CN111028246A - Medical image segmentation method and device, storage medium and electronic equipment - Google Patents
Medical image segmentation method and device, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN111028246A CN111028246A CN201911252001.4A CN201911252001A CN111028246A CN 111028246 A CN111028246 A CN 111028246A CN 201911252001 A CN201911252001 A CN 201911252001A CN 111028246 A CN111028246 A CN 111028246A
- Authority
- CN
- China
- Prior art keywords
- sample data
- medical image
- module
- image segmentation
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30016—Brain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Abstract
The invention discloses a medical image segmentation method, a medical image segmentation device, a storage medium and electronic equipment. The method comprises the following steps: acquiring a medical image to be processed, and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder on the basis of a multi-scale cavity convolution module; determining a segmentation target of the medical image based on an output result of the image segmentation model. According to the technical scheme of the embodiment, the image segmentation model with the scale-cavity convolution module is arranged, high-precision image segmentation is carried out on the medical image, and details and spatial information of the segmented target are reserved on the basis of obtaining the accurate segmented target.
Description
Technical Field
The embodiment of the invention relates to the technical field of image processing, in particular to a medical image segmentation method, a medical image segmentation device, a storage medium and electronic equipment.
Background
Brain tumors are one of the most common and aggressive primary brain tumors, seriously compromising human health. Based on the accurate segmentation result of the brain tumor, doctors can obtain various information of the tumor such as the shape, the size, the position and the like, and carry out quantitative analysis and tracking comparison on the information to master the development and the growth state of the tumor lesion.
The existing brain tumor segmentation method cannot accurately extract detail features and integrate multi-scale receptive field information, so that the brain tumor segmentation precision is poor.
Disclosure of Invention
The invention provides a medical image segmentation method, a medical image segmentation device, a storage medium and electronic equipment, and aims to improve the medical image segmentation precision.
In a first aspect, an embodiment of the present invention provides a medical image segmentation method, including:
acquiring a medical image to be processed, and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder on the basis of a multi-scale cavity convolution module;
determining a segmentation target of the medical image based on an output result of the image segmentation model, wherein the segmentation target is highlighted in the output result.
In a second aspect, an embodiment of the present invention further provides a medical image segmentation apparatus, including:
the medical image input module is used for acquiring a medical image to be processed and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder on the basis of a multi-scale hole convolution module;
a segmentation target determination module for determining a segmentation target of the medical image based on an output result of the image segmentation model, wherein the segmentation target is highlighted in the output result.
In a third aspect, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement an operation recommendation method according to any embodiment of the present invention.
In a fourth aspect, embodiments of the present invention further provide a storage medium containing computer-executable instructions, where the computer-executable instructions, when executed by a computer processor, implement an operation recommendation method according to any of the embodiments of the present invention.
According to the technical scheme provided by the embodiment of the invention, the image segmentation model with the scale cavity convolution module is arranged, the high-precision image segmentation is carried out on the medical image, and the detail information of the segmented target is reserved on the basis of obtaining the accurate segmented target.
Drawings
Fig. 1 is a schematic flow chart of a medical image segmentation method according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of an image segmentation model according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a multi-scale hole convolution module according to an embodiment of the present invention;
FIG. 4 is a comparison graph of the segmentation results of brain tumor images in different image segmentation modes according to an embodiment of the present invention;
fig. 5 is a flowchart illustrating a medical image segmentation method according to a second embodiment of the present invention;
fig. 6 is a schematic structural diagram of a medical image segmentation apparatus according to a third embodiment of the present invention;
fig. 7 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a medical image segmentation method according to an embodiment of the present invention, where the embodiment is applicable to a case where a medical image is accurately segmented, and the method may be implemented by a medical image segmentation apparatus provided in this embodiment, where the apparatus may be implemented by software and/or hardware, and the apparatus may be integrated in an electronic device such as a computer. The method specifically comprises the following steps:
s110, obtaining a medical image to be processed, and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder based on a multi-scale cavity convolution module.
And S120, determining a segmentation target of the medical image based on the output result of the image segmentation model, wherein the segmentation target is highlighted and displayed in the output result.
In this embodiment, the medical image may be a medical image acquired by a medical acquisition device at any detected position of the target object, for example, the medical image may be, but is not limited to, a magnetic resonance image, an ultrasound image, or a CT image, and further, the medical image may be, but is not limited to, a brain tumor image, a breast image, or an intervertebral disc image acquired by the medical acquisition device.
The medical image is automatically segmented through the pre-trained image segmentation model to obtain an accurate segmentation target, and the segmentation precision and the segmentation efficiency of the medical image are improved.
