CN114419068B - A medical image segmentation method, device, equipment and storage medium - Google Patents
A medical image segmentation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN114419068B CN114419068B CN202210060260.2A CN202210060260A CN114419068B CN 114419068 B CN114419068 B CN 114419068B CN 202210060260 A CN202210060260 A CN 202210060260A CN 114419068 B CN114419068 B CN 114419068B
- Authority
- CN
- China
- Prior art keywords
- image
- center point
- determining
- segmentation
- medical
- 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.)
- Active
Links
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
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
-
- 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/41—Medical
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
本发明公开了一种医学图像的分割方法、装置、设备及存储介质，该方法包括：获取待处理医学图像对应的初始分割图像中的感兴趣图像，并确定所述感兴趣图像中的中心点对应的第一中心点参数；基于预设缩放参数，确定所述感兴趣图像对应的缩放图像；基于所述第一中心点参数以及所述缩放图像中至少一个图像点分别对应的图像点参数，确定各所述图像点分别与所述中心点之间的图像距离；基于至少一个图像距离，确定更新后的感兴趣图像，并基于所述更新后的感兴趣图像，确定所述待处理医学图像对应的目标分割图像。本发明解决了现有的图像分割算法得到的分割图像的精确度较差的问题，完善了分割图像的后处理算法，提高了修正后的分割图像的分割精确度。
The present invention discloses a medical image segmentation method, device, equipment and storage medium, the method comprising: obtaining an image of interest in an initial segmented image corresponding to a medical image to be processed, and determining a first center point parameter corresponding to a center point in the image of interest; determining a scaled image corresponding to the image of interest based on a preset scaling parameter; determining image distances between each image point and the center point based on the first center point parameter and image point parameters corresponding to at least one image point in the scaled image; determining an updated image of interest based on at least one image distance, and determining a target segmented image corresponding to the medical image to be processed based on the updated image of interest. The present invention solves the problem of poor accuracy of segmented images obtained by existing image segmentation algorithms, improves the post-processing algorithm of segmented images, and improves the segmentation accuracy of the corrected segmented images.
Description
Technical Field
The present invention relates to the field of image segmentation technologies, and in particular, to a method, an apparatus, a device, and a storage medium for segmenting a medical image.
Background
Image segmentation is a technique and process of dividing an image into several specific regions with unique properties and presenting objects of interest, which is a key step from image processing to image segmentation. In particular, in the medical industry, accurate regions of interest (e.g., lesions) are extracted from examination images, which plays a critical role in the diagnosis of a physician and even in the treatment of the patient.
The image segmentation algorithms at present are many, such as a segmentation algorithm based on a threshold value, a segmentation algorithm based on a region, a neural network segmentation model and the like, and in a real application scene, the image segmentation results are mostly directly put into use, or some conventional post-processing operations are performed on the image segmentation results.
However, any image segmentation algorithm has a certain segmentation error, the conventional post-processing operation cannot effectively improve the accuracy of an image segmentation result, and particularly cannot meet the requirement of the medical industry on the accuracy of image segmentation, so that the practicability is poor.
Disclosure of Invention
The invention provides a medical image segmentation method, a device, equipment and a storage medium, which are used for solving the problem of poor accuracy of the existing image segmentation result.
According to an aspect of the present invention, there is provided a segmentation method of a medical image, comprising:
Acquiring an interested image in an initial segmentation image corresponding to a medical image to be processed, and determining a first center point parameter corresponding to a center point in the interested image;
determining a zoom image corresponding to the interested image based on a preset zoom parameter;
Determining image distances between each image point and the center point respectively based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaled image respectively;
And determining an updated image of interest based on at least one image distance, and determining a target segmentation image corresponding to the medical image to be processed based on the updated image of interest.
According to another aspect of the present invention, there is provided a segmentation apparatus of a medical image, including:
The first center point parameter determining module is used for acquiring an interested image in an initial segmentation image corresponding to the medical image to be processed and determining a first center point parameter corresponding to a center point in the interested image;
The scaling image determining module is used for determining a scaling image corresponding to the interested image based on preset scaling parameters;
the image distance determining module is used for determining the image distance between each image point and the center point respectively based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaled image respectively;
The target segmentation image determining module is used for determining an updated interested image based on at least one image distance and determining a target segmentation image corresponding to the medical image to be processed based on the updated interested image.
According to another aspect of the present invention, there is provided an electronic apparatus including:
at least one processor, and
A memory communicatively coupled to the at least one processor, wherein,
The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of segmentation of medical images according to any one of the embodiments of the present invention.
According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to execute a method of segmenting a medical image according to any of the embodiments of the present invention.
