Anterior visual pathway feature extraction method and device, electronic device, and storage medium
By rotating and projecting the three-dimensional image of the forward vision path into two dimensions, and extracting and fitting the skeleton control points, the problem of low accuracy in calculating the structural features of the forward vision path is solved, and a more accurate assessment of eye health status is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SOUTHERN UNIVERSITY OF SCIENCE AND TECHNOLOGY
- Filing Date
- 2024-07-19
- Publication Date
- 2026-06-30
Smart Images
Figure CN118968082B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of feature extraction technology, and in particular to a method and apparatus for front-view feature extraction, an electronic device, and a storage medium. Background Technology
[0002] The anterior visual pathway refers to the optic nerve, optic chiasm, and optic tract within the visual pathway. Changes in the structural features of the anterior visual pathway are related to the health of the eyes. However, most common methods for examining the structural features of the anterior visual pathway involve mapping a three-dimensional image of the anterior visual pathway onto a two-dimensional plane and then extracting the structural features from the two-dimensional plane. This method is prone to low accuracy in calculating the structural features of the anterior visual pathway, which can easily lead to misjudgments of the health of the eyes. Therefore, improving the accuracy of calculating the structural features of the anterior visual pathway has become an urgent technical problem to be solved. Summary of the Invention
[0003] The main objective of this application is to propose a method, apparatus, electronic device, and storage medium for extracting features of the forward-looking path, aiming to improve the accuracy of calculating the structural features of the forward-looking path.
[0004] To achieve the above objectives, a first aspect of this application proposes a forward-looking path feature extraction method, the method comprising:
[0005] A brain image is acquired, and the brain image is segmented to obtain a three-dimensional image of the forward visual path. A spatial rectangular coordinate system is constructed, wherein the spatial rectangular coordinate system includes a first axis, a second axis, and a third axis, and the first axis, the second axis, and the third axis are perpendicular to each other.
[0006] Based on the two-dimensional projection plane formed by the first axis and the second axis, the three-dimensional image of the front view path is rotated to obtain a rotated image of the front view path.
[0007] Based on the two-dimensional projection surface, the rotating image of the forward path is projected in two dimensions to obtain a two-dimensional image of the forward path;
[0008] Skeleton extraction is performed on the two-dimensional image of the forward-looking path to obtain two-dimensional skeleton points of the forward-looking path;
[0009] Based on the two-dimensional skeleton points of the forward-looking path and the third axis, determine the skeleton control points in the forward-looking path rotation image;
[0010] The coordinates of the skeleton control points are restored to obtain the forward path control points;
[0011] Based on the spatial rectangular coordinate system, the forward-looking path control points are fitted to obtain a three-dimensional skeleton image of the forward-looking path;
[0012] Based on the spatial rectangular coordinate system, feature extraction is performed on the three-dimensional skeleton image of the forward-looking path to obtain the feature data of the three-dimensional image of the forward-looking path.
[0013] In some embodiments, rotating the three-dimensional image of the forward-looking path based on the two-dimensional projection plane formed by the first axis and the second axis to obtain a rotated image of the forward-looking path includes:
[0014] Construct the reference plane of the three-dimensional image of the forward-looking path to obtain the forward-looking path reference plane;
[0015] Calculate the normal vector of the forward-looking reference plane to obtain the reference plane normal vector;
[0016] Based on the reference normal vector and the two-dimensional projection plane, the front-view path rotation matrix of the front-view path three-dimensional image is determined;
[0017] Calculate the inverse of the forward path rotation matrix to obtain the forward path rotation inverse matrix;
[0018] Based on the inverse rotation matrix of the forward view path, the three-dimensional image of the forward view path is rotated to obtain the rotated image of the forward view path.
[0019] In some embodiments, the forward-looking path 3D image includes multiple voxel points, and the step of constructing the reference plane of the forward-looking path 3D image to obtain the forward-looking path reference plane includes:
[0020] Calculate the number of non-zero voxels within a preset target range centered on each voxel point;
[0021] Based on the number of non-zero voxels, three target voxel points are selected from the plurality of voxels, wherein the number of non-zero voxels within the target range of the target voxel point is the smallest;
[0022] Based on the three target voxel points, the forward-looking path reference plane is constructed.
[0023] In some embodiments, determining the skeleton control points in the front-view rotation image based on the front-view two-dimensional skeleton points and the third axis includes:
[0024] Based on the two-dimensional skeleton points of the forward-looking path and the third axis, the three-dimensional skeleton axis of the forward-looking path is found from the rotated image of the forward-looking path. The three-dimensional skeleton axis of the forward-looking path is used to characterize the axis that forms the two-dimensional skeleton points of the forward-looking path when projected onto the two-dimensional projection plane in the rotated image of the forward-looking path.
[0025] The midpoint of the three-dimensional skeleton axis of the forward-looking path is used as the skeleton control point.
[0026] In some embodiments, fitting the forward-looking path control points based on the spatial rectangular coordinate system to obtain a three-dimensional skeleton image of the forward-looking path includes:
[0027] Based on the aforementioned spatial rectangular coordinate system, the coordinates of the forward-looking path control points are marked to obtain the control point coordinates;
[0028] Based on the coordinates of the control points, curve fitting is performed on the forward-looking path control points to obtain a three-dimensional skeleton image of the forward-looking path.
[0029] In some embodiments, the control point coordinates include a first value, a second value, and a third value. The step of performing curve fitting on the forward-looking path control points based on the control point coordinates to obtain the three-dimensional skeleton image of the forward-looking path includes:
[0030] Based on the first value, a Bezier curve is fitted to the forward look-ahead control point to obtain a first fitting curve, wherein the first fitting curve is used to characterize the fitting curve of the forward look-ahead control point in the first axis direction.
[0031] Based on the second value, a Bezier curve is fitted to the forward look-ahead control point to obtain a second fitting curve, wherein the second fitting curve is used to characterize the fitting curve of the forward look-ahead control point in the second axis direction.
[0032] Based on the third value, a Bezier curve is fitted to the forward-looking control point to obtain a third fitting curve, wherein the third fitting curve is used to characterize the fitting curve of the forward-looking control point in the third axis direction.
