map by using a convolutional neural network to obtain a liver feature map;
and performing three-dimensional image reconstruction based on the liver characteristic diagram, and outputting a reconstruction result, wherein the reconstruction result is a three-dimensional liver tumor image or a three-dimensional liver blood vessel image.
Optionally, the respectively extracting a global semantic feature map and a local visual feature map based on the CT image includes:
carrying out feature extraction and dimension reduction on the CT image to obtain a global semantic feature map;
cutting the CT image according to a preset standard to obtain a local characteristic image;
and extracting visual features in the local feature map to serve as a local visual feature map.
Optionally, the performing feature fusion on the global semantic feature map and the local visual map by using a convolutional neural network to obtain a liver feature map, includes:
fusing and recovering the global semantic feature map and the local visual feature map in a low dimension by using a residual error network to match with sampling;
and performing channel dimension fusion on the global semantic feature map and the local visual feature map with the same scale to obtain a liver feature map.
Optionally, the reconstructing a three-dimensional image based on the liver feature map and outputting a reconstruction result includes:
carrying out gray level processing on the liver characteristic map to determine a first liver characteristic map;
constructing a three-dimensional data field based on the first liver feature map, and selecting data with the highest gray value as an initial liver image;
processing the initial liver image by using morphological operation to determine a target liver image, wherein the morphological operation comprises at least one of dilation, erosion, opening operation and closing operation;
and performing three-dimensional image reconstruction based on the target liver image, and outputting a reconstruction result.
Optionally, the method further includes:
and storing the reconstruction result to a user database, wherein the file format of the reconstruction result in the user database is nrrd format.
In a second aspect, the present invention provides a liver image reconstruction system, including:
the system comprises an image acquisition module, a display module and a display module, wherein the image acquisition module is used for acquiring a CT image of a patient in a multi-scale receptive field, and the CT image is a liver tumor image or a liver blood vessel image of the patient;
the characteristic map determining module is used for respectively extracting a global semantic characteristic map and a local visual characteristic map based on the CT image;
the feature fusion module is used for performing feature fusion on the global semantic feature map and the local visual map by using a convolutional neural network to obtain a liver feature map;
and the three-dimensional reconstruction module is used for reconstructing a three-dimensional image based on the liver characteristic diagram and outputting a reconstruction result, wherein the reconstruction result is a three-dimensional liver tumor image or a three-dimensional liver blood vessel image.
Optionally, the feature map determining module includes:
the global feature acquisition unit is used for carrying out feature extraction and dimension reduction on the CT image to obtain a global semantic feature map;
the local feature acquisition unit is used for cutting the CT image according to a preset standard to obtain a local feature map; and extracting the visual features in the local feature map to be used as a local visual feature map.
Optionally, the feature fusion module includes:
the recovery unit is used for fusing and recovering the low-dimensional global semantic feature map and the local visual feature map by using residual error network matched sampling;
and the fusion unit is used for fusing the global semantic feature map and the local visual feature map with the same scale as a channel dimension to obtain a liver feature map.
Optionally, the three-dimensional reconstruction module includes:
the gray processing unit is used for carrying out gray processing on the liver characteristic map and determining a first liver characteristic map;
the image screening unit is used for constructing a three-dimensional data field based on the first liver characteristic diagram and initiating a liver image;
an operation unit for processing the initial liver image by using morphological operation to determine a target liver image, wherein the morphological operation comprises at least one of dilation, erosion, opening operation and closing operation;
and the three-dimensional image reconstruction unit is used for reconstructing a three-dimensional image based on the target liver image and outputting a reconstruction result.
Optionally, the method further includes:
and the user database is used for storing the reconstruction result, and the file format of the reconstruction result in the user database is the nrrd format.
The embodiment of the invention provides a liver image reconstruction method and a reconstruction system, which can acquire a CT image of a patient in a multi-scale receptive field, wherein the CT image is a liver tumor image or a liver blood vessel image of the patient; respectively extracting a global semantic feature map and a local visual feature map based on the CT image; performing feature fusion on the global semantic feature map and the local visual map by using a convolutional neural network to obtain a liver feature map; and reconstructing a three-dimensional image based on the liver characteristic diagram, and outputting a reconstruction result, wherein the reconstruction result is a three-dimensional liver tumor image or a three-dimensional hepatic blood vessel image. The invention uses the deep learning technology (convolution neural network) to complete the fusion of the global and local characteristics of the CT image of the patient, and the fused liver characteristic diagram contains more accurate liver tumor or liver blood vessel information of the patient.
