CN111127492B - Method, apparatus, and computer-readable storage medium for brain image registration - Google Patents
Method, apparatus, and computer-readable storage medium for brain image registration Download PDFInfo
- Publication number
- CN111127492B CN111127492B CN202010226114.3A CN202010226114A CN111127492B CN 111127492 B CN111127492 B CN 111127492B CN 202010226114 A CN202010226114 A CN 202010226114A CN 111127492 B CN111127492 B CN 111127492B
- Authority
- CN
- China
- Prior art keywords
- hyperbolic
- target
- brain image
- curved surface
- registered
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30016—Brain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30204—Marker
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention provides a method, a device and a computer readable storage medium for realizing brain image registration through a computer, wherein the brain image comprises a brain image to be registered and a target brain image, the method comprises the following steps of aiming at the cerebral cortex of the brain image to be registered: computing the surface having the plurality of boundaries to obtain a hyperbolic metric for the surface; according to the hyperbolic measurement and the number of interest curves, hyperbolic trousers decomposition is performed on the curved surface to obtain a plurality of curved surface pieces; decomposing each curved surface sheet into two hyperbolic hexagons; and calculating hyperbolic conformal mapping between the hyperbolic hexagon and the target hyperbolic hexagon to realize registration of the brain image. The scheme of the invention obviously improves the quality of brain image registration and ensures the differential homoblast mapping between the brain image to be registered and the target brain image.
Description
Technical Field
The present invention relates generally to the field of computer image processing. More particularly, the present invention relates to a method, apparatus, and computer-readable storage medium for achieving brain image registration by a computer.
Background
The automatic calculation of surface correspondences (or "registrations") by harmonic mapping is an active research area in computer vision, computer graphics and computational geometry. It can help record and understand physical and biological phenomena and has wide applications in biometric identification, medical imaging, and motion capture. Despite current efforts to study harmonic mappings, there is limited progress in computing a harmonic mapping of differential homoblastosis on general topological surfaces with sulcus landmark ("sulcal landmark") constraints.
Disclosure of Invention
The solution of the present invention solves this above problem by changing the Riemannian metric on the target surface to a hyperbolic metric, thereby ensuring that the harmonic mapping is differentially homomorphic under landmark constraints ("Landmark constraints").
In one aspect of the present invention, there is provided a method for achieving brain image registration by a computer, wherein the brain image includes a brain image to be registered and a target brain image, the method including:
for the cerebral cortex of the brain image to be registered, executing the following steps:
cutting along a plurality of interest curves of the cerebral cortex to obtain a curved surface with a plurality of boundaries;
computing the surface having the plurality of boundaries to obtain a hyperbolic metric for the surface;
according to the hyperbolic metric and the number of interest curves, hyperbolic trousers decomposition is performed on the curved surface to obtain a plurality of curved surface pieces;
further decomposing each of the plurality of patches into two hyperbolic hexagons;
and calculating hyperbolic conformal mapping between the hyperbolic hexagon aiming at the brain image to be registered and the target hyperbolic hexagon of the target brain image so as to realize the registration of the brain image to be registered to the target brain image.
In one embodiment, each of the curved patches is a curved surface having three boundaries and a genus of zero.
In one embodiment, the method further comprises performing the following operations for the cerebral cortex of the target brain image to obtain a target hyperbolic hexagon of the target brain image: cutting along a plurality of target interest curves of the cerebral cortex to obtain a target curved surface with a plurality of target boundaries; calculating the target curved surface with the plurality of target boundaries to obtain a target hyperbolic metric of the curved surface; performing hyperbolic trousers decomposition on the curved surface according to the target hyperbolic measurement and the number of the target interest curves to obtain a plurality of target curved surface pieces; and further decomposing each target surface patch of the plurality of target surface patches into a target hyperbolic hexagon.
In one embodiment, wherein the hyperbolic metric is calculated using a discrete hyperbolic reed flow algorithm.
In one embodiment, wherein prior to computing the hyperbolic conformal mapping, the method further comprises: embedding the hyperbolic hexagons and the target hyperbolic hexagons on a Poincare disc at equal intervals respectively; and realizing the registration of the brain image to be registered to the target brain image according to the hyperbolic hexagon embedded on the Poincare disc and the target hyperbolic hexagon.
