Point cloud generation method and device, electronic equipment and computer readable storage medium
By filtering and edge detection in medical image data to generate point clouds, the problem of unintuitive medical image data is solved, and the accurate presentation and intuitive observation of target objects are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JIANGYU KANGJIAN INNOVATION MEDICAL TECH CHENGDU CO LTD
- Filing Date
- 2024-12-12
- Publication Date
- 2026-06-19
Smart Images
Figure CN122244284A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of image processing, specifically to a point cloud generation method, apparatus, electronic device, and computer-readable storage medium. Background Technology
[0002] Medical imaging data is data obtained by imaging certain structures and functions of the human body using various medical imaging techniques. This data can present information about the inside of the human body in the form of images.
[0003] However, medical images such as computed tomography (CT) and magnetic resonance imaging (MRI) are mostly two-dimensional planar images, which are not intuitive. Summary of the Invention
[0004] In view of the above, embodiments of this application provide a point cloud generation method, apparatus, electronic device, and computer-readable storage medium that can intuitively represent certain structures of the human body.
[0005] In a first aspect, embodiments of this application provide a point cloud generation method, including: Acquire medical imaging data of the target object; In the medical image data, voxels of the target object are filtered based on Henle values to obtain a first voxel set; Edge detection is performed on the medical image data to obtain a second voxel set, which includes voxels at the contour of the target object. The point cloud of the target object is generated based on the first voxel set and the second voxel set.
[0006] After acquiring medical image data of the target object, this embodiment of the application filters out voxels of the target object to form a first voxel set. This helps to discard voxel information such as background tissues and surrounding environment that are irrelevant to the target object in the image data. This embodiment of the application also performs edge detection on the medical image data to obtain a second voxel set, which covers voxels at the outline of the target object. This makes the boundary of the target object clearly presented. Then, this embodiment of the application generates a point cloud based on the first voxel set and the second voxel set, which can accurately focus on the target object. This allows the generated point cloud to accurately present the shape of the target object, thereby enabling the user to intuitively observe the spatial structure and shape of the target object.
[0007] In some embodiments, the voxels of the target object in the medical image data are filtered based on the Henle value to obtain a first voxel set, including: Obtain the Henle value of each voxel in the medical image data; Among the voxels of the medical image data, voxels whose Henle values are within a preset Henle value range are selected to obtain the first voxel set. The preset Heinz value range is the Heinz value range of the target object voxels.
[0008] In some embodiments, edge detection is performed on the medical image data to obtain a second voxel set, including: The medical image data is sliced along the imaging plane of the medical image data to obtain multiple two-dimensional image data. Edge detection is performed on the multiple two-dimensional image data to obtain the edge regions of the multiple two-dimensional image data; The edge regions of the multiple two-dimensional image data are stacked to obtain the second voxel set.
[0009] In some embodiments, edge detection is performed on the plurality of two-dimensional image data to obtain the edge regions of the plurality of two-dimensional image data, including: The pixels of the two-dimensional image data are convolved with a preset horizontal convolution kernel to obtain the horizontal gradient of the two-dimensional image data. The pixels of the two-dimensional image data are convolved with a preset vertical convolution kernel to obtain the vertical gradient of the two-dimensional image data. The gradient magnitude of each pixel in the two-dimensional image data is determined based on the horizontal gradient and the vertical gradient. Based on the gradient magnitude, pixels located in the edge region are selected from the two-dimensional image data to obtain the edge region of the two-dimensional image data.
[0010] In some embodiments, generating a point cloud of the target object based on the first voxel set and the second voxel set includes: The intersection is obtained by performing a logical AND operation on the first voxel set and the second voxel set. A point cloud of the target object is generated based on the intersection.
[0011] In some embodiments, performing a logical AND operation on the first voxel set and the second voxel set to obtain the intersection includes: The first voxel set is morphologically dilated along the depth axis of the medical image data to obtain a third voxel set. The second voxel set is morphologically dilated along the depth axis to obtain the fourth voxel set; Perform a logical AND operation on the third voxel set and the fourth voxel set to obtain the intersection.
[0012] In some embodiments, generating a point cloud of the target object based on the intersection includes: In the intersection, connected regions formed by interconnected voxels are identified; The point cloud of the target object is generated based on the voxels that constitute the largest connected region.