Exemplarily, referring to fig. 2, fig. 2 is a schematic structural diagram of an image segmentation model according to an embodiment of the present invention. The image segmentation model in fig. 2 includes an encoder for extracting image features in a medical image by downsampling and a decoder for restoring feature information by upsampling, and a multi-scale hole convolution module is transversely disposed between the encoder and the decoder. The image segmentation module provided with the multi-scale cavity convolution module can improve the fusion of deep context semantic information and shallow detail texture information of the tumor image, realize detail characteristics and integrate multi-scale receptive field information, solve the problem that the traditional convolution layer has contradiction between characteristic resolution and receptive field for semantic segmentation tasks, increase the receptive field without reducing the resolution of the characteristics when the multi-scale cavity convolution module is applied, and improve the segmentation detail information of a segmentation target in a medical image.
For example, referring to fig. 2, the encoder includes 3 first residual modules connected in sequence, and the decoder includes 3 second residual modules connected in sequence. And a first residual error module in the encoder is transversely connected with a second residual error module in the decoder through a multi-scale hole convolution module. It should be noted that the number of residual modules in the encoder and the decoder can be adjusted according to the user requirement.
Optionally, for the first residual module or the second residual module, the method includes: the device comprises a first convolution layer and a second convolution layer, wherein a regularization layer and an activation function layer are respectively arranged behind the first convolution layer and the second convolution layer, an input end and an output end in a residual error module are in short circuit, convolution kernels of the first convolution layer and the second convolution layer can be 3 x 3, and the activation function layer can be a RELU function. In the encoder, any first residual error module is used for performing convolution processing on input sample data or output data of a last first residual error module, and outputting a processing result to a next first residual error module and a multi-scale cavity convolution module connected with the first residual error module. Illustratively, a first residual error module in the encoder performs feature extraction on input sample data, and outputs obtained feature information to a next first residual error module and a multi-scale cavity convolution module connected with the first residual error module. In the decoder, any one of the second residual error modules is used for performing convolution processing on the output result of the last second residual error module or the output result of the encoder and the output result of the multi-scale hole convolution module connected with the second residual error module.
In fig. 2, the encoder includes three first residual error modules, which sequentially obtain feature information of different levels, and the feature information of different levels is respectively processed by the multi-scale cavity convolution module to obtain multi-scale feature information of different levels, the decoder includes three second residual error modules, each of which respectively fuses the feature information of the previous residual error module and the multi-scale feature information obtained by processing by the multi-scale cavity convolution module to realize the fusion of the multi-scale feature information of different levels, wherein the feature information obtained by the fusion includes both the deep-level detail feature information and the shallow-level contour position feature information, thereby improving the integration capability of the multi-scale receptive field information.
Optionally, the multi-scale void convolution module includes at least two convolution layers connected in sequence and a fusion layer connected to the at least two convolution layers, respectively, where the number of channels of the at least two convolution layers is the same. Referring to fig. 3, fig. 3 is a schematic structural diagram of a multi-scale hole convolution module according to an embodiment of the present invention. The multi-scale void convolution module comprises four convolution layers which are connected in sequence. A first convolution layer consisting of convolutions with a cavity convolution coefficient of 1 and convolution kernel size of 1 × 1; the second convolution consists of convolution with a hole convolution coefficient of 12 and a convolution kernel of 3 x 3; the third convolution is composed of convolution with a cavity convolution coefficient of 24 and a convolution kernel of 3 multiplied by 3; the fourth convolution is composed of convolution with cavity convolution coefficient of 36 and convolution kernel of 3 × 3, and the channels of the four convolution layers are the same. The multi-scale cavity convolution module further comprises a fusion layer (Concat layer) which is respectively connected with the four convolution layers and used for fusing the output characteristics of the four convolution layers. The multi-scale cavity convolution module performs multi-scale feature extraction on input data through convolution layers of different cavity convolution coefficients and performs fusion through a fusion layer to realize fusion and superposition of multi-scale feature information.
Optionally, the image segmentation model further includes a softmax layer, configured to determine a prediction probability of each tag according to output information of the decoder.