According to the technical scheme, the first center point parameter corresponding to the center point in the image of interest in the initial segmentation image and the scaling image corresponding to the image of interest are obtained, the image distance between each image point and the center point is determined based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaling image, the updated image of interest is determined based on the at least one image distance, and the target segmentation image corresponding to the medical image to be processed is determined based on the updated image of interest, so that the problem that the accuracy of the segmentation image obtained by the existing image segmentation algorithm is poor is solved, the post-processing algorithm of the segmentation image is perfected, the segmentation accuracy of the corrected segmentation image and the practicability of the segmentation image are improved, and effective and accurate reference value is provided for subsequent diagnosis and treatment.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for segmenting medical images according to an embodiment of the present invention;
Fig. 2 is a flowchart of a method for segmenting a medical image according to a second embodiment of the present invention;
Fig. 3 is a flowchart of a specific example of a medical image segmentation method according to the second embodiment of the present invention;
Fig. 4 is a schematic structural diagram of a medical image segmentation apparatus according to a third embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," "initial," "target," and the like in the description and claims of the present invention and in the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example 1
Fig. 1 is a flowchart of a method for segmenting a medical image according to an embodiment of the present invention, where the method may be performed by a device for segmenting a medical image, the device for segmenting a medical image may be implemented in hardware and/or software, and the device for segmenting a medical image may be configured in an intelligent terminal such as a notebook computer, a mobile terminal, a desktop computer, and a tablet computer. As shown in fig. 1, the method includes:
S110, acquiring an image of interest in an initial segmentation image corresponding to the medical image to be processed, and determining a first center point parameter corresponding to a center point in the image of interest.
The medical image to be processed can be a three-dimensional image or a two-dimensional image. By way of example, the image type of the medical image to be processed may be a CT (Computed Tomography, electronic computer tomography) image, a DR (digital radiography ) image, an MRI (Magnetic Resonance Imaging, magnetic resonance imaging) image, an ultrasound image or a PET (positron emission tomography, positron emission computed tomography) image, or the like, without limitation of the image type of the medical image to be processed.
The embodiment of the invention is explained taking the type of medical image as an MRI image as an example. MRI imaging is an imaging method that uses magnetic resonance phenomena to acquire electromagnetic signals from a human body and reconstruct human body information. Specifically, when the detected object is a brain region, before the interested image in the initial segmentation image corresponding to the medical image to be processed is acquired, the method further comprises the step of preprocessing the medical image to be processed to obtain a corrected medical image to be processed. Wherein the preprocessing operation includes at least one of an image registration operation, a skull removal operation, and a denoising operation, as an example.
In one embodiment, the initial segmentation image may be a segmentation image obtained by segmenting the medical image to be processed using a preset segmentation algorithm, which includes, but is not limited to, a threshold-based segmentation algorithm, a region-based segmentation algorithm, an edge-based segmentation algorithm, a morphological watershed algorithm, a neural network segmentation model, and the like, as examples, and the preset segmentation algorithm is not limited herein.
Specifically, the image of interest is an image obtained by segmenting the image in the initial segmentation image based on a preset segmentation algorithm, and in the field of medical images, the initial segmentation image includes a background image and a focus image, wherein the focus image is the image of interest in the embodiment.
In one embodiment, optionally, the method further comprises acquiring a boundary point corresponding to the image of interest, and determining a center point in the image of interest based on boundary point coordinates of the boundary point. Illustratively, assuming that the medical image to be processed is a three-dimensional image, the coordinates (X, Y, Z) of the center point satisfy the formula:
In another embodiment, the method further comprises the steps of taking each image point in the image of interest as an image point to be detected in sequence, determining the sum of image distances between the image point to be detected and other image points based on the image point coordinates of each image point in the image of interest, and taking the image point to be detected with the minimum sum of the image distances as a center point. The type of the image distance may be, for example, a euclidean distance, a manhattan distance, a chebyshev distance, a mahalanobis distance, or a minkowski distance, etc., and the type of the image distance is not limited herein.
In one embodiment, optionally, the first center point parameter includes a first center point coordinate and a first center point gray value.
S120, determining a zoom image corresponding to the interested image based on a preset zoom parameter.
In one embodiment, optionally, the preset scaling parameters include a scaling number of turns and a scaling direction. Specifically, the zoom direction includes a zoom-out direction and an expansion direction. The zoom direction may be used to describe a relationship between the zoomed image and the image of interest, and illustratively, when the zoom direction is a zoom-out direction, the zoomed image belongs to a portion of the edge images in the image of interest, and when the zoom direction is an expansion direction, the zoomed image belongs to a portion of the edge images outside the image of interest.
In one embodiment, the scaling number of turns and the scaling direction can be set individually according to actual requirements, and the scaling number of turns is exemplified as 5 turns, the scaling direction is expanded, and specific parameter values of the scaling number of turns and the scaling parameter are not limited herein.
In another embodiment, the method further comprises the steps of obtaining the image size of the medical image to be processed, and determining the corresponding scaling circle number of the medical image to be processed based on a preset mapping relation, wherein the preset mapping relation is used for describing the mapping relation between the scaling circle number and the image size. Specifically, there may be a positive correlation between the number of scaling turns and the image size in the preset mapping relationship, that is, when the image size is larger, the number of scaling turns is larger, and when the image size is smaller, the number of scaling turns is smaller.
