Medical cross-sectional image correction and fusion method, device and system and storage medium

Abstract

The invention discloses a medical cross-sectional image correction and fusion method, device and system and a storage medium, and relates to the technical field of medical imaging. The method comprises the following steps: acquiring a Dicom medical image, reading an image sequence, processing to obtain volume data, calculating outer bounding box space physical data, judging whether a reference body and a correction body are intersected or not to carry out body space correction fusion, calculating a central point offset and a correction matrix of the correction body relative to the reference body, and calculating a sectioning position and a sectioning direction vector of the correction body. Generating a corrected body sectioning graph; adjusting the resolution to be consistent; and carrying out normalization processing and fusion processing on the corrected body sectioning graph and the reference body sectioning graph. Compared with the prior art, the medical cross-sectional image correction and fusion method provided by the embodiment of the invention has the advantages that accurate alignment of the correction body in a three-dimensional space is realized, and the consistency of the fused image on an anatomical structure is ensured; the fusion error caused by the resolution difference is further reduced, and the precision of the fused image is improved.

Description

Technical Field

[0001] The present application belongs to the field of medical imaging technology, and in particular, relates to a medical tomographic image correction and fusion method, device, system and storage medium. Background Art

[0002] Medical tomography is an important concept in the field of medical imaging. It mainly involves image data obtained by computed tomography (CT) and other tomography techniques. Medical tomography refers to the internal structure image of a specific layer reconstructed after multi-angle scanning of the human body or object through computed tomography (CT) or similar technologies. These images show the cross-section of the object in a certain direction, avoid the problem of image projection overlap, and provide clearer internal structure information.

[0003] Anatomical images have good spatial resolution and can provide the anatomical location of the lesion and the anatomical structure relationship of the surrounding normal tissues, but they are not sensitive to some lesions. Functional images can provide information about human physiological metabolism, but the image resolution is poor. In order to obtain more comprehensive diagnostic information, it is usually necessary to combine multiple imaging technologies. However, due to the different imaging principles and parameter settings of different imaging technologies, there are differences in spatial position, resolution, contrast, etc. between different images, which makes it difficult for doctors to diagnose. In addition, it is not easy to observe the overall lesion characteristics by only looking at the original scanning orientation.

[0004] The current mainstream correction and fusion methods mainly include the following methods: based on 2D image physical position matching, based on feature point matching, and based on deep learning.

[0005] Based on the 2D physical position matching method, slices with similar physical positions to the current image slice will be found, and fused and displayed with the image center point as the reference. However, this method only considers the two-dimensional space of the image, requiring that the scanning directions between the two sets of sequences are absolutely consistent or slightly misaligned. It has high requirements on scanning data, especially imaging data from multi-modal devices, which is basically difficult to meet, and it is impossible to present multi-angle fusion effects.

[0006] Based on the feature point matching method, the correction and fusion of the image is achieved by setting, detecting and matching the feature points in the image. This method is more accurate than the matching of the slice position, but it requires manual marking of feature points and setting of affine transformation matrix. The operation complexity and technical requirements are high. The representation of image features and the selection of similarity measurement are easily affected by human factors, and the algorithm is complex and inefficient. In addition, problems such as image noise and motion artifacts may affect the accuracy of the correction.

[0007] Based on deep learning, the complex patterns of changes between images are learned through training data, and the similarity measure between given image pairs is obtained to achieve automatic correction fusion. However, this method requires manually annotated data as the gold standard. Due to privacy issues, the samples are limited and uneven, and the annotation is time-consuming and costly. The training effect is easily affected by image noise, motion artifacts, etc. Only image features are considered, and label information such as the physical spatial position and scanning orientation of the Dicom image is not considered. Summary of the invention

[0008] The purpose of this application is to provide a medical tomographic image correction and fusion method, device, system and storage medium, which can improve the accuracy and diagnostic efficiency of medical images by correcting and fusing medical tomographic images of different sequences.

[0009] In order to achieve the above objectives, the solution of this application is:

[0010] In a first aspect, an embodiment of the present application provides a medical tomographic image correction and fusion method, comprising: acquiring two sets of Dicom medical images, respectively reading image sequences of the two sets of Dicom medical images, and obtaining a reference sequence and a correction sequence;

[0011] Processing the reference sequence and the correction sequence to obtain volume data, the volume data including a reference volume corresponding to the reference sequence and a correction volume corresponding to the correction sequence;

[0012] Calculating outer bounding box spatial physical data according to the volume data, the outer bounding box spatial physical data including a reference volume data outer bounding box corresponding to the reference volume and a correction volume data outer bounding box corresponding to the correction volume;

[0013] According to the outer bounding box of the reference volume data and the outer bounding box of the corrected volume data, it is determined whether the reference volume and the corrected volume have an intersection. If no intersection exists, the process ends; if an intersection exists, the reference volume and the corrected volume are corrected and fused in volume space. The volume space correction fusion method includes:

[0014] Obtaining a cross-sectional view of the reference body, calculating the center point offset and correction matrix of the correction body relative to the reference body; calculating the cross-sectional position and cross-sectional direction vector of the correction body according to the center point offset and correction matrix, and generating a cross-sectional view of the correction body;

[0015] Adjust the image resolution of the corrected body section image to be consistent with that of the reference body section image;

[0016] The corrected body section diagram and the reference body section diagram are normalized and fused.

[0017] According to the above method of the embodiment of the present application, the following additional technical features may also be provided:

[0018] Furthermore, the reference sequence and the correction sequence are processed to obtain volume data, including respectively reading the spatial physical information of the reference sequence and the correction sequence to obtain slices, the slices correspond to the Dicom medical image, and the slices are sorted from small to large according to the spatial physical information to obtain volume data. Median filtering is used to remove noise in the Dicom medical image and enhance the contrast of the Dicom medical image.