[0033] The first fitting curve, the second fitting curve, and the third fitting curve are combined to obtain the three-dimensional skeleton image of the front-view path.
[0034] In some embodiments, the forward-looking path 3D skeleton image includes multiple skeleton voxels. The feature extraction of the forward-looking path 3D skeleton image based on the spatial Cartesian coordinate system to obtain feature data of the forward-looking path 3D image includes:
[0035] Based on the aforementioned spatial rectangular coordinate system, generate the three-dimensional spatial coordinates of each of the skeleton voxels;
[0036] Based on the three-dimensional spatial coordinates, the horizontal diameter of the three-dimensional skeleton image of the forward-looking path is calculated to obtain the horizontal diameter of the forward-looking path;
[0037] Based on the three-dimensional spatial coordinates, the view path length of the three-dimensional skeleton image of the forward view path is calculated to obtain the forward view path length;
[0038] The feature data is obtained by combining the transverse diameter of the forward-looking path and the length of the forward-looking path.
[0039] To achieve the above objectives, a second aspect of this application provides a forward-looking path feature extraction apparatus, the apparatus comprising:
[0040] The forward visual path segmentation module is used to acquire brain images, segment the brain images to obtain a three-dimensional forward visual path image, and construct a spatial rectangular coordinate system, wherein the spatial rectangular coordinate system includes a first axis, a second axis, and a third axis, and the first axis, the second axis, and the third axis are perpendicular to each other.
[0041] A front-view rotation module is used to rotate the front-view three-dimensional image based on a two-dimensional projection plane formed by the first axis and the second axis to obtain a front-view rotated image.
[0042] A forward-looking path projection module is used to project the forward-looking path rotated image into a two-dimensional image based on the two-dimensional projection surface to obtain a forward-looking path two-dimensional image.
[0043] The forward-looking path 2D skeletonization module is used to extract the skeleton from the forward-looking path 2D image to obtain the forward-looking path 2D skeleton points.
[0044] The skeleton control point finding module is used to determine the skeleton control points in the front-view path rotation image based on the two-dimensional skeleton points of the front-view path and the third axis.
[0045] The control point recovery module is used to recover the coordinates of the skeleton control points to obtain the forward-looking path control points;
[0046] The three-dimensional skeleton fitting module is used to fit the forward-looking path control points based on the spatial rectangular coordinate system to obtain a three-dimensional skeleton image of the forward-looking path.
[0047] The forward-looking path feature extraction module is used to extract features from the three-dimensional skeleton image of the forward-looking path based on the spatial rectangular coordinate system, so as to obtain the feature data of the three-dimensional image of the forward-looking path.
[0048] To achieve the above objectives, a third aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect.
[0049] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.
[0050] The present application proposes a method, apparatus, electronic device, and storage medium for extracting features of the forward visual path. This method involves rotating a three-dimensional forward visual path image obtained by cutting out a brain image to obtain a rotated forward visual path image. The rotated forward visual path image is then projected onto a two-dimensional projection plane formed by the first and second axes. This improves the ease of projecting the three-dimensional forward visual path image onto a two-dimensional plane, thereby increasing the efficiency of forward visual path feature extraction. Furthermore, based on the two-dimensional skeleton points of the forward visual path in the two-dimensional image and the third axis of the constructed spatial rectangular coordinate system, the corresponding skeleton control points in the rotated forward visual path image are located. These skeleton control points are then rotated and restored to obtain the positions of the forward visual path control points in the three-dimensional forward visual path image. Finally, the forward visual path control points are fitted to obtain a three-dimensional skeleton image of the forward visual path. This achieves skeleton extraction from the three-dimensional forward visual path image, avoiding the direct extraction of forward visual path structural features from the two-dimensional forward visual path image, thus improving the accuracy of calculating forward visual path structural features. Attached Figure Description
[0051] Figure 1 This is a flowchart of the front-look path feature extraction method provided in the embodiments of this application;
[0052] Figure 2 yes Figure 1 The flowchart of step S102 in the document;
[0053] Figure 3 yes Figure 2 The flowchart of step S201 in the text;
[0054] Figure 4 yes Figure 1 The flowchart of step S105 in the process;
[0055] Figure 5 yes Figure 4 The flowchart of step S107 in the process;
[0056] Figure 6 yes Figure 5 The flowchart of step S502 in the document;
[0057] Figure 7 yes Figure 1 The flowchart of step S108 in the process;
[0058] Figure 8 This is a schematic diagram of the front-view path feature extraction device provided in the embodiments of this application;
[0059] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation
[0060] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0061] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.
[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.
[0063] First, let's analyze some of the terms used in this application:
[0064] The visual pathway, also known as the visual cortex, is a crucial pathway connecting the retina and the brain's visual center. It comprises the anterior and posterior visual pathways. The anterior visual pathway mainly consists of the retina, optic nerve, optic chiasm, and optic tract. The retina is the light-sensitive structure; after receiving light, it converts the light signals into neural signals, which are then transmitted to the brain via the optic nerve. The optic nerve is the nerve that transmits light signals from the eye to the brain; some of its fibers cross at the optic chiasm before entering different areas of the brain, thus forming a visual image. The integrity of the anterior visual pathway is essential for normal visual function. The posterior visual pathway mainly includes the lateral geniculate body, optic radiation, and visual cortex. At the lateral geniculate body, visual signals cross with other sensory signals and are then transmitted to the visual cortex via the optic radiation, where further processing and interpretation occur, ultimately forming visual perception and cognition.
[0065] The anterior visual pathway is a vital pathway connecting the retina and the visual center, playing a crucial role in transmitting visual signals. It consists of the optic nerve, optic chiasm, and optic tract, and is an essential part of vision formation. The lateral geniculate body, optic radiation, and visual cortex are classified as part of the posterior visual pathway. The integrity of the anterior visual pathway is essential for maintaining normal vision because many eye and brain diseases, such as glaucoma, multiple sclerosis, and orbital tumors, can lead to changes in its structure.