[0013] Secondly, embodiments of this application provide a point cloud generation apparatus, comprising: The data acquisition module is used to acquire medical image data of the target object; A voxel filtering module is used to filter voxels of the target object in the medical image data based on the Henle value to obtain a first voxel set; An edge detection module is used to perform edge detection on the medical image data to obtain a second voxel set, the second voxel set including voxels at the contour of the target object; The point cloud of the target object is generated based on the first voxel set and the second voxel set.
[0014] Thirdly, embodiments of this application provide an electronic device, which includes a processor and a memory. The memory is used to store instructions, and the processor is used to call the instructions in the memory to cause the electronic device to execute the point cloud generation method as described in the first aspect.
[0015] Fourthly, embodiments of this application provide a computer-readable storage medium that stores computer instructions that, when executed on an electronic device, cause the electronic device to perform the point cloud generation method as described in the first aspect. Attached Figure Description
[0016] Figure 1 A flowchart illustrating the steps of a point cloud generation method provided in an embodiment of this application.
[0017] Figure 2 This is a schematic diagram of voxel resampling provided in an embodiment of this application.
[0018] Figure 3 A flowchart of the sub-steps of step 103 provided in an embodiment of this application.
[0019] Figure 4 A three-dimensional coordinate diagram of medical imaging data provided in an embodiment of this application.
[0020] Figure 5 This is a schematic diagram of the edge detection results of a two-dimensional breast image provided in an embodiment of this application.
[0021] Figure 6This is a schematic diagram of a scenario for a point cloud generation method provided in an embodiment of this application.
[0022] Figure 7 This is a schematic diagram of the structure of a point cloud generation device provided in an embodiment of this application.
[0023] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0024] To better understand the above-mentioned objectives, features, and advantages of this application, the application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.
[0025] The following description sets forth many specific details to provide a full understanding of this application. The described embodiments are only some, not all, of the embodiments of this application.
[0026] 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 in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application.
[0027] It should be further noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0028] In this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and drawings of this application are used to distinguish similar objects, not to describe a specific order or sequence.
[0029] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.
[0030] The following describes the terminology used in the embodiments of this application: A voxel is short for volume pixel. A volume pixel is the smallest unit of image data segmentation in three-dimensional space and is used in fields such as 3D imaging, scientific data, and medical imaging.
[0031] Thoracic magnetic resonance imaging (MRI) is an examination technique that uses the magnetic resonance phenomenon to generate detailed images of the interior of the thoracic cavity. Thoracic MRI images, produced using thoracic MRI, can display information such as the thoracic bones, lungs, the examination table of the MRI scanner, and the outline of the breast.
[0032] The Henness value is a relative value used to measure the degree to which human tissue absorbs X-rays, reflecting the density differences between different tissues. The unit of the Henness value is the Hounsfield Unit (HU), which is the quantitative scale unit for radiometric absorption in the Hounsfield scale. The Henness value is commonly used in X-ray computed tomography (CT) scans and is therefore also known as the CT number.
[0033] Digital Imaging and Communications in Medicine (DICOM) is a standard for medical imaging and related information, describing medical image formats that can be used for data exchange to meet clinical needs. DICOM is widely used in radiology, cardiovascular imaging, and radiological diagnostic equipment.
[0034] The DICOM standard covers multiple aspects, including data format, information model, and communication protocol. Regarding data format, it specifies the storage structure of medical imaging data, such as the arrangement of pixel data and the format of header information (including patient information, device information, imaging parameters, etc.). In terms of information model, it defines the relationships between various entities in medical imaging data (such as patients, examinations, sequences, images, etc.). The communication protocol enables the secure and efficient transmission of medical imaging data between different devices.
[0035] Medical image data using the DICOM standard can be viewed as a collection that includes different types of data. For example, computed tomography (CT) images include image information but lack scanner parameter settings. DICOM can supplement CT images by adding these additional information through tags, making the CT images more complete.
[0036] A Modality Lookup Table (Modality LUT) is a lookup table in the DICOM specification used to convert digital signal values (such as raw data from imaging device detectors) into visually displayable image pixel values or as a reference table for interpreting image data. It is closely related to different imaging modalities (such as CT, MRI, and ultrasound) and defines how image data is displayed and its features are interpreted in each modality. For example, it converts data generated by different imaging devices such as CT, MRI, and ultrasound into a standard output range, which is typically in physically meaningful units, such as Henry's units or radiometric intensity, thus helping to more accurately reflect the data information captured by the imaging device.
