CN117830239A - Skull image processing method and device, electronic equipment and storage medium - Google Patents
Skull image processing method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN117830239A CN117830239A CN202311785307.2A CN202311785307A CN117830239A CN 117830239 A CN117830239 A CN 117830239A CN 202311785307 A CN202311785307 A CN 202311785307A CN 117830239 A CN117830239 A CN 117830239A
- Authority
- CN
- China
- Prior art keywords
- image
- skull
- head scanning
- region
- straight line
- 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
- 210000003625 skull Anatomy 0.000 title claims abstract description 235
- 238000003672 processing method Methods 0.000 title abstract description 13
- 238000000034 method Methods 0.000 claims abstract description 54
- 239000013598 vector Substances 0.000 claims abstract description 39
- 238000012217 deletion Methods 0.000 claims abstract description 26
- 230000037430 deletion Effects 0.000 claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 21
- 230000008439 repair process Effects 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 23
- 238000004590 computer program Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 7
- 239000007787 solid Substances 0.000 claims description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 abstract description 8
- 230000000295 complement effect Effects 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 14
- 238000002591 computed tomography Methods 0.000 description 8
- 238000001514 detection method Methods 0.000 description 7
- 238000013461 design Methods 0.000 description 5
- 206010018852 Haematoma Diseases 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 4
- 210000004556 brain Anatomy 0.000 description 4
- 230000007812 deficiency Effects 0.000 description 4
- 230000002950 deficient Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000010146 3D printing Methods 0.000 description 2
- 208000032843 Hemorrhage Diseases 0.000 description 2
- 208000000386 Hypertensive Intracranial Hemorrhage Diseases 0.000 description 2
- 208000030886 Traumatic Brain injury Diseases 0.000 description 2
- 230000000740 bleeding effect Effects 0.000 description 2
- 210000000988 bone and bone Anatomy 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000007428 craniotomy Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000000638 solvent extraction Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 230000037182 bone density Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000005429 filling process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002601 radiography Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 238000003325 tomography Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Abstract
The application provides a skull image processing method, a skull image processing device, electronic equipment and a storage medium, and relates to the technical field of image processing, wherein the method comprises the following steps: acquiring a head scanning CT image containing a skull region, and performing linear clustering classification on the head scanning CT image to obtain a linear clustering result; based on the straight line clustering result, calculating a direction vector of a space cutting plane corresponding to the head scanning CT image; judging whether an image deletion exists in a skull region in the head scanning CT image according to the direction vector; if yes, inputting the head scanning CT image into the pre-trained skull repairing model to obtain a complete skull region image corresponding to the head scanning CT image. According to the method and the device, on the premise that normal data are not affected, abnormal skull image data can be complemented to the greatest extent, and skull complement operation time is saved.
Description
Technical Field
The present disclosure relates to the field of image processing technologies, and in particular, to a method and an apparatus for processing a skull image, an electronic device, and a storage medium.
Background
In clinic, when scanning is performed by adopting a layer thickness of 5mm during bleeding identification, and when three-dimensional directional operation, robot minimally invasive hematoma puncture and neuroendoscopic hematoma removal operation are adopted after hypertensive cerebral hemorrhage is confirmed, high-precision electronic computer tomography (Computed Tomography, CT) data are required, and the layer thickness is generally less than or equal to 1mm. In practice, the operation of the image technicians in hospitals is mostly not standard, the initial plane is not the canthus ear surface, or the scanning is started from the neck, or the scanning is started from the chin, and the initial plane is limited by the memory capacity of the CT equipment for data processing, so that the CT images of the skull, which are obtained, can not be completely scanned to the top of the head, are stopped, and the obtained CT images of the skull are incomplete. However, when planning a surgical path, feature points need to be found based on the complete skull surface, and then an effective path design is performed. Based on the above problems, it is necessary to perform a repair operation on the skull, and then perform feature point extraction and path design work based on the bone surface.
The existing skull complement method based on artificial intelligence (Artificial Intelligence, AI) is mainly used for generating a skull part to be complemented by adopting the AI-based method aiming at the skull deficiency caused by brain trauma and the skull deficiency caused by craniotomy, and performing skull complement operation by combining 3D printing or other technologies. Due to the fact that the missing skull caused by the conditions is complex, the data volume required by model training is large, normal skull images cannot be distinguished from defective skull images, normal data can be caused to undergo skull complementation to consume operation time, and correct data can be caused to become abnormal with a certain probability.
Disclosure of Invention
The application provides a skull image processing method, a skull image processing device, electronic equipment and a storage medium, which can ensure that abnormal skull image data is complemented to the maximum extent on the premise of not influencing normal data, and save skull complement operation time.
In a first aspect, there is provided a skull image processing method, comprising:
acquiring a head scanning CT image containing a skull region, and performing linear clustering classification on the head scanning CT image to obtain a linear clustering result;
based on the straight line clustering result, calculating a direction vector of a space cutting plane corresponding to the head scanning CT image;
judging whether an image deletion exists in a skull region in the head scanning CT image according to the direction vector;
if yes, inputting the head scanning CT image into the pre-trained skull repairing model to obtain a complete skull region image corresponding to the head scanning CT image.