In one embodiment, wherein the registering according to the hyperbolic hexagon and the target hyperbolic hexagon comprises: calculating hyperbolic conformal mapping between the hyperbolic hexagons and the target hyperbolic hexagons by using a discrete surface Richardy flow algorithm and a hyperbolic embedding algorithm.
In one embodiment, the method further comprises: and taking the hyperbolic conformal mapping as an initial mapping to execute a nonlinear thermal diffusion algorithm so as to correct the initial mapping, and taking a corrected mapping result as a final mapping result of the curved surface and the target curved surface.
In one embodiment, the method further comprises: and carrying out color coding on the brain image to be registered and the target brain image according to the final mapping result so as to show the correspondence between the brain image to be registered and the target brain image through colors.
In another aspect of the present invention, there is provided an apparatus for brain image registration, including: at least one processor; and at least one memory for storing program instructions that, when loaded and executed by the at least one processor, cause the apparatus to perform the operations described in the above-described methods and embodiments thereof.
In yet another aspect of the present invention, a computer-readable storage medium is provided, in which program instructions are stored, the program instructions being adapted to be loaded by a processor and to perform the above-described method and operations described in its various embodiments.
According to the method, the device and the computer-readable storage medium, the problems existing in the brain image registration can be well solved by utilizing the Rich flow algorithm. The experimental result shows that the image registration of the invention always keeps differential homoembryo. In addition, the approach of the present invention can achieve relatively high performance when evaluated using some popular cortical surface registration evaluation criteria.
Drawings
The above and other objects, features and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description read in conjunction with the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are illustrated by way of example and not by way of limitation, and like reference numerals designate like or corresponding parts throughout the several views, in which:
FIG. 1 is a simplified flow diagram illustrating a method for brain image registration according to an embodiment of the present invention;
fig. 2 is a detailed flowchart illustrating a method for brain image registration according to an embodiment of the present invention;
FIG. 3 is a schematic view showing the trousers exploded according to an embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating hyperbolic hexagon matching according to an embodiment of the present invention;
FIG. 5 is a schematic diagram illustrating registration brain image and target brain image matching according to an embodiment of the present invention; and
fig. 6 is a schematic block diagram illustrating an apparatus for brain image registration according to an embodiment of the present invention.
Detailed Description
Embodiments will now be described with reference to the accompanying drawings. It will be appreciated that for simplicity and clarity of illustration, where considered appropriate, reference numerals may be repeated among the figures to indicate corresponding or analogous elements. In addition, this application sets forth numerous specific details in order to provide a thorough understanding of the embodiments described herein. However, it will be understood by those of ordinary skill in the art that the embodiments described herein may be practiced without these specific details. In other instances, well-known methods, procedures, and components have not been described in detail so as not to obscure the embodiments described herein. Moreover, this description is not to be taken as limiting the scope of the embodiments described herein.
Fig. 1 is a simplified flow diagram illustrating a method 100 for brain image registration according to an embodiment of the present invention. As previously mentioned, the method 100 may be implemented by various types of computing devices including, for example, a computer, and the brain images or brain image data involved therein may be acquired by, for example, Magnetic Resonance Imaging (MRI) techniques or devices. In one embodiment, the aforementioned brain image may include a brain image to be registered and a target brain image. For example, the brain image to be registered may be a brain image to be studied or detected (e.g., a brain image of a suspected patient), while the target brain image is a normal brain image that may be used for comparison or reference with the brain image to be registered.