The relationship type of the preset mapping relationship may be a linear relationship, an exponential relationship, a sinusoidal relationship, a discrete relationship, or the like, which is not limited herein.
In one embodiment, the preset mapping relationship may be used to describe a mapping relationship between the number of scaling turns and the overall image size of the medical image to be processed. For example, in the preset mapping relationship, the number of scaling turns corresponding to the overall image size of 512×512×200 is 10, and the number of scaling turns corresponding to the overall image size of 128×128×100 is 5. In this embodiment, the number of scaling turns of the scaled image in each coordinate axis dimension is the same.
In another embodiment, the preset mapping relationship may be used to describe a mapping relationship between the number of scaling turns and the image size corresponding to a single coordinate axis dimension of the medical image to be processed. For example, in the preset mapping relationship, the number of scaling turns corresponding to the image size [400,512] corresponding to the single coordinate axis dimension is 10, the number of scaling turns corresponding to the image size [128,200] corresponding to the single coordinate axis dimension is 5, the number of scaling turns corresponding to the image size [50,100] corresponding to the single coordinate axis dimension is 2, and the number of scaling turns on the X-axis, Y-axis and Z-axis is 10, 5 and 2, respectively, assuming that the overall image size of the medical image to be processed is 425×137×80. In this embodiment, the number of scaling turns of the scaled image in each coordinate axis dimension may be different.
In the medical technical field, an acquired medical image to be processed generally has a small image dimension in the Z-axis direction, if scaled images are determined on the basis of the same scaling number of times in all three coordinate axis dimensions, program errors (for example, the image dimension is 1028×1028×20, the scaling number of times is 30, and the scaling operation error in the Z-axis direction) are easy to occur, and the image segmentation effect is reduced. The scaling turns are set based on the image size of the single coordinate axis dimension, so that the situation of program error reporting can be avoided, and the image segmentation effect is improved.
S130, determining image distances between each image point and the center point respectively based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaled image.
In one embodiment, optionally, the image point parameters include image point coordinates and image point gray values of the image points. Illustratively, the first center point parameter is (x 0,y0,z0,g0) and the image point parameter is (x i,yi,zi,gi), where i represents the i-th image point in the scaled image.
The type of the image distance may be, for example, a euclidean distance, a manhattan distance, a chebyshev distance, a mahalanobis distance, or a minkowski distance, etc., and the type of the image distance is not limited herein.
S140, determining an updated interested image based on at least one image distance, and determining a target segmentation image corresponding to the medical image to be processed based on the updated interested image.
In one embodiment, optionally, determining the updated image of interest based on the at least one image distance includes using an image point in the scaled image corresponding to the image distance satisfying the preset distance range as a target image point, or based on the at least one image distance, sorting the image points in the scaled image, and determining the target image point in the scaled image according to the sorting result and the preset selection number, and determining the updated image of interest based on the target image point.
In one embodiment, the preset distance range may be, for example, less than a preset distance threshold. In another embodiment, optionally, the method further comprises determining the preset selection number based on an image size of the scaled image. The product of the number of image points in the scaled image and the image point size may be used as the image size of the scaled image, or the number of image points in the scaled image may be used as the image size of the scaled image. The manner of calculating the image size of the scaled image is not limited here. Specifically, the image size of the scaled image has a positive correlation with the preset selection number, that is, when the image size is larger, the preset selection number is larger, otherwise, when the image size is smaller, the preset selection number is smaller. The specific association relationship between the image size and the preset selected number is not limited, and the user can set according to actual requirements.
The method comprises the steps of determining an updated interested image based on target image points, wherein the method comprises the steps of dividing target image points in a scaled image into interested images when the scaling direction is an expanding direction to obtain the updated interested image, and dividing image points (which belong to part of the interested images) except the target image points in the scaled image into background images in an initial segmentation image when the scaling direction is a reducing direction to obtain the updated interested image.
In this embodiment, the updated initial segmented image corresponding to the image of interest is the target segmented image corresponding to the medical image to be processed.
According to the technical scheme, the first center point parameter corresponding to the center point in the image of interest in the initial segmentation image and the scaling image corresponding to the image of interest are obtained, the image distance between each image point and the center point is determined based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaling image, the updated image of interest is determined based on the at least one image distance, and the target segmentation image corresponding to the medical image to be processed is determined based on the updated image of interest, so that the problem that the accuracy of the segmentation image obtained by the existing image segmentation algorithm is poor is solved, the post-processing algorithm of the segmentation image is perfected, the segmentation accuracy of the corrected segmentation image and the practicability of the segmentation image are improved, and effective and accurate reference value is provided for subsequent diagnosis and treatment.