In a second aspect, there is provided a skull image processing apparatus comprising:
the dividing module is used for acquiring a head scanning CT image containing a skull region, and carrying out linear clustering type division on the head scanning CT image to obtain a linear clustering type result;
the calculating module is used for calculating the direction vector of the space cutting plane corresponding to the head scanning CT image based on the straight line clustering result;
The judging module is used for judging whether the skull region in the head scanning CT image has image deletion or not according to the direction vector;
and the input module is used for inputting the head scanning CT image into the pre-trained skull repair model if yes, so as to obtain a complete skull region image corresponding to the head scanning CT image.
In a third aspect, there is provided an electronic device comprising: a processor and a memory for storing a computer program, the processor being for invoking and running the computer program stored in the memory for performing the method as in the first aspect or in various implementations thereof.
In a fourth aspect, a computer-readable storage medium is provided for storing a computer program for causing a computer to perform the method as in the first aspect or in various implementations thereof.
According to the technical scheme provided by the application, after the head scanning CT image containing the skull region is obtained, the head scanning CT image can be subjected to linear clustering classification to obtain a linear clustering result; then, based on a straight line clustering result, calculating a direction vector of a space cutting plane corresponding to the head scanning CT image; judging whether the skull region in the head scanning CT image has image deletion or not according to the direction vector; when judging that the skull region in the head scanning CT image has image missing, inputting the head scanning CT image into a pre-trained skull repair model to obtain a complete skull region image corresponding to the head scanning CT image. According to the technical scheme, the head top skull cutting trace is automatically identified on the head scanning CT image through straight line clustering classification and tangential plane fitting, intelligent complementation is carried out on the image with the cutting trace, and the image without the cutting trace is automatically skipped, so that skull complementation operation time can be saved, and the maximum complementation of abnormal data can be ensured on the premise of not influencing normal data.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application. Additional features and advantages of the present application will be set forth in the detailed description which follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic diagram of an example of a CT scanning method according to an embodiment of the present application;
FIG. 2 is a schematic diagram illustrating an example contrast of a CT scanning method according to an embodiment of the present application;
fig. 3 is an application scenario diagram provided in an embodiment of the present application;
fig. 4 is a schematic flow chart of a skull image processing method according to an embodiment of the present application;
fig. 5 is a flowchart of a skull image processing method according to another embodiment of the present application;
FIG. 6 is a schematic diagram illustrating an example of a morphological hole filling process according to an embodiment of the present application;
FIG. 7 is a schematic view of an example of a cut-out of a cranium top region according to an embodiment of the present application;
fig. 8 is a 3D diagram of a detection effect of whole brain according to an embodiment of the present application;
fig. 9 is a schematic diagram of an example of a straight line clustering class according to an embodiment of the present application;
FIG. 10 is a schematic illustration of an example of a spatial cutting plane fit provided in an embodiment of the present application;
fig. 11 is a schematic structural diagram of a skull image processing device according to an embodiment of the present application;
fig. 12 is a schematic structural diagram of a skull image processing apparatus according to an embodiment of the present application;
fig. 13 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present application based on the embodiments herein.
It should be noted that the terms "first," "second," and the like in the description and claims of the present application and 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 embodiments of the present application 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 server that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In a specific application scenario, as shown in fig. 1, the standard CT scan method is to scan about 10cm from canthus ear line to top of head, and the scan data covers the whole head area.
In clinic, when scanning is performed by adopting a layer thickness of 5mm during bleeding identification, and when three-dimensional directional operation, robot minimally invasive hematoma puncture and neuroendoscopic hematoma removal operation are adopted after hypertensive cerebral hemorrhage is confirmed, high-precision electronic computer tomography (Computed Tomography, CT) data are required, and the layer thickness is generally less than or equal to 1mm. In practice, the operation of the image technicians in hospitals is mostly not standard, the initial plane is not the canthus ear plane, or the scanning is started from the neck, or the scanning is started from the chin, and the operation is limited by the memory capacity of the CT equipment for data processing, so that the operation can not be completely performed to the top of the head, the obtained skull CT image is incomplete, and the contrast diagram is shown in figure 2.
When the chord length method or the software is adopted for planning the operation path, the characteristic points are required to be searched based on the whole skull surface, and then the effective path design is carried out. A large clinical study is retrospective, with missing data occupying about 1/3 of the amount that cannot be used and that cannot require a second scan by the patient. Based on the above problems, it is necessary to perform a repair operation on the skull, and then perform feature point extraction and path design work based on the bone surface.
The existing skull complement method based on artificial intelligence (Artificial Intelligence, AI) is mainly used for generating a skull part to be complemented by adopting the AI-based method aiming at the skull deficiency caused by brain trauma and the skull deficiency caused by craniotomy, and performing skull complement operation by combining 3D printing or other technologies. Due to the fact that the missing skull caused by the conditions is complex, the data volume required by model training is large, normal skull images cannot be distinguished from defective skull images, normal data can be caused to undergo skull complementation to consume operation time, and correct data can be caused to become abnormal with a certain probability.
It should be understood that the technical solution of the present application may be applied to the following scenarios, but is not limited to:
in some implementations, fig. 3 is an application scenario diagram provided in an embodiment of the present application, where, as shown in fig. 3, an electronic device 110 and a network device 120 may be included in the application scenario. The electronic device 110 may establish a connection with the network device 120 through a wired network or a wireless network.
By way of example, the electronic device 110 may be, but is not limited to, a desktop computer, a notebook computer, a tablet computer, and the like. The network device 120 may be a terminal device or a server, but is not limited thereto. In one embodiment of the present application, the electronic device 110 may send a request message to the network device 120, where the request message may be used to request that a complete skull region image be acquired, and further, the electronic device 110 may receive a response message sent by the network device 120, where the response message includes the complete skull region image.