[0037] Medical imaging data is data obtained by imaging certain structures and functions of the human body using various medical imaging techniques. This data can present information about the internal structure of the human body in the form of images. For example, medical imaging data can include CT images and MRI images. However, most of the above-mentioned medical imaging data are two-dimensional planar images, which is not a very intuitive way of presenting them.
[0038] For example, when a breast cancer patient needs to have breast cancer cells removed, the shape of the breast will change. However, chest MRI images are difficult to show the shape of the breast. Doctors or patients need to construct the shape of the breast in their minds to imagine the changes in the shape of the breast before and after the removal of breast cancer cells, which is not intuitive.
[0039] In view of the above, embodiments of this application provide a point cloud generation method, apparatus, electronic device, and computer-readable storage medium.
[0040] The point cloud generation method of this application may include: acquiring medical image data of a target object; filtering voxels of the target object in the medical image data based on the Heinz value to obtain a first voxel set; performing edge detection on the medical image data to obtain a second voxel set, the second voxel set including voxels at the contour of the target object; and generating a point cloud of the target object based on the first voxel set and the second voxel set.
[0041] After acquiring medical image data of the target object, this embodiment of the application filters out voxels of the target object to form a first voxel set. This helps to discard voxel information such as background tissues and surrounding environment that are irrelevant to the target object in the image data. This embodiment of the application also performs edge detection on the medical image data to obtain a second voxel set, which covers voxels at the outline of the target object. This makes the boundary of the target object clearly presented. Then, this embodiment of the application generates a point cloud based on the first voxel set and the second voxel set, which can accurately focus on the target object. This allows the generated point cloud to accurately present the shape of the target object, thereby enabling the user to intuitively observe the spatial structure and shape of the target object.
[0042] For example, electronic devices can convert voxel data from chest MRI images into point clouds, which can then show patients the changes in breast shape after cutting out breast cancer cells.
[0043] The point cloud generation method of this application can be applied to one or more electronic devices. The electronic device is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions. Its hardware includes, but is not limited to, processors, microprogrammed control units (MCUs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc. The electronic device can be a medical device, a personal computer, a server, etc., but is not limited to these.
[0044] Figure 1 This is a flowchart illustrating the steps of an embodiment of the point cloud generation method of this application. Depending on different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted.
[0045] See Figure 1 As shown, the point cloud generation method may include the following steps.
[0046] Step 101: Obtain medical image data of the target object.
[0047] The target object is a human body part whose appearance is to be represented in point cloud form. For example, the target object can be a breast, a face, teeth, etc., but is not limited to these.
[0048] Medical imaging data is data obtained by imaging certain structures and functions of the human body using medical imaging techniques. When the target object is the breast, the medical imaging data may include images from a thoracic magnetic resonance imaging (MRI) scan.
[0049] Medical imaging data can include voxels, which are used to describe information about human tissue in three-dimensional space. A voxel has a certain spatial location and size, and includes relevant physical quantities or characteristic information of the tissue at that location.
[0050] Medical imaging data can be processed using the DICOM standard.
[0051] Step 102: In the medical image data, voxels of the target object are screened based on the Henle value to obtain the first voxel set.
[0052] In some embodiments, step 102 can be implemented in the following manner: First, the electronic device acquires the Henle values of each voxel in the medical image data.
[0053] For example, medical imaging data can use the DICOM standard. Medical imaging data using the DICOM standard will be referred to as DICOM data below. DICOM data includes the rescale slope attribute and rescale intercept attribute for each voxel.
[0054] After acquiring DICOM data, the electronic device can calculate the Henle value of each voxel in the medical image data based on the rescaling slope and rescaling intercept attributes. The calculation formula is shown below: output units= m×stored value+b; Wherein, output units refer to the Heinz value of a voxel, m refers to the rescaling slope attribute, b refers to the rescaling intercept attribute, and stored value refers to the original data value of the voxel, that is, the stored value comes from the original value acquired by the imaging device and stored in the DICOM data.
[0055] In some embodiments, the electronic device may also resample the voxel data according to the voxel dimensions recorded in the DICOM data, so that the size of each voxel is a preset size.
[0056] In DICOM data, voxel size is represented by the slice thickness attribute and the pixel spacing attribute. The slice thickness attribute describes the thickness of the slice, and the pixel spacing attribute describes the distance between pixels in all directions.
[0057] Assuming that the layer thickness and pixel pitch attributes in the DICOM data are both 2, that is, the voxel size is 2x2x2, in Figure 2 As shown, cube 21 is composed of eight 2×2×2 voxels 211. During the resampling process, the electronic device can resample the eight 2x2x2 voxels 211 using an interpolation algorithm, such as nearest neighbor interpolation, to obtain cube 22, which is composed of 64 1x1x1 voxels 221.
[0058] Since different imaging instruments may produce differences when scanning human tissue due to various parameter settings, the embodiments of this application resample voxels to convert DICOM data acquired under different parameter settings into new data with a unified standard, which helps to eliminate the differences in DICOM data caused by different parameter settings.
[0059] Furthermore, the embodiments of this application can make the voxel distribution more uniform through resampling, which is beneficial to retain the positional information between points in the subsequently generated point cloud and avoid point cloud collapse on the depth axis.
[0060] That is, in the embodiments of this application, after the electronic device acquires medical image data, it can calculate the Heinz value of each voxel in the medical image data and resample the voxels in the medical image data so that the size of each voxel meets the preset size.
[0061] Then, the electronic device can filter voxels from the voxels of the medical image data whose Henle values are within a preset Henle value range to obtain the first voxel set; wherein, the preset Henle value range is the Henle value range of the target object voxels.
[0062] For example, when the target object is a breast, which is mostly soft tissue such as skin, the range of the Heinz value can be between -50HU and -150HU. The electronic device can filter voxels with Heinz values between -50HU and -150HU from the voxels of the medical image data and add the voxel to the first voxel set.
[0063] Step 103: Perform edge detection on the medical image data to obtain the second voxel set.
[0064] The second set of voxels includes voxels at the outline of the target object.
[0065] In some embodiments, step 103 can be implemented in the following manner: the electronic device can input medical image data into a pre-trained edge detection model to obtain a second voxel set.
[0066] In other embodiments, reference is made to... Figure 3 As shown, the electronic device can also perform edge detection by executing the following steps 1031 to 1033 to obtain a second voxel set: Step 1031: Slice the medical image data along the imaging plane to obtain multiple two-dimensional image data.
[0067] In the three-dimensional spatial representation of medical images, the three coordinate axes include the horizontal axis (also called the transverse axis or X-axis), the vertical axis (also called the longitudinal axis or Y-axis), and the depth axis (also called the Z-axis). The imaging plane is the plane formed by the horizontal and vertical axes. This plane is generally perpendicular to the depth axis.
[0068] refer to Figure 4 As shown, in DICOM data, the horizontal axis (X-axis) is usually oriented from the right side of the human body to the left; the vertical axis (Y-axis) is oriented from the front to the back of the human body; and the depth axis (Z-axis) is oriented from the feet to the head of the human body.
[0069] Step 1032: Perform edge detection on multiple two-dimensional image data to obtain the edge regions of multiple two-dimensional image data.
[0070] In some embodiments, the electronic device can perform edge detection on each two-dimensional image data using edge detection algorithms such as the Sobel operator, Prewitt operator, Canny operator, or Laplacian operator.
[0071] The following example uses the Sobel operator to perform edge detection on a two-dimensional image dataset to illustrate the edge detection process.
[0072] Specifically, referring to steps a1 to a4 below, the edge detection process includes: Step a1: Perform convolution operation between the pixels of the two-dimensional image data and a preset horizontal convolution kernel to obtain the horizontal gradient of the two-dimensional image data.
[0073] For example, the horizontal convolution kernel can be denoted as... , It can be as follows: .
[0074] The above The value is only an example, and the actual application process can be set according to the requirements. This application embodiment does not limit this.
[0075] Electronic devices can compare the pixel values of two-dimensional image data pixels with... Perform convolution operations to obtain the horizontal gradient. .
[0076] Step a2: Perform convolution operation between the pixels of the two-dimensional image data and a preset vertical convolution kernel to obtain the vertical gradient of the two-dimensional image data.
[0077] For example, the vertical convolution kernel can be denoted as... , It can be as follows: .
[0078] The above The value is only an example, and the actual application process can be set according to the requirements. This application embodiment does not limit this.
[0079] Electronic devices can compare the pixel values of two-dimensional image data pixels with... Perform convolution operations to obtain the gradient in the vertical direction. .