As shown in fig. 1, for the cortex of the above-mentioned brain image to be registered, at step 102, the method 100 cuts along a plurality of interest curves of the cortex to obtain a curved surface having a plurality of boundaries. According to one application scenario, the interest curves herein may be a plurality of sulci on the cerebral cortex obtained by MRI techniques, and the deformation of these particular sulci may be used as an indicator for analyzing morphological abnormalities of the brain, such as for alzheimer's disease ("AD") and williams syndrome ("WS"), etc. By cutting the brain along these sulci loops, the cortex becomes a zero-deficit, multi-boundary surface that may include a plurality of triangular meshes with corresponding vertices ("vertex"). After obtaining a plurality of bounded surfaces, at step 104, the method 100 computes the surface having the plurality of bounds to obtain a measure of hyperbolic curvature of the surface. In one embodiment, the present invention uses a discrete hyperbolic-Rich flow algorithm to compute a hyperbolic metric for a surface, and when the input is a surface M with a triangular mesh and the output is the hyperbolic metric for the surface M, an exemplary computation process is described as follows:
step 1: at the vertexThe point distribution has a radiusThe circle of (a); for each edgeThe two circles intersect at an angleCalled edge weight;
And step 3: determining the angle associated with each corner by the current edge length using the inverse hyperbolic cosine law;
Step 6: repeating steps 2-5 until all verticesLess than user specified fault tolerance (e.g., less than 10)-8);
And 7: finally determining each vertexRadius of (2)And calculating the side length by using the step 2l ij As a measure of hyperbolic curve。
After obtaining the hyperbolic metric for the surface via the exemplary discrete hyperbolic reed flow algorithm, at step 106, the method 100 performs hyperbolic pant decomposition on the surface according to the hyperbolic metric and the number of interest curves to obtain a plurality of surface patches.
In order to better understand the solution of the present invention, the decomposition of the double curved trousers described above will be described below with reference to fig. 3. It should be noted that the description herein is merely exemplary and not limiting, and that the following decomposition of hyperbolic trousers can be modified by one skilled in the art with appropriate modifications to accommodate different application scenarios.
As previously mentioned, after cutting along the sulcus cerebri, the input surface of the present invention, i.e., having a plurality of boundary components, can be obtainedZero-genus (e.g., the surface in step 104). Further, the surface has a hyperbolic measure (e.g., determined via the discrete hyperbolic reed flow algorithm described above) and all boundaries are geodesic lines. On the above premise, the process of decomposing the hyperbolic trousers will now be described with reference to fig. 3:
firstly, two boundary rings can be arbitrarily selectedAndcalculate their product. If the product is homotopy withThen another pair of boundary rings is selected. Otherwise, assume thatDifferent from the arbitrary boundary ring, calculate its corresponding Mobius (M ö bius) transformAnd the fixing points thereofAnd. Hyperbola passing through these fixed points isA shaft of which isHomotopy geodesic. Cutting the grid along these geodesic lines and repeating the foregoing for each connected boundary componentProcessing until all connected components become pants. The left side of fig. 3 shows the trousers in an exploded process and the right side shows such trousers. The hyperbolic trousers decomposition algorithm is described below in an exemplary step-wise manner, where the input is a topological surface with B boundaries (e.g., a surface in a discrete hyperbolic metric richflow algorithm), and the output is a trousers decomposition of M:
step 2: if Q has a boundary less than 3, then end; otherwise, entering step 3;
And 4, step 4:、anddefining a pants patch ("pans patch"), removing the pants patch from M, and removing the pants patch from QAnd. Will be provided withPut into queue Q. Returning to the step 1 to continue the steps until the hyperbolic curve is finishedThe trousers of the measured curved surface M are decomposed, thereby decomposing the curved surface into a plurality of curved pieces.
After completing the decomposition of the hyperbolic pant as described above, the method 100 proceeds to step 108, where the method 100 further decomposes each of the plurality of patches into two hyperbolic hexagons. In order to decompose a curved surface patch into a hyperbolic hexagon, the shortest path between two boundary rings (or boundary closed curves) needs to be determined first. For this reason, assume that one hyperbolic pant M has three geodesic boundaries. In a general coverage space,Andare respectively lifted to hyperbolasAnd. Presence of reflectionAndthe symmetry axis of which isAndfollowed by Mobius (M ö bius) transformationAxis of corresponds toAndshortest geodesic line therebetweenWherein ""denotes the composition between the mappings. For ease of understanding, the decomposition of a curved patch 401 belonging to the image of the brain to be registered into two hyperbolic hexagons 402 and 403 is shown in fig. 4. Further shown in fig. 4 is the decomposition of the curved patch 404 of the target brain image compared to the brain image to be registered into two hyperbolic hexagons 405 and 406.
After obtaining the hyperbolic hexagon of the brain image, in step 110, the method 100 calculates hyperbolic conformal mapping between the hyperbolic hexagon for the brain image to be registered and the target hyperbolic hexagon for the target brain image, so as to realize registration of the brain image to be registered to the target brain image.