[0066] The anterior visual pathway refers to the optic nerve, optic chiasm, and optic tract within the visual pathway. Changes in the structural features of the anterior visual pathway are related to the health of the eyes. However, most common methods for examining the structural features of the anterior visual pathway involve mapping a three-dimensional image of the anterior visual pathway onto a two-dimensional plane and then extracting the structural features from the two-dimensional plane. This method is prone to low accuracy in calculating the structural features of the anterior visual pathway, which can easily lead to misjudgments of the health of the eyes. Therefore, improving the accuracy of calculating the structural features of the anterior visual pathway has become an urgent technical problem to be solved.
[0067] Based on this, embodiments of this application provide a method and apparatus for extracting front-view path features, an electronic device, and a storage medium, aiming to improve the accuracy of calculating front-view path structural features.
[0068] The front-look path feature extraction method, apparatus, electronic device, and storage medium provided in this application are specifically described through the following embodiments. First, the front-look path feature extraction method in this application is described.
[0069] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.
[0070] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.
[0071] The front-look-ahead feature extraction method provided in this application relates to the field of microwave engineering technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the front-look-ahead feature extraction method, but is not limited to the above forms.
[0072] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.
[0073] Figure 1 This is an optional flowchart of the front-look path feature extraction method provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S101 to S108.
[0074] Step S101: Obtain a brain image, perform image segmentation on the brain image to obtain a three-dimensional image of the forward visual path, and construct a spatial rectangular coordinate system, wherein the spatial rectangular coordinate system includes a first axis, a second axis and a third axis, and the first axis, the second axis and the third axis are perpendicular to each other;
[0075] Step S102: Based on the two-dimensional projection plane formed by the first axis and the second axis, rotate the three-dimensional image of the front view path to obtain the rotated image of the front view path.
[0076] Step S103: Based on the two-dimensional projection plane, perform two-dimensional projection on the rotated image of the front view path to obtain a two-dimensional image of the front view path.
[0077] Step S104: Extract the skeleton from the two-dimensional image of the forward-looking path to obtain the two-dimensional skeleton points of the forward-looking path.
[0078] Step S105: Based on the two-dimensional skeleton points of the forward-looking path and the third axis, determine the skeleton control points in the rotating image of the forward-looking path.
[0079] Step S106: Restore the coordinates of the skeleton control points to obtain the forward-looking path control points;
[0080] Step S107: Based on the spatial rectangular coordinate system, fit the control points of the forward-looking path to obtain a three-dimensional skeleton image of the forward-looking path.
[0081] Step S108: Based on the spatial rectangular coordinate system, feature extraction is performed on the three-dimensional skeleton image of the forward-looking path to obtain feature data of the three-dimensional image of the forward-looking path.
[0082] Steps S101 to S108, as illustrated in this embodiment, involve segmenting the acquired brain image to obtain a three-dimensional image of the forward visual path. A spatial rectangular coordinate system, including a first axis, a second axis, and a third axis, is arbitrarily constructed within the space containing this three-dimensional image. Furthermore, the plane formed by the first and second axes is used as a two-dimensional projection plane. Based on this two-dimensional projection plane, the three-dimensional image of the forward visual path is rotated so that the rotated image can be easily and conveniently projected onto the two-dimensional projection plane, thus obtaining a two-dimensional image of the forward visual path. The forward visual path is then extracted. The skeleton of the 2D image is obtained by first extracting the 2D skeleton points of the front-view path in the 2D image. Then, along the third axis of the spatial rectangular coordinate system, the skeleton control points corresponding to the 2D skeleton points of the front-view path are determined from the rotated image of the front-view path. These skeleton control points are then restored to the 3D image of the front-view path to obtain the front-view path control points. Finally, based on the positions of the front-view path control points in the spatial rectangular coordinate system, they are fitted to obtain the 3D skeleton image of the front-view path. Then, based on the positions of the front-view path control points in the spatial rectangular coordinate system, features are extracted from the 3D skeleton image of the front-view path to obtain the feature data of the 3D image of the front-view path. Therefore, this application improves the convenience of projecting the three-dimensional image of the anterior visual path obtained by cutting out brain images into a rotated image of the anterior visual path, and then projects the rotated image of the anterior visual path onto a two-dimensional projection plane formed by the first and second axes. This improves the efficiency of anterior visual path feature extraction. Furthermore, based on the two-dimensional skeleton points of the anterior visual path in the two-dimensional image of the anterior visual path and the third axis of the constructed spatial rectangular coordinate system, the corresponding skeleton control points in the rotated image of the anterior visual path are found. The skeleton control points are rotated and restored to obtain the position of the anterior visual path control points in the three-dimensional image of the anterior visual path. Then, the anterior visual path control points are fitted to obtain the three-dimensional skeleton image of the anterior visual path. This realizes the skeleton extraction of the three-dimensional image of the anterior visual path, avoiding the direct extraction of anterior visual path structural features using the two-dimensional image of the anterior visual path, thereby improving the accuracy of calculating the anterior visual path structural features.
[0083] In one application scenario, taking the calculation of the view length of the front-view path as an example, existing technologies typically use a two-dimensional image of the front-view path projected onto a two-dimensional plane to calculate the view length, without fully considering the structure of the front-view path in three-dimensional space. Therefore, the view length calculated using the two-dimensional image of the front-view path is not accurate enough. This application can obtain a two-dimensional front-view path plane by projecting a rotated image of the front-view path into a pre-constructed two-dimensional plane after rotating the three-dimensional image of the front-view path. Then, the two-dimensional front-view path plane is skeletonized to obtain the skeleton points of the front-view path in the two-dimensional plane. Based on these skeleton points, skeleton control points in the rotated image of the front-view path are found. Furthermore, after determining the skeleton control points, the skeleton control points are inversely rotated to obtain the position of the skeleton control points in the three-dimensional image of the front-view path, i.e., the three-dimensional control points of the front-view path. Finally, the three-dimensional front-view path and control points in the three-dimensional image of the front-view path are fitted to obtain a three-dimensional skeleton image of the front-view path, thereby providing a three-dimensional skeleton image of the front-view path for calculating the view length of the front-view path and improving the accuracy of the calculation of the view length of the front-view path.
[0084] In step S101 of some embodiments, the brain image can be a complete brain MRI scan image. By performing an MRI scan on the user's brain, a brain image of the user can be obtained. Then, a three-dimensional U-shaped network is used to segment or extract features from the obtained brain image to obtain a three-dimensional image of the user's forward vision path. After obtaining the three-dimensional image of the forward vision path, a spatial rectangular coordinate system containing a first axis, a second axis, and a third axis is constructed in the space where the three-dimensional image of the forward vision path is located, so as to facilitate the coordinate construction of the three-dimensional image of the forward vision path and reduce the difficulty of forward vision path feature extraction.
[0085] In step S102 of some embodiments, the two-dimensional projection plane refers to a plane that is on the same plane as the first axis and the second axis in the control rectangular coordinate system. The reference plane of the front-view path three-dimensional image is determined by a preset rule, and then the rotation angle of the front-view path three-dimensional image is determined according to the reference plane. The front-view path three-dimensional image is rotated according to the rotation angle to obtain the front-view path rotated image.
[0086] Please see Figure 2 In some embodiments, step S102 may include, but is not limited to, steps S201 to S205:
[0087] Step S201: Construct the reference plane of the three-dimensional image of the forward-looking path to obtain the forward-looking path reference plane;
[0088] Step S202: Calculate the normal vector of the forward-looking road reference plane to obtain the reference plane normal vector;
[0089] Step S203: Based on the reference normal vector, determine the front-view path rotation matrix of the front-view path 3D image;
[0090] Step S204: Calculate the inverse matrix of the forward path rotation matrix to obtain the forward path rotation inverse matrix;
[0091] Step S205: Based on the forward path rotation inverse matrix, rotate the forward path 3D image to obtain the forward path rotated image.
[0092] In step S201 of some embodiments, the reference plane refers to the plane formed by at least three points in the forward-looking 3D image. By identifying at least three points in the forward-looking 3D image, and then constructing the reference plane of the forward-looking 3D image based on the identified at least three points, the forward-looking reference plane is obtained.
[0093] Specifically, please refer to Figure 3 In some embodiments, step S201 may include, but is not limited to, steps S301 to S303:
[0094] Step S301: Calculate the number of non-zero voxels within a preset target range centered on each voxel point;
[0095] Step S302: Based on the number of non-zero voxels, select three target voxel points from multiple voxels, where the number of non-zero voxels within the target range of the target voxel point is the smallest.
[0096] Step S303: Construct the forward-looking road reference plane based on the three target voxel points.
[0097] In step S301 of some embodiments, a voxel is a point existing in the 3D image. The preset target range can be a cube with a side length of N×N×N centered on any voxel in the 3D image. For example, the preset range can be a cube with a side length of 3×3×3 centered on any voxel in the front-view 3D image. The number of non-zero voxels refers to the number of voxels in the 3D space containing the front-view 3D image. The number of non-zero voxels is obtained by calculating the number of non-zero voxels within the target range centered on each voxel in the front-view 3D image.
[0098] In step S302 of some embodiments, when the number of non-zero voxels within the target range of each voxel point in the forward-looking 3D image is obtained, the three voxel points with the smallest number of non-zero voxels are selected from all voxel points as target voxel points.
[0099] It should be noted that, based on the three-dimensional structural features of the forward-looking path, the three target voxel points in this application should be three of the four endpoints of the forward-looking path.
[0100] In step S303 of some embodiments, based on the principle that a plane can be determined by three points, the reference plane of the forward-looking path is determined according to the three target voxel points, and the reference plane of the forward-looking path is obtained.
[0101] In steps S301 to S303 of this embodiment, by calculating the number of non-zero voxels within the target range in the front-view path 3D image, three voxel points are selected from multiple voxel points to construct the front-view path reference plane. This selects a reference plane for the 2D projection of the front-view path 3D image, which can improve the efficiency of 2D projection of the front-view path 3D image, thereby improving the efficiency of front-view path feature extraction.
[0102] In steps S202 to S204 of some embodiments, the normal vector of the front-view reference plane can be calculated by the normal vector calculation formula to obtain the reference plane normal vector. Then, any straight line perpendicular to the reference plane normal vector is selected as the rotation axis of the front-view three-dimensional image. Based on the angle between the rotation axis and the two-dimensional projection plane formed by the first axis and the second axis in the spatial rectangular coordinate system, the front-view rotation matrix of the front-view three-dimensional image is determined. Then, the inverse matrix of the front-view rotation matrix is calculated to obtain the front-view rotation inverse matrix.
[0103] In step S205 of some embodiments, the obtained forward path rotation inverse matrix is used as the difference matrix, and the forward path three-dimensional image is rotated using the spatial difference method to obtain the forward path rotated image.
[0104] In steps S201 to S205 of this embodiment, a front-view path reference plane is constructed for the front-view path 3D image. Then, the reference plane normal vector of the front-view path reference plane is calculated. Based on the reference normal vector and the 2D projection plane, the front-view path rotation matrix of the front-view path 3D image is determined, thereby obtaining the front-view path rotation inverse matrix that can be rotated using the spatial difference method. Finally, the obtained front-view path rotation inverse matrix is used as the difference matrix, and the front-view path 3D image is rotated using the spatial difference method to obtain the front-view path rotated image. This makes it easier to perform 2D projection on the front-view path rotated image, reduces the complexity and difficulty of front-view path feature extraction, and improves the efficiency of front-view path feature extraction.
[0105] In step S103 of some embodiments, the third coordinate of each voxel point in the forward-looking path rotation image is converted to 0, thereby realizing the two-dimensional projection of the forward-looking path rotation image onto the two-dimensional projection plane and obtaining the forward-looking path two-dimensional image. For example, there are voxel points 1 (1, 2, 3) and voxel points (2, 3, 4) in the forward-looking path rotation image. Based on the two-dimensional projection plane formed by the first axis and the second axis, the two-dimensional projection of voxel points 1 and 2 can be obtained as (1, 2) and (2, 3) respectively.
[0106] In step S104 of some embodiments, the skeleton of the front-look path two-dimensional image can be extracted by methods such as refinement iterative algorithm, distance field transformation and central axis change, so as to obtain the front-look path two-dimensional skeleton points of the front-look path two-dimensional image.
[0107] In step S105 of some embodiments, after obtaining the two-dimensional skeleton points of the front-view path two-dimensional image, the skeleton control points in the front-view path rotated image are queried along the third axis in the spatial rectangular coordinate system.
[0108] Please see Figure 4 In some embodiments, step S105 may include, but is not limited to, steps S401 to S402:
[0109] Step S401: Based on the two-dimensional skeleton points and the third axis of the forward-looking path, find the three-dimensional skeleton axis of the forward-looking path from the rotating image of the forward-looking path. The three-dimensional skeleton axis of the forward-looking path is used to characterize the axis that forms the two-dimensional skeleton points of the forward-looking path when projected onto the two-dimensional projection plane in the rotating image of the forward-looking path.
[0110] Step S402: Use the midpoint of the three-dimensional skeleton axis of the forward-looking path as the skeleton control point.
[0111] In step S401 of some embodiments, when the front-view rotation matrix is projected onto the two-dimensional projection plane formed by the first axis and the second axis, vertical projection is used. Therefore, it is only necessary to determine the front-view two-dimensional skeleton points in the front-view two-dimensional image, and then the intersection line segment of the straight line of the third axis where the front-view two-dimensional skeleton points are located and the front-view rotated image can be found along the third axis. This line segment represents the front-view three-dimensional skeleton axis that forms the front-view two-dimensional skeleton points when projected onto the two-dimensional projection plane in the front-view rotated image.
[0112] In step S402 of some embodiments, the midpoint of the three-dimensional skeleton axis of the forward-looking path is calculated based on the length of the three-dimensional skeleton axis of the forward-looking path, and the midpoint is used as the skeleton control point of the forward-looking path rotation matrix.
[0113] In steps S401 to S402 of this embodiment, by finding the three-dimensional skeleton axis of the front-view path along the third axis, and then taking the midpoint of the three-dimensional skeleton axis of the front-view path as the skeleton control point, the position of the two-dimensional skeleton point of the front-view path in the front-view path rotation image can be obtained, thereby realizing the construction of the two-dimensional skeleton of the front-view path into three dimensions.
[0114] In step S106 of some embodiments, the forward-looking path rotated image is an image obtained by rotating the forward-looking path three-dimensional image based on the forward-looking path rotation inverse matrix. Therefore, the coordinates of the skeleton control points in the forward-looking path rotated image in the forward-looking path three-dimensional image can be calculated using the forward-looking path rotation inverse matrix, thereby obtaining the forward-looking path control points.
[0115] In step S107 of some embodiments, after obtaining the control points of the forward-looking path in the three-dimensional image of the forward-looking path, the control points of the forward-looking path are labeled with coordinates according to the constructed spatial rectangular coordinate system to obtain the spatial coordinates of the control points of the forward-looking path. Then, based on the spatial coordinates of the control points of the forward-looking path, the control points of the forward-looking path are curve-fitted to obtain the three-dimensional skeleton image of the forward-looking path.
[0116] Please see Figure 5 In some embodiments, step S107 may include, but is not limited to, steps S501 to S502:
[0117] Step S501: Based on the spatial rectangular coordinate system, mark the coordinates of the forward-looking control points to obtain the coordinates of the control points;
[0118] Step S502: Based on the coordinates of the control points, perform curve fitting on the control points of the forward-looking path to obtain a three-dimensional skeleton image of the forward-looking path.
[0119] In step S501 of some embodiments, coordinates are assigned to the forward-looking control points based on the spatial rectangular coordinate system to obtain the spatial coordinates of each forward-looking control point, i.e., the control point coordinates.
[0120] In step S502 of some embodiments, after obtaining the control point coordinates of each forward-looking control point, curve fitting is performed on the three directions represented by the control point coordinates to obtain three fitted curves. The three fitted curves are then merged to obtain a three-dimensional skeleton image of the forward-looking path.
[0121] Please see Figure 6 In some embodiments, step S502 may include, but is not limited to, steps S601 to S604:
[0122] Step S601: Based on the first value, perform Bézier curve fitting on the forward-looking control point to obtain a first fitting curve, wherein the first fitting curve is used to characterize the fitting curve of the forward-looking control point in the first axis direction.
[0123] Step S602: Based on the second value, perform Bézier curve fitting on the forward look-ahead control point to obtain a second fitting curve, wherein the second fitting curve is used to characterize the fitting curve of the forward look-ahead control point in the second axis direction.
[0124] Step S603: Based on the third value, perform Bézier curve fitting on the forward-looking control point to obtain the third fitting curve, wherein the third fitting curve is used to characterize the fitting curve of the forward-looking control point in the third axis direction.
[0125] Step S604: Merge the first fitting curve, the second fitting curve, and the third fitting curve to obtain the three-dimensional skeleton image of the front-view path.
[0126] In steps S601 to S603 of some embodiments, by performing Bézier curve fitting on the first value in the coordinates of each control point, a first fitting curve for all forward-looking control points in the first axis direction can be obtained. Similarly, by performing Bézier curve fitting on the second value in the coordinates of each control point, a second fitting curve is obtained. By performing Bézier curve fitting on the third value in the coordinates of each control point, a third fitting curve is obtained.
[0127] It is important to know that before performing curve fitting on the forward path control points, the forward path control points need to be numbered and marked to obtain their control point indices. Furthermore, to facilitate Bézier curve fitting on the forward path control points, the control points can be numbered starting from 0. For example, based on the positions of the four forward path control points in the 3D image of the forward path, the four forward path control points can be numbered and marked, resulting in the control point index of the starting forward path control point being 0, the control point index of the second forward path control point being 1, the control point index of the third forward path control point being 2, and the control point index of the last forward path control point being 3.
[0128] Specifically, the expressions for the first, second, and third fitted curves are as follows:
[0129]
[0130] in, This represents the first fitted curve. This represents the second fitted curve. This represents the third fitted curve, where i represents the control point index of the forward-looking control point, n represents the number of forward-looking control points, t represents the parameter value, and x represents the parameter value. i This represents the first value of the control point coordinates for the forward-looking path control point with control point index i, y. i The second value of z represents the coordinates of the control point at control point index i in the forward path. i This represents the third value of the control point coordinates for the forward-looking path control point with control point index i.
[0131] It should be noted that the parameter value t in the first, second, and third fitting curves mentioned above ranges from [0,1].
[0132] In step S604 of some embodiments, by merging the first fitting curve, the second fitting curve and the third fitting curve, the skeleton curve in the front-view path three-dimensional image can be obtained, that is, the front-view path three-dimensional skeleton image.
[0133] It is important to know that by dividing the control points of the anterior visual pathway into the left optic nerve region, right optic nerve region, optic chiasm region, left optic tract region, and right optic tract region, and then using the four endpoints of the control points of the anterior visual pathway as fixed anchor points for curve fitting of the left optic nerve region, right optic nerve region, left optic tract region, and right optic tract region, and using the two endpoints of the optic chiasm region as fixed anchor points for curve fitting of the optic chiasm region, the three-dimensional skeleton image of the anterior visual pathway can be made to better fit the actual anterior visual pathway skeleton structure.
[0134] In steps S601 to S604 of this embodiment, Bézier curves are fitted to the first, second, and third values of each control point coordinate to obtain a first fitted curve, a second fitted curve, and a third fitted curve. The first fitted curve, the second fitted curve, and the third fitted curve are then merged to obtain a three-dimensional skeleton image of the forward-looking path. This achieves the construction of the three-dimensional skeleton of the forward-looking path, making the structural features of the forward-looking path clearer and more consistent with the actual skeleton structure of the forward-looking path. This improves the effectiveness of the extracted forward-looking path features and the accuracy of the calculated structural features of the forward-looking path.
[0135] In steps S501 to S502 of this embodiment, the coordinates of the forward-looking control points are marked based on the constructed spatial rectangular coordinate system to obtain the control point coordinates of each forward-looking control point. Then, based on the control point coordinates, curve fitting is performed on the forward-looking control points in the three-dimensional image of the forward-looking path to obtain a three-dimensional skeleton image of the forward-looking path that fits the actual situation, thereby improving the accuracy of calculating the structural features of the forward-looking path.
[0136] In step S108 of some embodiments, after obtaining the three-dimensional skeleton image of the front-view path, the three-dimensional coordinates of each skeleton voxel point in the three-dimensional skeleton image of the front-view path can be generated according to the constructed spatial rectangular coordinate system. Then, based on the three-dimensional coordinates of each point, feature extraction can be performed on the three-dimensional skeleton image of the front-view path to obtain feature data of the three-dimensional skeleton image of the front-view path. The skeleton voxel points represent the voxel points in the three-dimensional skeleton image of the front-view path.
[0137] It is important to know that the feature data of the 3D image of the forward-looking path includes the transverse diameter of the forward-looking path and the length of the forward-looking path.
[0138] Please see Figure 7 In some embodiments, step S108 includes, but is not limited to, steps S701 to S704:
[0139] Step S701: Generate the three-dimensional spatial coordinates of each skeleton voxel point based on the spatial rectangular coordinate system;
[0140] Step S702: Based on the three-dimensional spatial coordinates, calculate the transverse diameter of the three-dimensional skeleton image of the forward-looking path to obtain the transverse diameter of the forward-looking path.
[0141] Step S703: Based on the three-dimensional spatial coordinates, calculate the path length of the three-dimensional skeleton image of the forward path to obtain the forward path length;
[0142] Step S704: Combine the transverse diameter and length of the forward-looking path to obtain feature data.
[0143] In step S701 of some embodiments, coordinates are assigned to the skeleton voxels in the three-dimensional skeleton image of the forward-looking path according to the constructed spatial rectangular coordinate system, so as to obtain the three-dimensional coordinates of the skeleton voxels, that is, the three-dimensional spatial coordinates.
[0144] In steps S702 and S703 of some embodiments, the transverse diameter refers to the length of the cross-section of the forward view path. The view path length refers to the length of the selected view segment in the forward view path. Based on the three-dimensional spatial coordinates of each skeleton voxel point, the transverse diameter and view path length of the three-dimensional skeleton image of the forward view path can be calculated, thereby obtaining the transverse diameter and forward view path length of the forward view path.
[0145] It is important to know that, in order to facilitate the calculation of the transverse diameter of the selected cross section and the length of the selected view segment in the 3D skeleton image of the forward path, it is necessary to number and label the skeleton voxels to obtain voxel indices. Since the skeleton voxels are 3D spatial points, they can be numbered and labeled horizontally according to the structural characteristics of 3D space and the constructed 3D Cartesian coordinate system to obtain horizontal voxel indices, and vertically according to the vertical plane to obtain vertical voxel indices. These horizontal and vertical voxel indices together constitute the voxel index. In this context, the horizontal plane refers to the plane formed by the first axis and the second axis, and the vertical plane refers to the plane formed by the first axis and the third axis. For example, if the voxel index of skeleton voxel point 1 with three-dimensional spatial coordinates (1,1,2) is (1,2), and 1 in the voxel index is the vertical voxel index of skeleton voxel point 1, and 2 in the voxel index is the horizontal voxel index of skeleton voxel point 1, then the voxel index of skeleton voxel point 2 with three-dimensional spatial coordinates (2,1,3) can be (1,3), and the voxel index of skeleton voxel point 3 with three-dimensional spatial coordinates (2,2,2) can be (2,2).
[0146] Specifically, the forward sight diameter and forward sight length can be calculated using the following formulas:
[0147]
[0148] Where r represents the transverse diameter of the forward look-through path, l represents the length of the forward look-through path, n represents the number of skeleton voxels in the cross-section of the forward look-through path in the 3D skeleton image, i represents the vertical voxel index of the skeleton voxel, j represents the horizontal voxel index of the skeleton voxel, and t i,jp represents the 3D spatial coordinates of a skeleton voxel point with voxel vertical index i and voxel horizontal index j. j The coordinates of the control points of the forward-looking path control points in the cross-section of the forward-looking path in the 3D skeleton image of the forward-looking path are indicated by m, which represents the number of forward-looking path control points in the selected view segment of the forward-looking path in the 3D skeleton image of the forward-looking path. i,j This represents the coordinates of the control point of the forward path control point with the horizontal index j on the cross section with the vertical index i of the voxel.
[0149] In step S704 of some embodiments, feature data of the three-dimensional image of the forward-looking path can be obtained by combining the transverse diameter and length of the forward-looking path.
[0150] In steps S701 to S704 of this embodiment, feature extraction is performed on the three-dimensional skeleton image of the front-view path. Compared with feature extraction directly on the two-dimensional skeleton image of the front-view path, the accuracy of the extracted front-view path features can be improved.
[0151] This application obtains a rotated forward visual path image by rotating a three-dimensional forward visual path image obtained from a segmented brain image. This rotated image is then projected onto a two-dimensional projection plane formed by the first and second axes. This improves the ease of projecting the three-dimensional forward visual path image onto a two-dimensional plane, thereby increasing the efficiency of forward visual path feature extraction. Furthermore, based on the two-dimensional skeleton points of the forward visual path in the two-dimensional image and the third axis of the constructed spatial rectangular coordinate system, the corresponding skeleton control points in the rotated forward visual path image are found. These control points are then rotated and restored to obtain the positions of the forward visual path control points in the three-dimensional image. Finally, the forward visual path control points are fitted to obtain a three-dimensional skeleton image of the forward visual path. This achieves skeleton extraction from the three-dimensional forward visual path image, avoiding the direct extraction of forward visual path structural features from the two-dimensional image, thus improving the accuracy of calculating forward visual path structural features.
[0152] Please see Figure 8 This application also provides a front-look path feature extraction device that can implement the above-described front-look path feature extraction method. The device includes:
[0153] The forward vision path cutting module 801 is used to acquire brain images, perform image cutting on the brain images to obtain a three-dimensional image of the forward vision path, and construct a spatial rectangular coordinate system, wherein the spatial rectangular coordinate system includes a first axis, a second axis and a third axis, and the first axis, the second axis and the third axis are perpendicular to each other;
[0154] The forward path rotation module 802 is used to rotate the three-dimensional image of the forward path based on the two-dimensional projection plane formed by the first axis and the second axis to obtain the forward path rotated image.
[0155] The forward path projection module 803 is used to perform two-dimensional projection on the forward path rotation image based on the two-dimensional projection plane to obtain a two-dimensional image of the forward path.
[0156] The front-view path 2D skeletonization module 804 is used to extract the skeleton from the front-view path 2D image to obtain the front-view path 2D skeleton points.
[0157] The skeleton control point search module 805 is used to determine the skeleton control points in the rotating image of the front view based on the two-dimensional skeleton points of the front view path and the third axis.
[0158] The control point recovery module 806 is used to recover the coordinates of the skeleton control points to obtain the forward-looking path control points;
[0159] The 3D skeleton fitting module 807 is used to fit the control points of the forward-looking path based on the spatial rectangular coordinate system to obtain the 3D skeleton image of the forward-looking path.
[0160] The forward-looking path feature extraction module 808 is used to extract features from the three-dimensional skeleton image of the forward-looking path based on the spatial rectangular coordinate system, so as to obtain the feature data of the three-dimensional image of the forward-looking path.
[0161] The specific implementation of this forward-looking path feature extraction device is basically the same as the specific implementation of the aforementioned forward-looking path feature extraction method, and will not be repeated here.
[0162] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned forward-looking path feature extraction method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.
[0163] Please see Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:
[0164] The processor 901 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.
[0165] The memory 902 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902 and is called and executed by the processor 901 using the front-look-through feature extraction method of the embodiments of this application.
[0166] The input / output interface 903 is used to implement information input and output;
[0167] The communication interface 904 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).
[0168] Bus 905 transmits information between various components of the device (e.g., processor 901, memory 902, input / output interface 903, and communication interface 904);
[0169] The processor 901, memory 902, input / output interface 903, and communication interface 904 are connected to each other within the device via bus 905.
[0170] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described forward path feature extraction method.
[0171] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
[0172] The anterior visual path feature extraction method, device, electronic device, and storage medium provided in this application embodiment rotate a three-dimensional anterior visual path image obtained by cutting out an image from a brain image to obtain a rotated anterior visual path image. This rotated anterior visual path image is then projected onto a two-dimensional projection plane formed by the first and second axes. This improves the convenience of projecting the three-dimensional anterior visual path image onto a two-dimensional plane, thereby increasing the efficiency of anterior visual path feature extraction. Furthermore, based on the two-dimensional skeleton points of the anterior visual path in the two-dimensional anterior visual path image and the third axis of the constructed spatial rectangular coordinate system, the corresponding skeleton control points in the rotated anterior visual path image are found. These skeleton control points are then rotated and restored to obtain the positions of the anterior visual path control points in the three-dimensional anterior visual path image. Finally, the anterior visual path control points are fitted to obtain a three-dimensional anterior visual path skeleton image. This achieves skeleton extraction from the three-dimensional anterior visual path image, avoiding the direct extraction of anterior visual path structural features from the two-dimensional anterior visual path image, thus improving the accuracy of calculating anterior visual path structural features.
[0173] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.
[0174] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.
[0175] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.
[0176] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.
[0177] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0178] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.
[0179] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. The coupling or direct coupling or communication connection between the shown or discussed units may be through some interfaces, or indirect coupling or communication connection between the apparatus or units, and may be electrical, mechanical, or other forms.
[0180] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0181] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0182] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0183] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.
Claims
1. A method for extracting features of a forward-looking path, characterized in that, The method includes: A brain image is acquired, and the brain image is segmented to obtain a three-dimensional image of the forward visual path. A spatial rectangular coordinate system is constructed, wherein the spatial rectangular coordinate system includes a first axis, a second axis, and a third axis, and the first axis, the second axis, and the third axis are perpendicular to each other. Based on the two-dimensional projection plane formed by the first axis and the second axis, the three-dimensional image of the front view path is rotated to obtain a rotated image of the front view path. Based on the two-dimensional projection surface, the rotating image of the forward path is projected in two dimensions to obtain a two-dimensional image of the forward path; Skeleton extraction is performed on the two-dimensional image of the forward-looking path to obtain two-dimensional skeleton points of the forward-looking path; Based on the two-dimensional skeleton points of the forward-looking path and the third axis, determine the skeleton control points in the forward-looking path rotation image; The coordinates of the skeleton control points are restored to obtain the forward path control points; Based on the spatial rectangular coordinate system, the forward-looking path control points are fitted to obtain a three-dimensional skeleton image of the forward-looking path; Based on the spatial rectangular coordinate system, feature extraction is performed on the three-dimensional skeleton image of the forward-looking path to obtain the feature data of the three-dimensional image of the forward-looking path.
2. The method according to claim 1, characterized in that, The process of rotating the three-dimensional image of the forward-looking path based on the two-dimensional projection plane formed by the first axis and the second axis to obtain a rotated image of the forward-looking path includes: Construct the reference plane of the three-dimensional image of the forward-looking path to obtain the forward-looking path reference plane; Calculate the normal vector of the forward-looking reference plane to obtain the reference plane normal vector; Based on the reference plane normal vector and the two-dimensional projection plane, the forward path rotation matrix of the forward path three-dimensional image is determined; Calculate the inverse of the forward path rotation matrix to obtain the forward path rotation inverse matrix; Based on the inverse rotation matrix of the forward view path, the three-dimensional image of the forward view path is rotated to obtain the rotated image of the forward view path.
3. The method according to claim 2, characterized in that, The forward-looking path 3D image includes multiple voxel points, and the reference plane for constructing the forward-looking path 3D image, to obtain the forward-looking path reference plane, includes: Calculate the number of non-zero voxels within a preset target range centered on each voxel point; Based on the number of non-zero voxels, three target voxels are selected from the plurality of voxel points, wherein the number of non-zero voxels within the target range of the target voxel points is the smallest; Based on the three target voxel points, the forward-looking path reference plane is constructed.
4. The method according to claim 1, characterized in that, The step of determining the skeleton control points in the front-view rotation image based on the two-dimensional skeleton points of the front-view path and the third axis includes: Based on the two-dimensional skeleton points of the forward-looking path and the third axis, the three-dimensional skeleton axis of the forward-looking path is found from the rotated image of the forward-looking path. The three-dimensional skeleton axis of the forward-looking path is used to characterize the axis that forms the two-dimensional skeleton points of the forward-looking path when projected onto the two-dimensional projection plane in the rotated image of the forward-looking path. The midpoint of the three-dimensional skeleton axis of the forward-looking path is used as the skeleton control point.
5. The method according to claim 1, characterized in that, The process of fitting the forward-looking path control points based on the spatial rectangular coordinate system to obtain a three-dimensional skeleton image of the forward-looking path includes: Based on the aforementioned spatial rectangular coordinate system, the coordinates of the forward-looking path control points are marked to obtain the control point coordinates; Based on the coordinates of the control points, curve fitting is performed on the forward-looking path control points to obtain a three-dimensional skeleton image of the forward-looking path.
6. The method according to claim 5, characterized in that, The control point coordinates include a first value, a second value, and a third value. The step of performing curve fitting on the forward-looking path control points based on the control point coordinates to obtain the three-dimensional skeleton image of the forward-looking path includes: Based on the first value, a Bezier curve is fitted to the forward look-ahead control point to obtain a first fitting curve, wherein the first fitting curve is used to characterize the fitting curve of the forward look-ahead control point in the first axis direction. Based on the second value, a Bezier curve is fitted to the forward look-ahead control point to obtain a second fitting curve, wherein the second fitting curve is used to characterize the fitting curve of the forward look-ahead control point in the second axis direction. Based on the third value, a Bezier curve is fitted to the forward-looking control point to obtain a third fitting curve, wherein the third fitting curve is used to characterize the fitting curve of the forward-looking control point in the third axis direction. The first fitting curve, the second fitting curve, and the third fitting curve are combined to obtain the three-dimensional skeleton image of the front-view path.
7. The method according to any one of claims 1, characterized in that, The forward-looking path 3D skeleton image includes multiple skeleton voxels. Based on the spatial Cartesian coordinate system, feature extraction is performed on the forward-looking path 3D skeleton image to obtain feature data of the forward-looking path 3D image, including: Based on the aforementioned spatial rectangular coordinate system, generate the three-dimensional spatial coordinates of each of the skeleton voxels; Based on the three-dimensional spatial coordinates, the horizontal diameter of the three-dimensional skeleton image of the forward-looking path is calculated to obtain the horizontal diameter of the forward-looking path; Based on the three-dimensional spatial coordinates, the view path length of the three-dimensional skeleton image of the forward view path is calculated to obtain the forward view path length; The feature data is obtained by combining the transverse diameter of the forward-looking path and the length of the forward-looking path.
8. A forward-looking path feature extraction device, characterized in that, The device includes: The forward visual path segmentation module is used to acquire brain images, segment the brain images to obtain a three-dimensional forward visual path image, and construct a spatial rectangular coordinate system, wherein the spatial rectangular coordinate system includes a first axis, a second axis, and a third axis, and the first axis, the second axis, and the third axis are perpendicular to each other. A front-view rotation module is used to rotate the front-view three-dimensional image based on a two-dimensional projection plane formed by the first axis and the second axis to obtain a front-view rotated image. A forward-looking path projection module is used to project the forward-looking path rotated image into a two-dimensional image based on the two-dimensional projection surface to obtain a forward-looking path two-dimensional image. The forward-looking path 2D skeletonization module is used to extract the skeleton from the forward-looking path 2D image to obtain the forward-looking path 2D skeleton points. The skeleton control point finding module is used to determine the skeleton control points in the front-view path rotation image based on the two-dimensional skeleton points of the front-view path and the third axis. The control point recovery module is used to recover the coordinates of the skeleton control points to obtain the forward-looking path control points; The three-dimensional skeleton fitting module is used to fit the forward-looking path control points based on the spatial rectangular coordinate system to obtain a three-dimensional skeleton image of the forward-looking path. The forward-looking path feature extraction module is used to extract features from the three-dimensional skeleton image of the forward-looking path based on the spatial rectangular coordinate system, so as to obtain the feature data of the three-dimensional image of the forward-looking path.
9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the forward-looking path feature extraction method according to any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the forward path feature extraction method according to any one of claims 1 to 7.