In addition, because the convolutional neural network is adopted in the scheme, the information of hepatic vessels, hepatic tumors and the like can be automatically identified after the convolutional neural network is trained, hands are liberated, the condition that a doctor manually adds the hepatic vessels, the hepatic tumors and the like is avoided, the whole process is quickly and automatically completed, the intervention of professionals is not needed, and meanwhile, the system is high in transportability, does not need to be specially installed, and is not limited by an operating system, computer configuration and regions.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
Referring to fig. 1, the present invention provides a liver image reconstruction method, which specifically includes:
s11, acquiring a CT image of a patient in a multi-scale receptive field, wherein the CT image is a liver tumor image or a liver blood vessel image of the patient;
the receptive field is one of the important concepts of the convolutional neural network, and the multi-scale receptive field refers to the different ranges of the receptive field. If the receptive field is too small, the convolutional neural network can only observe the local features of the image; if the receptive field is too large, although the global information is more well understood, it usually contains many invalid information. Therefore, in the embodiment of the invention, the CT image under the multi-scale receptive field is obtained to provide sufficient image basis for the subsequent global semantic feature map and the local visual feature map.
Optionally, the CT image is a CT DICOM format picture.
And S12, respectively extracting a global semantic feature map and a local visual feature map based on the CT image.
And S13, performing feature fusion on the global semantic feature map and the local visual map by using a convolutional neural network to obtain a liver feature map.
And S14, reconstructing a three-dimensional image based on the liver characteristic diagram, and outputting a reconstruction result, wherein the reconstruction result is a three-dimensional liver tumor image or a three-dimensional liver blood vessel image.
The invention relates to a liver image reconstruction method and a reconstruction system, which are used for acquiring a CT image of a patient in a multi-scale receptive field, wherein the CT image is a liver tumor image or a liver blood vessel image of the patient; respectively extracting a global semantic feature map and a local visual feature map based on the CT image; performing feature fusion on the global semantic feature map and the local visual map by using a convolutional neural network to obtain a liver feature map; and reconstructing a three-dimensional image based on the liver characteristic diagram, and outputting a reconstruction result, wherein the reconstruction result is a three-dimensional liver tumor image or a three-dimensional hepatic blood vessel image. The invention uses the deep learning technology (convolution neural network) to complete the fusion of the global and local characteristics of the CT image of the patient, and the fused liver characteristic diagram contains more accurate liver tumor or liver blood vessel information of the patient.
In addition, because the convolutional neural network is adopted in the scheme, the information of hepatic vessels, hepatic tumors and the like can be automatically identified after the convolutional neural network is trained, hands are liberated, the condition that a doctor manually adds the hepatic vessels, the hepatic tumors and the like is avoided, the whole process is quickly and automatically completed, the intervention of professionals is not needed, and meanwhile, the system is high in transportability, does not need to be specially installed, and is not limited by an operating system, computer configuration and regions.
The following describes a specific implementation manner in the embodiment of the present invention.
In an embodiment, the extracting the global semantic feature map and the local visual feature map based on the CT image respectively may specifically include:
and performing feature extraction and dimension reduction on the CT image to obtain a global semantic feature map.
Optionally, a convolutional neural network multi-layer multi-head self-attention mechanism may be used to learn the CT image, and then the convolutional neural network is used to perform feature extraction and dimension reduction on the semantic related feature map, and the scales obtained are: 128 × 128, 64 × 64, 32 × 32, 16 × 16, etc. for different scales.
Further, the obtaining of the local visual feature map based on the CT image may be:
cutting the CT image according to a preset standard to obtain a local characteristic map;
and extracting the visual features in the local feature map to be used as a local visual feature map.
In the embodiment, the image in the CT DICOM format based on the same patient is cut according to the preset standard, and the visual features in the local feature map are extracted as the local visual feature map. Optionally, the local visual feature map may also be a feature map under multiple different scale receptive fields, 128 × 128, 64 × 64, 32 × 32, 16 × 16.
In a further embodiment of the present invention, the liver feature map may also be determined based on the global semantic feature map and the local visual feature map.
The specific steps for determining the liver characteristic map comprise:
fusing and recovering the global semantic feature map and the local visual feature map in a low dimension by using a residual error network to match with sampling; and fusing the global semantic feature map and the local visual feature map with the same scale as a channel dimension to obtain a liver feature map.
The residual error network is one of the convolutional neural networks, and specifically comprises the following steps: and (3) skipping the connection of neurons in the next layer by certain layers of the neural network, and connecting interlayer layers to weaken the strong connection between each layer.
Optionally, in the embodiment of the present invention, a residual error network is used to cooperate with sampling to fuse and recover the low-dimensional global semantic feature map and the local visual feature map; and fusing the global semantic feature map and the local visual feature map with the same scale as a channel dimension, and ensuring that a module effectively learns the features with different scales to obtain the liver feature map.
Further, after obtaining the liver feature map, referring to fig. 2, the three-dimensional image reconstruction process may be:
step S21, carrying out gray level processing on the liver characteristic map, and determining a first liver characteristic map;
in addition, after the liver characteristic map is obtained and before the gray level processing, the Gaussian filtering setting threshold value can be used for filtering out irrelevant noise in the liver characteristic map.
S22, constructing a three-dimensional data field based on the first liver feature map, and determining an initial liver image;
optionally, after the three-dimensional data field, data with the highest gray value is selected as an initial liver image.
And step S23, processing the initial liver image by using morphological operation to determine a target liver image.
The basic operations in the morphological operations include dilation, erosion, open operations, close operations, skeleton extraction, limit erosion, hit-on-miss transformation, morphological gradients, top-hat transformation, particle analysis, watershed transformation, gray value morphological gradients, and the like.
And S24, reconstructing a three-dimensional image based on the target liver image and outputting a reconstruction result.
Optionally, the reconstruction result may be obtained after reconstruction using a Visualization Tool Kit (VTK), and output in a form of a window.
In an optional implementation manner, the reconstruction result may be further stored in a user database, and a file format of the reconstruction result in the user database is an nrrd format.
In a further embodiment of the invention, a liver image reconstruction system is also provided. Referring to fig. 3, the liver image reconstruction system specifically includes:
the image acquisition module 300 is configured to acquire a CT image of a patient in a multi-scale receptive field, where the CT image is a liver tumor image or a liver blood vessel image of the patient;
a feature map determination module 400, configured to extract a global semantic feature map and a local visual feature map based on the CT image;
the feature fusion module 500 is configured to perform feature fusion on the global semantic feature map and the local visual map by using a convolutional neural network to obtain a liver feature map;
and the three-dimensional reconstruction module 600 is configured to perform three-dimensional image reconstruction based on the liver feature map, and output a reconstruction result, where the reconstruction result is a three-dimensional liver tumor image or a three-dimensional liver blood vessel image.
Optionally, the feature map determining module 400 includes:
a global feature obtaining unit 410, configured to perform feature extraction and dimension reduction on the CT image to obtain a global semantic feature map;
the local feature obtaining unit 420 is configured to cut the CT image according to a preset standard to obtain a local feature map; and extracting visual features in the local feature map to serve as a local visual feature map.
Optionally, the feature fusion module 500 includes:
a recovering unit 510, configured to fuse and recover the low-dimensional global semantic feature map and the local visual feature map by using a residual network to match with sampling;
and a fusion unit 520, configured to fuse the global semantic feature map and the local visual feature map with the same scale by using channel dimensions to obtain a liver feature map.
Optionally, the three-dimensional reconstruction module 600 includes:
a gray processing unit 610, configured to perform gray processing on the liver feature map to determine a first liver feature map;
an image screening unit 620, configured to construct a three-dimensional data field based on the first liver feature map, and select data with a highest gray value as an initial liver image;
an operation unit 630, configured to process the initial liver image by using a morphological operation, wherein the morphological operation includes at least one of dilation, erosion, opening operation, and closing operation, to determine a target liver image;
and a three-dimensional image reconstruction unit 640, configured to perform three-dimensional image reconstruction based on the target liver image, and output a reconstruction result.
Optionally, the liver image reconstruction system further includes: and the user database 700 is used for storing the reconstruction result, and the file format of the reconstruction result in the user database is nrrd format.
The description of the liver image reconstruction method in this embodiment may refer to the description of the liver image reconstruction method, and will not be repeated here.
While various embodiments of the present invention have been described above, various alternatives described in the various embodiments can be combined and cross-referenced without conflict to extend the variety of possible embodiments that can be considered disclosed and disclosed in connection with the embodiments of the present invention.
Although the embodiments of the present invention have been disclosed, the present invention is not limited thereto. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.