In addition, fig. 3 illustrates one electronic device 110 and one network device 120, and may actually include other numbers of electronic devices and network devices, which is not limited in this application.
In other realizations, the technical solutions of the present application may also be executed by the electronic device 110, or the technical solutions of the present application may also be executed by the network device 120, which is not limited in this application.
After the application scenario of the embodiment of the present application is introduced, the following details of the technical solution of the present application will be described:
fig. 4 is a flowchart of a skull image processing method according to an embodiment of the present application, which may be performed by the electronic device 110 shown in fig. 3, but is not limited thereto. As shown in fig. 4, the method may include the steps of:
step 210, acquiring a head scanning CT image containing a skull region, and performing linear clustering classification on the head scanning CT image to obtain a linear clustering result.
The HU (Heat Unithu) value of the computerized tomography (Computed Tomography, CT) is used to reflect the density of the tissue and thus to diagnose the nature of the foreign body. The HU is a heat capacity unit of a bulb in medical equipment such as DR (Digital Radiography) and CT.
In a specific application scenario, after acquiring a head scan CT image containing a skull region, extraction of the skull region may be performed according to the HU value of the skull. The HU value of the normal human skull is 150-1000, so the area with the HU value larger than 150 can be taken as the skull area.
vBoneMask iSlice ＝1,vImage iSlice >150
In the formula, vBonneMask iSlice For the extracted skull region, vImage iSlice For a 1mm CT flat scan image, the iSlice single finger i Zhang Duanceng, such as a head image with a scan layer thickness of 1mm, has 165 sheets in total. The iSlice { i=1, 2,..165 }.
In a specific application scenario, since the existence of the cutting can lead to the existence of a large number of straight line segments in a skull region (such as the top of the skull), the straight line clustering result can be determined by performing straight line clustering classification on the head scanning CT image. So as to further judge whether the current skull region has image deletion according to the straight line clustering result, namely, distinguishing normal skull from defective skull. Only when judging that the image is missing, the completion operation of the defective skull region can be performed, so that the situation that the correct data are complemented into error data to cause skull deformation is avoided, and the calculation time consumed by skull completion is reduced.
And 220, calculating the direction vector of the head scanning CT image corresponding to the space cutting plane based on the straight line clustering result.
For the embodiment of the disclosure, at least one straight line clustering result is obtained in the straight line clustering classification process, and each straight line clustering result respectively comprises a preset number of initial straight lines. And then fitting the space cutting plane corresponding to each straight line clustering result according to the coordinate point value on the corresponding initial straight line, and further determining the direction vector of the space cutting plane corresponding to the head scanning CT image, wherein the direction vector is used for judging whether the skull region in the head scanning CT image has image missing or not.
Step 230, judging whether the skull region in the head scanning CT image has image deletion according to the direction vector.
For the embodiment of the disclosure, after determining to obtain the direction vector, for each coordinate point on the initial straight line in the cluster result, the vertical distance to the spatial cutting plane needs to be calculated according to the direction vector, and when the coordinate point is a certain point on the skull cutting straight line, the corresponding vertical distance should be as small as possible. After calculating the vertical distance from each coordinate point on the initial straight line to the space cutting plane in the straight line clustering result, accumulating all the vertical distances, comparing the vertical distance accumulated value with a preset distance threshold, and determining that the space cutting plane is a real fitting result when the vertical distance accumulated value is less than or equal to the preset distance threshold, namely that the skull region in the head scanning CT image has image deletion; and when the vertical distance accumulated value is judged to be larger than the preset distance threshold value, determining that no image deletion exists in the skull region in the head scanning CT image. The preset distance threshold may be set according to an actual application scenario, for example, may be 1mm.
Step 240, if it is determined that there is an image missing in the skull region in the head scan CT image, inputting the head scan CT image to the pre-trained skull repair model, and obtaining a complete skull region image corresponding to the head scan CT image.
The skull repairing model is a task model after skull repairing task training by utilizing a large number of sample images in advance. The skull repairing model may be any deep convolutional neural Network model, such as a Residual Network model (Residual Network), a lightweight model (MobileNet), or a Network model modified by some Network layers based on the above Network models, which is not limited in detail herein.
In a specific application scenario, in order to ensure the training accuracy of the skull repairing model, the sample image required by training can be determined in a sample data set in a targeted manner according to the actual repairing requirement (such as a cutting angle) of a head scanning CT image (for example, a cutting surface of 0-10 degrees is required to be repaired, a skull missing sample image comprising a cutting surface of 0-10 degrees can be generated by cutting the sample image after the sample image comprising a complete skull region is acquired, for example, a cutting surface of-45 degrees is required to be repaired, and a skull missing sample image comprising a cutting surface of-45 degrees can be generated by cutting the sample image after the sample image comprising the complete skull region is acquired). And then, according to the image restoration data corresponding to the skull missing region, a corresponding preset characteristic label is configured for the sample image. And then performing task training of skull repair on the skull repair model by using the skull missing sample image provided with the preset feature tag.
Correspondingly, the training method of the skull repair model comprises the following steps: generating a sample image configured with a preset feature tag, wherein the sample image is a pre-cut image containing a skull missing region, and the preset feature tag is image restoration data corresponding to the skull missing region; inputting the sample image into a skull repairing model, and performing repairing training of a skull missing region on the skull repairing model, wherein in repairing training of the skull missing region, the sample image is used as an input feature, a preset feature label is used as a training label, model parameters in the skull repairing model are iteratively updated until the accuracy of the skull repairing model on repairing of the skull missing region is greater than a preset accuracy threshold, and the completion of the skull repairing model training is judged.
For the embodiment of the disclosure, after judging that the image missing exists in the skull region in the head scanning CT image, three-dimensional image data corresponding to the head scanning CT image can be input into a pre-trained skull repair model, the skull repair model can carry out image repair on the head scanning CT image based on model parameters obtained by training, and a complete skull region image corresponding to the head scanning CT image is output after repair is finished.
In summary, according to the skull image processing method provided by the application, after a head scanning CT image containing a skull region is obtained, linear clustering classification can be performed on the head scanning CT image to obtain a linear clustering result; then, based on a straight line clustering result, calculating a direction vector of a space cutting plane corresponding to the head scanning CT image; judging whether the skull region in the head scanning CT image has image deletion or not according to the direction vector; when judging that the skull region in the head scanning CT image has image missing, inputting the head scanning CT image into a pre-trained skull repair model to obtain a complete skull region image corresponding to the head scanning CT image. According to the technical scheme, the head top skull cutting trace is automatically identified on the head scanning CT image through straight line clustering classification and tangential plane fitting, intelligent complementation is carried out on the image with the cutting trace, and the image without the cutting trace is automatically skipped, so that skull complementation operation time can be saved, and the maximum complementation of abnormal data can be ensured on the premise of not influencing normal data.
Based on the embodiment shown in fig. 4, as a refinement and extension of the above embodiment, for a complete description of a specific implementation procedure of the method of this embodiment, as shown in fig. 5, the method includes the following steps:
Step 310, acquiring a head scanning CT image containing a skull region, and performing linear clustering classification on the head scanning CT image to obtain a linear clustering result.
For the disclosed embodiments, after acquiring a head scan CT image containing a skull region, extraction of the skull region may be performed according to HU values of the skull. The HU value of the normal skull is 150-1000, so the area with HU value greater than 150 can be taken as the skull area, and the specific implementation process can be referred to the related description in the embodiment step 210, and the description is omitted here.
However, due to the differences in skull bone density of different human bodies, the skull bone region extracted in this way often has a hollow missing part, which is honeycomb-shaped, rather than solid in shape. Therefore, before performing the straight line clustering classification on the head scan CT image to obtain the straight line clustering result, the steps of the embodiment may further include: and carrying out morphological hole filling treatment on the head scanning CT image to obtain a solid head scanning CT image. As shown in fig. 6, solid skull regions under different angle views (e.g., xoy view, xoz view, yoz view) can be obtained by morphological hole filling.
In a specific application scenario, the presence of the cuts can result in a large number of straight line segments in the skull region (e.g., the top of the skull), so that straight line detection can be performed by scanning the CT image over the head. As shown in fig. 7, straight line detection and analysis can be performed based on the skull image below the xoz plane. In order to improve the operation efficiency, the upper half of the head (i.e., the cranium top region) may be taken as the target region for detecting the straight line. The target area is an area containing all cutting lines, when the target area to be intercepted is determined, the target area can be determined according to the cutting angle corresponding to the head scanning CT image, and if the cutting angle is 0 degrees, the target area can be selected as an upper image of one quarter of the head scanning CT image; if the cutting angle is 45 degrees, the target area can be selected as the upper image of one half of the head scanning CT image.
For the embodiment of the disclosure, when the straight line detection is performed on the head scanning CT image, the extraction of the straight line cluster may be performed based on the hough transform, and specifically, the parameters of the hough transform may be set based on the actual situation of the cutting. Here, hough transform is a common technique in the field of image processing, and will not be described here in detail. Illustratively, the parameters of the hough transform may be selected as follows: angle detection range [ -45,45], straight length >10mm, straight gap: 10mm.
As shown in the 3D diagram of the detection effect of the whole brain in fig. 8, during the detection of the upper half brain, other straight lines of non-cutting surfaces are detected, and particularly, a plurality of interferences are easily introduced in the bilateral ear regions of the cranium, which is determined by the specific physiological characteristics of the human brain, and the situation that the skull above the two ears is relatively flat exists. Further processing is required to screen the detected lines. In the application, a DBSCAN clustering method can be adopted, and the straight lines are clustered according to the characteristic that the coplanar straight lines have the same slope and intercept, and the slope and the intercept are taken as clustering characteristics. Firstly, clustering is carried out by adopting slopes, and linear clusters with different angles are distinguished. Since DBSCAN clustering also belongs to a common algorithm, the description is omitted. In order to reserve the target straight line to the maximum extent, the angle error of the clustering reserves a certain fault-tolerant space, and the clustering parameters are set as follows: the angle error eps=3°, the minimum number of samples min_samples=10 are clustered, and at least one straight line clustering result can be obtained after clustering.
