A tumor image automatic segmentation and three-dimensional reconstruction system

Abstract

The application discloses a tumor image automatic segmentation and three-dimensional reconstruction system and relates to the technical field of image processing. The system comprises a tumor image acquisition module, a region selection module, a deformation degree determination module, an image correction module and a three-dimensional model construction module. The tumor image is acquired through multiple perspectives; then, a reference region is selected through deformation information of a region in the image, and the tumor image is corrected based on the reference region under each perspective, so as to construct a tumor three-dimensional model. The scheme corrects the tumor position discontinuity and morphological distortion problems caused by respiratory motion, obtains more accurate corrected tumor images, and thus the accuracy of the generated tumor three-dimensional model can be improved.

Description

Technical Field

[0001] This application relates to the field of medical image processing technology, specifically to an automatic segmentation and three-dimensional reconstruction system for tumor images. Background Technology

[0002] Upper gastrointestinal tumors refer to malignant tumors occurring in the esophagus, stomach, and duodenum, with common types including esophageal cancer, gastric cancer, and duodenal cancer. Early symptoms are often subtle, with dysphagia, upper abdominal pain, and hematemesis being typical manifestations, requiring imaging-assisted diagnosis. Medical imaging technology plays a crucial role in the auxiliary diagnosis of upper gastrointestinal tumors, providing doctors with a wealth of valuable information. To achieve accurate and consistent tumor region identification, pre-trained deep neural networks are currently widely used to identify and segment tumor regions online. After acquiring segmented images of the tumor region, 3D reconstruction technology is used to convert the two-dimensional tumor segmentation data into a 3D model, thus providing powerful assistance to doctors in surgical planning and radiotherapy design.

[0003] Currently, deep learning methods are commonly used to automatically segment tumor images, and 3D reconstruction is performed by combining multi-view images. However, during the tumor image acquisition phase, the patient's respiratory movements cause dynamic deformation of the tumor region in different frames. This results in discontinuities and morphological distortions in the tumor region in upper gastrointestinal tumor images, leading to lower accuracy in the generated 3D tumor model. Summary of the Invention

[0004] To address the issue of low accuracy in 3D tumor modeling using tumor images due to patient respiration, this application provides an automatic tumor image segmentation and 3D reconstruction system, the specific technical solution of which is as follows:

[0005] This application provides an automatic segmentation and three-dimensional reconstruction system for tumor images, the system comprising:

[0006] The tumor image acquisition module is used to acquire initial tumor images from multiple acquisition perspectives. The initial tumor images include multiple image regions.

[0007] The region selection module is used to select a reference region from multiple image regions based on the biological tissue deformation information of the image region under each acquisition view during the target respiratory cycle. The reference region is the image region with the least degree of biological tissue deformation.

[0008] The deformation determination module is used to compare the tumor region in the image area under each acquisition view with the reference region to obtain the respiratory deformation of each pixel in the tumor region under each acquisition view.

[0009] The image correction module is used to correct the position of each pixel in the tumor region in the initial tumor image under each acquisition view based on the respiratory deformation of each pixel in the tumor region under each acquisition view, so as to obtain the corrected tumor image under each acquisition view.

[0010] The 3D model building module is used to construct a 3D model of the tumor based on various corrected tumor images.

[0011] In some possible implementations, the region selection module specifically includes the following units:

[0012] The image acquisition unit is used to acquire historical tumor images from various acquisition angles within the target respiratory cycle.

[0013] The image sorting unit is used to sort the historical tumor images and the corresponding initial tumor images from each acquisition view in chronological order to construct a tumor image sequence from each acquisition view.

[0014] The linkage determination unit is used to perform the following for each acquisition viewpoint: based on the tumor image sequence under the acquisition viewpoint, determine the respiratory linkage of each image region under the acquisition viewpoint.

[0015] The evaluation value determination unit is used to determine the respiratory deformation evaluation value of each image region based on the respiratory linkage of each image region under each acquisition view.

[0016] The region determination unit is used to determine the image region with the smallest respiratory deformation evaluation value as the reference region.

[0017] In some possible implementations, the linkage determination unit is specifically used for:

[0018] In the tumor image sequence under the acquisition perspective, the image frame difference and image matching degree of the target image region between adjacent tumor images are obtained, and the target image region is any image region.

[0019] By utilizing the image frame difference and image matching degree between adjacent tumor images of the target image region, the respiratory linkage of the target image region under the acquisition viewpoint can be determined.

[0020] In some possible implementations, the evaluation value determination unit is specifically used for:

[0021] The acquisition perspectives are sorted according to a preset traversal order to obtain the arrangement order of the acquisition perspectives;

[0022] The respiratory coordination difference is obtained by calculating the absolute value of the difference in respiratory coordination between target image regions under adjacent acquisition perspectives. The target image region is any image region.

[0023] By using the differences in respiratory linkage, the respiratory deformation evaluation value of the target image region is determined.

[0024] In some possible implementations, the deformation determination module specifically includes the following units:

[0025] The region filtering unit is used to filter out each tumor region from each image region under the target acquisition view. The target acquisition view can be any acquisition view.

[0026] The deformation ratio determination unit is used to determine the relative deformation ratio of each pixel in each tumor region based on the distance between each pixel in each tumor region and the centroid of the corresponding tumor region.

[0027] The displacement determination unit is used to determine the relative displacement of each pixel in each tumor region based on the distance between each pixel in each tumor region and the centroid of the reference region.

[0028] The deformation degree determination unit is used to determine the respiratory deformation degree of each pixel in each tumor region under the target acquisition view by using the relative deformation ratio and relative displacement degree of each pixel in each tumor region.