[0019] Further, according to the volume data, the spatial physical data of the outer bounding box is calculated, including obtaining Dicom tags, the Dicom tags including position tags, orientation tags and pixel tags;

[0020] Get the spatial matrix F of the first point in the upper left corner of the Dicom medical image according to the position label. The formula is as follows:

[0021] Get the direction cosine matrix D of the Dicom medical image according to the orientation label. The formula is as follows:

[0022]

[0023] The physical size of the pixel of the Dicom medical image is obtained according to the pixel label. The formula of the coordinate matrix P of any point on the Dicom medical image is as follows:

[0024]

[0025] According to the space matrix F, direction cosine matrix D and coordinate matrix P, the arbitrary physical space position P′ of the Dicom medical image is calculated by matrix transformation. The formula is as follows:

[0026]

[0027] According to the calculation formula of any physical space position P′ of the Dicom medical image, the corner points of the outer bounding box are calculated, and the minimum and maximum coordinate values ​​of the outer bounding box in the X, Y, and Z axis directions are obtained according to the corner points to obtain bounds = (xmin, xmax, ymin, ymax, zmin, zmax);

[0028] According to the outer bounding box of the reference body data and the outer bounding box of the corrected body data, it is judged whether the reference body and the corrected body have an intersection. The judgment formula is as follows:

[0029] XIntersects=bounds1[1]≥bounds2[0]&&bounds2[1]≥bounds1[0];

[0030] YIntersects=bounds1[3]≥bounds2[2]&&bounds2[3]≥bounds1[2];

[0031] ZIntersects=bounds1[5]≥bounds2[4]&&bounds2[5]≥bounds1[4];

[0032] When the judgment formulas of XIntersects, YIntersects and ZIntersects are satisfied at the same time, the reference body and the correction body have an intersection.

[0033] Further, the center point offset and correction matrix of the correction body relative to the reference body are calculated, including aligning the centers of the correction body and the reference body, and rotating the correction body to the same orientation as the reference body; the body center coordinates (X1, Y1, Z1) of the reference body and the body center coordinates (X2, Y2, Z2) of the correction body are calculated according to any physical space position P′ of the Dicom medical image; the center point offset of the correction body offset = {X1-X2, Y1-Y2, Z1-Z2} is obtained according to the body center coordinates (X1, Y1, Z1) of the reference body and the body center coordinates (X2, Y2, Z2) of the correction body;

[0034] The correction matrix M is set as the orientation matrix of the reference body, which is expressed as follows:

[0035]

[0036] Further, a reference body section diagram is obtained, and the center point offset and correction matrix of the correction body relative to the reference body are calculated; the section position and section direction vector of the correction body are calculated according to the center point offset and the correction matrix, and a correction body section diagram is generated, including obtaining the section position and section direction of the reference body, and calculating the section position P1 and section direction vector D1 of the correction body according to the section position P0 and section direction vector D0 of the reference body, to obtain the correction body section diagram; the formula of the section position P1 of the correction body is as follows:

[0037] P1.x=(P0.x*voxelspace0+offset.x) / voxelspace1

[0038] P1.y=(P0.y*voxelspace0+offset.y) / voxelspace1

[0039] P1.z=(P0.z*voxelspace0+offset.z) / voxelspace1

[0040] Among them, P0 represents the section position of the reference body, offset represents the center point offset of the correction body, voxelspace0 represents the physical size of the reference body voxel, and voxelspace1 represents the physical size of the correction body voxel; P0.x, P0.y, and P0.z represent the components of the section position of the reference body in the x, y, and z directions respectively; P1.x, P1.y, and P1.z represent the components of the section position of the correction body in the x, y, and z directions respectively; offset.x, offset.y, and offset.z represent the components of the center point offset of the correction body in the x, y, and z directions respectively;

[0041] The formula of the cutting direction vector D1 of the correction body is as follows:

[0042] D1=D0×M

[0043] Wherein, D0 represents the cutting direction vector of the reference body, and M represents the direction matrix of the reference body.

[0044] Furthermore, the image resolution of the corrected body section image is adjusted to be consistent with the image resolution of the reference body section image, including calculating a resolution adjustment coefficient, and the formula is as follows:

[0045] sacle=voxelspace1 / voxelspace0

[0046] Among them, voxelspaceO represents the physical size of the reference volume voxel, and voxelspace1 represents the physical size of the correction volume voxel. The image resolution of the correction volume section is compared with the image resolution of the reference volume section to obtain a low-resolution image, which is processed by cubic interpolation. The formula is as follows:

[0047] P(x)=a0+a1(x-x0)+a2(x-x0) 2 +a3(x-x0) 3

[0048] Among them, P(x) represents the pixel value of the interpolation point, x represents the position of the interpolation point, (a0, a1, a2, a3) represents the coefficients calculated based on the surrounding pixels, and x0 represents the reference corresponding pixel value.

[0049] Furthermore, the corrected body section diagram and the reference body section diagram are normalized and fused, including extracting two-dimensional images from the corrected body section diagram and the reference body section diagram, respectively, and defining the extracted two-dimensional images. The formula is as follows:

[0050] I'(i,j)=(I(i,j)-mmin)*(255 / (max-min)

[0051] Wherein, i and j represent the coordinate position of the two-dimensional image; i = 1, 2, 3, ..., R; j = 1, 2, 3, ..., C; I(i, j) represents the grayscale value of any pixel at a two-dimensional image, I'(i, j) represents the normalized grayscale value of any pixel at a two-dimensional image, (R, C) represents the size of the matrix I(i, j), and (min, max) represents the grayscale range of the two-dimensional image;

[0052] All pixels of the two-dimensional images in the cross-section of the correction body and the cross-section of the reference body are traversed respectively to obtain the image pixels of the correction body and the image pixels of the reference body; the fusion processing includes a pixel-based fusion method, and the formula is as follows:

[0053] C(i,j)=a*C1(i,j)+(1-a)*C2(i,j)

[0054] Wherein, C1 represents the image pixel of the reference body, C2 represents the image pixel of the correction body, and a represents transparency.

[0055] In a second aspect, an embodiment of the present application provides a medical tomographic image correction and fusion device, comprising:

[0056] A data acquisition module is configured to acquire two sets of Dicom medical images, read the image sequences of the two sets of Dicom medical images respectively, and obtain a reference sequence and a correction sequence;

[0057] The data calculation module is configured to process the reference sequence and the correction sequence to obtain volume data, the volume data including a reference volume corresponding to the reference sequence and a correction volume corresponding to the correction sequence; and calculate outer bounding box space physical data according to the volume data, the outer bounding box space physical data including a reference volume data outer bounding box corresponding to the reference volume and a correction volume data outer bounding box corresponding to the correction volume.

[0058] The data judgment module is configured to judge whether the reference body and the correction body have an intersection according to the outer bounding box of the reference body data and the outer bounding box of the correction body data, and if no intersection exists, the process ends; if there is an intersection, the reference body and the correction body are corrected and fused in volume space;

[0059] The correction fusion module is configured to calculate the center point offset and correction matrix of the correction body relative to the reference body; calculate the section position and section direction vector of the correction body according to the center point offset and the correction matrix, and generate a section view of the correction body; adjust the image resolution of the section view of the correction body to be consistent with the image resolution of the section view of the reference body; and perform normalization and fusion processing on the section view of the correction body and the section view of the reference body.

[0060] In a third aspect, an embodiment of the present application provides a medical tomographic image correction and fusion system, the system comprising a processor and a memory, the memory storing a computer program, the computer program being loaded and executed by the processor to implement the medical tomographic image correction and fusion method provided in the first aspect of the embodiment of the present application.

[0061] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it is used to implement the medical tomographic image correction and fusion method as described in the first aspect of the embodiment of the present application.

[0062] Compared with the prior art, the medical tomographic image correction and fusion method provided in the embodiment of the present application has the following beneficial technical effects:

[0063] The embodiment of the present application calculates the center point offset of the correction body relative to the reference body and the correction matrix to achieve precise alignment of the correction body in three-dimensional space, thereby ensuring the anatomical consistency of the fused image; using methods such as cubic interpolation to adjust the image resolution can further reduce the fusion error caused by resolution differences and improve the accuracy of the fused image.

[0064] The embodiment of the present application adopts methods such as median filtering in the preprocessing stage to remove noise and enhance image contrast, which helps to improve the visualization effect of the fused image and make the lesion area clearer and more identifiable; the pixel or feature-based fusion method can fully retain the useful information in the reference body and the correction body, avoid information loss, and ensure that the fused image has higher information richness and accuracy.

[0065] The embodiment of the present application can provide doctors with more comprehensive and accurate lesion information by correcting and fusing medical tomographic images, which is helpful for early detection and accurate diagnosis of diseases; in clinical applications such as surgical planning and radiotherapy, the fused images generated by this method can serve as an important reference to guide doctors to formulate more accurate treatment plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic diagram showing the flow of a medical tomographic image correction and fusion method according to an embodiment of the present application is shown;

[0067] Figure 2 A structural block diagram of a medical tomographic image correction and fusion device according to an embodiment of the present application is shown;

[0068] Figure 3 A structural block diagram of a computer device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0069] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are described in detail below in conjunction with the accompanying drawings. It is to be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only some structures related to the present application are shown in the accompanying drawings, rather than all structures. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0070] The terms "including" and "having" and any variations thereof in this application are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0071] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0072] Dicom, the full name of which is Digital Imaging and Communications in Medicine, is an international standard (ISO 12052) for medical images and related information, dedicated to creating the most extensive and fundamental digital medical imaging standard; it defines a medical image format that can be used for data exchange with quality that meets clinical needs. Dicom is not limited to the image or file itself, it is the sum of data transmission, storage and display protocols, and provides all the tools that enable medical imaging data to be accurately expressed and processed in diagnosis; all Dicom data (such as medical images, commands and reports) are packaged in Dicom objects, which are saved through Dicom files. Each Dicom object is a collection of data elements. Dicom data elements consist of tags, value representations (VRs) and values. Tags are a unique 16-bit group number and a 16-bit element number used to identify data elements; VRs define the type and format of values, while values ​​are the actual data content.

[0073] like Figure 1As shown, the embodiment of the present application provides a medical tomographic image correction and fusion method, comprising the following steps:

[0074] Step 1 0 1, obtain two groups of Dicom medical images, read the image sequences of the two groups of Dicom medical images respectively, and obtain a reference sequence and a correction sequence.

[0075] The embodiments of the present application load DICOM files by using image processing libraries or software that support the DICOM format. These libraries provide functions for reading DICOM files and parsing their metadata and image data, such as Python's pydicom, SimpleITK, or MATLAB's Image Processing Toolbox.

[0076] Each image in a DICOM file may belong to a series, which usually consists of images that perform the same scan or imaging process. Each series has a unique identifier (SeriesUID). If the DICOM files are scattered in folders, first traverse the folders to find all DICOM files and group the images into different series according to DICOM tags (such as SeriesUID).

[0077] When selecting sequences, users need to manually select reference and correction sequences based on file name, sequence description, or other metadata; if there are obvious differences between the sequences, such as timestamps and patient locations, algorithms can be written to automatically select.

[0078] For the selected reference sequence and correction sequence, use the DICOM library or software to read each image in the sequence in order and sort the images; when reading the images, you also need to pay attention to the orientation and position of the DICOM images, because this information is crucial for subsequent image processing and registration.

[0079] Specifically, a set of MR scan tomography images and a set of PET tomography images are read, where MR is a set of coronal scan data and PET is a set of axial scan data. The scan information is shown in the following table:

[0080] Table 1

[0081]

[0082] Step 102: Process the reference sequence and the correction sequence to obtain volume data, where the volume data includes a reference volume corresponding to the reference sequence and a correction volume corresponding to the correction sequence.

[0083] In order to obtain volume data from a reference sequence and a correction sequence, the embodiment of the present application processes each layer of the volume data as a two-dimensional image, and performs noise removal and contrast enhancement on the image. By reading the spatial physical information of the reference sequence and the correction sequence, the slices are sorted according to their actual positions in space to ensure the correct arrangement of the volume data in three-dimensional space.

[0084] The sorted slices are stacked to form three-dimensional volume data, which involves converting the two-dimensional image array into a three-dimensional array, with the third dimension being the depth of the slice, and using image processing techniques such as median filtering to remove noise from the volume data. Median filtering is a nonlinear filtering technique that is particularly suitable for removing salt and pepper noise while maintaining edge information. It can also enhance contrast through histogram equalization, contrast stretching, etc. These methods are designed to improve the visual effect of the image and make the lesion area more clearly identifiable.

[0085] The volume data construction will filter out some illegal data, such as positioning slices, images without pixel size, and images without image position information, and sort the remaining slices from small to large according to the spatial physical position. The redundant information in the images in the original scanning sequence will affect the fusion effect. It is necessary to traverse the images that constitute the volume data and perform median filtering on each image to remove background noise in the image.