In a specific application scenario, as a possible implementation manner, a more severe clustering parameter, such as a smaller angle error and/or a larger clustering minimum sample number, may be set according to an actual application scenario, and a straight line cluster corresponding to a skull missing area may be obtained by direct clustering through the clustering parameter, and the number of interference straight lines is limited in view of a certain angle error between the interference straight line and the straight line cluster in the skull missing area, so that the interference straight line cluster is ignored in the process of straight line clustering. Therefore, when a DBSCAN clustering method is adopted to cluster to obtain a linear cluster, the existence of image deletion in a skull region in the head scanning CT image can be judged; otherwise, when the linear clusters are not obtained through clustering, judging that no image deletion exists in the skull region in the head scanning CT image.
As a possible implementation manner, a looser clustering parameter can be set according to an actual application scene, for example, a larger angle error and/or a smaller clustering minimum sample number are set, at least one linear cluster can be obtained by direct clustering through the clustering parameter, and the at least one linear cluster can comprise a linear cluster corresponding to a skull missing region and a linear cluster corresponding to an interference straight line. In the following steps of the embodiments of the present disclosure, the technical solutions in the present disclosure will be described by taking such a clustering method as an example, but the present disclosure is not limited to the specific embodiments.
As shown in fig. 9, l 1 ,l 2 ,l 3 Is a detected candidate straight line, wherein l 1 ,l 2 For straight lines on the cutting face, l 3 To interfere with straight line l 1 ,l 2 Intercept in x-axis is x 1 ，l 3 Intercept in x-axis of x 2 Therefore, the intercept is adopted for clustering, two straight line clustering results can be obtained after clustering, and one straight line clustering result comprises a straight line l 1 ,l 2 The other straight line clustering result comprises a straight line l 3 。
Accordingly, for the embodiment of the present disclosure, when performing straight line clustering classification on a head scan CT image to obtain a straight line clustering result, the steps of the embodiment may include: intercepting a cranium top area of a head scanning CT image as a target area divided by straight line clustering; extracting an initial straight line and straight line characteristic information corresponding to the initial straight line in a target area based on Hough transformation, wherein the straight line characteristic information at least comprises a straight line slope and a straight line intercept; based on the linear characteristic information and preset clustering parameters, carrying out clustering division on the initial straight lines to obtain at least one straight line clustering result, wherein the preset clustering parameters at least comprise a clustering maximum angle error and a clustering minimum sample number.
And 320, collecting a coordinate point set of an initial straight line contained in the straight line clustering result, performing tangential plane fitting based on the coordinate point set, and calculating by SVD (singular value decomposition) in the tangential plane fitting process to obtain a direction vector of a spatial cutting plane corresponding to the head scanning CT image.
For the embodiment of the disclosure, for each straight line clustering result, a coordinate point set may be determined according to at least one initial straight line obtained by clustering, and then arranged in a matrix form according to the x, y, z coordinate sequence in the coordinate point set:
and then performing tangential plane fitting:
the spatial plane expression is: ax+by+cz+d=0, the unknown parameters are: A. b, C, D, the point set matrix is solved by SVD decomposition with the parameters A, B, C, D.
And then constructing a coefficient matrix:the coefficient matrix P is as follows:
SVD decomposition is expressed as follows:
P＝U∑V T
solving singular vectors corresponding to the minimum singular values, which correspond toIs a direction vector of a planeD＝-(A×x 0 +B×y 0 +C×z 0 ). A 3D view of the fitted spatial cutting plane is shown in fig. 10.
And 330, judging whether the skull region in the head scanning CT image has image deletion according to the direction vector.
For the embodiment of the disclosure, after determining to obtain the direction vector, for each coordinate point on the initial straight line in the cluster result, the vertical distance to the spatial cutting plane needs to be calculated according to the corresponding direction vector, and when the coordinate point is a certain point on the skull cutting straight line, the corresponding vertical distance should be as small as possible. After calculating the vertical distance from each coordinate point on the initial straight line to the space cutting plane in the straight line clustering result, accumulating all the vertical distances, comparing the vertical distance accumulated value with a preset distance threshold, and determining that the space cutting plane is a real fitting result when the vertical distance accumulated value is less than or equal to the preset distance threshold, namely that the skull region in the head scanning CT image has image deletion; and when the vertical distance accumulated value is judged to be larger than the preset distance threshold value, determining that no image deletion exists in the skull region in the head scanning CT image. The preset distance threshold may be set according to an actual application scenario, for example, may be 1mm.
Accordingly, for embodiments of the present disclosure, the embodiment steps may include: calculating a first distance between each coordinate point in the coordinate point set and the space cutting plane according to the direction vector; accumulating and calculating a first distance corresponding to each coordinate point in the coordinate point set to obtain a second distance; if the second distance is less than or equal to the preset distance threshold value, determining that the skull region in the head scanning CT image has image deletion; if the second distance is larger than the preset distance threshold value, determining that no image deletion exists in the skull region in the head scanning CT image.
And 340, if judging that the skull region in the head scanning CT image has image deletion, inputting the head scanning CT image into a pre-trained skull repair model to obtain a complete skull region image corresponding to the head scanning CT image.
The pre-trained skull repair model includes a feature extraction module operable to extract image features and a missing skull region generation module operable to determine a skull missing region based on image features already present in the image and to generate a pixel value for each coordinate point in the skull missing region. Accordingly, for embodiments of the present disclosure, the embodiment steps may include: inputting the head scanning CT image to a feature extraction module, and extracting regional features of a current skull region in the head scanning CT image by using the feature extraction module; and generating a skull missing region corresponding to the current skull region by using a missing skull generation module based on the region characteristics, and obtaining a complete skull region image corresponding to the head scanning CT image.
In summary, according to the technical scheme in the application, after a head scanning CT image containing a skull region is acquired, linear clustering classification can be performed on the head scanning CT image to obtain a linear clustering result; then, based on a straight line clustering result, calculating a direction vector of a space cutting plane corresponding to the head scanning CT image; judging whether the skull region in the head scanning CT image has image deletion or not according to the direction vector; when judging that the skull region in the head scanning CT image has image missing, inputting the head scanning CT image into a pre-trained skull repair model to obtain a complete skull region image corresponding to the head scanning CT image. According to the technical scheme, the head top skull cutting trace is automatically identified on the head scanning CT image through straight line clustering classification and tangential plane fitting, intelligent complementation is carried out on the image with the cutting trace, and the image without the cutting trace is automatically skipped, so that skull complementation operation time can be saved, and the maximum complementation of abnormal data can be ensured on the premise of not influencing normal data.
Based on the above detailed description of the skull image processing method provided in fig. 4 and 5, as shown in fig. 11, fig. 11 is a block diagram of a skull image processing apparatus according to an exemplary embodiment. As shown in fig. 11, the apparatus includes:
The dividing module 41 is configured to acquire a head scan CT image including a skull region, and perform linear clustering classification on the head scan CT image to obtain a linear clustering result;
the calculating module 42 is configured to calculate a direction vector of the spatial cutting plane corresponding to the head scan CT image based on the straight line clustering result;
a judging module 43, configured to judge whether there is an image missing in a skull region in the head scanning CT image according to the direction vector;
the input module 44 is configured to input the head scan CT image to the pre-trained skull repair model to obtain a complete skull region image corresponding to the head scan CT image if it is determined that there is an image defect in the skull region in the head scan CT image.
In some embodiments of the present application, the partitioning module 41 may be configured to intercept a cranium top region of a head scan CT image as a target region for straight line clustering class partitioning; extracting an initial straight line and straight line characteristic information corresponding to the initial straight line in a target area based on Hough transformation, wherein the straight line characteristic information at least comprises a straight line slope and a straight line intercept; based on the linear characteristic information and preset clustering parameters, carrying out clustering division on the initial straight lines to obtain at least one straight line clustering result, wherein the preset clustering parameters at least comprise a clustering maximum angle error and a clustering minimum sample number.
In some embodiments of the present application, the calculating module 42 may be configured to collect a set of coordinate points of an initial straight line included in the straight line clustering result; and performing tangential plane fitting based on the coordinate point set, and obtaining a direction vector of a spatial cutting plane corresponding to the head scanning CT image by utilizing SVD decomposition calculation in the tangential plane fitting process.
In some embodiments of the present application, the determining module 43 may be configured to calculate, according to the direction vector, a first distance between each coordinate point in the set of coordinate points and the spatial cutting plane; accumulating and calculating a first distance corresponding to each coordinate point in the coordinate point set to obtain a second distance; if the second distance is less than or equal to the preset distance threshold value, determining that the skull region in the head scanning CT image has image deletion; if the second distance is larger than the preset distance threshold value, determining that no image deletion exists in the skull region in the head scanning CT image.
In some embodiments of the present application, as shown in fig. 12, the apparatus further includes: a processing module 45;
the processing module 45 is configured to perform morphological hole filling processing on the head scanning CT image to obtain a solid head scanning CT image;
correspondingly, the dividing module 41 may be configured to perform linear clustering classification on the header scanning CT image after the morphological hole filling processing, so as to obtain a linear clustering result.
In some embodiments of the present application, the skull repair model includes a feature extraction module and a missing skull region generation module, an input module 44 operable to input a head scan CT image to the feature extraction module, with the feature extraction module extracting region features of a current skull region in the head scan CT image; and generating a skull missing region corresponding to the current skull region by using a missing skull generation module based on the region characteristics, and obtaining a complete skull region image corresponding to the head scanning CT image.
In some embodiments of the present application, as shown in fig. 12, the apparatus further includes: a training module 46;
the training module 46 is configured to generate a sample image configured with a preset feature tag, where the sample image is a pre-cut image including a skull missing region, and the preset feature tag is image repair data corresponding to the skull missing region; inputting the sample image into a skull repairing model, and performing repairing training of a skull missing region on the skull repairing model, wherein in repairing training of the skull missing region, the sample image is used as an input feature, a preset feature label is used as a training label, model parameters in the skull repairing model are iteratively updated until the accuracy of the skull repairing model on repairing of the skull missing region is greater than a preset accuracy threshold, and the completion of the skull repairing model training is judged.
The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.
In the embodiment of the application, after the head scanning CT image containing the skull region is acquired, the head scanning CT image can be subjected to linear clustering classification to obtain a linear clustering result; then, based on a straight line clustering result, calculating a direction vector of a space cutting plane corresponding to the head scanning CT image; judging whether the skull region in the head scanning CT image has image deletion or not according to the direction vector; when judging that the skull region in the head scanning CT image has image missing, inputting the head scanning CT image into a pre-trained skull repair model to obtain a complete skull region image corresponding to the head scanning CT image. According to the technical scheme, the head top skull cutting trace is automatically identified on the head scanning CT image through straight line clustering classification and tangential plane fitting, intelligent complementation is carried out on the image with the cutting trace, and the image without the cutting trace is automatically skipped, so that skull complementation operation time can be saved, and the maximum complementation of abnormal data can be ensured on the premise of not influencing normal data.
The skull image processing apparatus of the embodiment of the present invention is described above from the viewpoint of functional blocks with reference to the drawings. It should be understood that the functional module may be implemented in hardware, or may be implemented by instructions in software, or may be implemented by a combination of hardware and software modules. Specifically, each step of the embodiment of the skull image processing method in the embodiment of the present invention may be completed by an integrated logic circuit of hardware in a processor and/or an instruction in a software form, and the steps of the skull image processing method applied in connection with the embodiment of the present invention may be directly embodied as a hardware decoding processor for execution, or may be completed by a combination of hardware and software modules in the decoding processor for execution. Alternatively, the software modules may be located in a well-established storage medium in the art such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, and the like. The storage medium is located in the memory, and the processor reads the information in the memory and combines the hardware to complete the steps in the embodiment of the skull image processing method.
Fig. 13 is a schematic block diagram of an electronic device 700 in accordance with one embodiment of the present invention.
As shown in fig. 13, the electronic device 700 may include:
a memory 710 and a processor 720, the memory 710 being configured to store a computer program and to transfer the program code to the processor 720. In other words, the processor 720 may call and run a computer program from the memory 710 to implement the method in the embodiment of the present invention.
For example, the processor 720 may be configured to perform the above-described method embodiments according to instructions in the computer program.
In some embodiments of the invention, the processor 720 may include, but is not limited to:
a general purpose processor, digital signal processor (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like.
In some embodiments of the invention, the memory 710 includes, but is not limited to:
volatile memory and/or nonvolatile memory. The nonvolatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable EPROM (EEPROM), or a flash Memory. The volatile memory may be random access memory (Random Access Memory, RAM) which acts as an external cache. By way of example, and not limitation, many forms of RAM are available, such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (Double Data Rate SDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), and Direct memory bus RAM (DR RAM).
In some embodiments of the invention, the computer program may be partitioned into one or more modules that are stored in the memory 710 and executed by the processor 720 to perform the methods provided by the invention. The one or more modules may be a series of computer program instruction segments capable of performing the specified functions, the instruction segments describing the execution of the computer program in the controller.
As shown in fig. 13, the electronic device 700 may further include:
a transceiver 730, the transceiver 730 being connectable to the processor 720 or the memory 710.
The processor 720 may control the transceiver 730 to communicate with other devices, and in particular, may transmit data or data to other devices or receive data or data transmitted by other devices. Transceiver 730 may include a transmitter and a receiver. Transceiver 730 may further include antennas, the number of which may be one or more.
It will be appreciated that the various components in the electronic device are connected by a bus system that includes, in addition to a data bus, a power bus, a control bus, and a status signal bus.
The present invention also provides a computer storage medium having stored thereon a computer program which, when executed by a computer, enables the computer to perform the method of the above-described method embodiments. Alternatively, an embodiment of the present invention also provides a computer program product containing instructions which, when executed by a computer, cause the computer to perform the method of the method embodiment described above.
When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the invention, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy Disk, a hard Disk, a magnetic tape), an optical medium (e.g., a digital video disc (Digital Video Disc, DVD)), or a semiconductor medium (e.g., a Solid State Disk (SSD)), or the like.
Those of ordinary skill in the art will appreciate that the various illustrative modules and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, and for example, the division of the modules is merely a logical function division, and there may be additional divisions when actually implemented, for example, multiple modules or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or modules, which may be in electrical, mechanical, or other forms.
The modules illustrated as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, i.e., may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. For example, functional modules in the embodiments of the present application may be integrated into one processing module, or each module may exist alone physically, or two or more modules may be integrated into one module.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that changes and substitutions are within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A method of skull image processing, comprising:
acquiring a head scanning CT image containing a skull region, and performing linear clustering classification on the head scanning CT image to obtain a linear clustering result;
Based on the straight line clustering result, calculating a direction vector of a space cutting plane corresponding to the head scanning CT image;
judging whether the skull region in the head scanning CT image has image deletion or not according to the direction vector;
if yes, inputting the head scanning CT image into a pre-trained skull repair model to obtain a complete skull region image corresponding to the head scanning CT image.
2. The method of claim 1, wherein the performing a straight line clustering classification on the head scan CT image to obtain a straight line clustering result comprises:
intercepting a cranium top region of the head scanning CT image as a target region divided by straight line clustering;
extracting an initial straight line and straight line characteristic information corresponding to the initial straight line in the target area based on Hough transformation, wherein the straight line characteristic information at least comprises a straight line slope and a straight line intercept;
and carrying out clustering division on the initial straight lines based on the straight line characteristic information and preset clustering parameters to obtain at least one straight line clustering result, wherein the preset clustering parameters at least comprise a clustering maximum angle error and a clustering minimum sample number.
3. The method of claim 2, wherein the calculating a direction vector of the head scan CT image for a spatial cutting plane based on the straight line clustering result comprises:
collecting a coordinate point set of an initial straight line contained in the straight line clustering result;
and performing tangential plane fitting based on the coordinate point set, and obtaining a direction vector of a spatial cutting plane corresponding to the head scanning CT image by utilizing SVD decomposition calculation in the tangential plane fitting process.
4. A method according to claim 3, wherein said determining whether there is an image deletion of the skull region in the head scan CT image from the direction vector comprises:
according to the direction vector, calculating a first distance between each coordinate point in the coordinate point set and the space cutting plane;
accumulating and calculating the first distance corresponding to each coordinate point in the coordinate point set to obtain a second distance;
if the second distance is less than or equal to a preset distance threshold value, determining that the skull region in the head scanning CT image has image deletion;
and if the second distance is larger than the preset distance threshold, determining that no image deletion exists in the skull region in the head scanning CT image.
5. The method of any one of claims 1 to 4, wherein prior to performing a linear clustering classification on the head scan CT image to obtain a linear clustering result, the method further comprises:
performing morphological hole filling treatment on the head scanning CT image to obtain a solid head scanning CT image;
the performing straight line clustering classification on the head scanning CT image to obtain a straight line clustering result comprises the following steps:
and performing linear clustering classification on the head scanning CT image subjected to the morphological hole filling treatment to obtain a linear clustering result.
6. The method according to claim 1, wherein the skull repair model includes a feature extraction module and a missing skull region generation module, the inputting the head scan CT image into a pre-trained skull repair model, obtaining a complete skull region image corresponding to the head scan CT image, comprising:
inputting the head scanning CT image to the feature extraction module, and extracting regional features of a current skull region in the head scanning CT image by using the feature extraction module;
and generating a skull missing region corresponding to the current skull region by using the missing skull generation module based on the region characteristics, and obtaining a complete skull region image corresponding to the head scanning CT image.
7. A method according to claim 1, further comprising a training method of the skull repair model, comprising:
generating a sample image configured with a preset feature tag, wherein the sample image is a pre-cut image containing a skull missing region, and the preset feature tag is image restoration data corresponding to the skull missing region;
and inputting the sample image into a skull repairing model, and performing repairing training of a skull missing region on the skull repairing model, wherein in the repairing training of the skull missing region, the sample image is used as an input characteristic, the preset characteristic label is used as a training label, model parameters in the skull repairing model are iteratively updated until the accuracy of the skull repairing model on the skull missing region is greater than a preset accuracy threshold, and the skull repairing model training is judged to be completed.
8. A skull image processing apparatus, comprising:
the dividing module is used for acquiring a head scanning CT image containing a skull region, and carrying out linear clustering classification on the head scanning CT image to obtain a linear clustering result;
The calculating module is used for calculating the direction vector of the space cutting plane corresponding to the head scanning CT image based on the straight line clustering result;
the judging module is used for judging whether the skull region in the head scanning CT image has image deletion or not according to the direction vector;
and the input module is used for inputting the head scanning CT image into the pre-trained skull repair model if yes, so as to obtain a complete skull region image corresponding to the head scanning CT image.
9. An electronic device, comprising:
a processor and a memory for storing a computer program, the processor being for invoking and running the computer program stored in the memory to perform the method of any of claims 1-7.
10. A computer readable storage medium storing a computer program for causing a computer to perform the method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311785307.2A CN117830239A (en) | 2023-12-22 | 2023-12-22 | Skull image processing method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311785307.2A CN117830239A (en) | 2023-12-22 | 2023-12-22 | Skull image processing method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117830239A true CN117830239A (en) | 2024-04-05 |
Family
ID=90505297
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311785307.2A Pending CN117830239A (en) | 2023-12-22 | 2023-12-22 | Skull image processing method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117830239A (en) |
-
2023
- 2023-12-22 CN CN202311785307.2A patent/CN117830239A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101864380B1 (en) | Surgical image data learning system | |
| JP2021536057A (en) | Lesion detection and positioning methods, devices, devices, and storage media for medical images | |
| CN110326024B (en) | Method and apparatus for processing histological images captured by a medical imaging apparatus | |
| KR20190088375A (en) | Surgical image data learning system | |
| WO2015066565A1 (en) | System and method for a situation and awareness-based intelligent surgical system | |
| CN110717905B (en) | Brain image detection method, computer device, and storage medium | |
| CN110223279B (en) | Image processing method and device and electronic equipment | |
| CN110210519B (en) | Classification method, computer device, and storage medium | |
| CN112950569B (en) | Melanoma image recognition method, device, computer equipment and storage medium | |
| CN111667474A (en) | Fracture identification method, apparatus, device and computer readable storage medium | |
| CN111968130B (en) | Brain contrast image processing method, device, medium and electronic equipment | |
| CN110751187A (en) | Training method of abnormal area image generation network and related product | |
| CN111445451B (en) | Brain image processing method, system, computer device and storage medium | |
| CN110826557A (en) | Method and device for detecting fracture | |
| CN113344926B (en) | Method, device, server and storage medium for recognizing biliary-pancreatic ultrasonic image | |
| CN111932538B (en) | Method, device, computer equipment and storage medium for analyzing thyroid gland atlas | |
| CN111160441B (en) | Classification method, computer device, and storage medium | |
| CN117830239A (en) | Skull image processing method and device, electronic equipment and storage medium | |
| CN109087357A (en) | Scan orientation method, apparatus, computer equipment and computer readable storage medium | |
| CN112200780A (en) | Bone tissue positioning method, bone tissue positioning device, computer equipment and storage medium | |
| EP3712845B1 (en) | Image identifying method and image identifying device | |
| CN111968160A (en) | Image matching method and storage medium | |
| CN115330675A (en) | HoVer-Transformer-based breast tumor benign and malignant identification method and device | |
| CN110197722B (en) | AI-CPU system platform | |
| CN113298773A (en) | Heart view identification and left ventricle detection device and system based on deep learning |
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 |