[0029] In some possible implementations, the deformation ratio determining unit is specifically used for:

[0030] Based on the first distance between the j-th pixel and the centroid of the corresponding tumor region in the tumor image sequence under the target respiratory cycle acquisition perspective, the first average distance between the j-th pixel and the centroid of the tumor region is determined, where j is a positive integer;

[0031] Using the first distance and the corresponding first average distance of the j-th pixel in the second tumor region, the relative deformation ratio of the j-th pixel in the second tumor region under the target acquisition view is determined. The second tumor region is the first tumor region in the initial tumor image.

[0032] In some possible implementations, the displacement determination unit is specifically used for:

[0033] Based on the second distance between the j-th pixel in each first tumor region and the centroid of the reference region in the tumor image sequence under the target acquisition view within the target respiratory cycle, the second average distance between the j-th pixel and the centroid of the reference region is determined, where j is a positive integer;

[0034] Using the second distance of the j-th pixel in the second tumor region and the corresponding second average distance, the relative displacement of the j-th pixel in the second tumor region under the target acquisition view is determined. The second tumor region is the first tumor region in the initial tumor image.

[0035] In some possible implementations, the image correction module is specifically used for:

[0036] Based on the respiratory deformation degree of each pixel in the tumor region in the initial tumor image under each acquisition view, the deformation correction amount of each pixel in the tumor region in the initial tumor image under each acquisition view is determined respectively.

[0037] Based on the deformation correction amount of each pixel in the tumor region in the initial tumor image under each acquisition view, the position of each pixel in the tumor region in the initial tumor image under each acquisition view is corrected to obtain the corrected tumor image under each acquisition view.

[0038] In some possible implementations, the 3D model building module is specifically used for:

[0039] The modified tumor images were sorted according to the order of image acquisition to determine the spatial location of each modified tumor image.

[0040] Each corrected tumor image is subjected to consistency processing to obtain the corresponding consistent tumor image.

[0041] The images of the tumors are stacked according to their corresponding spatial locations to obtain the image stacking result;

[0042] A three-dimensional model of the tumor was constructed based on the image stacking results.

[0043] In some possible implementations, the tumor image acquisition module is specifically used for:

[0044] Send image acquisition signals to the tumor scanner so that the tumor scanner can scan the tumor of the target patient;

[0045] Receive raw signal data from the target patient collected by the tumor scanner;

[0046] The original signal data is converted into a two-dimensional image format by a target reconstruction algorithm, resulting in initial tumor images from multiple acquisition perspectives.

[0047] This application has the following beneficial effects:

[0048] This application provides an automatic tumor image segmentation and 3D reconstruction system. First, an initial tumor image is acquired from multiple viewing angles using a tumor image acquisition module, providing a comprehensive image data foundation for subsequent processing. Then, a region selection module selects the image region with the least deformation as a baseline region, finding a relatively stable reference standard. Next, a deformation degree determination module compares each pixel in the tumor region from each acquisition angle with the baseline region to obtain the respiratory deformation degree of each pixel, accurately quantifying the degree of deformation caused by respiration. Then, an image correction module corrects the position of each pixel based on its respiratory deformation degree from each acquisition angle, obtaining corrected tumor images for each acquisition angle. This directly addresses the discontinuity and morphological distortion of the tumor position caused by respiratory motion, making the corrected tumor images closer to the true shape and location of the tumor, greatly improving the accuracy of the image data. Finally, a 3D model construction module constructs a 3D tumor model based on the corrected tumor images. Thus, the embodiments of this application can correct the problems of discontinuity in tumor location and distortion in shape caused by respiratory motion, and obtain more accurate corrected tumor images, thereby improving the accuracy of the generated three-dimensional tumor model. Attached Figure Description

[0049] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A schematic diagram of a tumor image automatic segmentation and three-dimensional reconstruction system provided in one embodiment of this application;

[0051] Figure 2 This is a schematic diagram of the structure of a region selection module provided in one embodiment of this application;

[0052] Figure 3 This is a schematic diagram of the deformation determination module provided in one embodiment of this application. Detailed Implementation

[0053] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an automatic tumor image segmentation and three-dimensional reconstruction system proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0054] 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 pertains.

[0055] The following describes a specific embodiment of an automatic tumor image segmentation and three-dimensional reconstruction system provided in this application.

[0056] like Figure 1 The diagram shows a schematic of an automatic tumor image segmentation and 3D reconstruction system. This automatic tumor image segmentation and 3D reconstruction system 100 includes a tumor image acquisition module 110, a region selection module 120, a deformation determination module 130, an image correction module 140, and a 3D model construction module 150.

[0057] The tumor image acquisition module 110 is used to acquire initial tumor images from multiple acquisition perspectives, and the initial tumor images include multiple image regions.

[0058] In this embodiment, the acquisition perspective refers to the perspective from which tumor images are acquired from different angles or positions. For example, in medical imaging examinations, tumor images may be acquired through different probe positions, scanning directions, or imaging modes (such as different combinations of scanning slices in CT scans), and each different acquisition method corresponds to an acquisition perspective.

[0059] Initial tumor images refer to raw image data about a tumor acquired from multiple acquisition perspectives using specific imaging equipment (such as CT). These images contain information such as the morphology and structure of the tumor and its surrounding biological tissues, and each initial tumor image consists of multiple image regions.

[0060] An imaging region refers to a localized area with certain boundaries and characteristics that is segmented from the initial tumor image. These regions can be segmented based on image segmentation techniques according to the grayscale, texture, and other features of the tissue, and are used for subsequent detailed analysis of the tumor and its surrounding tissues. For example, a deep learning neural network (U-Net) trained on large-scale labeled data can be used to automatically segment the initial tumor image to obtain several imaging regions.

[0061] As an example, the tumor image acquisition module 110 acquires initial tumor images using various imaging devices or from different acquisition parameter settings. For example, for a CT scan, multiple scans can be performed using different slice thicknesses and scanning angles.

[0062] Then, the acquired initial tumor images are preprocessed, including denoising and image enhancement, to improve image quality and facilitate subsequent analysis.

[0063] The region selection module 120 is used to select a reference region from multiple image regions based on the biological tissue deformation information of the image region under each acquisition view during the target respiratory cycle. The reference region is the image region with the smallest degree of biological tissue deformation.

[0064] In this embodiment, the target respiratory cycle refers to a specific respiratory cycle (e.g., a complete inhalation and exhalation process) selected as the time range for analysis during the medical image acquisition process, taking into account the impact of respiratory movements on the morphology of the tumor and surrounding tissues. Within this cycle, the deformation information of biological tissues is studied.

[0065] Biological tissue deformation information describes the changes in shape and position of biological tissues (including tumors and surrounding normal tissues) during a target respiratory cycle. By using techniques such as image registration and optical flow, tumor images at different time points (different stages of the respiratory cycle) can be compared to calculate deformation parameters such as displacement and strain in various directions.

[0066] A baseline region is an image region with the least degree of biological tissue deformation identified through analysis and comparison among multiple image regions. This region exhibits relatively stable morphological changes during respiration and can serve as a reference standard for subsequent comparisons and corrections.

[0067] As an example, the region selection module 120 uses image registration technology for each acquisition viewpoint to register the initial tumor image within the target respiratory cycle with historical tumor images at different time points within the target respiratory cycle, and calculates the displacement field of each image region at each time point, thereby obtaining the deformation information of biological tissue.

[0068] By analyzing deformation information, deformation degree indices, such as average displacement and maximum strain, are calculated for each image region. The deformation degree indices of each image region are compared, and the image region with the smallest deformation degree is selected as the benchmark region.

[0069] The deformation determination module 130 is used to compare the tumor region in the image region under each acquisition view with the reference region to obtain the respiratory deformation of each pixel in the tumor region under each acquisition view.

[0070] In this embodiment, the tumor region refers to the region containing tumor cells that is clearly identified in the initial tumor image. It is usually segmented from the image region using medical image analysis techniques (such as image feature-based classification, machine learning algorithms, etc.).

[0071] Breathing deformation is an indicator used to quantify the degree of deformation of each pixel in a tumor region during respiration. By comparing the pixels in the tumor region with a baseline region from each acquisition viewpoint, the change in spatial position of each pixel is calculated, and this change is the breathing deformation.

[0072] As an example, the deformation determination module 130 first uses image segmentation technology to accurately segment the tumor region from the initial tumor images from each acquisition viewpoint.

[0073] Then, using image registration or feature matching, the tumor region is aligned with the reference region, and the spatial displacement of each pixel in the tumor region relative to the corresponding position in the reference region is calculated, thereby determining the respiratory deformation degree of that pixel.

[0074] The image correction module 140 is used to correct the position of each pixel in the tumor region in the initial tumor image under each acquisition view based on the respiratory deformation of each pixel in the tumor region under each acquisition view, so as to obtain the corrected tumor image under each acquisition view.

[0075] In this embodiment, the corrected tumor image is obtained by adjusting the position of each pixel in the original tumor region in the initial tumor image based on the respiratory deformation of each pixel in the tumor region under each acquisition viewpoint. The corrected image more accurately reflects the true shape and location of the tumor when it is not disturbed by respiratory motion.

[0076] As an example, the image correction module 140 adjusts the original coordinates of each pixel based on the calculated breathing deformation. For instance, if the breathing deformation of a pixel indicates a displacement of 2mm in the X direction, then the X coordinate of that pixel is added to or subtracted by 2mm during correction.

[0077] The above position correction operation is performed on all pixels in the tumor region of the initial tumor image to obtain the corrected tumor region, and then the corrected tumor image is generated from each acquisition view.

[0078] The 3D model building module 150 is used to build a 3D model of the tumor based on various corrected tumor images.

[0079] In this embodiment, the tumor three-dimensional model refers to a model constructed using various modified tumor images and three-dimensional reconstruction techniques (such as surface rendering, volume rendering, etc.) that can intuitively display the three-dimensional spatial structure, shape, and location of the tumor.

[0080] As an example, the 3D model building module 150 uses 3D reconstruction algorithms, such as Marching Cubes algorithm (surface rendering) or ray casting algorithm (volume rendering), to integrate and reconstruct tumor information from various corrected tumor images.

[0081] During the reconstruction process, it is necessary to consider the spatial relationship and grayscale information between the images to ensure that the constructed three-dimensional model can accurately reflect the true morphology and spatial structure of the tumor.

[0082] As an optional embodiment, such as Figure 2 As shown, the region selection module 120 specifically includes the following units:

[0083] The image acquisition unit 121 is used to acquire historical tumor images from various acquisition angles within the target respiratory cycle;

[0084] Image sorting unit 122 is used to sort historical tumor images and corresponding initial tumor images under each acquisition view in chronological order to construct a tumor image sequence under each acquisition view.

[0085] The linkage determination unit 123 is used to perform the following for each acquisition view: based on the tumor image sequence under the acquisition view, determine the respiratory linkage of each image region under the acquisition view.

[0086] The evaluation value determination unit 124 is used to determine the respiratory deformation evaluation value of each image region based on the respiratory linkage of each image region under each acquisition view.

[0087] The region determination unit 125 is used to determine the image region with the smallest respiratory deformation evaluation value as the reference region.

[0088] In this embodiment, historical tumor images are used to characterize tumor image data acquired from various acquisition perspectives at past moments within the target respiratory cycle. These images record information such as the morphology and location of the tumor at different time points.

[0089] A tumor imaging sequence is a collection of images formed by arranging historical tumor images from various acquisition perspectives and their corresponding initial tumor images in chronological order. It reflects the changes of the tumor over time within the target respiratory cycle.

[0090] Respiratory coherence describes the degree of correlation between an imaging region and respiratory movements during respiration. A high respiratory coherence occurs when the changes in an imaging region align with the trends of respiratory movements during respiration; conversely, a low respiratory coherence occurs when they do not.

[0091] The respiratory deformation rating is an indicator used to quantify the degree of deformation of an image region during respiration. This value comprehensively considers factors such as morphological changes and positional movement of the image region during the respiratory cycle; a higher rating indicates a greater degree of deformation.

[0092] As an example, the image acquisition unit 121 first determines the target respiratory cycle at the current time based on the current time corresponding to the initial tumor image. Then, it acquires the historical tumor images corresponding to each acquisition viewpoint within the target respiratory cycle.

[0093] Then, the image sorting unit 122 arranges the historical tumor images and the initial tumor images under each acquisition view in chronological order according to the acquisition timestamps of the historical tumor images and the initial tumor images, forming a tumor image sequence under each acquisition view.

[0094] Then, the linkage determination unit 123 analyzes the morphological and positional changes of each image region based on the tumor image sequence under the acquisition perspective, thereby determining the respiratory linkage of each image region under the acquisition perspective.

[0095] Then, the evaluation value determination unit 124 comprehensively considers the respiratory linkage of the image area under different acquisition angles, establishes a mathematical model or algorithm, and transforms the respiratory linkage into a respiratory deformation evaluation value. For example, the average respiratory linkage of the image area under different acquisition angles can be calculated to obtain the respiratory deformation evaluation value of the image area.

[0096] Finally, the region determination unit 125 compares and sorts the respiratory deformation evaluation values ​​of all image regions, and selects the image region with the smallest respiratory deformation evaluation value as the benchmark region.

[0097] This embodiment constructs tumor imaging sequences and analyzes the respiratory linkages within the imaging regions, enabling a more comprehensive understanding of the dynamic changes of the tumor during respiration, thereby allowing for a more accurate assessment of the tumor's characteristics and behavior. This, in turn, improves the accuracy of subsequent three-dimensional tumor models.

[0098] As an optional embodiment, the linkage determination unit 123 is specifically used for:

[0099] In the tumor image sequence under the acquisition perspective, the image frame difference and image matching degree of the target image region between adjacent tumor images are obtained, and the target image region is any image region.

[0100] By utilizing the image frame difference and image matching degree between adjacent tumor images of the target image region, the respiratory linkage of the target image region under the acquisition viewpoint can be determined.

[0101] In this embodiment, image frame difference is used to characterize the pixel value difference of the target image region between two adjacent tumor images in a tumor image sequence. Specifically, it can be obtained by calculating the difference between the gray values ​​or other feature values ​​of corresponding pixels in two tumor images, reflecting the morphological or gray-level changes of the target image region between adjacent time points.

[0102] Image matching degree is an indicator used to measure the similarity of target image regions in two adjacent tumor images. It is usually calculated using image registration algorithms. A higher matching degree indicates a greater similarity between the target image regions in the two images, meaning that the morphology and position of the target image regions change less between adjacent time points.

[0103] As an example, the linkage determination unit 123 determines the position of the target image region in each pair of adjacent tumor images in the tumor image sequence. Then, it calculates the difference in feature values ​​(such as gray values) of corresponding pixels of the target image region in the two tumor images, which can be done using methods such as absolute difference or squared difference. Finally, it performs statistical processing on the differences of all pixels (such as calculating the average value or root mean square value) to obtain the image frame difference between the target image region and the pair of adjacent tumor images.

[0104] Then, a suitable image registration algorithm is selected, such as feature-based registration algorithms (SIFT, SURF, etc.) or gray-level-based registration algorithms (mutual information, normalized cross-correlation, etc.). Using the target image region as the registration object, registration is performed between two adjacent tumor images to find the transformation parameters that best match the target image region in the two tumor images. Based on the registration results, matching degree indices are calculated, such as the similarity score of the matched target image region and the registration error, which are used as the quantitative value of the image matching degree.

[0105] Finally, based on the image frame difference and image matching degree between adjacent tumor images of the target image region, the respiratory linkage of the target image region under the acquisition viewpoint is determined using the following formula:

[0106] ;

[0107] In the formula, The function is used to characterize the respiratory linkage of the i-th image region under the acquisition viewpoint, and N is used to characterize the number of tumor images in the tumor image sequence, where N is a positive integer not less than 2. This is used to characterize the image frame difference between the i-th image region and the tumor image in the n-th and (n+1)-th frames. The exp is used to characterize the image matching degree between the i-th image region and the tumor image in the n-th and (n+1)-th frames, and to characterize the exponential operation of the base e of the natural logarithm.

[0108] The greater the image frame difference between the i-th image region and each adjacent tumor image, the greater the respiratory linkage of the i-th image region under the acquisition view; the greater the image matching degree between the i-th image region and each adjacent tumor image, the smaller the respiratory linkage of the i-th image region under the acquisition view.

[0109] This embodiment, by simultaneously considering image frame difference and image matching degree, enables a more comprehensive and accurate assessment of changes in the target image region during respiration. Image frame difference reflects the morphological or grayscale changes of the target image region, while image matching degree measures its similarity. Combining the two allows for a more detailed depiction of the linkage between the target image region and respiratory movements.

[0110] As an optional embodiment, the evaluation value determination unit 124 is specifically used for:

[0111] The acquisition perspectives are sorted according to a preset traversal order to obtain the arrangement order of the acquisition perspectives;

[0112] The respiratory coordination difference is obtained by calculating the absolute value of the difference in respiratory coordination between target image regions under adjacent acquisition perspectives. The target image region is any image region.

[0113] By using the differences in respiratory linkage, the respiratory deformation evaluation value of the target image region is determined.

[0114] In this embodiment, the preset traversal order is a pre-defined rule or order for sorting the acquisition perspectives. It can be based on a specific logic, such as an order determined by angle size, acquisition time, etc.