[0080] Step a3: Determine the gradient magnitude of each pixel in the two-dimensional image data based on the horizontal gradient and the vertical gradient.
[0081] In some embodiments, the square root of the sum of the squares of the horizontal and vertical gradients can be taken to obtain the gradient magnitude of the pixel. For example, the following formula can be used as an example: ; in, This represents the gradient magnitude of a pixel. This represents the gradient in the horizontal direction of that pixel. This represents the gradient magnitude in the vertical direction of the pixel.
[0082] Step a4: Based on the gradient magnitude, filter pixels in the edge region of the two-dimensional image data to obtain the edge region of the two-dimensional image data.
[0083] In some embodiments, the electronic device can standardize the gradient magnitude of a pixel and map the gradient magnitude to a preset interval to obtain a standardized gradient; if the standardized gradient falls within a preset edge region gradient range, it indicates that the pixel belongs to the edge region.
[0084] For example, the preset range can be 0 to 1, and the edge gradient range can be 0.1 to 1. When the normalized gradient is greater than 0.1, it means that the pixel belongs to the edge region. If the normalized gradient is less than 0.1, it means that the pixel does not belong to the edge region.
[0085] The above method can be used to find all pixels belonging to the edge region, thus obtaining the edge region of the two-dimensional image data. For example, refer to... Figure 5 As shown, Figure 5 The edge detection is performed based on two-dimensional image data, which is generated from image data slices based on thoracic magnetic resonance imaging.
[0086] The above steps a1 to a4 use the Sobel operator as an example to illustrate the process of edge detection for two-dimensional image data. Electronic devices can also use other edge detection methods, and this application embodiment does not limit this.
[0087] After acquiring the edge regions of multiple two-dimensional image data, step 1033 can be executed. Step 1033: Stack the edge regions of multiple two-dimensional image data to obtain a second voxel set.
[0088] For example, the edge regions of two-dimensional image data are stacked in slice order to obtain a second voxel set, which includes voxels at the outline of the target object, such as voxels at the outline of the breast.
[0089] Step 104: Generate the point cloud of the target object based on the first voxel set and the second voxel set.
[0090] In some embodiments, step 104 can be implemented as follows: the electronic device performs a logical AND operation on the first voxel set and the second voxel set to obtain an intersection; and generates a point cloud of the target object based on the intersection.
[0091] The embodiments of this application generate point clouds of target objects based on the intersection of the first voxel set and the second voxel set, which means that only voxels that meet the following two conditions will be included in the point cloud construction.
[0092] Condition 1: The voxel is located within the target object itself.
[0093] Condition 2: A voxel located on the outline of the target object.
[0094] In other words, the electronic device performs multiple rigorous screenings to further remove voxels that are suspected to belong to the target object but are not on its outline edge, as well as voxels that are on the suspected edge but do not actually belong to the target object. For example, voxels corresponding to the examination table of the magnetic array imaging machine can be removed, so that the final generated point cloud is highly focused on the most critical and representative part of the target object, which greatly improves the accuracy and effectiveness of the point cloud data.
[0095] Furthermore, in some embodiments, performing a logical AND operation on the first voxel set and the second voxel set to obtain the intersection may include: performing a logical AND operation on the first voxel set and the second voxel set to obtain the intersection.
[0096] In other embodiments, the electronic device performs a logical AND operation on the first voxel set and the second voxel set to obtain the intersection. This may include: performing morphological dilation on the first voxel set along the depth axis of the medical image data to obtain a third voxel set; performing morphological dilation on the second voxel set along the depth axis to obtain a fourth voxel set; and then performing a logical AND operation on the third voxel set and the fourth voxel set to obtain the intersection.
[0097] In the three-dimensional spatial representation of medical images, the depth axis is one of the three coordinate axes (typically including the horizontal, vertical, and depth axes). It is used to represent the spatial position of an imaged object (such as human tissue or organs) within a specific orientation of the imaging device; for example, as a reference... Figure 4 As shown, the depth axis is typically oriented from the feet towards the head.
[0098] Due to edge blurring or noise in medical image data, the target object region presented by the first voxel set and the edge contour region presented by the second voxel set may be disconnected. That is, gaps or holes may appear in the target object region or the edge contour region. Therefore, this embodiment performs morphological dilation on the first voxel set and the second voxel set to fill the gaps, making the target object and the edge contour region more complete and continuous.