The brain image registration scheme of the present invention is described above with reference to fig. 1, and by combining hyperbolic metric and conformal mapping, the scheme of the present invention significantly improves the quality of brain image registration, and ensures differential homoblast of mapping between the brain image to be registered and the target brain image. Further, the inventive solution also produces less curvature errors and significantly reduces local area distortions.
Fig. 2 is a detailed flowchart illustrating a method 200 for brain image registration according to an embodiment of the present invention. From reading what is shown in fig. 2, a person skilled in the art can understand that the method 200 is a further refinement of the method 100 of fig. 1, and therefore the technical details described in connection with fig. 1 also apply to the same or similar processing shown in fig. 2, and will not be described in detail below.
As shown in fig. 2, at step 202, the method 200 may receive a brain image to be registered having a cortex image and a target brain image. In one application scenario, the brain image to be registered may be a brain image of a potential patient and the target brain image is a brain image of a normal person. By registering the brain image of the potential patient with the target brain image, the brain changes of the potential patient can be observed and analyzed, and a judgment can be made on the brain abnormality or lesion of the patient.
At step 204, the method 200 cuts the image of the cortex along the curve of interest to obtain a surface having a plurality of boundaries. As mentioned above, the interest curve may be a sulcus of the cortex, and the curved surface may be a curved surface formed by a triangular mesh having three vertices. Next, at step 206, the method 200 computes the surface having a plurality of boundaries to obtain a measure of hyperbolic curve, such as the side length described above, for the surface. Thereafter, at step 208, the method 200 performs hyperbolic pant decomposition on the curved surface according to the hyperbolic metric and the number of interest curves to obtain a plurality of curved patches. Subsequently, the method 200 further decomposes each of the plurality of patches into two hyperbolic hexagons, such as the hyperbolic hexagons 402 and 403 shown in fig. 4. Since the operations in steps 204-210 are the same as or similar to the operations in steps 102-108 shown in fig. 1, the description of the operations in steps 102-108 is also applicable to the operations in steps 204-210, and is not repeated herein for brevity.
In one embodiment, the method 200 may further include embedding the hyperbolic hexagons and the target hyperbolic hexagons on a poincare disk equidistantly at step 212, respectively, so that the registration of the brain image to be registered to the target brain image may be achieved according to the hyperbolic hexagons and the target hyperbolic hexagons embedded on the poincare disk.
How to embed hyperbolic hexagons on a poincare disk is described below by way of example, it being understood that the embedding algorithm herein is merely exemplary and that other suitable algorithms may be employed to achieve such embedding by one skilled in the art in light of the present disclosure. For ease of description, assume that the input is a simply connected grid M with a single boundary and a measure of hyperbola, and the output is an M-hyperbola embedded onto a poincare disk.
Step 1: root surface selection ('root face')fHaving hyperbolic triangular networksThe length of the side isl i ,l j ,l k Is embedded byf
Step 2: will be mixed withfAll the surfaces of the shared edge are placed into a queue Q;
and step 3: in the case where Q is not empty, the following steps 4-8 are repeatedly performed:
step 6: calculating the intersection point of the two Euclidean circles;
and 7: selecting the point of intersection of the forming surfaces oriented counter-clockwise, which isThe coordinates of (a);
and 8: will be provided withfAre put into a queue, these adjacent faces are not yet put into a queueAnd (4) embedding.
Upon completion of the embedding operation described above, method 200 may further include calculating a hyperbolic conformal mapping between the hyperbolic hexagons and the target hyperbolic hexagons using a discrete-surface richflow algorithm and a hyperbolic embedding algorithm at step 214. Alternatively, in some embodiments, the initial hyperbolic hexagon match between the hyperbolic hexagon and the target hyperbolic hexagon may also be calculated by solving laplace's equation with dirichlet boundary conditions.
In utilizing the discrete surface reed flow algorithm, its inputs include: hyperbolic hexagons of a brain image to be registered, the hyperbolic hexagons being in a gridMThe form is embedded in a hyperbolic space; the curvature of the target hyperbolic hexagon is used as a target curvature; and a curvature error threshold. The output of this algorithm is the circle filling ("circle packing") metric (through iterative iterations of the edges and vertices of the mesh)M,Г,Ф) WhereinMIs a grid of wires,Гis a conformal factor, andФis a rounded intersection angle; by the three aforementioned quantities, the side lengths of all the grids can be obtained, thus obtaining the metric. Points on the hyperbolic hexagonal mesh are then mapped into the target hexagonal region by a hyperbolic embedding algorithm (e.g., using a poincare disk) to obtain a hyperbolic conformal mapping.