Example two
Fig. 2 is a flowchart of a medical image segmentation method according to a second embodiment of the present invention, where the step of determining, based on the updated image of interest, the target segmented image corresponding to the medical image to be processed in the foregoing embodiment is further refined, and optionally, the step of determining, based on the updated image of interest, the target segmented image corresponding to the medical image to be processed includes determining a second center point parameter corresponding to a center point in the updated image of interest, determining whether a second center point coordinate in the second center point parameter is the same as a previous center point coordinate in a previous center point parameter corresponding to the second center point parameter, if not, repeating an operation of determining, based on a preset zoom parameter, the target segmented image corresponding to the medical image to be processed based on the updated image of interest corresponding to the second center point parameter, and if yes, determining the target segmented image corresponding to the medical image to be processed based on the updated image of interest corresponding to the second center point parameter.
The specific implementation steps of the embodiment include:
S210, acquiring an image of interest in an initial segmentation image corresponding to the medical image to be processed, and determining a first center point parameter corresponding to a center point in the image of interest.
On the basis of the embodiment, the method optionally further comprises the steps of inputting the medical image to be processed into at least two target image segmentation models respectively, and determining an initial segmentation image corresponding to the medical image to be processed based on segmentation results and voting mechanisms which are output by the target image segmentation models respectively.
Wherein the target image segmentation model specifically comprises a neural network model trained based on a sample medical image, exemplary model types of the target image segmentation model include, but are not limited to, a full convolution network model (Fully Convolutional Network, FCN), a semantic segmentation model (SegNet), a cavity convolution model, and the like. Wherein in particular the model structure of the different object image segmentation models is different and/or the sample medical images used for training are different.
Wherein the voting mechanism comprises a soft voting mechanism or a hard voting mechanism.
In one embodiment, if the target image segmentation model outputs a result of segmenting the tag, i.e., 0 or 1, the corresponding voting mechanism is a hard voting mechanism. The hard voting mechanism is to select the most classification labels output by different target image segmentation models as the final classification labels.
In another embodiment, if the target image segmentation model outputs a probability of segmenting the tag, i.e., 0-1, the corresponding voting scheme is a soft voting scheme. The soft voting mechanism is to determine the final classification label based on the probabilities of the classification labels output by the different target image segmentation models. Specifically, the method can evaluate each target image segmentation model based on a test medical image, take an evaluation result as voting weights corresponding to each target image segmentation model respectively, and determine a final segmentation label based on the segmentation result and the voting weights.
On the basis of the embodiment, the method optionally further comprises the steps of obtaining a sample medical image, executing a dicing operation on the sample medical image based on at least two dicing size standards to obtain a plurality of dicing medical images, and training at least two initial image segmentation models based on the plurality of dicing medical images to obtain a trained target image segmentation model.
Wherein, in particular, when the image type of the sample medical image is a three-dimensional image, the dicing size criteria may be 4 x 4, 8 x 4, 16 x 1, etc. When the image type of the sample medical image is a two-dimensional image, the cut size criteria may be 4*4, 4*8, 32 x 6, etc.
Taking the image type of the sample medical image as a three-dimensional image as an example, classifying the segmented medical image, wherein the classification result comprises a three-dimensional data set of the three-dimensional segmented medical image and/or a two-dimensional data set of the two-dimensional segmented medical image. The z parameter in the dicing size standard corresponding to the dicing medical image in the three-dimensional data set is not 1, and the z parameter in the dicing size standard corresponding to the dicing medical image in the two-dimensional data set is 1. In this embodiment, the initial image segmentation model comprises a three-dimensional initial image segmentation model and/or a two-dimensional initial image segmentation model.
Further, the three-dimensional data set and/or the two-dimensional data set can be further divided based on the dicing size standard, so that at least one of a first three-dimensional sub-data set corresponding to the dicing medical image with the same size, a second three-dimensional sub-data set corresponding to the dicing medical image with the different size, a first two-dimensional sub-data set corresponding to the dicing medical image with the same size and a second two-dimensional sub-data set corresponding to the dicing medical image with the different size can be obtained. For example, the image sizes of the diced medical images in the first three-dimensional sub-data set may each be 4 x 2; the image size of the segmented medical image in the second three-dimensional subset includes 4 x 2, 4 x 3, 8 x 2; the image sizes of the diced medical images in the first two-dimensional sub-data set may each be 4 x 1 or 4*4, the image sizes of the diced medical images in the second two-dimensional sub-data set may include 4 x 1, 4 x 2x 1 and 8 x 1, or the image sizes of the diced medical images in the second two-dimensional sub-data set may include 4*4, 4*2 and 8 x 8. In this embodiment, the initial image segmentation model includes at least one of a first three-dimensional initial image segmentation model, a second three-dimensional initial image segmentation model, a first two-dimensional initial image segmentation model, and a second two-dimensional initial image segmentation model.
This has the advantage that the sample size of the sample image used for model training can be increased, thereby improving the training effect of the model.
S220, determining a zoom image corresponding to the interested image based on a preset zoom parameter.
On the basis of the embodiment, the method can be used for selectively inputting the predicted image into at least two target image segmentation models respectively, determining a predicted interested image corresponding to the predicted image based on segmentation results and voting mechanisms which are output by the target image segmentation models respectively, comparing the image size of the predicted interested image with the image size of the checked interested image corresponding to the checked image, and determining the scaling circle number and/or the scaling direction corresponding to the medical image to be processed according to the comparison results.
Specifically, if the image size of the predicted image of interest is smaller than the image size of the verification image of interest, the scaling direction is the expansion direction, and if the image size of the predicted image of interest is larger than the image size of the verification image of interest, the scaling direction is the contraction direction.
Wherein, concretely, the scaling circle number of the medical image to be processed is determined based on the size difference between the image size of the predicted image of interest and the image size of the checked image of interest. The size difference and the scaling number of turns have positive correlation, namely when the size difference is larger, the scaling number of turns is larger, otherwise, when the size difference is smaller, the scaling number of turns is smaller. The specific mapping relation between the size difference and the scaling circle is not limited, and the user can set according to the actual requirement.
S230, determining image distances between each image point and the center point respectively based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaled image.
S240, determining an updated interested image based on at least one image distance, and determining a second center point parameter corresponding to a center point in the updated interested image.
S250, judging whether the second center point coordinate is the same as the last center point coordinate, if so, executing S260, and if not, executing S220.
Specifically, when the number of iterations is 0, the last center point parameter corresponding to the second center point parameter determined in S240 is the first center point parameter, and when the number of iterations n (n is greater than or equal to 1), the last center point parameter corresponding to the second center point parameter determined in S240 is the second center point parameter corresponding to the n-1 th iteration.
S260, determining a target segmentation image corresponding to the medical image to be processed based on the updated interested image corresponding to the second center point parameter.
Fig. 3 is a flowchart of a specific example of a medical image segmentation method according to the second embodiment of the present invention, and fig. 3 is an example of a medical image as a nuclear magnetic resonance image of a brain. Specifically, the obtained small number of annotated brain nuclear magnetic resonance images are used as sample medical images, and preprocessing operations are performed on the brain nuclear magnetic resonance images, and exemplary preprocessing operations include registration operations, skull removal operations, denoising operations and the like. Based on at least two dicing size standards, dicing the brain nuclear magnetic resonance image to obtain a plurality of dicing medical images, classifying the dicing medical images, specifically, the 3D data set comprises a three-dimensional dicing nuclear magnetic resonance image, the 2D data set comprises a two-dimensional dicing nuclear magnetic resonance image, further, the 3D data set and the 2D data set are respectively classified to obtain a 3D sub-data set with the same size, a 3D sub-data set with different sizes obtained based on mixed sampling, a 2D sub-data set with the same size and a 2D sub-data set with different sizes obtained based on mixed sampling. And training the initial image segmentation model corresponding to each sub-data set based on each sub-data set to obtain a plurality of target image segmentation models.
The method comprises the steps of respectively inputting a medical image to be processed into a plurality of target image segmentation models, determining focus images (namely interested images) in an initial segmentation model corresponding to the medical image to be processed based on output results and voting mechanisms, determining zoom images corresponding to the focus images based on preset zoom parameters, expanding or shrinking the focus images based on image distances between image points in the zoom images and center points in the focus images to obtain updated focus images, repeatedly executing the operation of expanding or shrinking the focus images based on the updated focus images, and obtaining accurately segmented focus images when center point coordinates of the updated focus images are identical with center point coordinates of a previous focus image, so as to obtain target segmentation images corresponding to the medical image to be processed.
According to the technical scheme, through determining the second center point parameter corresponding to the center point in the updated interested image, whether the second center point coordinate in the second center point parameter is the same as the last center point coordinate in the last center point parameter corresponding to the second center point parameter or not is judged, if so, the operation of determining the scaled image corresponding to the interested image based on the preset scaling parameter is repeatedly executed based on the second center point parameter until the second center point coordinate is the same as the last center point coordinate, a target segmented image is obtained, the problem that the accuracy of the segmented image obtained by single post-processing operation is not high is solved, and the segmentation accuracy of the finally obtained segmented image and the practicability of the segmented image are further improved.
Example III
Fig. 4 is a schematic structural diagram of a medical image segmentation apparatus according to a third embodiment of the present invention. As shown in fig. 4, the apparatus includes a first center point parameter determination module 310, a scaled image determination module 320, an image distance determination module 330, and a target segmentation image determination module 340.
The first central point parameter determining module 310 is configured to obtain an image of interest in an initial segmented image corresponding to a medical image to be processed, and determine a first central point parameter corresponding to a central point in the image of interest;
a zoomed image determining module 320, configured to determine a zoomed image corresponding to the image of interest based on a preset zoomed parameter;
an image distance determining module 330, configured to determine an image distance between each image point and the center point based on the first center point parameter and an image point parameter corresponding to at least one image point in the scaled image;
the target segmented image determining module 340 is configured to determine an updated image of interest based on at least one image distance, and determine a target segmented image corresponding to the medical image to be processed based on the updated image of interest.
According to the technical scheme, the first center point parameter corresponding to the center point in the image of interest in the initial segmentation image and the scaling image corresponding to the image of interest are obtained, the image distance between each image point and the center point is determined based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaling image, the updated image of interest is determined based on the at least one image distance, and the target segmentation image corresponding to the medical image to be processed is determined based on the updated image of interest, so that the problem that the accuracy of the segmentation image obtained by the existing image segmentation algorithm is poor is solved, the post-processing algorithm of the segmentation image is perfected, the segmentation accuracy of the corrected segmentation image and the practicability of the segmentation image are improved, and effective and accurate reference value is provided for subsequent diagnosis and treatment.
On the basis of the above embodiment, optionally, the first center point parameter includes a first center point coordinate and a first center point gray value, and the image point parameter includes an image point coordinate and an image point gray value of the image point.
Based on the above embodiment, optionally, the target segmentation image determination module 340 includes:
the target segmentation image determining unit is used for determining a second center point parameter corresponding to a center point in the updated interested image;
judging whether the second center point coordinates in the second center point parameters are the same as the last center point coordinates in the last center point parameters corresponding to the second center point parameters;
If not, repeating the operation of determining the scaled image corresponding to the interested image based on the preset scaling parameter based on the second center point parameter;
If so, determining a target segmentation image corresponding to the medical image to be processed based on the updated interested image corresponding to the second center point parameter.
Based on the above embodiment, optionally, the target segmentation image determination module 340 includes:
The image updating unit is used for taking image points corresponding to the image distances meeting the preset distance range in the scaled image as target image points, or sorting the image points in the scaled image based on at least one image distance, and determining the target image points in the scaled image according to the sorting result and the preset selection number;
An updated image of interest is determined based on the target image point.
On the basis of the above embodiment, optionally, the apparatus further includes:
The initial segmentation image determining module is used for respectively inputting the medical image to be processed into at least two target image segmentation models;
And determining an initial segmentation image corresponding to the medical image to be processed based on the segmentation result and the voting mechanism which are respectively output by each target image segmentation model.
On the basis of the above embodiment, optionally, the apparatus further includes:
the target image segmentation model training module is used for acquiring a sample medical image, and executing a dicing operation on the sample medical image based on at least two dicing size standards to obtain a plurality of dicing medical images;
Based on the plurality of segmented medical images, training at least two initial image segmentation models respectively to obtain a target image segmentation model after training.
On the basis of the above embodiment, optionally, the preset scaling parameters include a scaling number of turns and a scaling direction, and the apparatus further includes:
the preset scaling parameter determining module is used for respectively inputting the verification images into at least two target image segmentation models, and determining predicted interested images corresponding to the verification images based on segmentation results and voting mechanisms which are respectively output by the target image segmentation models;
comparing the image size of the predicted image of interest with the image size of the checked image of interest corresponding to the checked image, and determining the corresponding scaling circle number and/or the scaling direction of the medical image to be processed according to the comparison result.
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 the corresponding functional modules and beneficial effects of the execution method.
Example IV
Fig. 5 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention. The electronic device 10 in FIG. 5 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 5, the electronic device 10 includes at least one processor 11, and a memory, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, etc., communicatively connected to the at least one processor 11, in which the memory stores a computer program executable by the at least one processor, and the processor 11 may perform various appropriate actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 may also be stored. The processor 11, the ROM 12 and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
Various components in the electronic device 10 are connected to the I/O interface 15, including an input unit 16, such as a keyboard, mouse, etc., an output unit 17, such as various types of displays, speakers, etc., a storage unit 18, such as a magnetic disk, optical disk, etc., and a communication unit 19, such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the segmentation method of medical images.
In some embodiments, the method of segmentation of medical images may be implemented as a computer program tangibly embodied on a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the above described medical image segmentation method may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the segmentation method of the medical image by any other suitable means (e.g. by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be a special or general purpose programmable processor, operable to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.
The computer program for implementing the medical image segmentation method of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
Example five
The fifth embodiment of the present invention also provides a computer-readable storage medium storing computer instructions for causing a processor to execute a method for segmenting a medical image, the method comprising:
Acquiring an interested image in an initial segmentation image corresponding to a medical image to be processed, and determining a first center point parameter corresponding to a center point in the interested image;
determining a zoom image corresponding to the interested image based on a preset zoom parameter;
Determining image distances between each image point and the center point respectively based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaled image respectively;
and determining an updated image of interest based on the at least one image distance, and determining a target segmentation image corresponding to the medical image to be processed based on the updated image of interest.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on 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.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user, for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback), and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a Local Area Network (LAN), a Wide Area Network (WAN), a blockchain network, and the Internet.