[0099] Furthermore, this embodiment takes into account that if morphological dilation is performed on the horizontal or vertical axis of medical image data, it may cause tissues to connect together. For example, it may cause the skin of the breast to connect with its internal tissues. Therefore, this embodiment performs morphological dilation along the depth axis to avoid tissues connecting together. That is, when performing morphological dilation, the kernel used for morphological dilation can be a cuboid that only has a value on the depth axis.
[0100] In some embodiments, generating a point cloud of the target object based on the intersection includes: identifying connected regions formed by interconnected voxels in the intersection; and generating a point cloud of the target object based on the voxel constituting the largest connected region (denoted as the target voxel).
[0101] For example, in medical image data, electronic devices can use connected-component labeling to find all connected regions formed by voxels in the above intersection; then determine the voxels included in the largest connected region, and generate a point cloud based on these voxels.
[0102] Generating a point cloud of the target object based on the voxels constituting the largest connected region may include: setting the CT values of all voxels except the target voxel to the minimum CT value among the multiple voxels included in the medical image data to obtain updated medical image data; then, obtaining a point cloud based on the updated medical image data and the CT values of each voxel it includes, for example, by using a Marching Cube search algorithm to convert voxels into point clouds.
[0103] The Marching Cubes algorithm has important applications in processing three-dimensional data, such as medical images like CT and MRI. Its key function is to extract isosurfaces from these complex three-dimensional voxel data.
[0104] In medical imaging data, isosurfaces can be understood as surfaces with the same physical quantities (such as CT values, MRI signal intensity, etc., which correspond to a specific value). By extracting isosurfaces, the internal structure, which was originally presented in voxel form, can be displayed in a more intuitive triangular mesh surface form, which is helpful for subsequent visualization, analysis, and various application scenarios.
[0105] In the updated medical imaging data, the electronic device can set the data of 8 voxels as the 8 vertices of a cube.
[0106] The electronic device can then determine whether each vertex is above or below a preset isosurface threshold. If a point connected to an edge of a cube is above the isosurface on one side and below the isosurface threshold on the other side, then a vertex on that edge can form a triangular face of the mesh. Ultimately, these points can be combined to form point cloud data.
[0107] The aforementioned isosurface threshold can be configured to -150HU. In practical applications, the isosurface threshold can also be set according to requirements, and this application embodiment does not limit this.
[0108] The following example, using the breast as the target object and medical imaging data including thoracic magnetic array imaging data, illustrates the process of generating breast point clouds.
[0109] refer to Figure 6 As shown, the electronic device can acquire DICOM data, which can be image data from a thoracic magnetic array angiography using the DICOM standard.
[0110] Electronic devices can calculate voxel data based on DICOM data, resulting in voxel data 1. Voxel data 1 includes all voxels in the image data and the corresponding CT value for each voxel. For example, electronic devices can calculate the CT value of each voxel in medical image data based on the rescaling slope and rescaling intercept attributes in the DICOM data.
[0111] After acquiring voxel data 1, the electronic device can resample the voxel data based on a preset interpolation algorithm to obtain voxel data 2. In this embodiment, resampling not only ensures that the voxel data has a unified standard, thereby eliminating differences in voxel data such as voxel size caused by different imaging instrument parameter settings, but also makes the voxel distribution more uniform. This is beneficial for preserving the positional information between points in the subsequently generated point cloud and preventing the point cloud from collapsing along the depth axis.
[0112] Next, the electronic device can perform Heinz value detection based on voxel data 2. That is, voxels with Heinz values within a preset Heinz value range are filtered from voxel data 2 to obtain the first voxel set. When the target object is a breast, this method is beneficial for filtering voxels belonging to soft tissues such as skin.
[0113] Electronic devices can also perform edge detection based on voxel data 2 to obtain a second voxel set. When the target object is a breast, this method is helpful in filtering out the edge contour region of the breast.
[0114] After calculating the first voxel set, the electronic device can perform morphological dilation on the first voxel set along the depth axis to obtain the third voxel set. After calculating the second voxel set, the electronic device can also perform morphological dilation on the second voxel set along the depth axis to obtain the fourth voxel set.
[0115] This embodiment performs morphological dilation on the first and second voxel sets to fill gaps, resulting in greater integrity and continuity of the target object and its edge contours. Furthermore, along the depth axis, it prevents the thoracic skin from connecting with its internal tissues, improving the accuracy of subsequent point cloud processing.