In this embodiment, a medical image is input into an image segmentation model as shown in fig. 2, multi-scale feature extraction and fusion are performed on the medical image, the input medical image is identified based on the extracted multi-scale feature information, and a segmentation target is output. It should be noted that the output result of the image segmentation model may include only the segmentation target, or may be to highlight the segmentation target in the medical image, for example, the segmentation target may be highlighted by adjusting the color of the segmentation target in the medical image, and illustratively, the segmentation target of the medical image is black, and other areas are white; illustratively, when the segmentation target in the medical image includes multiple classifications, the pixel points of different classifications are set to different colors for differential display. It should be noted that fig. 2 is only an example, in other embodiments, the size of some input images is not limited to 128 × 128 × 4, but may be other sizes, such as 240 × 240 × 4, and the like, and the input medical image may be a two-dimensional image, such as an image with a size of 128 × 128 or 240 × 240, and the like.
In some embodiments, the medical image is a brain tumor image, and includes four different magnetic sequence representations, i.e., Flair modality, T1 modality, T2 modality, and T1c modality, the image of each modality may be a two-dimensional image, for example, 240 × 240 image, and accordingly, the input image is 240 × 240 × 4 image, where 4 represents Flair modality, T1 modality, T2 modality, and T1c modality. The image segmentation model processes the input image to obtain a segmentation result of the brain tumor region.
When the acquired brain tumor image is a three-dimensional image, the three-dimensional brain tumor image is sliced to obtain a two-dimensional brain tumor image, and the two-dimensional brain tumor image is image-segmented based on the image segmentation model.
Exemplarily, referring to fig. 4, fig. 4 is a comparison graph of the segmentation results of the brain tumor images according to different image segmentation modes provided by the embodiment of the present invention. As can be seen from fig. 4, the brain tumor segmentation result obtained by the image segmentation model provided in this embodiment includes detail information of the brain tumor edge, so that the accuracy of brain tumor segmentation is improved. Referring to table 1, table 1 shows the parameter comparison of the tumor image segmentation result based on ResU-Net and the image segmentation model provided in the embodiments of the present application (i.e., the multi-scale cavity network model). The tumor image segmentation result parameter in table 1 is the Dice score on the BraTS 2017 data set.
TABLE 1
| Method of producing a composite material | Complete tumor | Core tumor | Tumor enhancement |
| ResU-Net | 0.8675 | 0.7646 | 0.6884 |
| Multi-scale cavity network model (our) | 0.8754 | 0.7669 | 0.6967 |
As is apparent from table 1, the image segmentation model provided in the embodiment of the present application has high-precision segmentation on image segmentation of a complete tumor, a core tumor, and an enhanced tumor.
According to the technical scheme provided by the embodiment, the image segmentation model of the configured scale-cavity convolution module is arranged, high-precision image segmentation is carried out on the medical image, and the detail information of the segmented target is reserved on the basis of obtaining the accurate segmented target.
Example two
Fig. 5 is a schematic flow chart of a medical image segmentation method provided by a second embodiment of the present invention, which is optimized on the basis of the second embodiment, where the method includes:
and S210, acquiring sample data and a label of the sample data.
And the sample data and the label of the sample data are used for training to obtain an image segmentation model. Optionally, the obtaining of the sample data and the label of the sample data includes: acquiring initial sample data and an initial label of an initial sample image; sliding on the initial sample data based on a preset sliding window, and intercepting at least one sample data, wherein the sample data comprises a target to be segmented; and determining the label of the intercepted at least one sample data based on the corresponding relation between the initial sample data and the initial label.
For example, taking a brain tumor image as an example, the initial sample data may be the brain tumor image acquired by the magnetic resonance imaging apparatus, and the initial label includes classification identifiers of pixels in the initial sample image, where the initial sample image at least includes a segmentation target region and a background region, correspondingly, the pixels in different regions belong to different types, different types of pixels are respectively provided with different classification identifiers, and one pixel has only one classification identifier, and for example, the classification identifier of the pixel in the segmentation target region may be 1, and the classification identifier of the pixel in the background region may be 0. Taking the brain tumor image as an example, the initial sample image may include a background region, a necrotic and non-enhanced region, an edema region, and an enhanced tumor region, and accordingly, the classification identifier of the pixel point in the background region may be 0, the classification identifier of the pixel point in the necrotic and non-enhanced region may be 1, the classification identifier of the pixel point in the edema region may be 2, and the classification identifier of the pixel point in the enhanced tumor region may be 4.
The method comprises the steps of sliding on initial sample data through a preset sliding window to obtain at least one sample data including a target to be segmented, obtaining a plurality of sample data on the basis of one initial sample data, providing a large amount of sample data for training an image segmentation model, and reducing the difficulty in obtaining the sample data. The preset sliding window may be determined according to a required size of sample data, and may be, for example, 128 × 128. The initial sample data and the classification marks in the initial label are in one-to-one correspondence, and the label of the intercepted sample data can be determined from the initial label according to the relative position of the intercepted sample data in the initial sample data.
Optionally, the initial sample data is three-dimensional sample data, where after the initial sample data is acquired, the method further includes: and slicing the three-dimensional sample data along the axial direction of the segmentation target to obtain a plurality of two-dimensional initial sample data. For example, the size of three-dimensional sample data may be 152 × 192 × 146, and slicing along the axial direction of the segmentation target may result in 146 two-dimensional initial sample data with the size of 152 × 192. In this embodiment, the image segmentation model segments the two-dimensional image, which can improve the processing efficiency of the image segmentation model on the medical image.
In this embodiment, before the three-dimensional sample data is sliced, it may be further determined whether large-area invalid pixels exist in the three-dimensional sample data, and if so, the invalid pixels are cut from the three-dimensional sample data, so as to avoid interference of the large-area invalid pixels on image segmentation.
And S220, preprocessing the sample data to generate input data.
In this embodiment, the preprocessing of the sample data may be a normalization processing of the sample data. Different sample data can be obtained based on different users through different image acquisition devices, and can be obtained based on different dosages in the medical image acquisition process, so that the difference of different sample data is relatively large. The difference between different sample data is reduced through the normalization processing.
Optionally, the preprocessing the sample data may include: determining a pixel mean value and a pixel standard deviation of sample data; and normalizing the sample data based on the pixel value, the pixel mean value and the pixel standard deviation of each pixel point of the sample data to obtain input data.
For example, the input data may be obtained according to the following formula:
S230, inputting input data into an image segmentation model to be trained, and determining a loss function based on the label of the sample data and the prediction result of the image segmentation model.
In this embodiment, the Loss function may be composed of a Dice function and an entry Loss function, where the Dice function is
Where y is a label of sample data, yl is a prediction result of the image segmentation model, i is each class in the label, and c is a sum of classes.
The Encopy Loss function is
Wherein c is a label category, w is a weight of each type, f (x, theta) is a method prediction result, theta is an activation function, and x is an input image.
In this embodiment, the Loss function may be the sum of a Dice function and a weighted entry Loss function.
And S240, training the image segmentation model to be trained based on the loss function.
In this embodiment, the loss function is reversely input into the image segmentation model to be trained, and the network is continuously optimized by a random gradient descent method until a preset number of iterative training is completed or the image segmentation model reaches a preset segmentation precision, thereby completing training of the image segmentation model.
Optionally, after the training of the image segmentation model is completed, the image segmentation model obtained by the training may be tested through test sample data, when the segmentation precision of the image segmentation model in the test result meets the requirement, it is determined that the image segmentation model is successfully trained, and when the segmentation precision of the image segmentation model in the test result does not meet the requirement, the image segmentation model is retrained again.
The test sample data and the label of the test sample data are obtained in the same manner as the training sample data, and are not described herein again.
S250, acquiring a medical image to be processed, inputting the medical image into a pre-trained image segmentation model, and determining a segmentation target of the medical image based on an output result of the image segmentation model, wherein the segmentation target is highlighted and displayed in the output result.
For example, for the segmentation of the brain tumor image, the output result of the image segmentation model may be to set the background region to be white or black, and the necrotic and non-enhanced regions, the edema region, and the enhanced tumor region are displayed in three colors other than black or white, respectively, so as to achieve accurate segmentation and highlighting of the brain tumor image.
According to the technical scheme, the image segmentation model with the capability of extracting the detail features and integrating the multi-scale receptive field information is obtained through end-to-end training, the obtained medical image is segmented, the segmentation target with the detail information is obtained, the segmentation precision of the medical image is improved, and the auxiliary improvement of the diagnosis accuracy of the medical image is facilitated.
EXAMPLE III
Fig. 6 is a schematic structural diagram of a medical image segmentation apparatus provided in a third embodiment of the present invention, where the apparatus includes:
the medical image input module 310 is configured to acquire a medical image to be processed, and input the medical image into a pre-trained image segmentation model, where the image segmentation model includes an encoder and a decoder, and each first residual module in the encoder is laterally connected to each second residual module in the decoder based on a multi-scale hole convolution module;
and a segmentation target determination module 0, configured to determine a segmentation target of the medical image based on an output result of the image segmentation model, where the segmentation target is displayed in a highlighted manner in the output result.
On the basis of the above embodiment, the apparatus further includes:
the sample data acquisition module is used for acquiring sample data and a label of the sample data;
the preprocessing module is used for preprocessing the sample data to generate input data;
the loss function determining module is used for inputting input data into the image segmentation model to be trained and determining a loss function based on the label of the sample data and the prediction segmentation result of the image segmentation model to be trained;
and the model training module is used for training the image segmentation model to be trained based on the loss function.
Optionally, the sample data obtaining module includes:
the initial sample data acquisition unit is used for acquiring initial sample data and an initial label of an initial sample image, wherein the initial label comprises a classification identifier of each pixel point in the initial sample image;
the system comprises a sample data intercepting unit, a target segmentation unit and a display unit, wherein the sample data intercepting unit is used for sliding on initial sample data based on a preset sliding window and intercepting at least one sample data, and the sample data comprises a target to be segmented;
and the label determining unit is used for determining the label of the intercepted at least one sample data based on the corresponding relation between the initial sample data and the initial label.
Optionally, the initial sample data is three-dimensional sample data.
Correspondingly, the sample data obtaining module further comprises:
and the initial sample data slicing unit is used for slicing the three-dimensional sample data along the axial direction of the segmentation target after the initial sample data is obtained, so as to obtain a plurality of two-dimensional initial sample data.
Optionally, the preprocessing module is configured to:
determining a pixel mean value and a pixel standard deviation of sample data;
and normalizing the sample data based on the pixel value, the pixel mean value and the pixel standard deviation of each pixel point of the sample data to obtain input data.
Optionally, the first residual module or the second residual module includes: the device comprises a first convolution layer and a second convolution layer, wherein a regularization layer and an activation function layer are respectively arranged behind the first convolution layer and the second convolution layer.
Optionally, the multi-scale void convolution module includes at least two convolution layers connected in sequence and a fusion layer connected to the at least two convolution layers, respectively, where the number of channels of the at least two convolution layers is the same.
Optionally, any first residual error module is configured to perform convolution processing on input sample data or output data of a previous first residual error module, and output a processing result to a next first residual error module and a multi-scale cavity convolution module connected to the first residual error module;
and any second residual error module is used for performing convolution processing on the output result of the last second residual error module or the output result of the encoder and the output result of the multi-scale cavity convolution module connected with the second residual error module.
Optionally, the medical image is a brain tumor image, and the segmented target is a brain tumor region.
The medical image segmentation device provided by the embodiment of the invention can execute the medical image segmentation method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects for executing the medical image segmentation method.
Example four
Fig. 7 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention. FIG. 7 illustrates a block diagram of an exemplary electronic device 12 suitable for use in implementing embodiments of the present invention. The electronic device 12 shown in fig. 7 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiment of the present invention.
As shown in FIG. 7, electronic device 12 is embodied in the form of a general purpose computing device. The components of electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 and the processing unit 16.
The system memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)30 and/or cache memory. The electronic device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 7, and commonly referred to as a "hard drive"). Although not shown in FIG. 7, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. Memory 28 may include at least one program product having a set of program modules that are configured to carry out the functions of embodiments of the invention.
A program/utility 44 having a set of program modules 46 may be stored, for example, in memory 28, such program modules 46 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 46 generally carry out the functions and/or methodologies of the described embodiments of the invention.
The processing unit 16 executes various functional applications and data processing by executing programs stored in the system memory 28, for example, to implement a medical image segmentation method provided by an embodiment of the present invention, the method including:
acquiring a medical image to be processed, and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder on the basis of a multi-scale cavity convolution module;
determining a segmentation target of the medical image based on an output result of the image segmentation model, wherein the segmentation target is highlighted in the output result.
The processing unit 16 executes programs stored in the system memory 28 to perform various functional applications and data processing, such as implementing a medical image segmentation method provided by an embodiment of the present invention.
Of course, those skilled in the art will understand that the processor may also implement the technical solution of the medical image segmentation method provided by any embodiment of the present invention.
EXAMPLE five
An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements a medical image segmentation method provided in an embodiment of the present invention, where the method includes:
acquiring a medical image to be processed, and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder on the basis of a multi-scale cavity convolution module;
determining a segmentation target of the medical image based on an output result of the image segmentation model, wherein the segmentation target is highlighted in the output result.
Of course, the computer-readable storage medium stored thereon according to the embodiments of the present invention is not limited to the above method operations, and may also perform related operations in a medical image segmentation method according to any embodiment of the present invention.
Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, or device.
A computer readable signal medium may include a video clip, feature encoding of a second video, feature encoding of respective video clips, etc., having computer readable program code embodied therein. Such forms of the broadcast video clip, feature encoding of the second video, feature encoding of each video clip, and the like. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, or the like, as well as conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
It should be noted that, in the embodiment of the video processing apparatus, the modules included in the embodiment are only divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (12)
1. A medical image segmentation method, comprising:
acquiring a medical image to be processed, and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder on the basis of a multi-scale cavity convolution module;
determining a segmentation target of the medical image based on an output result of the image segmentation model.
2. The method of claim 1, further comprising:
acquiring sample data and a label of the sample data;
preprocessing the sample data to generate input data;
inputting the input data into an image segmentation model to be trained, and determining a loss function based on a label of the sample data and a prediction result of the image segmentation model;
and training the image segmentation model to be trained based on the loss function.
3. The method of claim 2, wherein obtaining sample data and a tag for the sample data comprises:
acquiring initial sample data and an initial label of the initial sample image, wherein the initial label comprises a classification identifier of each pixel point in the initial sample image;
sliding on the initial sample data based on a preset sliding window, and intercepting at least one sample data, wherein the sample data comprises a target to be segmented;
and determining the label of the intercepted at least one sample data based on the corresponding relation between the initial sample data and the initial label.
4. The method according to claim 3, wherein the initial sample data is three-dimensional sample data, and before sliding on the initial sample data based on a preset sliding window, the method further comprises:
and slicing the three-dimensional sample data along the axial direction of the segmentation target to obtain a plurality of two-dimensional initial sample data.
5. The method of claim 2, wherein pre-processing the sample data, generating input data comprises:
determining a pixel mean value and a pixel standard deviation of the sample data;
and normalizing the sample data based on the pixel value, the pixel mean value and the pixel standard deviation of each pixel point of the sample data to obtain the input data.
6. The method of claim 1, wherein the first residual module or the second residual module comprises: the device comprises a first convolution layer and a second convolution layer, wherein a regularization layer and an activation function layer are respectively arranged behind the first convolution layer and the second convolution layer.
7. The method of claim 1, wherein the multi-scale hole convolution module comprises at least two convolution layers connected in sequence and a fusion layer connected to the at least two convolution layers respectively, wherein the at least two convolution layers have the same number of channels.
8. The method according to claim 1, wherein any of the first residual modules is configured to perform convolution processing on input sample data or output data of a previous first residual module, and output a processing result to a next first residual module and a multi-scale hole convolution module connected to the first residual module;
and any one second residual error module is used for performing convolution processing on the output result of the last second residual error module or the output result of the encoder and the output result of the multi-scale cavity convolution module connected with the second residual error module.
9. The method according to any one of claims 1-8, wherein the medical image is a brain tumor image and the segmented object is a brain tumor region.
10. A medical image segmentation apparatus, characterized by comprising:
the medical image input module is used for acquiring a medical image to be processed and inputting the medical image into a pre-trained image segmentation model, wherein the image segmentation model comprises an encoder and a decoder, and each first residual error module in the encoder is transversely connected with each second residual error module in the decoder on the basis of a multi-scale hole convolution module;
a segmentation target determination module for determining a segmentation target of the medical image based on an output result of the image segmentation model, wherein the segmentation target is highlighted in the output result.
11. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements a method of medical image segmentation as claimed in any one of claims 1 to 9 when executing the computer program.
12. A storage medium containing computer-executable instructions, which when executed by a computer processor implement the medical image segmentation method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911252001.4A CN111028246A (en) | 2019-12-09 | 2019-12-09 | Medical image segmentation method and device, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911252001.4A CN111028246A (en) | 2019-12-09 | 2019-12-09 | Medical image segmentation method and device, storage medium and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111028246A true CN111028246A (en) | 2020-04-17 |
Family
ID=70206135
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911252001.4A Pending CN111028246A (en) | 2019-12-09 | 2019-12-09 | Medical image segmentation method and device, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111028246A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111915597A (en) * | 2020-08-07 | 2020-11-10 | 温州医科大学 | Focal image detection method and device |
| CN112102251A (en) * | 2020-08-20 | 2020-12-18 | 上海壁仞智能科技有限公司 | Method and device for segmenting image, electronic equipment and storage medium |
| CN112530554A (en) * | 2020-12-17 | 2021-03-19 | 上海联影医疗科技股份有限公司 | Scanning positioning method and device, storage medium and electronic equipment |
| CN112635060A (en) * | 2020-12-29 | 2021-04-09 | 北京航空航天大学合肥创新研究院 | Viability evaluation method and device, viability evaluation equipment and storage medium |
| CN112950652A (en) * | 2021-02-08 | 2021-06-11 | 深圳市优必选科技股份有限公司 | Robot and hand image segmentation method and device thereof |
| CN113362409A (en) * | 2021-05-28 | 2021-09-07 | 北京百度网讯科技有限公司 | Image coloring method and device, image coloring model training method and device, electronic equipment and storage medium |
| CN113538394A (en) * | 2021-07-26 | 2021-10-22 | 泰康保险集团股份有限公司 | Image segmentation method and device, electronic equipment and storage medium |
| CN113538359A (en) * | 2021-07-12 | 2021-10-22 | 北京曙光易通技术有限公司 | System and method for finger vein image segmentation |
| CN113592771A (en) * | 2021-06-24 | 2021-11-02 | 深圳大学 | Image segmentation method |
| CN115409990A (en) * | 2022-09-28 | 2022-11-29 | 北京医准智能科技有限公司 | Medical image segmentation method, device, equipment and storage medium |
| CN115700762A (en) * | 2022-12-28 | 2023-02-07 | 苏州国科康成医疗科技有限公司 | Target element segmentation method, model and electronic equipment for medical image |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635882A (en) * | 2019-01-23 | 2019-04-16 | 福州大学 | Salient object detection method based on multi-scale convolution feature extraction and fusion |
| CN110110617A (en) * | 2019-04-22 | 2019-08-09 | 腾讯科技（深圳）有限公司 | Medical image dividing method, device, electronic equipment and storage medium |
| CN110197709A (en) * | 2019-05-29 | 2019-09-03 | 广州瑞多思医疗科技有限公司 | A kind of 3-dimensional dose prediction technique based on deep learning Yu priori plan |
| CN110490878A (en) * | 2019-07-29 | 2019-11-22 | 上海商汤智能科技有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN110503052A (en) * | 2019-08-27 | 2019-11-26 | 西南交通大学 | A kind of image, semantic dividing method based on improvement U-NET network |
| US20190370972A1 (en) * | 2018-06-04 | 2019-12-05 | University Of Central Florida Research Foundation, Inc. | Capsules for image analysis |
-
2019
- 2019-12-09 CN CN201911252001.4A patent/CN111028246A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190370972A1 (en) * | 2018-06-04 | 2019-12-05 | University Of Central Florida Research Foundation, Inc. | Capsules for image analysis |
| CN109635882A (en) * | 2019-01-23 | 2019-04-16 | 福州大学 | Salient object detection method based on multi-scale convolution feature extraction and fusion |
| CN110110617A (en) * | 2019-04-22 | 2019-08-09 | 腾讯科技（深圳）有限公司 | Medical image dividing method, device, electronic equipment and storage medium |
| CN110197709A (en) * | 2019-05-29 | 2019-09-03 | 广州瑞多思医疗科技有限公司 | A kind of 3-dimensional dose prediction technique based on deep learning Yu priori plan |
| CN110490878A (en) * | 2019-07-29 | 2019-11-22 | 上海商汤智能科技有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN110503052A (en) * | 2019-08-27 | 2019-11-26 | 西南交通大学 | A kind of image, semantic dividing method based on improvement U-NET network |
Non-Patent Citations (1)
| Title |
|---|
| JIANXIN ZHANG 等: ""SDResU-net：Separable and dilated residual U-net for mri Brain tumor segmentation"", 《CURRENT MEDICAL IMAGING》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111915597A (en) * | 2020-08-07 | 2020-11-10 | 温州医科大学 | Focal image detection method and device |
| CN112102251A (en) * | 2020-08-20 | 2020-12-18 | 上海壁仞智能科技有限公司 | Method and device for segmenting image, electronic equipment and storage medium |
| CN112102251B (en) * | 2020-08-20 | 2023-10-31 | 上海壁仞智能科技有限公司 | Method and device for dividing image, electronic equipment and storage medium |
| CN112530554B (en) * | 2020-12-17 | 2023-08-18 | 上海联影医疗科技股份有限公司 | Scanning positioning method and device, storage medium and electronic equipment |
| CN112530554A (en) * | 2020-12-17 | 2021-03-19 | 上海联影医疗科技股份有限公司 | Scanning positioning method and device, storage medium and electronic equipment |
| WO2022127318A1 (en) * | 2020-12-17 | 2022-06-23 | 上海联影医疗科技股份有限公司 | Scanning positioning method and apparatus, storage medium and electronic device |
| CN112635060A (en) * | 2020-12-29 | 2021-04-09 | 北京航空航天大学合肥创新研究院 | Viability evaluation method and device, viability evaluation equipment and storage medium |
| CN112635060B (en) * | 2020-12-29 | 2023-11-14 | 北京航空航天大学合肥创新研究院 | Viability assessment method, device, assessment equipment and storage medium |
| WO2022143811A1 (en) * | 2020-12-29 | 2022-07-07 | 北京航空航天大学合肥创新研究院 | Survivability evaluation method and apparatus, evaluation device, and storage medium |
| CN112950652A (en) * | 2021-02-08 | 2021-06-11 | 深圳市优必选科技股份有限公司 | Robot and hand image segmentation method and device thereof |
| CN112950652B (en) * | 2021-02-08 | 2024-01-19 | 深圳市优必选科技股份有限公司 | Robot and hand image segmentation method and device thereof |
| CN113362409A (en) * | 2021-05-28 | 2021-09-07 | 北京百度网讯科技有限公司 | Image coloring method and device, image coloring model training method and device, electronic equipment and storage medium |
| CN113362409B (en) * | 2021-05-28 | 2023-10-31 | 北京百度网讯科技有限公司 | Image coloring and model training method and device, electronic equipment and storage medium |
| CN113592771A (en) * | 2021-06-24 | 2021-11-02 | 深圳大学 | Image segmentation method |
| CN113592771B (en) * | 2021-06-24 | 2023-12-15 | 深圳大学 | Image segmentation method |
| CN113538359A (en) * | 2021-07-12 | 2021-10-22 | 北京曙光易通技术有限公司 | System and method for finger vein image segmentation |
| CN113538359B (en) * | 2021-07-12 | 2024-03-01 | 北京曙光易通技术有限公司 | System and method for finger vein image segmentation |
| CN113538394B (en) * | 2021-07-26 | 2023-08-08 | 泰康保险集团股份有限公司 | Image segmentation method and device, electronic equipment and storage medium |
| CN113538394A (en) * | 2021-07-26 | 2021-10-22 | 泰康保险集团股份有限公司 | Image segmentation method and device, electronic equipment and storage medium |
| CN115409990A (en) * | 2022-09-28 | 2022-11-29 | 北京医准智能科技有限公司 | Medical image segmentation method, device, equipment and storage medium |
| CN115700762A (en) * | 2022-12-28 | 2023-02-07 | 苏州国科康成医疗科技有限公司 | Target element segmentation method, model and electronic equipment for medical image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111028246A (en) | Medical image segmentation method and device, storage medium and electronic equipment | |
| CN110111313B (en) | Medical image detection method based on deep learning and related equipment | |
| CN111524106B (en) | Skull fracture detection and model training method, device, equipment and storage medium | |
| CN110059697B (en) | Automatic lung nodule segmentation method based on deep learning | |
| US10853409B2 (en) | Systems and methods for image search | |
| CN110276408B (en) | 3D image classification method, device, equipment and storage medium | |
| CN112085714B (en) | Pulmonary nodule detection method, model training method, device, equipment and medium | |
| US20230052133A1 (en) | Medical image processing method and apparatus, device, storage medium, and product | |
| An et al. | Medical image segmentation algorithm based on multilayer boundary perception-self attention deep learning model | |
| US20200372654A1 (en) | Sampling latent variables to generate multiple segmentations of an image | |
| CN112465834A (en) | Blood vessel segmentation method and device | |
| CN112151179A (en) | Image data evaluation method, device, equipment and storage medium | |
| CN112614144A (en) | Image segmentation method, device, equipment and storage medium | |
| CN114820584A (en) | Lung focus positioner | |
| CN110135428B (en) | Image segmentation processing method and device | |
| CN115601299A (en) | Intelligent liver cirrhosis state evaluation system and method based on images | |
| CN111369524A (en) | Nodule recognition model training method, nodule recognition method and device | |
| CN110992310A (en) | Method and device for determining partition where mediastinal lymph node is located | |
| CN113192031B (en) | Vascular analysis method, vascular analysis device, vascular analysis computer device, and vascular analysis storage medium | |
| CN111523578B (en) | Image classification method and device and neural network model training method and device | |
| CN113781387A (en) | Model training method, image processing method, device, equipment and storage medium | |
| CN116309465B (en) | Tongue image detection and positioning method based on improved YOLOv5 in natural environment | |
| CN112927235A (en) | Brain tumor image segmentation method based on multi-scale superpixel and nuclear low-rank representation | |
| CN112750124B (en) | Model generation method, image segmentation method, model generation device, image segmentation device, electronic equipment and storage medium | |
| CN112530554B (en) | Scanning positioning method and device, storage medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200417 |