After obtaining the hyperbolic conformal mapping between the brain image to be registered and the target brain image by using the above operations, the method 200 performs a non-linear thermal diffusion algorithm using the hyperbolic conformal mapping as an initial mapping to correct the initial mapping at step 216, and uses a corrected mapping result as a final mapping result between the curved surface of the brain image to be registered and the curved surface of the target brain image, thereby implementing registration of the brain image. In order to better understand the solution of the present invention, the following non-linear heat diffusion algorithm will be described by way of example and not limitation.
In the nonlinear heat diffusion algorithm, the inputs are assumed to be: two curved surface models (or called grid)MAndNwhich have hyperbolic measurements on Poincare disks, respectivelyC M AndC N one-to-one registration (,) And a threshold value. Here, theAre the vertices of the mesh M and,is the 3D coordinate on grid N; the output is: novel microagglomeration (,). An exemplary step of the algorithm is as follows:
step 1: for theMEach vertex of (1)Embedding its neighboring domains on a Poincare disk, whereinHaving coordinates(ii) a To pairThe same operation is performed and its coordinates on the poincare disk are marked;
With regard to equation (3), it can be applied to the following scenario:
suppose thatIs the initial mapping, g1And g2Is a hyperbolic metric. ComputingS 1 AndS 2 the conformal atlas of (1). For theS 1 AndS 2 selecting local conformal parameterszAndw,fwith partial representationOr simply byThen the nonlinear diffusion can be given by the equation (3) where. Suppose thatv i Is selected asS 1 Has a local representation of the vertexz i Then, after diffusion, a local representation of its image can be obtained。
And 4, step 4: using updatedTo calculateNUp to new 3D coordinatesP i And repeatedly performing the above-mentioned processIs proceeded untilIs less than。
Through the nonlinear thermal diffusion algorithm, the scheme of the invention corrects the initial mapping, realizes the final mapping result of the curved surface of the brain image to be registered and the target curved surface, and obtains a new differential homoembryo.
Additionally or alternatively, the method 200 of the present invention may further perform color coding on the brain image to be registered and the target brain image according to the final mapping result in step 218, so as to show the correspondence ("correspondance") of the brain image to be registered and the target brain image by color. For example, fig. 5 shows the color-coded results of the brain image to be registered and the target brain image at 514 and 516, respectively, in which the cerebral cortex is colored in progressive dark green, light yellow, and dark yellow so that the user can see the correspondence of the two brain images. Furthermore, the colored to-be-registered brain image and the target brain image can also respectively comprise mark points with the same color, so that the mark points with the same color further show the correspondence of the two brain images, and a professional can compare and analyze the difference of the two brain images conveniently.
The brain image registration scheme of the present invention is described above with reference to the method flow of fig. 2, and the differential homoblast of brain image registration is realized by the scheme of the present invention. In addition, by using a nonlinear thermal diffusion method and a Rich flow algorithm, conformal mapping is realized on the basis of hyperbolic measurement, and the defect that differential homoblast cannot be guaranteed in image registration in the prior art is overcome. In addition, although the method of fig. 2 also relates to processing of the target brain image, in some application scenarios, hyperbolic hexagon data about the target brain image may be stored in advance as the target hyperbolic hexagon so as to be compared with the hyperbolic hexagon of the brain image to be registered, so as to achieve efficient brain image registration.
Fig. 5 is a schematic diagram illustrating matching of a registered brain image and a target brain image according to an embodiment of the present invention, in which the upper row illustrates an image processing procedure of a brain image to be registered, and the lower row illustrates an image processing procedure of a target brain image, which is the main steps of the registration procedure described in conjunction with fig. 1 to 4.
As shown in fig. 5, the cortex M of the brain image to be registered is shown at 502, and the cortex N of the target brain image is shown at 504, and both have been cut along the sulcus to form a boundary. At 506, the embedding on the poincare disk is performed on the cortex M, and at 508, the embedding on the poincare disk is performed on the cortex N. Next, at 510 and 512, trousers decomposition is performed on the cerebral cortex M and N, respectively, to obtain a plurality of trousers, and each of the trousers may be further decomposed into two hyperbolic hexagons, such that a one-to-one mapping may be formed between corresponding portions of the cerebral cortex M and N. Finally, at 514 and 516, the obtained initial mapping may be modified (e.g., by the non-linear thermal diffusion algorithm described previously) and then the registration results for the cortex M and N may be image-coded to show correspondence therebetween. It is noted that fig. 5 is only a summary and schematic illustration of the preceding solution, and reference may be made to the description in connection with fig. 1-4 for further processing details.
Fig. 6 is a schematic block diagram illustrating an apparatus 600 for brain image registration according to an embodiment of the present invention.
As shown in fig. 6, the apparatus 600 for brain image registration may include a CPU 6011, which may be a general-purpose CPU, a dedicated CPU, or an execution unit on which other information processing and programs run. Further, the apparatus 600 may further include a mass storage 6012 and a read only memory ROM6013, wherein the mass storage 6012 may be configured to store various types of data including various brain image data, algorithm data, intermediate result results, and various programs required to operate the apparatus 600, and the ROM6013 may be configured to store power-on self-test for the apparatus 600, initialization of various functional modules in the system, a driver for basic input/output of the system, and data required to boot the operating system.
Further, the apparatus 600 also includes other hardware platforms or components, such as a TPU (tensor processing unit) 6014, a GPU (graphics processing unit) 6015, an FPGA (field programmable gate array) 6016, and an M L U (machine learning unit) 6017 as shown it is understood that although various hardware platforms or components are shown in the apparatus 600, this is merely exemplary and not limiting, and those skilled in the art may add or remove corresponding hardware according to actual needs.
The inventive device 600 further comprises a communication interface 6018, such that it can be connected via this communication interface 6018 to a local area network/wireless local area network (L AN/W L AN) 605, which in turn can be connected via L AN/W L AN to a local server 606 or to the Internet ("Internet") 607 alternatively or additionally, the inventive device 600 can also be connected via the communication interface 6018 directly to the Internet or a cellular network based on wireless communication technology, for example based on third generation ("3G"), fourth generation ("4G") or 5 th generation ("5G") wireless communication technology.
The peripheral devices of the apparatus 600 may include a display device 602, an input device 603, and a data transmission interface 604. In one embodiment, the display device 602 may, for example, include one or more speakers and/or one or more visual displays configured to provide voice prompts and/or visual displays of the operational procedures or end results of the brain image registration of the present invention. Input device 603 may include, for example, a keyboard, mouse, microphone, gesture capture camera, or other input buttons or controls configured to receive input of test data or user instructions. The data transfer interface 604 may include, for example, a serial interface, a parallel interface, or a universal serial bus interface ("USB"), a small computer system interface ("SCSI"), serial ATA, FireWire ("FireWire"), PCI Express, and a high-definition multimedia interface ("HDMI"), which are configured for data transfer and interaction with other devices or systems. In accordance with aspects of the present invention, the data transfer interface 604 may receive brain image data from the MRI device and transmit the brain image data or various other types of data and results to the apparatus 600.
The above-mentioned CPU 6011, mass storage 6012, read only memory ROM6013, TPU6014, GPU 6015, FPGA 6016, M L U6017, and communication interface 6018 of the apparatus 600 of the present invention may be connected to each other through a bus 6019, and implement data interaction with peripheral devices through the bus 6019, in one embodiment, the CPU 6011 may control other hardware components and peripheral devices thereof in the apparatus 600 through the bus 6019.
An apparatus that may be used to perform the brain image registration of the present invention is described above in connection with fig. 6. It is to be understood that the device configurations herein are merely exemplary, and that the implementations and entities of the invention are not limited thereto, but may be varied without departing from the spirit of the invention.
It should also be appreciated that any module, unit, component, server, computer, terminal, or device executing instructions of the examples of the invention may include or otherwise access a computer-readable medium, such as a storage medium, computer storage medium, or data storage device (removable) and/or non-removable) such as a magnetic disk, optical disk, or magnetic tape. Computer storage media may include volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information, such as computer readable instructions, data structures, program modules or other data.
Based on the foregoing, the present invention also discloses a computer readable storage medium having stored therein program instructions adapted to be loaded and executed by a processor: aiming at the cerebral cortex of the brain image to be registered, the following steps are executed: cutting along a plurality of interest curves of the cerebral cortex to obtain a curved surface with a plurality of boundaries; computing the surface having the plurality of boundaries to obtain a hyperbolic metric for the surface; according to the hyperbolic measurement and the number of interest curves, hyperbolic trousers decomposition is performed on the curved surface to obtain a plurality of curved surface pieces; further decomposing each of the plurality of patches into two hyperbolic hexagons; and calculating hyperbolic conformal mapping between the hyperbolic hexagon aiming at the brain image to be registered and the target hyperbolic hexagon of the target brain image so as to realize the registration of the brain image to be registered to the target brain image. Preferably or additionally, the computer readable storage medium may further comprise program instructions to take the obtained hyperbolic conformal mapping as an initial mapping and to modify the initial mapping using, for example, the aforementioned non-linear thermal diffusion algorithm. In summary, the computer readable storage medium includes program instructions for performing the processing operations described in connection with fig. 1-5.
It should be understood that the terms "first" or "second," etc. in the claims, description, and drawings of the present disclosure are used for distinguishing between different objects and not for describing a particular order. The terms "comprises" and "comprising," when used in the specification and claims of this disclosure, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only, and is not intended to be limiting of the invention disclosed. As used in the specification and claims of this disclosure, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should be further understood that the term "and/or" as used in this disclosure and in the claims refers to any and all possible combinations of one or more of the associated listed items and includes such combinations.
Although the embodiments of the present invention are described above, the descriptions are only examples for facilitating understanding of the present invention, and are not intended to limit the scope and application scenarios of the present invention. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (8)
1. A method for achieving brain image registration by a computer, wherein the brain image includes a brain image to be registered and a target brain image, comprising:
for the cerebral cortex of the brain image to be registered, executing the following steps:
cutting along a plurality of interest curves of the cerebral cortex to obtain a curved surface with a plurality of boundaries;
calculating the curved surface with the plurality of boundaries by using a discrete hyperbolic Rich flow algorithm to obtain a hyperbolic metric of the curved surface;
according to the hyperbolic metric and the number of interest curves, hyperbolic trousers decomposition is performed on the curved surface to obtain a plurality of curved surface pieces;
further decomposing each of the plurality of patches into two hyperbolic hexagons;
embedding the hyperbolic hexagons of the brain images to be registered and the target hyperbolic hexagons of the target brain images on a Poincare disc at equal intervals respectively; and
and calculating hyperbolic conformal mapping between the hyperbolic hexagon of the brain image to be registered and the target hyperbolic hexagon of the target brain image according to the hyperbolic hexagon and the target hyperbolic hexagon which are embedded into the Poincare disc so as to realize registration of the brain image to be registered to the target brain image.
2. The method of claim 1, wherein each said curved surface patch is a curved surface having three boundaries and a genus of zero.
3. The method of claim 1, further comprising, for a cerebral cortex of the target brain image, performing the following operations to obtain a target hyperbolic hexagon of the target brain image:
cutting along a plurality of target interest curves of the cerebral cortex to obtain a target curved surface with a plurality of target boundaries;
calculating the target curved surface with the plurality of target boundaries to obtain a target hyperbolic metric of the curved surface;
performing hyperbolic trousers decomposition on the curved surface according to the target hyperbolic measurement and the number of the target interest curves to obtain a plurality of target curved surface pieces; and
and further decomposing each target curved surface slice in the plurality of target curved surface slices into a target hyperbolic hexagon.
4. The method of claim 1, wherein implementing the registration according to the hyperbolic hexagon and the target hyperbolic hexagon comprises:
calculating hyperbolic conformal mapping between the hyperbolic hexagons and the target hyperbolic hexagons by using a discrete surface Richardy flow algorithm and a hyperbolic embedding algorithm.
5. The method of claim 1, further comprising:
and taking the hyperbolic conformal mapping as an initial mapping to execute a nonlinear thermal diffusion algorithm so as to correct the initial mapping, and taking a corrected mapping result as a final mapping result of the curved surface and the target curved surface.
6. The method of claim 5, further comprising:
and carrying out color coding on the brain image to be registered and the target brain image according to the final mapping result so as to show the correspondence between the brain image to be registered and the target brain image through colors.
7. An apparatus for brain image registration, comprising:
at least one processor; and
at least one memory for storing program instructions that, when loaded and executed by the at least one processor, cause the apparatus to perform the method of any of claims 1-6.
8. A computer readable storage medium having stored therein program instructions adapted to be loaded by a processor and to perform the method according to any of claims 1-6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010226114.3A CN111127492B (en) | 2020-03-27 | 2020-03-27 | Method, apparatus, and computer-readable storage medium for brain image registration |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010226114.3A CN111127492B (en) | 2020-03-27 | 2020-03-27 | Method, apparatus, and computer-readable storage medium for brain image registration |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111127492A CN111127492A (en) | 2020-05-08 |
CN111127492B true CN111127492B (en) | 2020-07-17 |
Family
ID=70493934
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010226114.3A Active CN111127492B (en) | 2020-03-27 | 2020-03-27 | Method, apparatus, and computer-readable storage medium for brain image registration |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111127492B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117853320A (en) * | 2024-03-07 | 2024-04-09 | 电子科技大学成都学院 | Image mapping method, system and storage medium based on multimedia control |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1781111A (en) * | 2002-11-06 | 2006-05-31 | 几何信息学股份有限公司 | Analysis of geometric surfaces by comformal structure |
US10262414B2 (en) * | 2015-07-29 | 2019-04-16 | University Of Louisville Research Foundation, Inc. | Computer aided diagnostic system for mapping of brain images |
CN109377561B (en) * | 2018-10-26 | 2024-01-19 | 北京工商大学 | Conformal geometry-based digital-analog surface grid generation method |
-
2020
- 2020-03-27 CN CN202010226114.3A patent/CN111127492B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN111127492A (en) | 2020-05-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
ES2685977T3 (en) | Method and system to analyze image data | |
Samavati et al. | A hybrid biomechanical intensity based deformable image registration of lung 4DCT | |
US20140375631A1 (en) | Image generation apparatus, method, and medium with image generation program recorded thereon | |
US10706575B2 (en) | Unsupervised asymmetry detection | |
Ballester et al. | Segmentation and measurement of brain structures in MRI including confidence bounds | |
CN111127492B (en) | Method, apparatus, and computer-readable storage medium for brain image registration | |
JP2021529622A (en) | Method and computer program for segmentation of light interference tomography images of the retina | |
Meng et al. | Learning‐based subject‐specific estimation of dynamic maps of cortical morphology at missing time points in longitudinal infant studies | |
Acosta et al. | Cortical surface mapping using topology correction, partial flattening and 3D shape context-based non-rigid registration for use in quantifying atrophy in Alzheimer's disease | |
WO2016132367A1 (en) | Chest radiograph (cxr) image analysis | |
Roy et al. | An accurate and robust skull stripping method for 3-D magnetic resonance brain images | |
Maillard et al. | A deep residual learning implementation of metamorphosis | |
JP6987721B2 (en) | Image processing equipment, methods and programs | |
US20170132791A1 (en) | Cutting three-dimensional image | |
CN112381822B (en) | Method for processing images of focal zones of the lungs and related product | |
CN112750110A (en) | Evaluation system for evaluating lung lesion based on neural network and related products | |
CN112927235A (en) | Brain tumor image segmentation method based on multi-scale superpixel and nuclear low-rank representation | |
Lafitte et al. | Accelerating multi-modal image registration using a supervoxel-based variational framework | |
WO2015111308A1 (en) | Three-dimensional medical image display control device, operation method thereof, and three-dimensional medical image display control program | |
CN112785562B (en) | System for evaluating based on neural network model and related products | |
Wang et al. | A heat kernel based cortical thickness estimation algorithm | |
Cheng et al. | Ground truth delineation for medical image segmentation based on Local Consistency and Distribution Map analysis | |
US20220108525A1 (en) | Patient-specific cortical surface tessellation into dipole patches | |
Shen et al. | Statistical shape analysis for brain structures | |
US20150199840A1 (en) | Shape data generation method and apparatus |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |