Estimation method of brain gullet depth, electronic equipment and readable storage medium

Abstract

The invention discloses a brain gullet depth estimation method, electronic equipment and a readable storage medium, and belongs to the field of medical image processing. The estimation method comprises the following steps: performing cerebral cortex reconstruction based on magnetic resonance imaging data of a subject to obtain a cerebral cortex polygon mesh; based on the cerebral cortex polygon mesh, a cerebral cortex surface reference surface is obtained through a cerebral cortex boundary surface construction method; and based on the cerebral cortex polygonal mesh and the cerebral cortex surface reference surface, obtaining the distance between each vertex in the cerebral cortex polygonal mesh and the cerebral cortex surface reference surface, wherein the distance is the cerebral furrow depth at the corresponding vertex. According to the method for estimating the cerebral sulcus depth, the electronic equipment and the readable storage medium provided by the invention, the reference plane capable of being used for estimating the sulcus depth is constructed by adopting the Alpha shape method, and the cerebral cortex sulcus depth calculated according to the reference plane is more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical image processing, in particular to a method for estimating sulcal depth, an electronic device and a readable storage medium. BACKGROUND

[0002] The sulcal depth of the cerebral cortex surface is a measure of the complexity of the cerebral cortex surface morphology, which reflects the degree of protrusion or depression of the sulci and gyri of the cerebral cortex surface. There are various methods for calculating the sulcal depth of the cerebral cortex surface, one of which is widely used is the FreeSurfer software package. In FreeSurfer, the calculation result of the sulcal depth of the cerebral cortex surface is saved as lh.sulc and rh.sulc files, and the "sulc" file refers to the sulcal depth file. In the "sulc" file, the sulcal depth value of the cerebral cortex surface, the positive value indicates a deeper sulcus, the negative value indicates a shallower sulcus, and the value close to zero indicates a relatively flat area.

[0003] The sulcal depth of the cerebral cortex surface calculated by FreeSurfer shows a phenomenon of reduced accuracy when approaching the edge of the gyrus, so in the scene where the accuracy of the sulcal depth information is more stringent, it cannot be used or reliable results cannot be obtained after use because it cannot distinguish the sulcal depth finely.

[0004] Therefore, it is necessary to provide a method for estimating the sulcal depth of the cerebral cortex, an electronic device and a readable storage medium which can improve the accuracy of the sulcal depth estimation of the cerebral cortex. SUMMARY

[0005] In order to solve at least one aspect of the above problems and defects in the prior art, the present application provides a method for estimating the sulcal depth of the cerebral cortex, an electronic device and a readable storage medium, which can at least partially improve the accuracy of the sulcal depth estimation of the cerebral cortex. The technical method is as follows:

[0006] According to one aspect of the present application, a method for estimating the sulcal depth of the cerebral cortex is provided. The estimation method comprises the following steps:

[0007] reconstructing the cerebral cortex based on the magnetic resonance imaging data of the subject to obtain a polygonal mesh of the cerebral cortex;

[0008] obtaining a reference surface of the cerebral cortex surface by a cerebral cortex boundary surface construction method based on the polygonal mesh of the cerebral cortex;

[0009] The distance between each vertex in the cerebral cortex polygonal mesh and the cerebral cortex surface reference plane is obtained based on the cerebral cortex polygonal mesh and the cerebral cortex surface reference plane, and the distance is the cerebral cortex surface sulcal depth at the corresponding vertex.

[0010] Specifically, the cerebral cortex surface reference plane is obtained by a cerebral cortex boundary plane construction method based on the cerebral cortex polygonal mesh, comprising:

[0011] The vertex data of each vertex in the cerebral cortex polygonal mesh is input into the Alpha Shape method, and the Alpha Shape method is fitted to the vertex to generate the cerebral cortex surface reference plane.

[0012] Further, the cerebral cortex surface reference plane is located close to the contour of the cerebral cortex polygonal mesh.

[0013] Further, the cerebral cortex surface reference plane wraps all vertices in the cerebral cortex polygonal mesh and is a closed surface.

[0014] Specifically, the vertex data of each vertex in the cerebral cortex polygonal mesh is input into the Alpha Shape method, and the Alpha Shape method is fitted to the vertex to generate the cerebral cortex surface reference plane, comprising:

[0015] The outer surface of the cerebral cortex polygonal mesh is polyhedronally divided to form a plurality of polyhedron units on the outer surface of the cerebral cortex polygonal mesh;

[0016] Based on the vertex data of each polyhedron unit in the plurality of polyhedron units, the circumscribed circle of each polyhedron unit and the corresponding circumscribed circle radius are obtained;

[0017] The circumscribed circle radius of each polyhedron unit is compared with the radius threshold value respectively to screen the polygons on the outer surface of the cerebral cortex polygonal mesh;

[0018] The polygons in the screened cerebral cortex polygonal mesh are counted to retain the polygons belonging to the outer surface of the cerebral cortex polygonal mesh, and the curved surface formed by connecting the triangles belonging to the outer surface of the cerebral cortex polygonal mesh is the cerebral cortex surface reference plane.

[0019] Specifically, the circumscribed circle radius of each polyhedron unit is compared with the radius threshold value respectively, and when the circumscribed circle radius of the polyhedron unit is less than the radius threshold value, all edges and faces in the polyhedron unit are retained;

[0020] When the circumscribed circle radius of the polyhedron unit is greater than the radius threshold value, all edges and faces in the polyhedron unit are deleted.

[0021] Specifically, the polygons in the cerebral cortex polygonal mesh after screening are counted to retain the polygons belonging to the outer surface of the cerebral cortex, including:

[0022] The polygons in the cerebral cortex polygonal mesh are counted to determine the polygons located on the outer surface of the cerebral cortex and retained.

[0023] Preferably, in the Alpha shape method, the radius threshold is 10-30 mm.

[0024] Specifically, the vertex data at least includes vertex position.

[0025] Further, the cerebral cortex surface reference surface is obtained by the cerebral cortex boundary surface construction method based on the cerebral cortex polygonal mesh, and further includes:

[0026] The cerebral cortex polygonal mesh is smoothed before inputting the cerebral cortex boundary surface construction method, and the vertex data of each vertex in the smoothed cerebral cortex polygonal mesh is input into the Alpha shape method for fitting.

[0027] Specifically, the smoothing method is Laplace smoothing, and the relaxation coefficient is set to 0.03-0.05.

[0028] Specifically, the cerebral cortex is reconstructed based on the magnetic resonance imaging data of the subject to obtain the cerebral cortex polygonal mesh, including:

[0029] The magnetic resonance imaging data of the subject is preprocessed;

[0030] The preprocessed magnetic resonance imaging data is subjected to cerebral cortex surface reconstruction by the marching cubes method to obtain the cerebral cortex polygonal mesh.

[0031] Further, the cerebral cortex is the pia mater layer,

[0032] The cerebral cortex polygonal mesh is a cerebral cortex surface triangular mesh or a cerebral cortex surface quadrilateral mesh.

[0033] According to another aspect of the present application, an electronic device is provided. Wherein the electronic device includes a memory and at least one processor, the memory is in communication connection with the at least one processor, the memory stores programs or instructions, and the programs or instructions are executed by the at least one processor, and the electronic device is used for the above-mentioned any one of the cerebral sulcus depth estimation method.

[0034] According to still another aspect of the present application, a readable storage medium is provided. Wherein, the readable storage medium stores programs or instructions, which are executed by a processor to perform the method for estimating sulcal depth of brain as any one of the above.

[0035] The method for estimating sulcal depth of brain, the electronic device and the readable storage medium according to the embodiments of the present application have at least one of the following advantages:

[0036] (1) The method for estimating sulcal depth of brain, the electronic device and the readable storage medium according to the present application construct a reference plane for sulcal depth estimation by using the Alpha shape method, and the sulcal depth of the cerebral cortex calculated according to the reference plane is more accurate, especially the sulcal depth at the edge region of the gyrus is more accurate;

[0037] (2) The method for estimating sulcal depth of brain, the electronic device and the readable storage medium according to the present application use the three-dimensional mesh file representing the pia mater layer to estimate the sulcal depth of the brain, which is closer to the real sulcal depth of the brain;

[0038] (3) The method for estimating sulcal depth of brain, the electronic device and the readable storage medium according to the present application estimate the sulcal depth of the cerebral cortex surface by calculating the distance between each vertex of the pia mater layer polygon mesh and the reference surface of the cerebral cortex surface, which avoids the computational complexity of calculating the cumulative of the normal direction in the existing calculation, reduces the computer overhead, and improves the image processing speed. BRIEF DESCRIPTION OF DRAWINGS

[0039] These and / or other aspects and advantages of the present application will become apparent and readily understood from the following description, including the appended drawings, in which:

[0040] Figure 1 is a flow chart of the method for estimating sulcal depth of brain according to one embodiment of the present application;

[0041] Figure 2 is a flow chart of the Alpha shape method shown in Figure 1

[0042] Figure 3 is a schematic diagram of the ordered vertex point set extracted when the cerebral cortex polygon mesh is a cerebral cortex triangle mesh shown in Figure 2

[0043] Figure 4A is a schematic diagram of the cerebral cortex triangle mesh after tetrahedral subdivision shown in Figure 3

[0044] Figure 4B Figure 4A ​​​​the cross-sectional view along the A-A line shown in the figure;

[0045] Figure 4C is Figure 4A the schematic diagram of the cerebral cortex triangular mesh after screening by tetrahedral unit shown in the figure;

[0046] Figure 4D is Figure 4C the cross-sectional view along the B-B line shown in the figure;

[0047] Figure 5A is the visualization diagram of the sulcal depth estimated by using the FreeSurfer software;

[0048] Figure 5B is the visualization diagram of the sulcal depth obtained according to the estimation method shown in the figure. Figure 1 DETAILED DESCRIPTION

[0049] The technical solutions of the present application will be further specifically described below by examples in combination with the accompanying drawings. In the description, the same or similar reference numerals indicate the same or similar components. The following description of the embodiments of the present application with reference to the accompanying drawings is intended to explain the overall inventive concept of the present application, and should not be understood as a limitation of the present application.

[0050] The term "data" used herein should be understood broadly as continuous values such as sound, image, etc., referred to as analog data; or discrete, such as symbols, characters, etc., referred to as digital data.

[0051] Referring to Figure 1 , the flow of the sulcal depth estimation method of the brain according to one embodiment of the present application is shown. The estimation method comprises the following steps:

[0052] reconstructing the cerebral cortex based on the magnetic resonance imaging data of the subject to obtain a cerebral cortex polygonal mesh;

[0053] obtaining a cerebral cortex surface reference surface by a cerebral cortex boundary surface construction method based on the cerebral cortex polygonal mesh;

[0054] obtaining the distance between each vertex in the cerebral cortex polygonal mesh and the cerebral cortex surface reference surface based on the cerebral cortex polygonal mesh and the cerebral cortex surface reference surface, the distance being the sulcal depth of the cerebral cortex surface at the corresponding vertex (i.e. the sulcal depth of the brain).

[0055] ​In one example, the cerebral cortex is a pia mater layer, and the cerebral cortex polygonal mesh is a pia mater layer polygonal mesh. In the prior art, the interface between the white matter and the gray matter is usually used as the input for the reconstruction of the cerebral cortex, and because the interface between the white matter and the gray matter is more inward than the pia mater layer, the sulcal depth calculated based on the interface is less accurate than the calculation result based on the pia mater layer.

[0056] In one example, the magnetic resonance imaging data includes at least one of a T1 weighted imaging (T1WI) image of magnetic resonance, a T2 weighted imaging (T2WI) image of magnetic resonance, image data (e.g., image digital data, content related to the image expressed in a computer language) of magnetic resonance imaging, etc.

[0057] In one example, the cerebral cortex polygonal mesh is a cerebral cortex surface triangle mesh, a cerebral cortex surface quadrilateral mesh, or other convex polygonal mesh. The cerebral cortex polygonal mesh is a three-dimensional polygonal mesh of the cerebral cortex, such as a cerebral cortex surface triangle mesh, a cerebral cortex surface quadrilateral mesh, or other convex polygonal mesh, which are all three-dimensional meshes.

[0058] In one example, the cerebral cortex boundary surface construction method is an Alpha-Shape algorithm, and of course, other existing boundary surface construction methods can also be selected by those skilled in the art to construct the reference plane. The present example is only an illustrative example, and those skilled in the art should not understand it as a limitation on the present application.

[0059] In one example, the cerebral cortex reconstruction based on the magnetic resonance imaging data of the subject is to obtain a cerebral cortex polygonal mesh, including:

[0060] Pretreating the magnetic resonance imaging data of the subject;

[0061] Reconstructing the cerebral cortex surface of the pretreated magnetic resonance imaging data by a marching cubes method to obtain a cerebral cortex polygonal mesh.

[0062] In one example, when the magnetic resonance imaging data is a magnetic resonance imaging image, the T1WI (T1 weighted imaging) image has the characteristics of high contrast and small deformation, so that the image is easy to segment and register, and therefore the magnetic resonance imaging image can be preferably a T1WI image.

[0063] In one example, the magnetic resonance imaging image can be preferably a T2 weighted imaging (T2WI) image, since the T2WI image can highlight different brain tissues.

[0064] In one example, the pre-processing can include sampling the magnetic resonance imaging image and transforming the image intensity values to the interval of 0-255.

[0065] For example, the magnetic resonance imaging image can be denoised by a filter, and different methods can be taken to remove artifacts according to different artifact types, such as compensating methods or changing phase and frequency encoding. After denoising and removing artifacts, the three-dimensional size of the image can be normalized or resampled to, for example, 256x256x256, and the spatial resolution of the head MRI image can also be normalized or resampled to, for example, 1mmx1mmx1mm (i.e., the size of each voxel is 1mmx1mmx1mm). After normalizing or resampling the spatial resolution of the head MRI image, the image intensity also needs to be gray normalized (or voxel intensity normalized), that is, the gray value of each voxel is normalized (or voxel intensity normalized) to the range of 0-255. The head MRI image processed by size normalization and gray normalization can adjust the three-dimensional size, spatial resolution and gray range of the original head T1 weighted image to a predetermined size, so as to reduce the complexity of the subsequent processing steps, thereby improving the robustness of the entire processing flow, while reducing or even eliminating the influence of magnetic susceptibility artifacts and RF field inhomogeneity on the image, thereby improving the clarity and accuracy of the image processed in the subsequent steps.

[0066] In one example, the pre-processing can also be implemented by using the mri_convert command in the FreeSurfer software package.

[0067] In one example, the pre-processed magnetic resonance imaging image can be reconstructed by a moving cuber method to obtain a brain cortex polygonal mesh. Of course, the brain cortex surface reconstruction can also be performed by, for example, FreeSurfer, FastCSR, DeepCSR, etc., to obtain a brain cortex polygonal mesh.

[0068] For example, the brain cortex surface reconstruction is performed using FreeSurfer, and the specific command is reconall -s $SUBJECT -all, where $SUBJECT is the subject identifier, and -all is a processing parameter. After the magnetic resonance imaging image is processed by FreeSurfer, the processing result is obtained, and the processing result includes $SUBJECT / surf / lh.pial, which is the dura mater layer (for example, the interface between the brain cortex gray matter and the cerebrospinal fluid) triangular mesh reconstruction result of the left brain, and $SUBJECT / surf / rh.pial, which is the dura mater layer triangular mesh reconstruction result of the right brain.

[0069] In one example, in order to reduce the adverse effects of abnormal spikes or noise that may exist in the image on subsequent calculations, the brain cortex polygonal mesh can be smoothed before inputting the brain cortex boundary surface construction method (for example, the Alpha shape method), and then the vertex data of each vertex in the smoothed brain cortex polygonal mesh is input into the Alpha shape method for fitting.

[0070] In one example, the brain cortex polygonal mesh can be processed by a method such as Gaussian smoothing or local neighborhood smoothing, which helps to improve the continuity and consistency of the data, so that the subsequent Alpha-Shape calculation is more stable and accurate.

[0071] In one example, Laplace smoothing can be used. For example, the number of smoothing iterations can be set to 100, and the value range of the relaxation coefficient is set to 0.03-0.05, and preferably, the relaxation coefficient is set to 0.04. The present example is only an illustrative example, and those skilled in the art should not understand it as a limitation of the present application.

[0072] In one example, the brain cortex surface reference surface is obtained based on the brain cortex polygonal mesh by the Alpha shape method, including:

[0073] The vertex data of each vertex in the brain cortex polygonal mesh is input into the Alpha shape method (i.e., the Alpha-Shape algorithm) to fit all the vertices in the brain cortex polygonal mesh to generate the brain cortex surface reference surface.

[0074] In one example, the brain cortex surface reference surface is located close to the contour of the brain cortex polygonal mesh. Alternatively, the brain cortex surface is wrapped around the contour of the brain cortex polygonal mesh, so that all the vertices (or nodes) on the brain cortex polygonal mesh are covered by a closed curved surface.

[0075] In one example, the vertex data includes at least vertex position. Of course, the vertex data can also include normal vector at the vertex, image gray value, texture of the image at the vertex, number of polygon faces, number of vertices, vertex ordering, etc. The present example is merely an illustrative example, and those skilled in the art should not understand it as a limitation to the present application.

[0076] In one example, the vertex normal vector is used to represent the direction of each vertex in the cerebral cortex polygon mesh.

[0077] In one example, as shown in Figure 2 the vertex data of each vertex in the cerebral cortex polygon mesh is input into the Alpha Shape method to fit each vertex to generate the cerebral cortex surface reference surface, including:

[0078] The cerebral cortex polygon mesh is polyhedronally divided to form a plurality of polyhedral units on the cerebral cortex polygon mesh;

[0079] Based on the vertex data of each polyhedral unit in the plurality of polyhedral units, the circumscribed circle of each polyhedral unit and the corresponding circumscribed circle radius are obtained;

[0080] The circumscribed circle radius of all polyhedral units is compared with the radius threshold value respectively to perform cerebral cortex polygon mesh screening;

[0081] The polygons in the screened cerebral cortex polygon mesh are counted to retain the polygons belonging to the outer surface of the cerebral cortex, and the curved surface connected by the polygons belonging to the outer surface of the cerebral cortex is the cerebral cortex surface reference surface.

[0082] In one example, the circumscribed circle radius of all polyhedral units is compared with the radius threshold value respectively:

[0083] When the circumscribed circle radius of the polyhedral unit is less than the radius threshold value, all edges and faces in the polyhedral unit are retained;

[0084] When the circumscribed circle radius of the polyhedral unit is greater than the radius threshold value, all edges and faces in the polyhedral unit are deleted.

[0085] In one example, in the Alpha Shape method, the value range of the radius threshold value is 10-30 mm. Preferably, the radius threshold value is 20 mm.

[0086] In one example, the polygons in the screened cerebral cortex polygon mesh are counted to retain the polygons belonging to the outer surface of the cerebral cortex, including:

[0087] The polyhedrons to which the polygons in the cerebral cortex polygonal mesh belong are counted, and the polygons located in the same polyhedron are determined to be located on the outer surface of the cerebral cortex and are retained.

[0088] In one example, the cerebral cortex polygonal mesh is a cerebral cortex triangular mesh, which is composed of multiple triangular facets. Of course, those skilled in the art will understand that when the cerebral cortex polygonal mesh is a cerebral cortex quadrilateral mesh, it is typically composed of multiple quadrilateral facets, and similarly, it can be composed of other polygonal facets. Each facet has vertices and multiple edges, which can also be called nodes. In a triangular facet, three ordered vertices and three ordered edges constitute a triangular facet.

[0089] For example, the first step is to extract the vertices of the cerebral cortex triangle mesh to obtain an ordered vertex point set, such as Figure 3 As shown. The two cerebral cortex reconstruction files $SUBJECT / surf / lh.pial and $SUBJECT / surf / rh.pial stored in the python package NiBabel can be read and parsed using the nib.freesurfer.read_geometry method to extract the vertices of the cerebral cortex triangle mesh to obtain an ordered vertex point set. All vertex data (i.e., ordered vertex point sets) are used as input to the Alpha shape method. That is, the faces in the reconstructed cerebral cortex triangle mesh (i.e., faces containing image texture features and color features) have been removed from the image, leaving only vertex information (such as vertex position, vertex sorting, vertex normal vector, etc.). As shown Figure 3 As shown, the edges and faces in the cerebral cortex triangular mesh are now composed of individual vertices in an ordered vertex set, and no longer have the original texture features and color features. Of course, those skilled in the art will understand that when the cerebral cortex polygonal mesh is a polygonal mesh other than a triangular mesh, the same principles or methods can be used to extract the ordered vertex set. This example is merely illustrative and should not be construed by those skilled in the art as limiting the present invention.

[0090] Step 2: The brain polygon mesh formed by the ordered vertex point set is converted into a tetrahedral cerebral cortex polygon mesh with a surface composed of multiple tetrahedral units using Delaunay tetrahedron (that is, multiple tetrahedral units are formed on the outer surface of the cerebral cortex triangle mesh, such as Figure 4A (as shown in the figure), while also capturing the proximity relationships between vertices. For example, you can use the scipy.spatial.Delaunay method in the Python package scipy or the qdelaunay command in the Qhull package to process this.

[0091] In Figure 4A The surface of the brain cortex is covered with triangles of various sizes, which are the outer sides of the tetrahedrons at the outer surface. The color of the triangles is related to the normal direction of the plane in which the triangle is located. Figure 4B The cross-sectional view is obtained by cutting along the A-A tangent in Figure 4A The cross-sectional view is obtained by cutting along the A-A tangent in The surface of the brain cortex is covered with triangles of various sizes, which are the outer sides of the tetrahedrons at the outer surface. The color of the triangles is related to the normal direction of the plane in which the triangle is located.

[0092] Although the brain cortex triangular mesh satisfies the convex hull property through Delaunay tetrahedral partitioning, the concave part of the brain cortex, such as the front part of the lateral fissure, is lifted up and cannot well reflect the reference plane of the sulcus and gyrus. Therefore, in order to obtain the ideal sulcus and gyrus depth reference plane, the brain cortex mesh is further adjusted on the basis of the Delaunay tetrahedral partitioning, as follows:

[0093] Step 3: Filter the triangles (i.e., all edges and faces) of the cerebral cortex polygonal mesh with multiple tetrahedral units (i.e., the tetrahedralized cerebral cortex polygonal mesh). In three-dimensional space, a tetrahedron has a unique circumscribed circle. Traverse all tetrahedral units in the tetrahedralized cerebral cortex polygonal mesh, solve the circumscribed circle radius of each tetrahedral unit, and compare each circumscribed circle radius with the radius threshold (for example, the predefined parameter alpha value, which can be set to 20mm). If the circumscribed circle radius is less than the alpha value, retain all triangles in the tetrahedral unit (i.e., all edges and faces composed of vertices). Otherwise, delete the triangles in the tetrahedral unit (i.e., all edges and faces composed of vertices). Triangle filtering will delete larger tetrahedral units and retain smaller tetrahedral units, so that the outer surface of the modified cerebral cortex triangular mesh shows more detailed structures, thus achieving tetrahedral unit filtering. Those skilled in the art will appreciate that the same principles and methods can be used to process polygonal meshes of other types of cerebral cortex when performing polygon screening on polyhedral units (e.g., tetrahedrons, pentahedrons, etc.). This example is merely illustrative and should not be construed by those skilled in the art as limiting the present invention.

[0094] Step 4: Extract the outer boundary surface of the modified cerebral cortex polygonal mesh

[0095] The outer surface of the modified cerebral cortex polygonal mesh has a more refined three-dimensional mesh, and its surface infrastructure still has tetrahedral units with convex hull properties. Its interior still contains a large number of internal structures represented by three-dimensional triangle meshes (that is, it is still a polyhedron with internal solid structure). Since only the outer surface is needed as a reference plane for calculating sulcus depth, the internal structure is not required. Therefore, it is necessary to count and filter the triangles in the modified cerebral cortex polygonal mesh.

[0096] Since only the surface of the cerebral cortex triangular mesh is tetrahedralized in the tetrahedron partitioning, and the interior is not tetrahedralized, by counting the number of tetrahedrons to which each triangle belongs, it is possible to determine whether the triangle is located on the outer surface, that is, a triangle located on the outer surface will only belong to one tetrahedron. Through this statistical screening, only the triangular meshes located on the outer surface are retained, and the surface formed by connecting these retained triangular meshes is the reference surface of the cerebral cortex surface (such as Figure 4C As shown in FIG, the outer boundary surface of the modified cerebral cortex polygonal mesh is extracted. Figure 4D For the Figure 4CThe cross-sectional view obtained by cutting the B-B tangent line in the above figure can be seen that the complex internal structure is removed, and only the outer surface is reserved. At this point, the construction of the cerebral cortex surface reference plane for calculating the sulcal depth is completed, and the visualization image is referred to Figure 5B .

[0097] In one example, the cerebral cortex surface reference plane can be obtained as follows: (1) select an optional point on the outer surface of the reconstructed cerebral cortex polygonal mesh as an initial point. Take a search radius of 2 times the radius threshold (for example, 40 mm) as the search radius, search all the vertices within the search radius, and form an ordered vertex point set. (2) Take a new point from the ordered vertex point set, and form a circle with the initial point, that is, the new point and the initial point are located on the circumference, and obtain the center of the circle. Traverse the distance from all vertices in the ordered vertex point set to the center of the circle, and when the distance is greater than the radius threshold (for example, 20 mm), define the initial point and the new point as contour points. Traverse all the ordered vertex points to obtain all the contour points. When the distance is less than the radius threshold, repeat steps (1)-(2) with the next point in the ordered vertex point set until all points are traversed and determined. Connect the ordered vertices that are contour points to form a curved surface, which is the cerebral cortex surface reference plane.

[0098] Those skilled in the art can understand that the above two examples can both obtain the cerebral cortex surface reference plane, and those skilled in the art can select according to actual conditions. The present example is only an illustrative example, and those skilled in the art should not understand it as a limitation of the present application.

[0099] After the cerebral cortex surface reference plane is constructed, the distance from each vertex in the cerebral cortex surface polygonal mesh to the cerebral cortex reference plane is calculated point by point. The distance is the distance between the vertex and the cerebral cortex surface reference plane, and the number of distances corresponds to the number of vertices. The distance reflects the sulcal depth at the vertex.

[0100] In one example, when the distance from each vertex in the cerebral cortex surface polygonal mesh to the cerebral cortex reference plane is calculated point by point, the vertex data is the data before smoothing the cerebral cortex surface polygonal mesh, that is, the vertex data when the cerebral cortex surface polygonal mesh is reconstructed by the cerebral cortex.

[0101] In one example, the cerebral cortex mesh is heavily smoothed before inputting the alpha shape method to counter the influence of possible errors (e.g. obviously abnormal small protrusions) on the calculation results to resist the cortex reconstruction mesh, so as to maintain the stability of the calculation results. Due to such processing, a closed spherical and relatively smooth reference plane is formed on the cerebral cortex surface polygon mesh, or a closed spherical and relatively smooth reference plane is enclosed outside the cerebral cortex surface polygon mesh. At this time, the vertex data located in the smooth reference plane is assigned a positive value, and the vertex data located outside the smooth reference plane (for example, the position still sharply protruding after smoothing) is assigned a negative value. Through the addition of positive and negative values here, the cerebral cortex surface sulcus depth obtained by the above method has positive and negative meanings.

[0102] In one example, the positive and negative values of the cerebral cortex surface sulcus depth can also be determined by the following method: the cerebral cortex reference plane is a closed hollow sphere, when the vertex (i.e. vertex data before smoothing) in the cerebral cortex surface polygon mesh is located inside the hollow sphere, the distance from the vertex to the cerebral cortex reference plane obtained by the vertex is set to a negative value. When the vertex (i.e. vertex data before smoothing) in the cerebral cortex surface polygon mesh is located outside the hollow sphere, the distance from the vertex to the cerebral cortex reference plane obtained by the vertex is set to a positive value.

[0103] In one example, as Figure 5A a visualization diagram of the sulcus depth estimated by the method according to the prior art using FreeSurfer is shown, Figure 5B a visualization diagram of the sulcus depth estimated by the method according to one embodiment of the present application is shown.

[0104] From Figure 5A and Figure 5B it can be seen that for the same sulcus depth value, Figure 5B the consistency of the sulcus depth in the precentral gyrus is better, and the sulcus depth at the edge of the gyrus is also better. Taking the middle part of the precentral gyrus as a reference, the sulcus depth calculated by FreeSurfer is obviously deeper than the sulcus depth calculated by the method according to the present application (as shown by the positions circled in Figure 5A the middle frontal gyrus region of the sulcus depth calculated by the method according to the present application is obviously deeper (as shown by the positions circled in Figure 5A ), and at the posterior part of the middle frontal gyrus, the sulcus depth calculated by FreeSurfer changes more slowly than the actual sulcus depth at the transition from the gyrus to the sulcus. In Figure 5B , the sulcus depth at the two positions is similar, and at the posterior part of the middle frontal gyrus, the sulcus depth changes in accordance with the actual sulcus depth at the transition from the gyrus to the sulcus (as shown by the positions circled in Figure 5B ).

[0105] The above differences are caused by the following aspects:

[0106] (1) The used brain cortex reconstruction files are different. The FreeSurfer method uses lh.white file as input, which represents the interface between white matter and gray matter. The present application uses lh.pial file, which represents the interface between gray matter and cerebrospinal fluid (i.e. the pia mater layer);

[0107] (2) The sulcal depth calculation methods are different. The sulcal depth calculation of the FreeSurfer method is obtained from the cumulative deformation distance of the normal direction in the process of smoothing the cortex to the inflated cortex. In the specific calculation process, an explicit reference plane is not used, but is calculated by iterative accumulation. The present application uses the AlphaShape method to construct an explicit reference plane (i.e. the brain cortex surface reference plane), and calculates the sulcal depth based on the reference plane. The physical meaning is more clear and accurate, and the estimation has a reference standard, so that the estimation result is more accurate, and it is easy to check if there is an abnormal situation.

[0108] In one example, the estimated sulcal depth of the brain cortex surface obtained by the method provided by the present application can be used in the following scenarios:

[0109] (1) Image registration: The sulcal depth of the brain cortex surface can be used as a reference feature for registration. By comparing the sulcal depth maps of the brain cortex surface of different individuals, the brain cortex surfaces of different individuals can be aligned, thereby realizing the registration between individuals, and the accuracy and efficiency of multi-modal image registration, image registration at different time points, and statistical analysis in population studies can be improved;

[0110] (2) Brain region localization and target planning: The sulcal depth of the brain cortex surface can help locate specific brain regions and anatomical structures. By analyzing a large amount of surface sulcal depth data of individuals, a brain region template can be established to automatically or semi-automatically identify and locate specific brain regions in individual brain images. Thus, the accuracy and efficiency of target planning for electroencephalographic stimulation, magnetoencephalographic stimulation, deep brain stimulation and other treatment methods can be improved;

[0111] (3) Brain image analysis and visualization: The sulcal depth of the brain cortex surface can be combined with other brain image data to provide detailed description and visualization of brain structure. By combining surface sulcal depth information with structural features such as functional connectivity and brain network topology, the relationship between brain structure and function can be better understood, and more accurate anatomical markers can be provided in brain image analysis and visualization.

[0112] In one example, an electronic device is provided according to another embodiment of the present application. The electronic device (not shown) includes a processor (not shown) and a memory (not shown). The memory has stored thereon a program which, when executed by the processor, implements the method of estimating the depth of a sulcus of the brain according to any of the examples described above.

[0113] In one example, the processor can be a microprocessor, such as a general purpose processor, e.g., a graphics processing unit (GPU), a central processing unit (CPU), a digital signal processor (DSP), etc. In one example, the processor can also be a microprocessor core implemented by hardware circuitry, such as a microprocessor core implemented by reconfigurable logic in a hardware logic component, e.g., a field programmable gate array (FPGA), a complex programmable logic device (CPLD), an application specific integrated circuit (ASIC), an application specific standard product (ASSP), a system on a chip (SOC), etc.

[0114] In one example, the processor can also be a virtual processor, which can be a virtual processor with Intel x86 processor features, or a virtual processor with PowerPC processor features. Preferably, the processor is a graphics processing unit. In one example, the processor can be a single core processor, or a multi-core processor.

[0115] In one example, the memory includes a volatile memory (i.e., random access memory) and a non-volatile memory. The volatile memory includes a main memory, a cache, etc., and the non-volatile memory includes a secondary memory, etc. In one example, the memory can be configured as a remote memory, which can be connected to the processor via a network (wired or wireless). The network includes, but is not limited to, a wide area network, a local area network, a metropolitan area network, a personal area network, the Internet, a satellite communication network, and any combination thereof.

[0116] In one example, the processor executes a program obtained from the memory to create a corresponding task thread and execute the thread. In one example, the processor obtains a program from the external memory based on a read instruction in the memory to create a corresponding task thread and execute the thread. The program is used to implement the control method for tracking a target object.

[0117] Although the subject matter described herein is provided in the general context of computer-executable instructions of a program module being executed by a computer system on a computer-readable medium, those skilled in the art will recognize that the subject matter described herein also can be implemented in combination with other program modules. Generally, program modules include routines, programs, components, data structures, and other types of structures that perform particular tasks or implement particular abstract data types. Those skilled in the art will appreciate that the method steps described in any one of the examples herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. The implementation depends on the particular application and the

[0118] When the method steps are implemented in software functions and sold or used as an independent product, they can be stored in a computer-readable storage medium. Therefore, the technical solution of the present application or the part of the original technology that makes an essential contribution can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various examples of the present application.

[0119] In one example, a readable storage medium according to still another embodiment of the present application is provided. The "readable storage medium" of the embodiments of the present application refers to any medium involved in providing a program or instructions to a processor for execution. The medium can take various forms, including but not limited to non-volatile medium, volatile medium, and transmission medium. Non-volatile medium includes, for example, optical disks or magnetic disks, such as storage devices. Volatile medium includes dynamic memory, such as main memory. Transmission medium includes coaxial cables, copper wires, and optical fibers, including wires comprising buses. Transmission medium can also take the form of acoustic or light waves, such as acoustic waves or light waves generated during radio frequency (RF) and infrared (IR) data communications. Common forms of readable storage medium include, for example, floppy disks, flexible disks, hard disks, magnetic tapes, any other magnetic medium, CD-ROMs, DVDs, any other optical medium, punch cards, paper tapes, any other physical medium with patterns of holes, RAM, PROM, and EPROM, FLASH-EPROM, any other memory chip or cartridge, a carrier wave as described below, or any other medium from which a computer can read.

[0120] The readable storage medium stores programs or instructions, which are executed by the processor to perform the above-mentioned sulcus depth estimation method.

[0121] The method for estimating the depth of cerebral sulcus, the electronic device and the readable storage medium according to the embodiments of the present application have at least one of the following advantages:

[0122] (1) The method for estimating the depth of cerebral sulcus, the electronic device and the readable storage medium provided by the present application construct a reference plane that can be used for sulcus depth estimation by using the Alpha shape method, and the sulcus depth of the cerebral cortex calculated according to the reference plane is more accurate, especially the sulcus depth at the edge region of the cerebral sulcus is more accurate;

[0123] (2) The method for estimating the depth of cerebral sulcus, the electronic device and the readable storage medium provided by the present application estimate the sulcus depth by using a three-dimensional mesh file representing the pia mater layer, which is closer to the real sulcus depth;

[0124] (3) The method for estimating the depth of cerebral sulcus, the electronic device and the readable storage medium provided by the present application estimate the sulcus depth of the cerebral cortex surface by calculating the distance between each vertex of the pia mater layer polygon mesh and the reference surface of the cerebral cortex surface, which avoids the computational complexity of calculating the cumulative of the normal direction in the existing calculation, reduces the computer overhead, and improves the image processing speed.

[0125] Although some embodiments of the present general inventive concept have been shown and described, it would be understood by those of ordinary skill in the art that changes might be made therein without departing from the principles and spirit of the present general inventive concept, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for estimating sulcal depth of a brain, the method comprising the steps of: reconstructing a cerebral cortex based on magnetic resonance imaging data of a subject to obtain a cerebral cortex polygonal mesh; constructing a cerebral cortex surface reference based on the cerebral cortex polygonal mesh by a cerebral cortex boundary surface construction method; obtaining distances between each vertex in the cerebral cortex polygonal mesh and the cerebral cortex surface reference based on the cerebral cortex polygonal mesh and the cerebral cortex surface reference, the distances being sulcal depth of the cerebral cortex surface at the corresponding vertices.

2. The method of claim 1, wherein, the constructing of the cerebral cortex surface reference based on the cerebral cortex polygonal mesh by the cerebral cortex boundary surface construction method comprises: inputting vertex data of each vertex in the cerebral cortex polygonal mesh into an Alpha shape method to fit the each vertex to generate the cerebral cortex surface reference.

3. The method of claim 2, wherein, the cerebral cortex surface reference is located close to the contour of the cerebral cortex polygonal mesh.

4. The method of claim 3, wherein, the cerebral cortex surface reference wraps all vertices in the cerebral cortex polygonal mesh and is a closed surface.

5. The method of any one of claims 2-4, wherein, the inputting of the vertex data of each vertex in the cerebral cortex polygonal mesh into the Alpha shape method to fit the each vertex to generate the cerebral cortex surface reference comprises: polyhedral partitioning an outer surface of the cerebral cortex polygonal mesh to form a plurality of polyhedral elements on the outer surface of the cerebral cortex polygonal mesh; obtaining an inscribed circle and a corresponding inscribed circle radius of each polyhedral element in the plurality of polyhedral elements based on vertex data of the each polyhedral element; comparing the inscribed circle radius of each polyhedral element with a radius threshold to screen polygons on the outer surface of the cerebral cortex polygonal mesh; counting the polygons in the screened cerebral cortex polygonal mesh to retain polygons belonging to the outer surface of the cerebral cortex polygonal mesh, the curved surface formed by connecting the triangles belonging to the outer surface of the cerebral cortex polygonal mesh being the cerebral cortex surface reference.

6. The method of claim 5, wherein, the comparing of the inscribed circle radius of each polyhedral element with the radius threshold, when the inscribed circle radius of a polyhedral element is less than the radius threshold, retains all edges and faces in the polyhedral element; when the inscribed circle radius of a polyhedral element is greater than the radius threshold, deletes all edges and faces in the polyhedral element.

7. The method of claim 5, wherein, the counting of the polygons in the screened cerebral cortex polygonal mesh to retain polygons belonging to the outer surface of the cerebral cortex comprises: counting the polyhedrons to which each polygon in the cerebral cortex polygonal mesh belongs, and determining polygons located in the same polyhedron as being located on the outer surface of the cerebral cortex and retaining.

8. The method of claim 5, wherein, In the Alpha shape method, the radius threshold is in the range of 10-30 mm.

9. The estimation method of any one of claims 2-4, wherein, The vertex data at least includes vertex position.

10. The estimation method of any one of claims 1-4, wherein, The cerebral cortex surface reference surface is obtained by a cerebral cortex boundary surface construction method based on the cerebral cortex polygonal mesh, and further comprising: Before inputting the cerebral cortex boundary surface construction method, the cerebral cortex polygonal mesh is smoothed, and the vertex data of each vertex in the smoothed cerebral cortex polygonal mesh is input into the Alpha shape method for fitting.

11. The estimation method of claim 10, wherein, The smoothing method is Laplace smoothing, and the relaxation coefficient is in the range of 0.03-0.

05.

12. The estimation method of any one of claims 1-4, wherein, The cerebral cortex is reconstructed based on the magnetic resonance imaging data of the subject to obtain the cerebral cortex polygonal mesh, comprising: Pretreating the magnetic resonance imaging data of the subject; The pretreated magnetic resonance imaging data is reconstructed by a moving cube method to obtain the cerebral cortex polygonal mesh.

13. The estimation method of claim 1, wherein, The cerebral cortex is the pia mater layer, The cerebral cortex polygonal mesh is a cerebral cortex surface triangular mesh or a cerebral cortex surface quadrilateral mesh.

14. An electronic device, characterized in that, The electronic device comprises a memory and at least one processor, the memory is in communication connection with the at least one processor, the memory stores programs or instructions, and the programs or instructions are executed by the at least one processor, and the electronic device is used to implement the estimation method of the depth of the cerebral sulcus as claimed in any one of claims 1-13.

15. A readable storage medium, characterized in that, The readable storage medium stores programs or instructions, and the programs or instructions are executed by a processor to execute the estimation method of the depth of the cerebral sulcus as claimed in any one of claims 1-13.

Images