[0086] Specifically, MR is set as a reference body and PET is set as a correction body, and volume data are constructed for MR and PET respectively. Here, the median filtering method is used to denoise the image and enhance the contrast information of the image.

[0087] Step 103: Calculate outer bounding box space physical data according to the volume data. The outer bounding box space physical data includes a reference volume data outer bounding box corresponding to the reference volume and a correction volume data outer bounding box corresponding to the correction volume.

[0088] The embodiment of this application first needs to read the key tags in the Dicom file, including:

[0089] Position label (Image Position (Patient)): This label is an array of three floating-point numbers, which indicates the position of the first voxel in the upper left corner of the image in the patient coordinate system. The spatial matrix F of the first point in the upper left corner of the Dicom medical image is obtained according to the position label. The formula is as follows:

[0090]

[0091] Orientation label (Image Orientation (Patient)): This label is an array of six floating-point numbers, which represent the direction cosines of the rows and columns of the image in the patient coordinate system. These six numbers are usually divided into two groups, and each group of three numbers represents the direction vectors of the rows and columns of the image. The direction cosine matrix D of the Dicom medical image is obtained according to the orientation label. The formula is as follows:

[0092]

[0093] Pixel Spacing: This label is the size of each pixel (in a two-dimensional image) or voxel (in a three-dimensional image) in the image in physical space (usually millimeters). For three-dimensional images, slice thickness may also need to be considered. The physical size of the pixel of the Dicom medical image is obtained according to the pixel label. The formula of the coordinate matrix P of any point on the Dicom medical image is as follows:

[0094]

[0095] The two-dimensional coordinates and slice positions are combined and transformed using the direction cosine matrix to obtain the coordinates in the three-dimensional space. According to the space matrix F, the direction cosine matrix D and the coordinate matrix P, the arbitrary physical space position P′ of the Dicom medical image is calculated by matrix transformation. The formula is as follows:

[0096]

[0097] Traverse all voxels and find the minimum and maximum coordinates on the X, Y, and Z axes. These values ​​will define the eight corner points of the outer bounding box. Once the corner points are determined, the minimum and maximum coordinate values ​​of the outer bounding box on the X, Y, and Z axes can be directly extracted from these corner points. According to the calculation formula of any physical space position P′ of the Dicom medical image, the corner points of the outer bounding box are calculated, and the minimum and maximum coordinate values ​​of the outer bounding box in the X, Y, and Z axis directions are obtained according to the corner points, and bounds = (xmin, xmax, ymin, ymax, zmin, zmax).

[0098] Specifically, the outer bounding box of the MR volume data is (-255.868, 192.987, -77.3257, 97.2682, -267.318, 172.695); the outer bounding box of the PET volume data is (-324.723, 275.277, -299.219, 300.781, -106.451, 138.189).

[0099] Step 104, based on the outer bounding box of the reference body data and the outer bounding box of the correction body data, determine whether the reference body and the correction body have an intersection, if not, then end; if there is an intersection, then perform volume space correction fusion on the reference body and the correction body, the volume space correction fusion method includes obtaining a reference body section view, calculating the center point offset and correction matrix of the correction body relative to the reference body; calculating the section position and section direction vector of the correction body according to the center point offset and the correction matrix, and generating a correction body section view.

[0100] In medical image processing, especially when processing DICOM images for volume space correction and fusion, it is first necessary to determine whether there is an intersection between the reference volume and the correction volume, which is achieved by comparing the bounding boxes of the volume data of the two. The bounding box is the smallest cube that can completely contain the volume data, which defines the position and size of the volume data in three-dimensional space.

[0101] For the reference volume and the correction volume, calculate their outer bounding boxes respectively, traverse all points in the volume data, find the minimum and maximum values ​​on the X, Y, and Z axes, and compare the two outer bounding boxes. If their intervals on any coordinate axis do not overlap (that is, the minimum value of one box is greater than the maximum value of the other box, or vice versa), it can be determined that the two have no intersection. If there is overlap, proceed to the next step.

[0102] Finding the coordinates of the center of the reference and corrected bodies is done by calculating the center points of their respective bounding boxes. Then, moving the center of the corrected body to align with the center of the reference body involves calculating a translation vector (i.e., center point offset) to move the corrected body.

[0103] Next, the correction body needs to be rotated to the same orientation as the reference body, which involves calculating a rotation matrix that transforms the direction cosine matrix of the correction body (obtained from the DICOM tag) to match the reference body. This process may involve the calculation of Euler angles or quaternions, depending on the rotation representation method used.

[0104] Determine the sectioning position and sectioning direction of the reference body, which is usually based on specific image processing requirements or doctor's instructions. The reference body sectioning diagram can be generated by selecting any angle for sectioning. For example, you can choose to section parallel to the Z axis to generate an axial sectioning diagram, parallel to the X axis to generate a coronal sectioning diagram, and parallel to the Y axis to generate a sagittal sectioning diagram; the correction body sectioning diagram is generated based on the reference body. The reference body sectioning diagram must be obtained before the correction body sectioning diagram; both the reference body and the correction body sectioning diagrams are generated by the principle of planar reconstruction.

[0105] Then, based on the cutting position and cutting direction vector of the reference body, as well as the center point offset and correction matrix calculated previously, the cutting position and cutting direction vector of the correction body at the same cutting position and direction are calculated, which involves converting the cutting information of the reference body into the coordinate system of the correction body through an inverse correction matrix (including inverse operations of translation and rotation).

[0106] Finally, the calculated sectioning position and sectioning direction vector of the correction body are used to extract the corresponding sectioning image from the volume data of the correction body, which involves slicing operations of the three-dimensional volume data and possible interpolation processing to generate smooth sectioning images.

[0107] In the embodiment of the present application, based on the outer bounding box of the reference volume data and the outer bounding box of the corrected volume data, it is determined whether the reference volume and the corrected volume have an intersection. The determination formula is as follows:

[0108] XIntersects=bounds1[1]≥bounds2[0]&&bounds2[1]≥bounds1[0];

[0109] YIntersects=bounds1[3]≥bounds2[2]&&bounds2[3]≥bounds1[2]; ZIntersects=bounds1[5]≥bounds2[4]&&bounds2[5]≥bounds1[4];

[0110] When the judgment formulas of XIntersects, YIntersects and ZIntersects are satisfied at the same time, the reference body and the correction body have an intersection.

[0111] Further, the center point offset and correction matrix of the correction body relative to the reference body are calculated, including aligning the centers of the correction body and the reference body, and rotating the correction body to the same orientation as the reference body; the body center coordinates (X1, Y1, Z1) of the reference body and the body center coordinates (X2, Y2, Z2) of the correction body are calculated according to any physical space position P′ of the Dicom medical image; the center point offset of the correction body offset = {X1-X2, Y1-Y2, Z1-Z2} is obtained according to the body center coordinates (X1, Y1, Z1) of the reference body and the body center coordinates (X2, Y2, Z2) of the correction body;

[0112] The correction matrix M is set as the orientation matrix of the reference body, which is expressed as follows:

[0113]

[0114] Further, the section position and section direction vector of the correction body are calculated according to the center point offset and the correction matrix to obtain the section view of the correction body, including obtaining the section position and section direction of the reference body, calculating the section position P1 and section direction vector D1 of the correction body according to the section position P0 and section direction vector D0 of the reference body, and obtaining the section view of the correction body; the formula of the section position P1 of the correction body is as follows:

[0115] P1.x=(P0.x*voxelspace0+offset.x) / voxelspace1

[0116] P1.y=(P0.y*voxelspace0+offset.y) / voxelspace1

[0117] P1.z=(P0.z*voxelspace0+offset.z) / voxelspace1

[0118] Among them, P0 represents the section position of the reference body, offset represents the center point offset of the correction body, voxelspace0 represents the physical size of the reference body voxel, and voxelspace1 represents the physical size of the correction body voxel; P0.x, P0.y, and P0.z represent the components of the section position of the reference body in the x, y, and z directions respectively; P1.x, P1.y, and P1.z represent the components of the section position of the correction body in the x, y, and z directions respectively; offset.x, offset.y, and offset.z represent the components of the center point offset of the correction body in the x, y, and z directions respectively;

[0119] The formula of the cutting direction vector D1 of the correction body is as follows:

[0120] D1=D0×M

[0121] Wherein, D0 represents the cutting direction vector of the reference body, and M represents the direction matrix of the reference body.

[0122] Specifically, according to the outer bounding box calculated in step 103, it is determined whether the MR and PET volume spaces have an intersection, XIntersects=192.987>-324.723&&275.277>-255.868; YIntersects=97.2682>-299.219&&300.781>-77.3257: ZIntersects=172.695>-106.451&&138.189>-267.318.

[0123] XIntersects, YIntersects, and ZIntersects are all true, indicating that the MR and PET volumes intersect.

[0124] The physical coordinates of the MR body center (-32.0397, 7.50629, -46.8817) and the physical coordinates of the PET body center (-25.504, 0, 15.869) are calculated according to the formula in step 104, and offset = (-6.53572, 7.50629, -62.7507) is obtained.

[0125] Set the original orientation of MR to the correction matrix of PET:

[0126]

[0127] The embodiment of the present application verifies the effects before and after correction through cross-sectional views in different orientations.

[0128] The method based on the physical position matching of 2D images can only be used for matching original slices. By selecting the specified slice in MR, the slice corresponding to the location is found in PET. However, the corresponding matching layer cannot be found for data with inconsistent posture, so this method cannot be used for this set of MR and PET data.

[0129] For this kind of multi-slice data, finding the corresponding feature points based on the feature point matching method is too time-consuming and requires very clear image information features.

[0130] The correction method of the present application does not require the positioning and matching of 2D slices and the selection of feature points. In a unified coordinate system, the coordinate transformation of MR and PET is performed to achieve the coincidence of the three-dimensional reconstructed body, thereby avoiding the subjectivity and errors of human operation and having a wider applicability.

[0131] By setting multiple MR cutting positions as a benchmark, the corresponding slices before and after PET correction were compared to verify the correction effect.

[0132] 1. Coronal section

[0133] The cutting position of MR coronal slices is:

[0134] PO=(0,0,0)

[0135] Cutting direction:

[0136] As a benchmark for PET correction,

[0137] The position of the uncorrected PET coronal slices:

[0138] P1=(0,0,0)

[0139] Cutting direction:

[0140] It can be seen that the cutting positions and anatomical tissues of uncorrected PET coronal slices and MR slices are different.

[0141] Corrected PET coronal slice cutting position:

[0142] P1.x=(0*0.859+-6.53572) / 1.5625=-4.18286

[0143] P1.y=(0*0.859+7.50629) / 1.5625=4.80403

[0144] P1.z=(0*0.859-62.7507) / 1.5625=-40.160

[0145] Cutting direction:

[0146]

[0147] The corrected PET coronal slices can be obtained with a high degree of matching between the position and anatomical tissue of the MR slices.

[0148] 2. Sagittal section

[0149] Cutting position for MR sagittal slices:

[0150] PO=(0,0,0)

[0151] Cutting direction:

[0152] As a benchmark for PET correction,

[0153] Cutting positions for uncorrected PET sagittal slices:

[0154] P1=(0,0,0)

[0155] Cutting direction:

[0156] It can be seen that the cutting positions and anatomical tissues of uncorrected PET sagittal slices and MR slices are different.

[0157] Cutting position for corrected PET sagittal slices:

[0158] P1.x=(0*0.859+-6.53572) / 1.5625=-4.18286

[0159] P1.y=(0*0.859+7.50629) / 1.5625=4.80403

[0160] P1.z=(0*0.859-62.7507) / 1.5625=-40.160

[0161] Cutting direction:

[0162]

[0163] The corrected PET sagittal slices can be obtained with a high degree of matching between the position and anatomical tissue of the MR slices.

[0164] 3. Axial section

[0165] The cutting position of MR axial slices is as follows:

[0166] P0=(0,0,0)

[0167] Cutting direction:

[0168]

[0169] As a benchmark for PET correction,

[0170] The position of the uncorrected PET axial slices:

[0171] P1=(0,0,0)

[0172] Cutting direction:

[0173] It can be seen that the cutting positions and anatomical tissues of uncorrected PET axial slices and MR slices are different.

[0174] Cutting position for corrected PET axial slices:

[0175] P1.x=(0*0.859+-6.53572) / 1.5625=-4.18286

[0176] P1.y=(0*0.859+7.50629) / 1.5625=4.80403

[0177] P1.z=(0*0.859-62.7507) / 1.5625=-40.160

[0178] Cutting direction:

[0179]

[0180] The corrected PET axial slices can be obtained with a high degree of matching between the position and anatomical tissue of the MR slices.

[0181] Step 105, adjusting the image resolution of the corrected body section image to be consistent with the image resolution of the reference body section image.

[0182] In order to adjust the image resolution of the corrected body section image to make it consistent with the image resolution of the reference body section image, it is necessary to obtain the resolution information of the two images. The resolution is usually expressed as dots per inch (DPI) or the aspect ratio per pixel. The low-resolution image is processed by cubic interpolation (also called bicubic interpolation). Cubic interpolation is an advanced interpolation method for image enlargement in two-dimensional space. It takes into account the influence of 16 surrounding pixels to generate a smoother image.

[0183] The embodiment of the present application adjusts the image resolution of the corrected body section image to be consistent with the image resolution of the reference body section image, including calculating the resolution adjustment coefficient sacle=voxelspace1 / voxelspace0, where voxelspace0 represents the physical size of the reference body voxel, and voxelspace1 represents the physical size of the corrected body voxel; comparing the image resolution of the corrected body section image with the image resolution of the reference body section image to obtain a low-resolution image, and processing the low-resolution image by cubic interpolation, the formula is as follows:

[0184] P(x)=a0+a1(x-x0)+a2(x-x0) 2 +a3(x-x0) 3 ;

[0185] Among them, P(x) represents the pixel value of the interpolation point, x represents the position of the interpolation point, (a0, a1, a2, a3) represents the coefficients calculated based on the surrounding pixels, and x0 represents the reference corresponding pixel value.

[0186] Specifically, for the subsequent fusion of images, it is necessary to ensure that the physical size of the pixels of the image after adjusting the resolution is consistent. By comparison, it is found that the resolution of MR is larger than that of PET. In order to maximize the preservation of MR image information, the low-resolution PET image is enlarged by bicubic spline interpolation. The size of the enlarged PET is as follows:

[0187] The MR voxel size is 0.8594, the PET voxel size is 1.5625, and the PET magnification factor is calculated as scale = 1.5625 / 0.8594 = 1.8181.

[0188] The size of the MR coronal image is 512*512, the size of the PET coronal image before adjustment is 384*158, and the size of the PET coronal image after adjustment is 698*287.

[0189] The size of the MR sagittal image is 174*512, the size of the PET sagittal image before adjustment is 384*158, and the size of the PET sagittal image after adjustment is 698*287.

[0190] The size of the MR axial image is 384*384, the size of the PET axial image before adjustment is 384*158, and the size of the PET sagittal image after adjustment is 698*698.

[0191] Step 106: normalize and fuse the corrected body section image and the reference body section image.

[0192] Normalization is the process of scaling data so that it falls into a small specific interval (usually [0, 1] or [-1, 1]). In image processing, this helps eliminate differences between different images due to factors such as brightness and contrast, making it easier to process them later.

[0193] For the normalization processing of the cross-section diagram of the correction body and the cross-section diagram of the reference body, the cross-section diagrams of the correction body and the reference body are firstly read from the file. These images are usually two-dimensional grayscale images or color images.

[0194] For grayscale images, we need to find the minimum and maximum values ​​in the image.

[0195] For color images, each color channel (red, green, and blue) needs to be normalized separately, or converted to grayscale and then normalized.

[0196] Fusion processing is the process of merging two or more images into one image. In medical image processing, it is usually used to combine image information from different imaging modalities or different time points to provide more comprehensive information.

[0197] The pixel-based fusion method defines how to combine the pixel values ​​of the two images according to application requirements. For the normalized corrected body section image and the reference body section image, each corresponding pixel of them is traversed, and the fused pixel value of each position is calculated according to the defined fusion rules. The fused pixel values ​​are combined into a new image, which is the fused image.

[0198] The embodiment of the present application performs normalization and fusion processing on the corrected body section diagram and the reference body section diagram, including extracting two-dimensional images from the corrected body section diagram and the reference body section diagram, respectively, and defining the extracted two-dimensional images. The formula is as follows:

[0199] I'(i,j)=(I(i,j)-min)*(255 / (max-min)

[0200] Among them, i and j represent the coordinate position of the two-dimensional image; i=1, 2, 3, ..., R; j=1, 2, 3, ..., C; I(i, j) represents the grayscale value of any pixel at a two-dimensional image, I'(i, j) represents the normalized grayscale value of any pixel at a two-dimensional image, (R, C) represents the size of the matrix I(i, j), and (min, max) represents the grayscale range of the two-dimensional image; image pixels are the category of digital images, and a two-dimensional image has a total of R*C points, which means there are R*C pixels, where i and represent the coordinate position of the pixels of the two-dimensional image, and each pixel value can be obtained.

[0201] All pixels of the two-dimensional images in the cross-section of the correction body and the cross-section of the reference body are traversed respectively to obtain the image pixels of the correction body and the image pixels of the reference body; the fusion processing includes a pixel-based fusion method, and the formula is as follows:

[0202] C(i,j)=a*C1(i,j)+(1-a)*C2(τ,j)

[0203] Wherein, C1 represents the image pixel of the reference body, C2 represents the image pixel of the correction body, and a represents transparency.

[0204] Specifically, since the effective value of MR image pixels is generally between zero and several thousand, while the effective value of PET image pixels is generally between zero and several hundred thousand, the pixel intervals are not in the same range, and fusion is meaningless. The pixel values ​​need to be uniformly mapped to the range of 0 to 255. By traversing each pixel for mapping, it can be obtained that the anatomical tissue correspondence between MR and PET in the fusion image before correction is incorrect, while the tissue correspondence after correction is correct.

[0205] like Figure 2 As shown, the embodiment of the present application provides a medical tomographic image correction and fusion device, including a data acquisition module 201, a data calculation module 202, a data judgment module 203 and a correction and fusion module 204, wherein:

[0206] The data acquisition module 201 is configured to acquire two sets of Dicom medical images, read the image sequences of the two sets of Dicom medical images respectively, and obtain a reference sequence and a correction sequence;

[0207] The data calculation module 202 is configured to process the reference sequence and the correction sequence to obtain volume data, the volume data including a reference volume corresponding to the reference sequence and a correction volume corresponding to the correction sequence; and calculate outer bounding box space physical data according to the volume data, the outer bounding box space physical data including a reference volume data outer bounding box corresponding to the reference volume and a correction volume data outer bounding box corresponding to the correction volume.

[0208] The data judgment module 203 is configured to judge whether the reference body and the correction body have an intersection according to the outer bounding box of the reference body data and the outer bounding box of the correction body data, and if no intersection exists, the process ends; if there is an intersection, the reference body and the correction body are corrected and fused in volume space;

[0209] The correction fusion module 204 is configured to calculate the center point offset and correction matrix of the correction body relative to the reference body; calculate the section position and section direction vector of the correction body according to the center point offset and the correction matrix, and generate a section view of the correction body; adjust the image resolution of the section view of the correction body to be consistent with the image resolution of the section view of the reference body; and perform normalization and fusion processing on the section view of the correction body and the section view of the reference body.

[0210] The medical tomographic image correction and fusion device of the embodiment of the present application can be a computer device, or a component in the computer device, such as an integrated circuit or a chip. The computer device can be a terminal, or it can be other devices other than a terminal. Exemplarily, the computer device can be a mobile phone, a tablet computer, a laptop computer, a PDA, a vehicle-mounted computer device, a mobile Internet device (Mobile Internet Device, MID), an ultra-mobile personal computer (Ultra-Mobile Personal Computer, UMPC), a netbook or a personal digital assistant (Personal Digital Assistant, PDA), etc. It can also be a server, a network attached storage (Network Attached Storage, NAS), a personal computer (Personal Computer, PC), etc., which is not specifically limited in the embodiment of the present application.

[0211] The medical tomographic image correction and fusion device provided in the embodiment of the present application can achieve Figure 1 To avoid repetition, the various processes implemented in the embodiment of the medical tomographic image correction and fusion method are not described here.

[0212] The present application also provides a computer device, such as Figure 3 As shown, the computer device includes a processor 301 and a memory 302. The memory 302 stores programs or instructions that can be run on the processor 301. When the program or instruction is executed by the processor 301, the various steps of the above-mentioned medical tomographic image correction and fusion method are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0213] It should be noted that the computer device in the embodiment of the present application includes the above-mentioned mobile computer device and non-mobile computer device.

[0214] The memory 302 can be used to store software programs and various data. The memory 302 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instructions required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory 302 may include a volatile memory or a non-volatile memory, or the memory 302 may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM). The memory 302 of the embodiment of the present application includes but is not limited to these and any other suitable types of memory.

[0215] The processor 301 may include one or more processing units; optionally, the processor 301 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor 301.

[0216] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, each process of the above-mentioned medical tomographic image correction and fusion method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0217] An embodiment of the present application also provides a chip, which includes a processor and a communication interface, wherein the communication interface and the processor are coupled, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned medical tomographic image correction and fusion method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0218] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0219] An embodiment of the present application also provides a computer program product, which is stored in a storage medium. The program product is executed by at least one processor to implement the various processes of the above-mentioned medical tomographic image correction and fusion method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0220] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0221] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. A medical tomographic image correction and fusion method, characterized in that: The method comprises: Obtain two sets of Dicom medical images, read the image sequences of the two sets of Dicom medical images respectively, and obtain a reference sequence and a correction sequence; Processing the reference sequence and the correction sequence to obtain volume data, wherein the volume data includes a reference volume corresponding to the reference sequence and a correction volume corresponding to the correction sequence; Calculating outer bounding box space physical data according to the volume data, the outer bounding box space physical data including a reference volume data outer bounding box corresponding to the reference volume and a correction volume data outer bounding box corresponding to the correction volume; According to the outer bounding box of the reference body volume data and the outer bounding box of the correction body volume data, it is determined whether the reference body and the correction body have an intersection, and if no intersection exists, the process ends; if an intersection exists, the reference body and the correction body are subjected to volume space correction fusion, and the volume space correction fusion method includes obtaining a reference body section view, calculating a center point offset and a correction matrix of the correction body relative to the reference body; calculating a section position and a section direction vector of the correction body according to the center point offset and the correction matrix, and generating a correction body section view; Adjusting the image resolution of the corrected body section image to be consistent with the image resolution of the reference body section image; The corrected body section image and the reference body section image are normalized and fused.

2. The medical tomographic image correction and fusion method according to claim 1, characterized in that: The processing of the reference sequence and the correction sequence to obtain volume data includes: The spatial physical information of the reference sequence and the correction sequence are read respectively to obtain slices, wherein the slices correspond to the Dicom medical image, and the slices are sorted from small to large according to the spatial physical information to obtain volume data, and the noise in the Dicom medical image is removed by median filtering to enhance the contrast of the Dicom medical image.

3. The medical tomographic image correction and fusion method according to claim 1, characterized in that: The step of calculating the outer bounding box space physical data according to the volume data includes: Obtain a Dicom tag, wherein the Dicom tag includes a position tag, an orientation tag, and a pixel tag; According to the position tag, the spatial matrix F of the first point in the upper left corner of the Dicom medical image is obtained, and the formula is as follows: The direction cosine matrix D of the Dicom medical image is obtained according to the orientation label, and the formula is as follows: The physical size of the pixel of the Dicom medical image is obtained according to the pixel label, and the formula of the coordinate matrix P of any point on the Dicom medical image is as follows: According to the space matrix F, the direction cosine matrix D and the coordinate matrix P, the arbitrary physical space position P′ of the Dicom medical image is calculated by matrix transformation, and the formula is as follows: According to the calculation formula of the arbitrary physical space position P′ of the Dicom medical image, the corner points of the outer bounding box are calculated, and the minimum coordinate value and the maximum coordinate value of the outer bounding box in the X, Y, and Z axis directions are obtained according to the corner points to obtain bounds=(xmin, xmax, ymin, ymax, zmin, zmax); According to the outer bounding box of the reference volume data and the outer bounding box of the corrected volume data, it is determined whether the reference volume and the corrected volume have an intersection, and the determination formula is as follows: XIntersects=bounds1[1]≥bounds2[0]&&bounds2[1]≥bounds1[0]; YIntersects=bounds1[3]≥bounds2[2]&&bounds2[3]≥bounds1[2]; ZIntersects=bounds1[5]≥bounds2[4]&&bounds2[5]≥bounds1[4]; When the judgment formulas of XIntersects, YIntersects and ZIntersects are satisfied at the same time, the reference body and the correction body have an intersection.

4. The medical tomographic image correction and fusion method according to claim 3, characterized in that: The calculating the center point offset of the correction body relative to the reference body and the correction matrix comprises: Aligning the centers of the correction body and the reference body, and rotating the correction body to be in the same orientation as the reference body; Calculate the body center coordinates (X1, Y1, Z1) of the reference body and the body center coordinates (X2, Y2, Z2) of the correction body according to any physical space position P′ of the Dicom medical image; Obtaining the center point offset of the correction body offset={X1-X2, Y1-Y2, Z1-Z2} according to the body center coordinates (X1, Y1, Z1) of the reference body and the body center coordinates (X2, Y2, Z2) of the correction body; The correction matrix M is set as the direction matrix of the reference body, which is expressed as follows:

5. The medical tomographic image correction and fusion method according to claim 4, characterized in that: The obtaining of the cross-sectional view of the reference body, and the calculation of the center point offset of the correction body relative to the reference body and the correction matrix; Calculating the sectioning position and sectioning direction vector of the correction body according to the center point offset and the correction matrix to generate a sectioning diagram of the correction body includes: The sectioning position and sectioning direction of the reference body are obtained, and the sectioning position P1 and the sectioning direction vector D1 of the correction body are calculated according to the sectioning position P0 of the reference body and the sectioning direction vector D0 of the reference body to generate a sectioning diagram of the correction body; the formula of the sectioning position P1 of the correction body is as follows: P1.x=(P0.x*voxelspace0+offset.x) / voxelspace1 P1.y=(P0.y*voxelspace0+offset.y) / voxelspace1 P1.z=(P0.z*voxelspace0+offset.z) / voxelspace1 Among them, P0 represents the section position of the reference body, offset represents the center point offset of the correction body, voxelspace0 represents the physical size of the reference body voxel, and voxelspace1 represents the physical size of the correction body voxel; P0.x, P0.y, and P0.z represent the components of the section position of the reference body in the x, y, and z directions respectively; P1.x, P1.y, and P1.z represent the components of the section position of the correction body in the x, y, and z directions respectively; offset.x, offset.y, and offset.z represent the components of the center point offset of the correction body in the x, y, and z directions respectively; The formula of the cutting direction vector D1 of the correction body is as follows: D1=D0×M Wherein, D0 represents the cutting direction vector of the reference body, and M represents the direction matrix of the reference body.

6. The medical tomographic image correction and fusion method according to claim 1, characterized in that: The step of adjusting the image resolution of the corrected body section image to be consistent with the image resolution of the reference body section image comprises: Calculate the resolution adjustment factor, the formula is as follows: sacle=voxelspace1 / voxelspace0 Among them, voxelspace0 represents the physical size of the reference volume voxel, and voxelspace1 represents the physical size of the correction volume voxel; Compare the image resolution of the corrected body section image with the image resolution of the reference body section image to obtain a low-resolution image, and process the low-resolution image by cubic interpolation. The formula is as follows: <h2 style=";text-align:left;direction:ltr">P(x) = a0 + a1(x-x0) + a2(x-x0)<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +a3(x-x0)<h2 style=";text-align:left;direction:ltr"> 3 Among them, P(x) represents the pixel value of the interpolation point, x represents the position of the interpolation point, (a0, a1, a2, a3) represents the coefficients calculated based on the surrounding pixels, and x0 represents the reference corresponding pixel value.

7. The medical tomographic image correction and fusion method according to claim 1, characterized in that: The normalizing and fusing of the corrected body section image and the reference body section image comprises: The two-dimensional images in the correction body section view and the reference body section view are extracted respectively, and the extracted two-dimensional images are defined as follows: I'(i,j)=(I(i,j)-mmin)*(255 / (max-min) Wherein, i and j represent the coordinate position of the two-dimensional image; i = 1, 2, 3, ..., R; j = 1, 2, 3, ..., C; I(i, j) represents the grayscale value of any pixel at a two-dimensional image, I'(i, j) represents the normalized grayscale value of any pixel at a two-dimensional image, (R, C) represents the size of the matrix I(i, j), and (min, max) represents the grayscale range of the two-dimensional image; Traversing all pixels of the two-dimensional images in the cross-sectional view of the correction body and the cross-sectional view of the reference body respectively to obtain image pixels of the correction body and image pixels of the reference body; The fusion process includes a pixel-based fusion method, and the formula is as follows: C(i,j)=a*C1(i,j)+(1-a)*C2(i,j) Wherein, C1 represents the image pixel of the reference body, C2 represents the image pixel of the correction body, and a represents transparency.

8. A medical tomographic image correction and fusion device, characterized in that: The device comprises: A data acquisition module is configured to acquire two sets of Dicom medical images, read the image sequences of the two sets of Dicom medical images respectively, and obtain a reference sequence and a correction sequence; a data calculation module configured to process the reference sequence and the correction sequence to obtain volume data, wherein the volume data includes a reference volume corresponding to the reference sequence and a correction volume corresponding to the correction sequence; and calculate outer bounding box space physical data according to the volume data, wherein the outer bounding box space physical data includes a reference volume data outer bounding box corresponding to the reference volume and a correction volume data outer bounding box corresponding to the correction volume. A data judgment module is configured to judge whether the reference body and the correction body have an intersection according to the outer bounding box of the reference body data and the outer bounding box of the correction body data, and if no intersection exists, the process ends; if an intersection exists, the reference body and the correction body are corrected and fused in volume space; The correction and fusion module is configured to calculate the center point offset and correction matrix of the correction body relative to the reference body; calculate the section position and section direction vector of the correction body according to the center point offset and the correction matrix to generate a section view of the correction body; adjust the image resolution of the section view of the correction body to be consistent with the image resolution of the section view of the reference body; and perform normalization and fusion processing on the section view of the correction body and the section view of the reference body.

9. A medical tomographic image correction and fusion system, the system comprising a processor and a memory, wherein a computer program is stored in the memory, characterized in that: The computer program is loaded and executed by the processor to implement the medical tomographic image correction and fusion method according to any one of claims 1 to 7.

10. A computer-readable storage medium, wherein a computer program is stored in the storage medium, characterized in that: When the computer program is executed by a processor, it is used to implement the medical tomographic image correction and fusion method as described in any one of claims 1 to 7.

Images