[0116] Next, the electronic device can perform a logical AND operation on the third voxel set and the fourth voxel set to obtain their intersection.
[0117] This embodiment uses logical AND operations to remove voxels that are suspected to belong to the target object but are not on its outline edge, as well as voxels that are on the suspected edge but do not actually belong to the target object. For example, voxels corresponding to the examination table of the magnetic resonance imaging (MRI) machine can be removed, so that the final generated point cloud is highly focused on the most critical and representative part of the target object, which greatly improves the accuracy and effectiveness of the point cloud data.
[0118] After obtaining the intersection, the electronic device can identify connected regions formed by interconnected voxels within the intersection; and generate a point cloud of the target object based on the voxel constituting the largest connected region (denoted as the target voxel).
[0119] This embodiment can identify voxels that constitute the largest connected region, thereby eliminating interference from connected regions corresponding to other tissues. For example, when the target object is a breast, interference from other soft tissues besides the skin of the breast contour can be eliminated, improving the accuracy of the identified breast contour. That is, this embodiment can accurately segment voxels belonging to the skin from the intersection, further reducing interfering voxels, thereby making the breast shape presented by the point cloud more accurate.
[0120] After determining the target voxel, the electronic device can set the CT value of other voxels besides the target voxel to the minimum value in voxel data 2 (i.e., the resampled voxel data) to obtain updated voxel data, so as to mask the other voxels during the point cloud generation process.
[0121] Then, using the updated voxel data and point cloud generation algorithms, such as the iso-cube search algorithm, the point cloud of the target object is generated. For example, it can generate... Figure 6 The breast dot cloud shown is 61.
[0122] The embodiments of this application can accurately segment the voxels corresponding to the target object from medical image data, and generate point clouds based on the voxels of the target object to intuitively present the shape of the target object.
[0123] For example, the outer contour of the breast, i.e., the voxels belonging to the breast skin, can be accurately segmented from the image data of thoracic MRI, and point clouds can be generated based on this to help show the breast shape to the user.
[0124] Based on the same idea as the point cloud generation method in the above embodiments, this application also provides a point cloud generation apparatus, which can be used to execute the above point cloud generation method. For ease of explanation, the structural schematic diagram of the point cloud generation apparatus embodiment only shows the parts related to the embodiments of this application. Those skilled in the art will understand that the illustrated structure does not constitute a limitation on the apparatus, and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.
[0125] like Figure 7 As shown, the point cloud generation device includes a data acquisition module 701, a voxel screening module 702, an edge detection module 703, and a point cloud generation module 704. In some embodiments, the above modules can be programmable software instructions stored in memory and executable by a processor. It is understood that in other embodiments, the above modules can also be program instructions or firmware embedded in the processor.
[0126] The data acquisition module 701 is used to acquire medical image data of the target object.
[0127] The voxel filtering module 702 is used to filter the voxels of the target object in the medical image data based on the Henle value to obtain a first voxel set.
[0128] The edge detection module 703 is used to perform edge detection on the medical image data to obtain a second voxel set, the second voxel set including voxels at the contour of the target object.
[0129] The point cloud generation module 704 is used to generate a point cloud of the target object based on the first voxel set and the second voxel set.
[0130] Figure 8 This is a schematic diagram of an embodiment of the electronic device of this application.
[0131] The electronic device 100 includes a memory 20, a processor 30, and a computer program 40 stored in the memory 20 and executable on the processor 30. When the processor 30 executes the computer program 40, it implements the steps described in the above-described point cloud generation method embodiments, for example... Figure 1 Steps 101 to 104 are shown.
[0132] For example, computer program 40 can also be divided into one or more modules / units, which are stored in memory 20 and executed by processor 30. The one or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 40 in electronic device 100. For example, it can be divided into... Figure 7 The data acquisition module 701, voxel screening module 702, edge detection module 703, and point cloud generation module 704 are shown.
[0133] Those skilled in the art will understand that the schematic diagram is merely an example of the electronic device 100 and does not constitute a limitation on the electronic device 100. It may include more or fewer components than shown in the diagram, or combine certain components, or different components. For example, the electronic device 100 may also include input / output devices, network access devices, buses, etc.
[0134] Processor 30 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors, single-chip microcomputers, or any conventional processor.
[0135] The memory 20 can be used to store computer programs 40 and / or modules / units. The processor 30 implements various functions of the electronic device 100 by running or executing the computer programs and / or modules / units stored in the memory 20 and by calling data stored in the memory 20. The memory 20 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device 100 (such as audio data), etc. In addition, the memory 20 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.
[0136] If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.
[0137] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the electronic device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and other division methods may be used in actual implementation.
[0138] Furthermore, the functional units in the various embodiments of this application can be integrated into the same processing unit, or each unit can exist physically separately, or two or more units can be integrated into the same unit. The integrated units described above can be implemented in hardware or in the form of hardware plus software functional modules.
[0139] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered exemplary and not restrictive in all respects. Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or electronic devices recited in the electronic device claims may also be implemented by the same unit or electronic device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any particular order.
[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this application without departing from the spirit and scope of the technical solutions of this application.
Claims
1. A point cloud generation method, characterized in that, include: Acquire medical imaging data of the target object; In the medical image data, voxels of the target object are filtered based on Henle values to obtain a first voxel set; Edge detection is performed on the medical image data to obtain a second voxel set, which includes voxels at the contour of the target object. The point cloud of the target object is generated based on the first voxel set and the second voxel set.
2. The point cloud generation method as described in claim 1, characterized in that, In the medical image data, voxels of the target object are filtered based on the Henle value to obtain a first voxel set, including: Obtain the Henle value of each voxel in the medical image data; Among the voxels of the medical image data, voxels whose Henle values are within a preset Henle value range are selected to obtain the first voxel set. The preset Heinz value range is the Heinz value range of the target object voxels.
3. The point cloud generation method as described in claim 1, characterized in that, The process of edge detection on the medical image data to obtain a second voxel set includes: The medical image data is sliced along the imaging plane of the medical image data to obtain multiple two-dimensional image data. Edge detection is performed on the multiple two-dimensional image data to obtain the edge regions of the multiple two-dimensional image data; The edge regions of the multiple two-dimensional image data are stacked to obtain the second voxel set.
4. The point cloud generation method as described in claim 3, characterized in that, The step of performing edge detection on the plurality of two-dimensional image data to obtain the edge regions of the plurality of two-dimensional image data includes: The pixels of the two-dimensional image data are convolved with a preset horizontal convolution kernel to obtain the horizontal gradient of the two-dimensional image data. The pixels of the two-dimensional image data are convolved with a preset vertical convolution kernel to obtain the vertical gradient of the two-dimensional image data. The gradient magnitude of each pixel in the two-dimensional image data is determined based on the horizontal gradient and the vertical gradient. Based on the gradient magnitude, pixels located in the edge region are selected from the two-dimensional image data to obtain the edge region of the two-dimensional image data.
5. The point cloud generation method according to any one of claims 1 to 4, characterized in that, The process of generating a point cloud of the target object based on the first voxel set and the second voxel set includes: The intersection is obtained by performing a logical AND operation on the first voxel set and the second voxel set. A point cloud of the target object is generated based on the intersection.
6. The point cloud generation method as described in claim 5, characterized in that, The step of performing a logical AND operation on the first voxel set and the second voxel set to obtain the intersection includes: The first voxel set is morphologically dilated along the depth axis of the medical image data to obtain a third voxel set. The second voxel set is morphologically dilated along the depth axis to obtain the fourth voxel set; Perform a logical AND operation on the third voxel set and the fourth voxel set to obtain the intersection.
7. The point cloud generation method as described in claim 5, characterized in that, The process of generating the point cloud of the target object based on the intersection includes: In the intersection, connected regions formed by interconnected voxels are identified; The point cloud of the target object is generated based on the voxels that constitute the largest connected region.
8. A point cloud generation device, characterized in that, include: The data acquisition module is used to acquire medical image data of the target object; A voxel filtering module is used to filter voxels of the target object in the medical image data based on the Henle value to obtain a first voxel set; An edge detection module is used to perform edge detection on the medical image data to obtain a second voxel set, the second voxel set including voxels at the contour of the target object; The point cloud generation module is used to generate a point cloud of the target object based on the first voxel set and the second voxel set.
9. An electronic device, the electronic device comprising a processor and a memory, characterized in that, The memory is used to store instructions, and the processor is used to call the instructions in the memory to cause the electronic device to execute the point cloud generation method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed on an electronic device, cause the electronic device to perform the point cloud generation method as described in any one of claims 1 to 7.