The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.
The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (7)
1. A method of segmenting a medical image, comprising:
Acquiring an interested image in an initial segmentation image corresponding to a medical image to be processed, and determining a first center point parameter corresponding to a center point in the interested image;
determining a zoom image corresponding to the interested image based on a preset zoom parameter;
Determining image distances between each image point and the center point respectively based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaled image respectively;
determining an updated image of interest based on at least one image distance, and determining a target segmentation image corresponding to the medical image to be processed based on the updated image of interest;
Wherein the determining, based on the updated image of interest, a target segmentation image corresponding to the medical image to be processed includes:
determining a second center point parameter corresponding to a center point in the updated image of interest;
judging whether a second center point coordinate in the second center point parameter is the same as a last center point coordinate in a last center point parameter corresponding to the second center point parameter;
If not, repeating the operation of determining the scaled image corresponding to the interested image based on the preset scaling parameter based on the second center point parameter;
If so, determining a target segmentation image corresponding to the medical image to be processed based on the updated image of interest corresponding to the second center point parameter;
wherein the method further comprises:
Respectively inputting the medical images to be processed into at least two target image segmentation models, wherein the model structures of different target image segmentation models are different and/or sample medical images for training are different;
Determining an interested image in the initial segmentation image corresponding to the medical image to be processed based on the segmentation result and the voting mechanism which are respectively output by each target image segmentation model;
wherein the determining an updated image of interest based on the at least one image distance comprises:
Taking image points corresponding to the image distances meeting the preset distance range in the scaled image as target image points, or sorting the image points in the scaled image based on at least one image distance, and determining the target image points in the scaled image according to a sorting result and a preset selection number;
The updated image of interest is determined based on the target image point.
2. The method of claim 1, wherein the first center point parameter comprises a first center point coordinate and a first center point gray value, and wherein the image point parameter comprises an image point coordinate and an image point gray value, respectively.
3. The method according to claim 1, wherein the method further comprises:
Acquiring a sample medical image, and executing a dicing operation on the sample medical image based on at least two dicing size standards to obtain a plurality of dicing medical images;
Based on the plurality of segmented medical images, training at least two initial image segmentation models respectively to obtain a target image segmentation model after training.
4. The method of claim 1, wherein the preset scaling parameters include a number of scaling turns and a scaling direction, and wherein the method further comprises:
Respectively inputting the verification images into at least two target image segmentation models, and determining a predicted interested image corresponding to the verification images based on segmentation results and voting mechanisms which are respectively output by the target image segmentation models;
Comparing the image size of the predicted interested image with the image size of the checking interested image corresponding to the checking image, and determining the corresponding scaling circle number and/or the scaling direction of the medical image to be processed according to the comparison result.
5. A medical image segmentation apparatus, comprising:
The first center point parameter determining module is used for acquiring an interested image in an initial segmentation image corresponding to the medical image to be processed and determining a first center point parameter corresponding to a center point in the interested image;
The scaling image determining module is used for determining a scaling image corresponding to the interested image based on preset scaling parameters;
the image distance determining module is used for determining the image distance between each image point and the center point respectively based on the first center point parameter and the image point parameter corresponding to at least one image point in the scaled image respectively;
The target segmentation image determining module is used for determining an updated interested image based on at least one image distance and determining a target segmentation image corresponding to the medical image to be processed based on the updated interested image;
wherein the target segmentation image determination module comprises:
The target segmentation image determining unit is used for determining a second center point parameter corresponding to a center point in the updated interested image, judging whether a second center point coordinate in the second center point parameter is the same as a last center point coordinate in a last center point parameter corresponding to the second center point parameter, if not, repeatedly executing an operation of determining a scaled image corresponding to the interested image based on the preset scaling parameter based on the second center point parameter, and if so, determining a target segmentation image corresponding to the medical image to be processed based on the updated interested image corresponding to the second center point parameter;
wherein the apparatus further comprises:
The system comprises an initial segmentation image determining module, an interested image determining module, a voting mechanism, a processing module and a processing module, wherein the initial segmentation image determining module is used for respectively inputting the medical image to be processed into at least two target image segmentation models, wherein the model structures of different target image segmentation models are different and/or sample medical images used for training are different;
wherein, the target segmentation image determining module further comprises:
And the interested image updating unit is used for taking image points corresponding to the image distances meeting the preset distance range in the scaled image as target image points or sorting the image points in the scaled image based on at least one image distance, determining the target image points in the scaled image according to a sorting result and a preset selection number, and determining the updated interested image based on the target image points.
6. An electronic device, the electronic device comprising:
at least one processor, and
A memory communicatively coupled to the at least one processor, wherein,
The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of segmentation of medical images according to any one of claims 1-4.
7. A computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions for causing a processor to implement the medical image segmentation method according to any one of claims 1-4 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210060260.2A CN114419068B (en) | 2022-01-19 | 2022-01-19 | A medical image segmentation method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210060260.2A CN114419068B (en) | 2022-01-19 | 2022-01-19 | A medical image segmentation method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114419068A CN114419068A (en) | 2022-04-29 |
| CN114419068B true CN114419068B (en) | 2024-12-10 |
Family
ID=81274845
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210060260.2A Active CN114419068B (en) | 2022-01-19 | 2022-01-19 | A medical image segmentation method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114419068B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115760877B (en) * | 2022-11-17 | 2024-10-18 | 平生医疗科技（昆山）有限公司 | Experiment cabin image segmentation method and device, electronic equipment and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109363698A (en) * | 2018-10-16 | 2019-02-22 | 杭州依图医疗技术有限公司 | A kind of method and device of breast image sign identification |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU7853900A (en) * | 1999-10-04 | 2001-05-10 | A.F.A. Products Group, Inc. | Improved image segmentation processing by user-guided image processing techniques |
| GB2456487A (en) * | 2007-01-09 | 2009-07-22 | Sony Uk Ltd | Image processing using RGB local mean and mapping of candidate colour components onto a possible dynamic range |
| CN103426169B (en) * | 2013-07-26 | 2016-12-28 | 西安华海盈泰医疗信息技术有限公司 | A kind of dividing method of medical image |
| JP7027248B2 (en) * | 2018-05-17 | 2022-03-01 | キヤノン株式会社 | Image processing equipment, programs, image processing methods |
| CN110197492A (en) * | 2019-05-23 | 2019-09-03 | 山东师范大学 | A kind of cardiac MRI left ventricle dividing method and system |
| CN110516754B (en) * | 2019-08-30 | 2022-11-01 | 大连海事大学 | Hyperspectral image classification method based on multi-scale superpixel segmentation |
| CN111428812B (en) * | 2020-04-14 | 2024-03-08 | 东软医疗系统股份有限公司 | Medical image data construction method and device |
| CN113160176B (en) * | 2021-04-23 | 2022-10-18 | 歌尔股份有限公司 | Defect detection method and device |
| CN113744201B (en) * | 2021-08-12 | 2024-10-15 | 珠海格力电器股份有限公司 | CT slice image processing method, device and equipment and CT scanner |
-
2022
- 2022-01-19 CN CN202210060260.2A patent/CN114419068B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109363698A (en) * | 2018-10-16 | 2019-02-22 | 杭州依图医疗技术有限公司 | A kind of method and device of breast image sign identification |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114419068A (en) | 2022-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Dilated-inception net: multi-scale feature aggregation for cardiac right ventricle segmentation | |
| US11996198B2 (en) | Determination of a growth rate of an object in 3D data sets using deep learning | |
| CN115018805B (en) | Segmentation model training method, image segmentation method, device, equipment and medium | |
| CN115409990B (en) | Medical image segmentation method, device, equipment and storage medium | |
| CN116188485A (en) | Image processing method, device, computer equipment and storage medium | |
| CN113362314B (en) | Medical image recognition method, recognition model training method and device | |
| CN116091490A (en) | Lung nodule detection method based on YOLOv4-CA-CBAM-K-means++ -SIOU | |
| CN114494215A (en) | Transformer-based thyroid nodule detection method | |
| CN116245832B (en) | Image processing method, device, equipment and storage medium | |
| CN114419068B (en) | A medical image segmentation method, device, equipment and storage medium | |
| CN117333487B (en) | Acne classification method, device, equipment and storage medium | |
| CN117746189A (en) | Aneurysm identification model training method, device, equipment and medium | |
| CN113344890B (en) | Medical image recognition method, recognition model training method and device | |
| Zhang et al. | A fast algorithm for automatic segmentation and extraction of a single object by active surfaces | |
| CN115482261A (en) | Blood vessel registration method, device, electronic equipment and storage medium | |
| Li et al. | 3D Multiple-Contextual ROI-Attention Network for Efficient and Accurate Volumetric Medical Image Segmentation | |
| CN114782768A (en) | Training method of pre-training network model, medical image processing method and equipment | |
| CN115147359A (en) | Lung lobe segmentation network model training method and device, electronic equipment and storage medium | |
| CN115578564B (en) | Training method and device for instance segmentation model, electronic equipment and storage medium | |
| AU2019334755B2 (en) | Determination of a growth rate of an object in 3D data sets using deep learning | |
| Wan et al. | A Heatmap Regression Approach for Right Ventricle Anatomical Landmarks Localization on Echocardiography | |
| CN115908296B (en) | Medical image class activation mapping evaluation method, device, computer equipment and storage medium | |
| CN112862804B (en) | System and method for processing retina blood vessel image | |
| Mohammed et al. | Detection Lung Nodules Using Medical CT Images Based on Deep learning techniques | |
| CN116894848A (en) | Image segmentation method, device, equipment and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |