CN113781547B

CN113781547B - Head symmetry axis identification method and device, storage medium and computer equipment

Info

Publication number: CN113781547B
Application number: CN202110897360.6A
Authority: CN
Inventors: 付璐; 才品嘉; 李戈
Original assignee: Neusoft Medical Systems Co Ltd
Current assignee: Neusoft Medical Systems Co Ltd
Priority date: 2021-08-05
Filing date: 2021-08-05
Publication date: 2024-05-31
Anticipated expiration: 2041-08-05
Also published as: CN113781547A

Abstract

The application discloses a head symmetry axis identification method and device, a storage medium and computer equipment, wherein the method comprises the following steps: identifying a brain tissue edge chain code and a chain code cutting point corresponding to a brain tissue region in a target head medical image, and extracting a front upper sagittal Dou Lian code and a rear upper sagittal Dou Lianma in the brain tissue edge chain code according to the position relation between the brain tissue edge chain code and the chain code cutting point; determining front segment candidate corner points according to front segment curvature values of each point in the front segment sagittal sinus chain code, and determining rear segment candidate corner points according to rear segment curvature values of each point in the rear segment sagittal sinus chain code; detecting a first position characteristic of a front segment candidate corner in a front segment sagittal sinus chain code, determining a front segment target corner based on the first position characteristic, detecting a second position characteristic of a rear segment candidate corner in a rear segment sagittal sinus chain code, and determining a rear segment target corner based on the second position characteristic; and determining a head symmetry axis corresponding to the brain tissue region according to the connecting line of the front section target corner and the rear section target corner.

Description

Head symmetry axis identification method and device, storage medium and computer equipment

Technical Field

The application relates to the technical field of medical images, in particular to a head symmetry axis identification method and device, a storage medium and computer equipment.

Background

Brain diseases greatly affect the quality of life of patients, so diagnosis and treatment thereof are hot spot problems in the global health field. The identification of the symmetry axis of the brain is an indispensable key step in the diagnosis, treatment and other aspects of brain diseases. For example, the method has important significance for cerebral hemorrhage edema, cerebral tumor occupation and other diseases, cerebral midline deviation, cerebral ischemia on left and right sides, quantitative calculation of bleeding volume, and pre-judging brain injury of head radiotherapy.

The basis for determining the symmetry axis of the brain is mainly the characteristics of the brain morphology and tissue structure. The current method for identifying the symmetry axis of the head mainly comprises the following steps: a three-dimensional least square fitting method and a brain tissue structure symmetry recognition method. The three-dimensional least square fitting method relies on complete and normal sickle brain structures and high-resolution image quality, but patients with treatment often have diseases of the brain, and the imaging quality of different hospital equipment is difficult to ensure. The brain tissue structure symmetry recognition method depends on the normal brain tissue internal structure and the head position correction condition during scanning, and in the technology depending on symmetry, brain tissue pathological changes and head position non-ideal during scanning can influence the symmetry of a recognition target (such as eyes). In summary, the accuracy of the head symmetry axis identification needs to be improved at present.

Disclosure of Invention

In view of the above, the present application provides a head symmetry axis identification method and apparatus, a storage medium, and a computer device, which are helpful for improving accuracy of head symmetry axis identification.

According to an aspect of the present application, there is provided a head symmetry axis recognition method including:

Identifying a brain tissue edge chain code and a chain code cutting point corresponding to a brain tissue region in a target head medical image, and extracting a front-section upper sagittal Dou Lian code and a rear-section upper sagittal Dou Lianma in the brain tissue edge chain code according to the position relation between the brain tissue edge chain code and the chain code cutting point;

determining front segment candidate corner points according to front segment curvature values of each point in the front segment upper sagittal sinus chain code, and determining rear segment candidate corner points according to rear segment curvature values of each point in the rear segment upper sagittal sinus chain code;

Detecting a first position feature of the anterior segment candidate corner in the anterior segment upper sagittal sinus chain code, determining an anterior segment target corner based on the first position feature, detecting a second position feature of the posterior segment candidate corner in the posterior segment upper sagittal sinus chain code, and determining a posterior segment target corner based on the second position feature;

and determining a head symmetry axis corresponding to the brain tissue region according to the connecting line of the front section target angular point and the rear section target angular point.

Optionally, extracting the anterior upper sagittal Dou Lian code and the posterior upper sagittal Dou Lian code in the brain tissue edge chain code according to the positional relationship between the brain tissue edge chain code and the chain code cutting point specifically includes:

determining the parts above and below the horizontal line corresponding to the chain code cutting point in the brain tissue edge chain code as the anterior segment upper sagittal Dou Lianma code and the posterior segment upper sagittal Dou Lian code respectively, wherein the chain code cutting point comprises the centroid of the brain tissue region;

and respectively cutting the middle parts of the front section upper sagittal Dou Lian codes and the rear section upper sagittal Dou Lian codes according to a preset proportion.

Optionally, the determining the anterior segment candidate corner according to the anterior segment curvature value of each point in the anterior segment upper sagittal sinus chain code, and determining the posterior segment candidate corner according to the posterior segment curvature value of each point in the posterior segment upper sagittal sinus chain code specifically includes:

acquiring a plurality of preset front-section chord length parameters and a plurality of preset rear-section chord length parameters;

Respectively calculating a plurality of front segment curvature values of each point in the sagittal sinus chain code on the front segment according to the plurality of preset front segment chord length parameters, and determining the front segment candidate corner points based on curvature products corresponding to the plurality of front segment curvature values;

And respectively calculating a plurality of back-section curvature values of each point in the sagittal sinus chain code on the back section according to the plurality of preset back-section chord length parameters, and determining the back-section candidate corner points based on curvature products corresponding to the plurality of back-section curvature values.

Optionally, the detecting a first position feature of the anterior segment candidate corner in the anterior segment upper sagittal sinus chain code, determining an anterior segment target corner based on the first position feature, and detecting a second position feature of the posterior segment candidate corner in the posterior segment upper sagittal sinus chain code, and determining a posterior segment target corner based on the second position feature specifically includes:

Respectively determining a preset anterior segment distance corresponding to each preset anterior segment chord length parameter and a preset posterior segment distance corresponding to each preset posterior segment chord length parameter;

selecting a plurality of groups of first reference points which are separated from two sides of any anterior segment candidate angular point by the preset anterior segment distance on the anterior segment upper sagittal sinus chain code, wherein each group of first reference points comprises two first reference points;

Based on a plurality of groups of first reference points, respectively establishing a first straight line and a second straight line which passes through any front segment candidate angular point and is perpendicular to the first straight line, and determining that the first position characteristic of any front segment candidate angular point is a salient point characteristic when the intersection point of the first straight line and the second straight line corresponding to at least one group of first reference points is in the brain tissue region;

after deleting the candidate corner points of the front section with the salient point characteristics, screening out the target corner points of the front section with the largest curvature product;

selecting a plurality of groups of second reference points which are separated from two sides of any posterior segment candidate angular point by the preset posterior segment distance on the posterior segment upper sagittal sinus chain code, wherein each group of second reference points comprises two;

based on a plurality of groups of second reference points, respectively establishing a third straight line and a fourth straight line which passes through any back-section candidate corner point and is perpendicular to the third straight line, and determining that the second position characteristic of any back-section candidate corner point is a bump characteristic when the intersection point of the third straight line and the fourth straight line corresponding to at least one group of second reference points is in the brain tissue region;

and after deleting the back-end candidate corner points with the salient point characteristics, screening back-end target corner points with the largest curvature product.

Optionally, the target head medical image comprises images of a plurality of faults; the determining a head symmetry axis corresponding to the brain tissue region according to the connecting line of the front section target corner and the rear section target corner specifically comprises the following steps:

Determining a first connecting line between a first front section target corner corresponding to the first target layer and a first rear section target corner and a second connecting line between a second front section target corner corresponding to the second candidate group and a second rear section target corner;

Establishing a symmetry axis identification surface according to the first connecting line and the second connecting line;

and acquiring intersection lines of the symmetry axis identification surface and brain tissue areas corresponding to the faults, and determining the intersection lines as head symmetry axes of the faults.

Optionally, the first target layer comprises multiple layers, and the second target layer comprises multiple layers; before the identifying of the brain tissue edge chain code corresponding to the brain tissue region in the target head medical image and the chain code cutting point, the method further comprises:

Acquiring the target head medical image, respectively carrying out contour recognition on the image of each fault, and determining brain tissue areas of a plurality of faults, wherein the target head medical image comprises a CT image and/or an MR image;

Acquiring a first candidate fault with the largest brain tissue area in a plurality of faults, and taking at least two faults which are respectively adjacent to the front and the rear of the first candidate fault and the first candidate fault as a first target layer;

Acquiring a second candidate fault with the smallest difference between the area of brain tissue areas in the faults and a preset area, and taking at least two faults which are respectively adjacent to the front and the rear of the second candidate fault and the second candidate fault as a second target layer;

correspondingly, identifying the brain tissue edge chain code and the chain code cutting point corresponding to the brain tissue region in the target head medical image specifically comprises the following steps:

And identifying brain tissue edge chain codes and chain code cutting points of brain tissue areas corresponding to the first target layer and the second target layer.

Optionally, before the construction of the symmetry axis identification surface, the method further includes:

According to the relative positions of all front section target corner points in the first target layer in the corresponding front section sagittal sinus chain codes, filtering the front section target corner points, and according to the relative positions of all rear section target corner points in the first target layer in the corresponding rear section sagittal sinus chain codes, filtering the rear section target corner points, so that the relative position difference of all front section target corner points in the filtered first target layer and the relative position difference of all rear section target corner points are smaller than a preset difference value;

According to the relative positions of all front section target corner points in the corresponding front section upper sagittal sinus chain codes in the second target layer, filtering the front section target corner points, and according to the relative positions of all rear section target corner points in the corresponding rear section upper sagittal sinus chain codes in the second target layer, filtering the rear section target corner points, so that the relative position difference of all front section target corner points and the relative position difference of all rear section target corner points in the filtered second target layer are smaller than a preset difference value;

Determining a first front-stage target corner point of the first target layer according to the curvature value of each front-stage target corner point in the first target layer, and determining a first rear-stage target corner point of the first target layer according to the curvature value of each rear-stage target corner point in the first target layer;

Determining a second front-stage target corner of the second target layer according to the curvature value of each front-stage target corner in the second target layer, and determining a second rear-stage target corner of the second target layer according to the curvature value of each rear-stage target corner in the second target layer.

According to another aspect of the present application, there is provided a head symmetry axis recognition device including:

The brain tissue identification module is used for identifying brain tissue edge chain codes and chain code cutting points corresponding to brain tissue areas in the target head medical image, and extracting front-section upper sagittal Dou Lian codes and rear-section upper sagittal Dou Lianma in the brain tissue edge chain codes according to the position relation between the brain tissue edge chain codes and the chain code cutting points;

The candidate corner determining module is used for determining front segment candidate corners according to front segment curvature values of each point in the front segment upper sagittal sinus chain code and determining rear segment candidate corners according to rear segment curvature values of each point in the rear segment upper sagittal sinus chain code;

The target angular point determining module is used for detecting a first position characteristic of the front segment candidate angular point in the front segment upper sagittal sinus chain code, determining a front segment target angular point based on the first position characteristic, detecting a second position characteristic of the rear segment candidate angular point in the rear segment upper sagittal sinus chain code, and determining a rear segment target angular point based on the second position characteristic;

and the symmetry axis identification module is used for determining the symmetry axis of the head corresponding to the brain tissue region according to the connecting line of the front section target angular point and the rear section target angular point.

Optionally, the brain tissue identification module is specifically configured to:

Optionally, the candidate corner determining module is specifically configured to:

Optionally, the target corner determining module is specifically configured to:

Optionally, the target head medical image comprises images of a plurality of faults; the symmetry axis identification module is specifically configured to:

Optionally, the first target layer comprises multiple layers, and the second target layer comprises multiple layers; the apparatus further comprises:

The target layer acquisition module is used for acquiring the target head medical image before the identification of a brain tissue edge chain code and a chain code cutting point corresponding to a brain tissue region in the target head medical image, respectively carrying out contour identification on images of each fault, and determining brain tissue regions of a plurality of faults, wherein the target head medical image comprises a CT image and/or an MR image; and

Acquiring a first candidate fault with the largest brain tissue area in a plurality of faults, and taking at least two faults which are respectively adjacent to the front and the rear of the first candidate fault and the first candidate fault as a first target layer; and

correspondingly, the brain tissue identification module is specifically used for:

Optionally, the apparatus further comprises:

The object corner screening module is used for filtering the front segment object corner according to the relative position of each front segment object corner in the corresponding front segment sagittal sinus chain code and filtering the rear segment object corner according to the relative position of each rear segment object corner in the corresponding rear segment sagittal sinus chain code before the symmetry axis identification surface is constructed, so that the relative position difference of each front segment object corner and the relative position difference of each rear segment object corner in the filtered first object layer are smaller than the preset difference value; and

According to the relative positions of all front section target corner points in the corresponding front section upper sagittal sinus chain codes in the second target layer, filtering the front section target corner points, and according to the relative positions of all rear section target corner points in the corresponding rear section upper sagittal sinus chain codes in the second target layer, filtering the rear section target corner points, so that the relative position difference of all front section target corner points and the relative position difference of all rear section target corner points in the filtered second target layer are smaller than a preset difference value; and

Determining a first front-stage target corner point of the first target layer according to the curvature value of each front-stage target corner point in the first target layer, and determining a first rear-stage target corner point of the first target layer according to the curvature value of each rear-stage target corner point in the first target layer; and

According to still another aspect of the present application, there is provided a storage medium having stored thereon a computer program which, when executed by a processor, implements the head symmetry axis identification method described above.

According to still another aspect of the present application, there is provided a computer apparatus including a storage medium, a processor and a computer program stored on the storage medium and executable on the processor, the processor implementing the head symmetry axis identification method described above when executing the program.

By means of the technical scheme, the head symmetry axis identification method, the head symmetry axis identification device, the storage medium and the computer equipment provided by the application are used for identifying brain tissue edge chain codes and chain code cutting points from brain tissue areas in target head medical images, cutting the brain tissue edge chain codes into front-section upper sagittal Dou Lian codes and rear-section upper sagittal Dou Lian codes according to the chain code cutting points, further finding out front-section candidate corner points and rear-section candidate corner points according to curvature values of the front-section upper sagittal Dou Lian codes and rear-section upper sagittal Dou Lianma points, screening concave points in the candidate corner points according to concave position characteristics of the front-section candidate corner points in the front-section upper sagittal sinus chain codes and the rear-section candidate corner points in the rear-section upper sagittal sinus chain codes, respectively serving as front-section target corner points and rear-section target corner points, and finally determining head symmetry axes corresponding to the brain tissue areas according to connecting lines of the front-section target corner points and the rear-section target corner points. According to the embodiment of the application, the upper sagittal sinus of the anterior segment and the upper sagittal Dou Lianxian of the posterior segment are utilized as the characteristics of the symmetry axis of the head, the brain tissue edge chain code of the brain tissue region outline which is not easily affected by brain tissue lesions and image quality is selected, the upper sagittal sinus characteristics reflected by the brain tissue edge chain code are utilized to identify the anterior segment target angular point and the posterior segment target angular point, so that the determination of the symmetry axis is realized, the problem of inaccurate symmetry axis identification caused by the influence of the brain tissue lesions or the image quality on the tissue definition is solved, the accuracy and the robustness of the head symmetry axis identification are improved, and the angular point identification is performed on one-dimensional data, namely the chain code, and the identification speed is high and the efficiency is high.

The foregoing description is only an overview of the present application, and is intended to be implemented in accordance with the teachings of the present application in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present application more readily apparent.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:

Fig. 1 is a schematic flow chart of a head symmetry axis identification method according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a brain tissue region image provided by an embodiment of the present application;

fig. 3 is a schematic flow chart of another method for identifying symmetry axes of a head according to an embodiment of the present application;

fig. 4 is a schematic structural diagram of a head symmetry axis recognition device according to an embodiment of the present application.

Detailed Description

The application will be described in detail hereinafter with reference to the drawings in conjunction with embodiments. It should be noted that, without conflict, the embodiments of the present application and features of the embodiments may be combined with each other.

In this embodiment, there is provided a head symmetry axis identification method, as shown in fig. 1, including:

step 101, identifying a brain tissue edge chain code and a chain code cutting point corresponding to a brain tissue region in a target head medical image, and extracting an anterior segment upper sagittal Dou Lian code and a posterior segment upper sagittal Dou Lianma in the brain tissue edge chain code according to the position relation between the brain tissue edge chain code and the chain code cutting point;

The target head medical image in the embodiment of the application can be a head CT image or an MR image of a patient, particularly can be a CT plain scan (NCCT), a CT perfusion image (CTP) or a CT vascular image (CTA), can also be an MR perfusion image (PWI), an MR diffusion image (DWI) and the like, and can be used for avoiding the poor recognition accuracy of a head symmetry axis caused by the influence of the pathological changes of brain tissues or the image quality of the patient on the tissue definition.

In the above embodiment, first, after obtaining a target head medical image, identifying the outline of a brain tissue region in the image, obtaining a brain tissue edge chain code corresponding to the outline of the brain tissue region, and identifying a chain code cutting point in the brain tissue region, where the chain code cutting point may specifically be the centroid or the center of the brain tissue region, and the chain code cutting point is used to cut the brain tissue edge chain code. In a specific application scenario, a horizontal line passing through a chain code cutting point can be drawn, and a brain tissue edge chain code is cut into two parts through the horizontal line, wherein each part comprises an upper sagittal sinus, the upper half part is used as an anterior segment upper sagittal Dou Lian code, and the lower half part is used as a posterior segment upper sagittal Dou Lian code. The target head medical image is generally obtained by scanning a plurality of faults on the brain of a patient, that is, the target head medical image generally comprises a plurality of fault images, the brain tissue region can be a brain tissue region corresponding to any fault which needs to be identified by a symmetry axis, the head symmetry axis of the brain tissue region is determined by analyzing a brain tissue edge chain code corresponding to the fault, or the brain tissue region can be brain tissue regions corresponding to a plurality of specified faults, and the symmetry axis identification of each fault brain tissue region is realized by analyzing the brain tissue edge chain code corresponding to the plurality of specified faults.

102, Determining front segment candidate corner points according to front segment curvature values of each point in the front segment sagittal sinus chain code, and determining rear segment candidate corner points according to rear segment curvature values of each point in the rear segment sagittal sinus chain code;

And secondly, for the sagittal Dou Lian code on the anterior segment and the sagittal Dou Lian code on the posterior segment of a certain fault, respectively finding out candidate angular points on the anterior segment, the sagittal Dou Lian code on the anterior segment and the sagittal sinus chain code on the posterior segment, so as to conveniently use the candidate angular points to carry out symmetry axis identification. Taking the determination of the candidate corner points of the front section as an example, the curvature value of each point in the sagittal sinus chain code on the front section, namely the front section curvature value, can be calculated according to a corner point detection method, so that the candidate corner points of the front section are screened according to the front section curvature value of each point. When the front section curvature value is calculated, one preset front section chord length parameter can be selected, and a plurality of preset front section chord length parameters can also be selected. If one is selected, a plurality of points with larger values (or points with values exceeding a preset value) can be used as front segment candidate corner points according to the values of the front segment curvatures of the points; if a plurality of points are selected, the curvature product of each point can be obtained by multiplying a plurality of front-section curvature values of each point, and a plurality of points (or points with values exceeding a certain value) with larger curvature products are used as front-section candidate corner points. By the method, the point with larger bending degree in the anterior segment sagittal sinus chain code, namely the anterior segment candidate corner point, can be found, and as shown in fig. 2, the point with the maximum bending degree of the anterior segment sagittal sinus edge and the point with the maximum bending degree of the posterior segment sagittal sinus edge are positioned on the straight line of the symmetry axis of the head according to the human head structure. And then, finding a posterior segment candidate corner point on the sagittal sinus chain code on the posterior segment by using a similar method so as to screen a anterior segment target corner point and a posterior segment target corner point from the anterior segment candidate corner point and the posterior segment candidate corner point respectively, thereby carrying out symmetry axis identification.

Step 103, detecting a first position feature of the anterior segment candidate corner in the anterior segment sagittal sinus chain code, determining an anterior segment target corner based on the first position feature, detecting a second position feature of the posterior segment candidate corner in the posterior segment sagittal sinus chain code, and determining a posterior segment target corner based on the second position feature;

Further, after determining the front candidate corner and the rear candidate corner for a certain fault brain tissue region, a front target corner and a rear target corner which are finally used for determining the symmetry axis are selected from the front candidate corner and the rear candidate corner respectively. In a specific application scenario, taking a front segment candidate corner as an example, the concave-convex characteristic of the front segment candidate corner at the position of the front segment sagittal Dou Lian code, namely the first position characteristic, namely whether the front segment candidate corner is a concave point or a convex point on the front segment sagittal sinus contour, as shown in fig. 2, screening the front segment candidate corner according to the characteristic that the intersection point of a symmetry axis and the front segment sagittal Dou Bianyuan and the corner point of the symmetry axis and the rear segment sagittal Dou Bianyuan should be the concave point, removing the convex point, only retaining the concave point, and screening the front segment candidate corner belonging to the concave point as the front segment target corner. If the candidate front-segment corner points screened in the step comprise a plurality of candidate front-segment corner points, one candidate front-segment corner point is selected from the plurality of corner points according to a preset rule to serve as a target front-segment corner point, for example, a corner point which is located near the center position in a front-segment upper sagittal sinus chain code is selected to serve as a target front-segment corner point, and for example, a corner point with the maximum curvature value is selected to serve as a target front-segment corner point. In addition, a posterior target corner may be selected among the posterior Hou Xuanjiao points in a similar manner as described above.

Step 104, determining a head symmetry axis corresponding to the brain tissue region according to the connecting line of the front section target corner point and the rear section target corner point.

Finally, the connecting line of the front section target angular point and the rear section target angular point corresponding to a certain fault can be used as the head symmetry axis corresponding to the fault, in addition, in order to realize rapid identification of the head symmetry axes corresponding to a plurality of faults, in a specific application scene, the symmetry axis identification surface can be fitted by using the connecting line of the front section target angular point and the rear section target angular point corresponding to a plurality of faults, so that the identification of the head symmetry axes of all faults is realized by using the identification surface.

By applying the technical scheme of the embodiment, a brain tissue edge chain code and a chain code cutting point are identified from a brain tissue region in a target head medical image, the brain tissue edge chain code is cut into a front-section upper sagittal Dou Lian code and a rear-section upper sagittal Dou Lian code according to the chain code cutting point, a front-section candidate corner and a rear-section candidate corner are further found according to curvature values of each point of the front-section upper sagittal Dou Lian code and the rear-section upper sagittal Dou Lianma code, concave points in the candidate corner are screened according to concave-convex position characteristics of the front-section candidate corner in the front-section upper sagittal sinus chain code and the rear-section candidate corner in the rear-section upper sagittal sinus chain code, the concave points in the candidate corner are respectively used as a front-section target corner and a rear-section target corner, and finally, a head symmetry axis corresponding to the brain tissue region is determined according to a connecting line of the front-section target corner and the rear-section target corner. According to the embodiment of the application, the upper sagittal sinus of the anterior segment and the upper sagittal Dou Lianxian of the posterior segment are utilized as the characteristics of the symmetry axis of the head, the brain tissue edge chain code of the brain tissue region outline which is not easily affected by brain tissue lesions and image quality is selected, the upper sagittal sinus characteristics reflected by the brain tissue edge chain code are utilized to identify the anterior segment target angular point and the posterior segment target angular point, so that the determination of the symmetry axis is realized, the problem of inaccurate symmetry axis identification caused by the influence of the brain tissue lesions or the image quality on the tissue definition is solved, the accuracy and the robustness of the head symmetry axis identification are improved, and the angular point identification is performed on one-dimensional data, namely the chain code, and the identification speed is high and the efficiency is high.

Further, as a refinement and extension of the foregoing embodiment, for fully explaining the implementation procedure of the present embodiment, another method for identifying a symmetry axis of a head is provided, as shown in fig. 3, which includes:

Step 201, acquiring the target head medical image, respectively carrying out contour recognition on the image of each fault, and determining brain tissue areas of a plurality of faults, wherein the target head medical image comprises a CT image and/or an MR image;

in the above embodiment, CT/MR images of a patient are acquired, and CT/MR imaging scanning is generally performed according to a plurality of faults, and specifically, CT plain scan (NCCT), CT perfusion imaging (CTP), CT vascular imaging (CTA), MR perfusion imaging (PWI), MR diffusion imaging (DWI) and the like may be selected, where the images need to include sagittal sinus structures on the brain, which is a main basis for determining the symmetry axis. The brain tissue mask, namely, a mask obtained by removing the skull of the original head portrait and the external tissues of the skull in a certain layer of original image can be extracted through an active contour model algorithm, specifically, the pixels occupied by the tissues in the skull can be marked as 1 through the active contour model algorithm, the rest part of the tissues in the skull can be marked as 0, and the pixels marked as 1 are taken as brain tissue areas.

Step 202, obtaining a first candidate fault with the largest brain tissue area in a plurality of faults, and taking at least two faults which are respectively adjacent to the front and the rear of the first candidate fault and the first candidate fault as a first target layer; acquiring a second candidate fault with the smallest difference between the area of brain tissue areas in the faults and a preset area, and taking at least two faults which are respectively adjacent to the front and the rear of the second candidate fault and the second candidate fault as a second target layer;

in the above embodiment, in order to quickly identify the symmetry axis of each fault, the first target layer and the second target layer are selected, so that the symmetry axis identification plane is determined in a manner of analyzing the brain tissue edge chain codes corresponding to the first target layer and the second target layer, and head symmetry axis identification is realized. In a specific application scene, calculating the area of a brain tissue region corresponding to each fault, taking a layer in which the brain tissue region with the largest area is as a first candidate fault, and taking the first candidate fault and the front layer and the rear layer of the first candidate fault as a first target layer so as to conveniently select target corner points in chain codes corresponding to a plurality of first target layers, thereby realizing the identification of a symmetry axis. In addition, a second candidate fault with the smallest difference from the preset area is selected from the multiple faults, the second candidate fault and the front layer and the rear layer of the second candidate fault are taken as a second target layer, the preset area can be specifically determined by counting the brain tissue area of each fault, for example, the fault with the area of which the size is ranked at the top 10% is taken as the second candidate layer, the preset area can also be determined by multiplying the area of the largest brain tissue area by a specific coefficient, for example, the preset area is 90% of the area of the largest brain tissue area, and the preset area can also be determined according to experience or historical data, and is not limited herein.

Step 203, identifying a brain tissue edge chain code and a chain code cutting point of a brain tissue region corresponding to the first target layer and the second target layer, and extracting a front upper sagittal Dou Lian code and a rear upper sagittal Dou Lianma in the brain tissue edge chain code according to the positional relationship between the brain tissue edge chain code and the chain code cutting point;

Optionally, the determining manner of the anterior-segment upper sagittal Dou Lian code and the posterior-segment upper sagittal Dou Lian code in the brain tissue edge chain code of each target layer in step 203 may specifically include: determining the parts above and below the horizontal line corresponding to the chain code cutting point in the brain tissue edge chain code as the anterior segment upper sagittal Dou Lianma code and the posterior segment upper sagittal Dou Lian code respectively, wherein the chain code cutting point comprises the centroid of the brain tissue region; and respectively cutting the middle parts of the front section upper sagittal Dou Lian codes and the rear section upper sagittal Dou Lian codes according to a preset proportion.

In the above embodiment, for each fault in the first target layer and the second target layer, the centroid of the brain tissue region is first determined, the centroid is used as the chain code cutting point of the layer, a horizontal line is drawn based on the position of the chain code cutting point, the part above the horizontal line in the brain tissue edge chain code is used as the anterior segment upper sagittal Dou Lian code, the part below the horizontal line in the brain tissue edge chain code is used as the posterior segment upper sagittal Dou Lian code, further, since the anterior segment target corner for finally identifying the symmetry axis is generally not located at two ends of the anterior segment upper sagittal Dou Lian code, and similarly, the posterior segment target corner is generally not located at two ends of the posterior segment upper sagittal Dou Lian code, in order to reduce the calculation amount, the anterior segment upper sagittal sinus chain code can be intercepted according to a preset proportion, the middle part is reserved, for example, the middle 70% part is reserved as the effective part, and the middle part of the posterior segment upper sagittal Dou Lian code is intercepted in the same way.

Step 204, obtaining a plurality of preset front-section chord length parameters and a plurality of preset rear-section chord length parameters; respectively calculating a plurality of front segment curvature values of each point in the sagittal sinus chain code on the front segment according to the plurality of preset front segment chord length parameters, and determining the front segment candidate corner points based on curvature products corresponding to the plurality of front segment curvature values; respectively calculating a plurality of back-section curvature values of each point in the sagittal sinus chain code on the back section according to the plurality of preset back-section chord length parameters, and determining the back-section candidate corner points based on curvature products corresponding to the plurality of back-section curvature values;

In the above embodiment, since the chord length parameters have a large influence on the selection of the subsequent corner points, in order to avoid that the incorrect chord length parameter selection has a large influence on the subsequent symmetry axis identification, a plurality of chord length parameters may be selected for the anterior segment and the posterior segment upper sagittal Dou Lianma, for example, 3 preset anterior segment chord length parameters and 3 preset posterior segment chord length parameters may be selected, the anterior segment preset chord length parameters may be Lf1, lf2, lf3, respectively, and specifically, lf 1=10, lf 2=15, lf 3=20, and the posterior segment preset chord length parameters may be Lb1, lb2, lb3, respectively. Specifically, lb1=4, lb2=8, and lb3=12 can be selected. And respectively obtaining front-segment curvature values Hf1, hf2 and Hf3 of each point on the front-segment sagittal sinus chain code corresponding to different front-segment preset chord length parameters and back-segment curvature values Hb1, hb2 and Hb3 of each point on the back-segment sagittal sinus chain code corresponding to different back-segment preset field parameters by using a CPDA angular point detection method, so as to calculate curvature products hf=hf 1×hf2×hf3 of each point on the front-segment sagittal sinus chain code and curvature products hb=hb 1×hb2×hb3 of each point on the back-segment sagittal sinus chain code. And further screening front segment candidate corner points on the front segment sagittal sinus chain code and rear segment candidate corner points on the rear segment sagittal sinus chain code according to Hf and Hb respectively.

Step 205, detecting a first position feature of the anterior segment candidate corner in the anterior segment sagittal sinus chain code, determining an anterior segment target corner based on the first position feature, detecting a second position feature of the posterior segment candidate corner in the posterior segment sagittal sinus chain code, and determining a posterior segment target corner based on the second position feature;

Optionally, step 205 may specifically include:

Step 205-1, respectively determining a preset anterior segment distance corresponding to each preset anterior segment chord length parameter and a preset posterior segment distance corresponding to each preset posterior segment chord length parameter;

step 205-2, selecting a plurality of groups of first reference points on the sagittal sinus chain code on the anterior segment, wherein the first reference points are separated from two sides of any anterior segment candidate corner by the preset anterior segment distance, and each group of first reference points comprises two first reference points;

Step 205-3, based on a plurality of groups of first reference points, respectively establishing a first straight line and a second straight line which passes through any front segment candidate corner point and is perpendicular to the first straight line, and determining that the first position characteristic of any front segment candidate corner point is a bump characteristic when the intersection point of the first straight line corresponding to at least one group of first reference points and the second straight line is in the brain tissue region;

Step 205-4, after deleting the candidate corner points of the front section with the salient point characteristics, screening out the target corner points of the front section with the largest curvature product;

In the above steps 205-1 to 205-4, candidate corner points are screened according to the convexity of each candidate corner point corresponding to each layer of image. For a certain anterior segment candidate corner point on the anterior segment upper sagittal sinus chain code, respectively selecting first reference points at the positions with preset anterior segment distances from two sides of the anterior segment candidate corner point, wherein the preset anterior segment distances are products of preset anterior segment chord length parameters and preset coefficients, for example, the preset anterior segment distances are 1/2 of preset anterior segment chord length parameters, the preset anterior segment chord length parameters correspond to the preset anterior segment distances, and accordingly multiple groups of first reference points can be obtained. For each set of first reference points, a first straight line L1 passing through the two first reference points is found, and a straight line L2 passing through the candidate corner point of the front section and perpendicular to the straight line L1 is continued to be found. If the intersection of L1 and L2 is within the brain tissue region, the first location feature of the anterior segment candidate corner is marked as a bump feature. For any front segment candidate corner point, deleting the front segment candidate corner point as long as the first position feature of the front segment candidate corner point calculated according to a certain preset front segment chord length parameter is a salient point feature, further selecting one with the largest curvature product from the rest front segment candidate corner points as a front segment target corner point, and determining the point with the largest concave point position and the largest concave degree in the sagittal sinus chain code on the front segment corresponding to each layer.

For example, a point 1/2 from the chord length parameter of a certain candidate point on the chain code is selected on both sides of the candidate point, a straight line L1 passing through the two points is obtained, and a straight line L2 passing through the candidate point and perpendicular to the straight line L1 is continuously obtained. If the intersection point of the straight lines L1 and L2 falls in the brain tissue region of the layer, namely the concave-convex mark of the candidate point under the current chord length parameter is-1, otherwise, the concave-convex mark is +1. At least one concave-convex mark is-1 under different chord length parameters, namely the candidate point is convex, and otherwise, the candidate point is concave. Since the superior sagittal sinus should be characterized as pits on the edge chain code, pits are preserved according to the mark. And selecting a concave point with the largest curvature product from front and rear chain code points of each layer as a front and rear upper sagittal Dou Duiying corner point of the layer.

Step 205-5, selecting a plurality of groups of second reference points on the sagittal sinus chain code on the posterior segment, wherein the second reference points are separated from two sides of any posterior segment candidate corner by the preset posterior segment distance, and each group of second reference points comprises two;

step 205-6, based on a plurality of groups of second reference points, respectively establishing a third straight line and a fourth straight line which passes through any back-stage candidate corner point and is perpendicular to the third straight line, and determining that the second position characteristic of any back-stage candidate corner point is a bump characteristic when the intersection point of the third straight line and the fourth straight line corresponding to at least one group of second reference points is in the brain tissue region;

And 205-7, after deleting the candidate corner points of the rear section with the salient point characteristics, screening out the target corner points of the rear section with the largest curvature product.

The manner in which the rear-stage target corner points are determined in steps 205-5 to 205-7 is similar to that in steps 205-2 to 205-4, and will not be described in detail herein.

Step 206, filtering the front-section target corner points according to the relative positions of the front-section target corner points in the corresponding front-section sagittal sinus chain codes in the first target layer, and filtering the rear-section target corner points according to the relative positions of the rear-section target corner points in the corresponding rear-section sagittal sinus chain codes in the first target layer, so that the relative position difference of the front-section target corner points and the relative position difference of the rear-section target corner points in the filtered first target layer are smaller than a preset difference value; according to the relative positions of all front section target corner points in the corresponding front section upper sagittal sinus chain codes in the second target layer, filtering the front section target corner points, and according to the relative positions of all rear section target corner points in the corresponding rear section upper sagittal sinus chain codes in the second target layer, filtering the rear section target corner points, so that the relative position difference of all front section target corner points and the relative position difference of all rear section target corner points in the filtered second target layer are smaller than a preset difference value;

Step 207, determining a first front-stage target corner of the first target layer according to the curvature value of each front-stage target corner in the first target layer, and determining a first rear-stage target corner of the first target layer according to the curvature value of each rear-stage target corner in the first target layer; determining a second front-section target corner point of the second target layer according to the curvature value of each front-section target corner point in the second target layer, and determining a second rear-section target corner point of the second target layer according to the curvature value of each rear-section target corner point in the second target layer;

In steps 206 to 207, taking determining the first front-stage target corner corresponding to the first target layer as an example, respectively calculating the relative positions of the front-stage target corners corresponding to the faults included in the first target layer in the located chain codes, and filtering the front-stage target corner according to the relative positions of all the front-stage target corners corresponding to the first target layer. The relative position of the front-section target corner points which are relatively close to each other is reserved, so that the relative position difference of the relative position differences of the front-section target corner points in the reserved first target layer is smaller than the preset difference value. And further taking the largest curvature product (the curvature product is the product of curvature values corresponding to chord length parameters of each front section) in the rest front section target corner points as a first front section target corner point af corresponding to the first target layer. Correspondingly, a first back-end target corner ab corresponding to the first target layer, and a second front-end target corner bf and a second back-end target corner bb corresponding to the second target layer are calculated in a similar manner as described above.

Step 208, determining a first connecting line between a first front section target corner corresponding to the first target layer and a first rear section target corner and a second connecting line between a second front section target corner corresponding to the second candidate group and a second rear section target corner; establishing a symmetry axis identification surface according to the first connecting line and the second connecting line; and acquiring intersection lines of the symmetry axis identification surface and brain tissue areas corresponding to the faults, and determining the intersection lines as head symmetry axes of the faults.

In the above embodiment, the first connecting line La of the first front-segment target angular point af and the first rear-segment target angular point ab, and the second connecting line Lb of the second front-segment target angular point bf and the second rear-segment target angular point bb are determined, and the planes Lb and La under standard image input are located in the same plane, so that a plane can be determined, and in some special application scenarios, if Lb and La are not coplanar, the planes can be located on the same plane by rotating Lb, and the planes are taken as symmetry axis identification planes, and the intersecting lines of different fault brain tissue regions and the symmetry axis identification planes are the head symmetry axes corresponding to the layer brain tissue regions. Therefore, two groups of target angular points (namely a first front section target angular point and a first rear section target angular point, and a second front section target angular point and a second rear section target angular point) are determined through analysis of brain tissue areas of a small number of faults, so that the recognition of head symmetry axes of all faults is realized, and the symmetry axis recognition efficiency is improved.

Further, as a specific implementation of the method of fig. 1, an embodiment of the present application provides a head symmetry axis recognition device, as shown in fig. 4, including:

Optionally, the target corner determining module is specifically configured to:

Optionally, the apparatus further comprises:

It should be noted that, for other corresponding descriptions of each functional unit related to the head symmetry axis recognition device provided by the embodiment of the present application, reference may be made to corresponding descriptions in the methods of fig. 1 to 3, and no further description is given here.

Based on the above-mentioned methods shown in fig. 1 to 3, correspondingly, the embodiment of the present application further provides a storage medium, on which a computer program is stored, which when executed by a processor, implements the above-mentioned head symmetry axis identification method shown in fig. 1 to 3.

Based on such understanding, the technical solution of the present application may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (may be a CD-ROM, a U-disk, a mobile hard disk, etc.), and includes several instructions for causing a computer device (may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective implementation scenario of the present application.

Based on the method shown in fig. 1 to 3 and the virtual device embodiment shown in fig. 4, in order to achieve the above objective, the embodiment of the present application further provides a computer device, which may specifically be a personal computer, a server, a network device, etc., where the computer device includes a storage medium and a processor; a storage medium storing a computer program; a processor for executing a computer program to implement the head symmetry axis identification method as described above and shown in fig. 1 to 3.

Optionally, the computer device may also include a user interface, a network interface, a camera, radio Frequency (RF) circuitry, sensors, audio circuitry, WI-FI modules, and the like. The user interface may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (e.g., bluetooth interface, WI-FI interface), etc.

It will be appreciated by those skilled in the art that the architecture of a computer device provided in the present embodiment is not limited to the computer device, and may include more or fewer components, or may combine certain components, or may be arranged in different components.

The storage medium may also include an operating system, a network communication module. An operating system is a program that manages and saves computer device hardware and software resources, supporting the execution of information handling programs and other software and/or programs. The network communication module is used for realizing communication among all components in the storage medium and communication with other hardware and software in the entity equipment.

Through the description of the above embodiments, it can be clearly understood by those skilled in the art that the present application may be implemented by means of software and necessary general hardware platform, or may be implemented by hardware to identify a brain tissue edge chain code and a chain code cutting point from a brain tissue region in a target head medical image, and cut the brain tissue edge chain code into a front upper sagittal Dou Lian code and a rear upper sagittal Dou Lian code according to the chain code cutting point, further find a front candidate corner and a rear candidate corner according to curvature values of points on the front upper sagittal Dou Lian code and the rear upper sagittal Dou Lianma code, and screen concave points in the candidate corner as front target corner and rear target corner according to the concave position features of the front candidate corner in the front upper sagittal sinus chain code and the rear candidate corner in the rear upper sagittal sinus chain code, and finally determine a head symmetry axis corresponding to the brain tissue region according to the connecting line of the front target corner and the rear target corner. According to the embodiment of the application, the upper sagittal sinus of the anterior segment and the upper sagittal Dou Lianxian of the posterior segment are utilized as the characteristics of the symmetry axis of the head, the brain tissue edge chain code of the brain tissue region outline which is not easily affected by brain tissue lesions and image quality is selected, the upper sagittal sinus characteristics reflected by the brain tissue edge chain code are utilized to identify the anterior segment target angular point and the posterior segment target angular point, so that the determination of the symmetry axis is realized, the problem of inaccurate symmetry axis identification caused by the influence of the brain tissue lesions or the image quality on the tissue definition is solved, the accuracy and the robustness of the head symmetry axis identification are improved, and the angular point identification is performed on one-dimensional data, namely the chain code, and the identification speed is high and the efficiency is high.

Those skilled in the art will appreciate that the drawing is merely a schematic illustration of a preferred implementation scenario and that the modules or flows in the drawing are not necessarily required to practice the application. Those skilled in the art will appreciate that modules in an apparatus in an implementation scenario may be distributed in an apparatus in an implementation scenario according to an implementation scenario description, or that corresponding changes may be located in one or more apparatuses different from the implementation scenario. The modules of the implementation scenario may be combined into one module, or may be further split into a plurality of sub-modules.

The above-mentioned inventive sequence numbers are merely for description and do not represent advantages or disadvantages of the implementation scenario. The foregoing disclosure is merely illustrative of some embodiments of the application, and the application is not limited thereto, as modifications may be made by those skilled in the art without departing from the scope of the application.

Claims

1. A head symmetry axis identification method, comprising:

Determining a plurality of front segment candidate corner points according to front segment curvature values of each point in the front segment upper sagittal sinus chain code, and determining a plurality of rear segment candidate corner points according to rear segment curvature values of each point in the rear segment upper sagittal sinus chain code, wherein the bending degree of each front segment candidate corner point determined by each point in the front segment upper sagittal sinus chain code at the position is larger than that of the rest point in the front segment upper sagittal sinus chain code, and the bending degree of each rear segment candidate corner point determined by each point in the rear segment upper sagittal sinus chain code at the position is larger than that of the rest point in the rear segment upper sagittal sinus chain code;

Detecting a first position feature of the anterior segment candidate corner in the anterior segment sagittal sinus chain code, determining an anterior segment target corner based on the first position feature, detecting a second position feature of the posterior segment candidate corner in the posterior segment sagittal sinus chain code, and determining a posterior segment target corner based on the second position feature, wherein the first position feature is a concave-convex feature of a position of the anterior segment candidate corner in the anterior segment sagittal Dou Lian code, the anterior segment target corner is a anterior segment candidate corner belonging to a concave point, the second position feature is a concave-convex feature of a position of the posterior segment candidate corner in the posterior segment sagittal Dou Lian code, and the posterior segment target corner is a posterior segment candidate corner belonging to a concave point;

2. The method according to claim 1, wherein the extracting the anterior superior sagittal Dou Lian code and the posterior superior sagittal Dou Lian code of the brain tissue edge chain code according to the positional relationship between the brain tissue edge chain code and the chain code cutting point specifically comprises:

determining the parts above and below the horizontal line corresponding to the chain code cutting point in the brain tissue edge chain code as the front section upper sagittal Dou Lianma code and the rear section upper sagittal Dou Lian code respectively, wherein the chain code cutting point comprises the mass center of the brain tissue region;

3. The method according to claim 1, wherein the determining the anterior segment candidate corner according to the anterior segment curvature value of each point in the anterior segment upper sagittal sinus chain code and the determining the posterior segment candidate corner according to the posterior segment curvature value of each point in the posterior segment upper sagittal sinus chain code specifically comprises:

4. A method according to claim 3, wherein said detecting a first location feature of said anterior segment candidate corner in said anterior segment sagittal sinus chain code and determining an anterior segment target corner based on said first location feature, and detecting a second location feature of said posterior segment candidate corner in said posterior segment sagittal sinus chain code and determining a posterior segment target corner based on said second location feature, in particular comprises:

5. The method of any one of claims 1 to 4, wherein the target head medical image comprises an image of a plurality of faults; the determining a head symmetry axis corresponding to the brain tissue region according to the connecting line of the front section target corner and the rear section target corner specifically comprises the following steps:

6. The method of claim 5, wherein the first target layer comprises a plurality of layers and the second target layer comprises a plurality of layers; before the identifying of the brain tissue edge chain code corresponding to the brain tissue region in the target head medical image and the chain code cutting point, the method further comprises:

7. The method of claim 6, wherein prior to establishing the symmetry axis identification plane, the method further comprises:

8. A head symmetry axis recognition device, comprising:

The candidate corner point determining module is used for determining a plurality of front segment candidate corner points according to front segment curvature values of each point in the front segment sagittal sinus chain code, and determining a plurality of rear segment candidate corner points according to rear segment curvature values of each point in the rear segment sagittal sinus chain code, wherein the bending degree of each front segment candidate corner point determined by each point in the front segment sagittal sinus chain code at the front segment is greater than that of the rest point in the front segment sagittal sinus chain code, and the bending degree of each rear segment candidate corner point determined by each point in the rear segment sagittal sinus chain code at the rear segment is greater than that of the rest point in the rear segment sagittal sinus chain code.

The target corner determining module is used for detecting a first position feature of the anterior segment candidate corner in the anterior segment upper sagittal sinus chain code, determining an anterior segment target corner based on the first position feature, detecting a second position feature of the posterior segment candidate corner in the posterior segment upper sagittal sinus chain code, and determining a posterior segment target corner based on the second position feature, wherein the first position feature is an concave-convex feature of the anterior segment candidate corner at the position of the anterior segment upper sagittal Dou Lian code, the anterior segment target corner is an anterior segment candidate corner belonging to a concave point, the second position feature is an concave-convex feature of the posterior segment candidate corner at the position of the posterior segment upper sagittal Dou Lian code, and the posterior segment target corner is a posterior segment candidate corner belonging to a concave point;

9. A storage medium having stored thereon a computer program, which when executed by a processor, implements the method of any of claims 1 to 7.

10. A computer device comprising a storage medium, a processor and a computer program stored on the storage medium and executable on the processor, characterized in that the processor implements the method of any one of claims 1 to 7 when executing the computer program.