[0115] The sorting order is the position number of each acquisition view after sorting the acquisition view according to the preset traversal order, which is used to clarify the sequential relationship between acquisition view.

[0116] The respiratory coordination difference is the absolute value of the difference in respiratory coordination between target image regions under adjacent acquisition viewpoints, which is obtained by calculating the respiratory coordination difference between adjacent acquisition viewpoints.

[0117] As an example, the evaluation value determination unit 124 first clarifies the specific rules for the preset traversal order, such as sorting by the angle between the acquisition viewpoint and a certain reference direction from smallest to largest, or sorting by the order of acquisition time, etc. Then, it sorts all acquisition views according to the rules and assigns a sorting order to each acquisition viewpoint, such as the first acquisition viewpoint being sorted by order 1, the second by order 2, and so on.

[0118] Then, for the sorted acquisition viewpoints, select two adjacent acquisition viewpoints in sequence. Obtain the respiratory coordination of the target image region under these two adjacent acquisition viewpoints. Calculate the difference between these two respiratory coordination values, i.e., the respiratory coordination difference.

[0119] Finally, based on the differences in respiratory coordination, the respiratory deformation evaluation value of the target image region is determined using the following formula:

[0120] ;

[0121] In the formula, The respiratory deformation evaluation value is used to characterize the i-th image region, and M is used to characterize the number of acquisition angles. Used to characterize the respiratory response of the i-th image region at the m-th acquisition viewpoint. The function exp is used to characterize the respiratory linkage of the i-th image region under the (m+1)-th acquisition viewpoint, and exp is used to characterize the exponential operation of the base e of the natural logarithm.

[0122] in, The smaller the value, the smaller the difference in respiratory linkage between adjacent acquisition perspectives in the i-th image region, and the larger the respiratory deformation evaluation value of the i-th image region.

[0123] This embodiment considers the changes in respiratory coordination in the target imaging region under different acquisition perspectives, enabling a more comprehensive assessment of the impact of respiratory motion on the target imaging region. Furthermore, different acquisition perspectives can provide information about tumor movement in different directions, thus more accurately reflecting the actual deformation of the target imaging region.

[0124] As an optional embodiment, such as Figure 3 As shown, the deformation determination module 130 specifically includes the following units:

[0125] The region filtering unit 131 is used to filter out each tumor region from each image region under the target acquisition view, wherein the target acquisition view is any acquisition view.

[0126] The deformation ratio determination unit 132 is used to determine the relative deformation ratio of each pixel in each tumor region based on the distance between each pixel in each tumor region and the centroid of the corresponding tumor region.

[0127] The displacement determination unit 133 is used to determine the relative displacement of each pixel in each tumor region based on the distance between each pixel in each tumor region and the centroid of the reference region.

[0128] The deformation determination unit 134 is used to determine the respiratory deformation of each pixel in each tumor region under the target acquisition view by using the relative deformation ratio and relative displacement of each pixel in each tumor region.

[0129] In this embodiment, the centroid is the geometric center of a region. For a tumor region, its centroid can be obtained by calculating the average coordinates of all pixels within the tumor region, representing the approximate center location of the tumor region.

[0130] The relative deformation ratio is used to measure the degree of deformation of each pixel within the tumor region relative to the centroid of the tumor region, reflecting the relative positional changes of each pixel within the tumor region.

[0131] Relative displacement is used to represent the degree of positional movement of each pixel within the tumor region relative to the centroid of the reference region, reflecting the spatial positional changes of each pixel during respiration.

[0132] As an example, the region filtering unit 131 can use image segmentation algorithms, such as threshold-based segmentation, region growing, edge detection, etc., to mark the parts of the image region belonging to the tumor in the target acquisition view based on the feature differences between the tumor and the surrounding tissue (such as gray value, texture, etc.) to obtain each tumor region.

[0133] Then, the deformation ratio determination unit 132 first calculates the coordinates of the centroid of each tumor region, and then calculates the distance between each pixel in the tumor region and the centroid of the tumor region. Based on the distribution of distances from all pixels to the centroid of the tumor region, the relative deformation ratio is determined. For example, the distance of each pixel can be compared with the average or maximum distance of all pixels in the tumor region to the centroid of the tumor region to obtain a relative ratio value as the corresponding relative deformation ratio.

[0134] Then, the displacement determination unit 133 first determines the coordinates of the centroid of the reference region, then calculates the distance from each pixel in each tumor region to the centroid of the reference region, and then determines the relative displacement based on the distribution of the distances from all pixels to the centroid of the reference region. For example, the distance of each pixel can be compared with a reference value (such as the distance from the pixel to the centroid of the reference region in the initial state) to obtain the relative displacement.

[0135] Finally, using the relative deformation ratio and relative displacement of each pixel in each tumor region, the breathing deformation of each pixel in each tumor region under the target acquisition view is determined by the following formula:

[0136] ;

[0137] In the formula, Used to characterize the respiratory deformation of the j-th pixel in the O-th tumor region of the initial tumor image under the target acquisition view. Used to characterize the relative deformation ratio of the j-th pixel in the O-th tumor region of the initial tumor image under the target acquisition view. Used to characterize the relative displacement of the j-th pixel in the O-th tumor region of the initial tumor image under the target acquisition view.

[0138] The greater the relative deformation ratio of a pixel, or the greater the relative displacement of a pixel, the greater the breathing deformation of the pixel.

[0139] This embodiment, by comprehensively considering the relative deformation ratio and relative displacement of pixels within the tumor region, can more accurately assess the impact of respiratory motion on the tumor region. The relative deformation ratio reflects the deformation within the tumor region, while the relative displacement reflects the overall movement of the tumor region. Combining the two can more comprehensively characterize tumor deformation during respiration.

[0140] As an optional embodiment, the deformation ratio determining unit 132 is specifically used for:

[0141] Based on the first distance between the j-th pixel and the centroid of the corresponding tumor region in the tumor image sequence under the target respiratory cycle acquisition perspective, the first average distance between the j-th pixel and the centroid of the tumor region is determined, where j is a positive integer;

[0142] Using the first distance and the corresponding first average distance of the j-th pixel in the second tumor region, the relative deformation ratio of the j-th pixel in the second tumor region under the target acquisition view is determined. The second tumor region is the first tumor region in the initial tumor image.

[0143] In this embodiment, the first distance is used to characterize the spatial distance between the j-th pixel in the second tumor region and the corresponding centroid of the tumor region under the target acquisition view. The second tumor region is the first tumor region in the initial tumor image.

[0144] The first average distance is used to characterize the average spatial distance between the j-th pixel in the first tumor region of all tumor images in the tumor image sequence under the target acquisition view within the target respiratory cycle and the corresponding centroid of the tumor region.

[0145] As an example, the relative deformation ratio is determined using the following formula:

[0146] ;

[0147] In the formula, Used to characterize the relative deformation ratio of the j-th pixel in the O-th tumor region of the initial tumor image under the target acquisition view. Used to characterize the first distance of the j-th pixel in the 0-th tumor region of the initial tumor image under the target acquisition view. The first average distance is used to characterize the j-th pixel in the O-th tumor region of the tumor image sequence under the target acquisition view within the target respiratory cycle. Evaluation value used to characterize the respiratory deformation of the Oth tumor region The respiratory deformation evaluation value is used to characterize the baseline region, and exp is used to characterize the exponential operation of the base e of the natural logarithm.

[0148] in, Used to characterize the Oth tumor region in the initial tumor image. The ratio of the deformation of the nth pixel within the target respiratory cycle to the respiratory deformation evaluation value of the Oth tumor region represents the value of the nth tumor region. The deformation ratio of the _th pixel within the tumor region. A larger value indicates that the _th pixel in the _O_th tumor region of the initial tumor image represents the _th_ pixel. The greater the relative deformation ratio of each pixel, the better.

[0149] This embodiment quantifies the degree of deformation of each pixel within the tumor region relative to the centroid by calculating the relative deformation ratio, thus providing more intuitive information about tumor deformation. This helps in subsequent correction of the initial tumor image and improves the accuracy of the 3D tumor model.

[0150] As an optional embodiment, the displacement determination unit 133 is specifically used for:

[0151] Based on the second distance between the j-th pixel in each first tumor region and the centroid of the reference region in the tumor image sequence under the target acquisition view within the target respiratory cycle, the second average distance between the j-th pixel and the centroid of the reference region is determined, where j is a positive integer;

[0152] Using the second distance of the j-th pixel in the second tumor region and the corresponding second average distance, the relative displacement of the j-th pixel in the second tumor region under the target acquisition view is determined. The second tumor region is the first tumor region in the initial tumor image.

[0153] In this embodiment, the second distance is used to characterize the spatial distance between the j-th pixel in the second tumor region and the centroid of the corresponding reference region under the target acquisition view. The second tumor region is the first tumor region in the initial tumor image.

[0154] The second average distance is used to characterize the average spatial distance between the j-th pixel in the first tumor region of all tumor images in the tumor image sequence under the target acquisition view within the target respiratory cycle and the centroid of the corresponding reference region.

[0155] As an example, the relative displacement is determined using the following formula:

[0156] ;

[0157] In the formula, Used to characterize the relative displacement of the j-th pixel in the O-th tumor region of the initial tumor image under the target acquisition view. The second distance is used to characterize the j-th pixel in the O-th tumor region of the initial tumor image from the target acquisition viewpoint. The second average distance is used to characterize the j-th pixel in the O-th tumor region of the initial tumor image under the target acquisition view. Used to characterize normalization operations.

[0158] in, Used to characterize the Oth tumor region in the initial tumor image. The positional change of the nth pixel relative to the reference region. A larger value indicates a higher positional change within the 0th tumor region of the initial tumor image. The more significant the change in position of a pixel relative to the reference region, the greater the relative displacement of the j-th pixel in the O-th tumor region in the initial tumor image.

[0159] This embodiment quantifies the deformation of each pixel within the tumor region relative to a reference region by calculating relative displacement, thus providing more intuitive information about tumor deformation. This helps in subsequent correction of the initial tumor image and improves the accuracy of the 3D tumor model.

[0160] As an optional embodiment, the image correction module 140 is specifically used for:

[0161] Based on the respiratory deformation degree of each pixel in the tumor region in the initial tumor image under each acquisition view, the deformation correction amount of each pixel in the tumor region in the initial tumor image under each acquisition view is determined respectively.

[0162] Based on the deformation correction amount of each pixel in the tumor region in the initial tumor image under each acquisition view, the position of each pixel in the tumor region in the initial tumor image under each acquisition view is corrected to obtain the corrected tumor image under each acquisition view.

[0163] In this embodiment, the deformation correction amount is a value calculated based on respiratory deformation and used to correct the position or shape of each pixel in the tumor region of the initial tumor image. It represents the degree to which each pixel needs to be adjusted to eliminate the deformation effect caused by respiratory motion.

[0164] The corrected tumor image is obtained by correcting the pixel positions of the tumor region in the initial tumor images from each acquisition viewpoint. This tumor image eliminates the influence of respiratory deformation and more accurately reflects the true morphology of the tumor.

[0165] As an example, the image correction module 140 first determines the deformation correction amount using the following formula:

[0166] ;

[0167] In the formula, This is used to characterize the deformation correction amount of the j-th pixel in the O-th tumor region of the initial tumor image under the m-th acquisition view. Used to characterize the respiratory deformation of the j-th pixel in the O-th tumor region of the initial tumor image under the m-th acquisition view. The average value of the respiratory deformation of the j-th pixel in the O-th tumor region in the initial tumor image under each acquisition viewpoint.

[0168] Then, the image correction module 140, for each initial tumor image acquired from a specific viewpoint, iterates through each pixel in the tumor region and adjusts the pixel position based on the calculated deformation correction amount. For example, if the original coordinates of a pixel are... Its deformation correction amount is The corrected coordinates are .

[0169] Finally, all the corrected pixels are recombined to form the corrected tumor images from each acquisition viewpoint. During the correction process, it is important to keep parameters such as image resolution and grayscale range constant to ensure the quality of the corrected image.

[0170] This embodiment eliminates the deformation caused by respiratory motion, allowing the corrected tumor images to more accurately reflect the true shape and location of the tumor. Therefore, creating a 3D tumor model based on the corrected tumor images improves the accuracy of the generated 3D tumor model.

[0171] As an optional embodiment, the 3D model building module 150 is specifically used for:

[0172] The modified tumor images were sorted according to the order of image acquisition to determine the spatial location of each modified tumor image.

[0173] Each corrected tumor image is subjected to consistency processing to obtain the corresponding consistent tumor image.

[0174] The images of the tumors are stacked according to their corresponding spatial locations to obtain the image stacking result;

[0175] A three-dimensional model of the tumor was constructed based on the image stacking results.

[0176] In this embodiment, spatial position refers to a relative positional information in three-dimensional space assigned to each corrected tumor image according to the image acquisition order. It reflects the spatial distribution of the tumor from different perspectives during image acquisition and serves as the basis for subsequent image stacking and three-dimensional model construction.

[0177] Consistency processing is a series of operations performed on corrected tumor images to eliminate differences in brightness, contrast, resolution, etc., between images, so that the images are consistent in visual and data features, in order to better stack images and build 3D models.

[0178] Consistent tumor images are corrected tumor images that have undergone consistency processing, possessing similar image features and enabling them to be fused together more accurately.

[0179] Image stacking results are obtained by superimposing consistent tumor images according to their corresponding spatial locations. It contains comprehensive information about the tumor from different perspectives and is key data for constructing a three-dimensional model.

[0180] As an example, the 3D model building module 150 first sorts the corrected tumor images according to the image acquisition sequence. Simultaneously, it determines the spatial location of each image based on the geometric relationships during acquisition (such as the angle and position of the imaging equipment) and clinical needs. For instance, if the acquisition is a circular scan around the patient's body, the position of each image in 3D space can be determined based on the scanning angle and distance.

[0181] Then, the average brightness and contrast of each corrected tumor image are calculated. Histogram equalization, grayscale stretching, and other methods are used to adjust the brightness and contrast of all corrected tumor images to a similar range. For example, a global histogram equalization algorithm can be used to enhance the overall contrast of the images, reducing the differences in brightness and contrast between different images. Simultaneously, the resolution of each corrected tumor image is checked. If differences exist, interpolation algorithms (such as bilinear interpolation, cubic spline interpolation, etc.) are used to unify the resolution of the corrected tumor images to the same level. For example, the resolution of all corrected tumor images is adjusted to 512×512 pixels. Finally, filtering algorithms (such as Gaussian filtering, median filtering, etc.) are used to remove noise from the corrected tumor images, improving image quality. For example, Gaussian filtering is used to smooth the corrected tumor images, reducing the impact of noise on subsequent analysis.

[0182] Then, based on the determined spatial locations, each consistent tumor image is placed in its corresponding position in three-dimensional space. This process can be achieved using image processing software or a dedicated medical image processing platform. During the stacking process, overlapping areas between images need to be considered, and appropriate fusion algorithms (such as weighted average fusion, maximum value fusion, etc.) should be used to process the overlapping areas to ensure a natural image transition. For example, for pixel values ​​in overlapping areas, a weighted average method can be used to calculate the final pixel value based on the pixel's position and weight in different images.

[0183] Finally, the tumor's contour information is extracted from the image stacking results, and a 3D surface model of the tumor is constructed using a surface reconstruction algorithm (such as the MarchingCubes algorithm). This algorithm generates the tumor's surface model by extracting isosurfaces from the 3D data field, which can visually display the tumor's shape. Alternatively, the image stacking results can be viewed as a 3D voxel data field, with each voxel corresponding to a pixel in the image. Based on the voxel's grayscale values ​​or other feature information, a voxel model of the tumor is constructed, which can include information about the tumor's internal structure.

[0184] This embodiment, through image stacking and 3D model construction, can integrate tumor image information from different acquisition perspectives, providing complete morphological and structural information of the tumor in three-dimensional space. This helps doctors to have a more comprehensive understanding of the characteristics of the tumor, including its size, shape, location, and relationship with surrounding tissues.

[0185] As an optional embodiment, the tumor image acquisition module 110 is specifically used for:

[0186] Send image acquisition signals to the tumor scanner so that the tumor scanner can scan the tumor of the target patient;

[0187] Receive raw signal data from the target patient collected by the tumor scanner;

[0188] The original signal data is converted into a two-dimensional image format by a target reconstruction algorithm, resulting in initial tumor images from multiple acquisition perspectives.

[0189] In this embodiment, the image acquisition signal is a command signal used to trigger the tumor scanner to start scanning. It can be an electrical signal, digital signal, etc., in a specific format, containing parameter information required for scanning, such as scanning range, scanning slice thickness, scanning mode, etc., to guide the tumor scanner to scan in a preset manner.

[0190] A tumor scanner is a medical device specifically designed for the detection and imaging of tumors. Common types include computed tomography (CT) scanners, magnetic resonance imaging (MRI) scanners, and positron emission tomography-computed tomography (PET) scanners. They utilize different imaging principles to acquire information about tumors inside the human body.

[0191] Raw signal data refers to the unprocessed data collected by the tumor scanner during the scanning of the target patient. This data is a digital representation of various physical signals received by the scanner detectors (such as X-ray attenuation signals, magnetic resonance signals, gamma-ray signals generated by positron annihilation, etc.), containing structural and functional information about the tumor and other tissues in the patient's body.

[0192] Target reconstruction algorithms are algorithms specifically designed to convert raw signal data into a two-dimensional image format. Different imaging devices have different imaging principles and data characteristics, therefore, appropriate reconstruction algorithms are needed to process the raw signal data in order to generate clear and accurate two-dimensional images.

[0193] Two-dimensional image formats are image data formats that are obtained by processing raw signal data and displayed in a two-dimensional plane. Common two-dimensional image formats include medical digital imaging and communication formats, which facilitate doctors to view, analyze, and diagnose images on a computer.

[0194] As an example, the tumor image acquisition module 110 determines the necessary scanning parameters based on the target patient's specific situation (such as disease condition and examination site) and clinical needs. These parameters include the scanning range (determining the body parts and areas to be scanned), slice thickness (the thickness of each slice), and scanning mode (such as plain scan, contrast-enhanced scan, etc.). The determined scanning parameters are then encoded according to the signal format required by the tumor scanner to generate an image acquisition signal. For example, for a CT scanner, the parameter information needs to be encapsulated into command signals using a specific communication protocol. The generated image acquisition signal is then sent to the tumor scanner via wired or wireless means. Common transmission methods include Ethernet connections and fiber optic transmission, ensuring that the signal reaches the scanner accurately and promptly.

[0195] Then, while sending image acquisition signals, a data receiving channel is established with the tumor scanner. Specifically, this typically involves setting parameters such as data transmission protocols and port numbers to ensure correct reception of data sent by the scanner. When the tumor scanner completes the scan and begins sending raw signal data, the tumor image acquisition module 110 receives the data according to the set protocol and caches it in memory or on the hard drive. During the reception process, data integrity checks are performed to ensure that no data is lost or corrupted.

[0196] Finally, a suitable target reconstruction algorithm is selected based on the type of tumor scanner and the characteristics of the raw signal data. For example, for raw CT scan data, commonly used reconstruction algorithms include filtered backprojection and iterative reconstruction algorithms. The selected target reconstruction algorithm is then implemented using a programming language, and the received raw signal data is input into the algorithm for processing. The algorithm performs a series of mathematical operations and transformations on the raw data, such as filtering, backprojection, and Fourier transform, to extract image information. After processing, the obtained image data is converted into a two-dimensional image format. Simultaneously, based on the acquisition perspective information during the scan, each two-dimensional image is assigned a corresponding perspective label, resulting in initial tumor images from multiple acquisition perspectives.

[0197] This embodiment acquires initial tumor images from multiple acquisition perspectives, which can more comprehensively and accurately display the morphology, structure, and location information of the tumor, helping doctors to more accurately diagnose the type, stage, and severity of the tumor.

[0198] 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 them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

[0199] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A system for automatic segmentation and three-dimensional reconstruction of tumor images, characterized in that, The system includes: A tumor image acquisition module is used to acquire initial tumor images from multiple acquisition perspectives, wherein the initial tumor images include multiple image regions. The region selection module is used to select a reference region from the multiple image regions based on the biological tissue deformation information of the image regions under each acquisition viewpoint during the target respiratory cycle. The reference region is the image region with the smallest degree of biological tissue deformation. The deformation determination module is used to compare the tumor region in the image region under each acquisition view with the reference region to obtain the breathing deformation of each pixel in the tumor region under each acquisition view. The image correction module is used to correct the position of each pixel in the tumor region in the initial tumor image under each acquisition view based on the respiratory deformation of each pixel in the tumor region under each acquisition view, so as to obtain the corrected tumor image under each acquisition view. A 3D model construction module is used to construct a 3D model of the tumor based on the modified tumor images described above. The region selection module includes the following units: An image acquisition unit is used to acquire historical tumor images from each acquisition perspective within the target respiratory cycle; The image sorting unit is used to sort the historical tumor images and the corresponding initial tumor images under each acquisition view in chronological order to construct a tumor image sequence under each acquisition view. The linkage determination unit is used to perform the following for each acquisition viewpoint: based on the tumor image sequence under the acquisition viewpoint, determine the respiratory linkage of each image region under the acquisition viewpoint. The evaluation value determination unit is used to determine the respiratory deformation evaluation value of each of the image regions based on the respiratory linkage of each of the image regions under each acquisition view. The region determination unit is used to determine the image region with the smallest respiratory deformation evaluation value as the reference region.

2. The automatic segmentation and three-dimensional reconstruction system for tumor images according to claim 1, characterized in that, The linkage determination unit is specifically used for: In the tumor image sequence under the acquisition perspective, the image frame difference and image matching degree of the target image region between adjacent tumor images are obtained, wherein the target image region is any one of the image regions; By utilizing the image frame difference and image matching degree between the target image region and each of the adjacent tumor images, the respiratory linkage of the target image region under the acquisition viewpoint is determined.

3. The automatic segmentation and three-dimensional reconstruction system for tumor images according to claim 1, characterized in that, The evaluation value determination unit is specifically used for: The acquisition perspectives are sorted according to a preset traversal order to obtain the arrangement order of the acquisition perspectives. The respiratory coordination difference is obtained by calculating the absolute value of the difference between the respiratory coordination of the target image regions under the acquisition viewpoints of adjacent sequences, where the target image region is any one of the image regions. The respiratory deformation evaluation value of the target image region is determined by using the respiratory linkage difference values.

4. The automatic segmentation and three-dimensional reconstruction system for tumor images according to claim 1, characterized in that, The deformation determination module includes the following units: A region filtering unit is used to filter each tumor region under the target acquisition view from each of the image regions under the target acquisition view, wherein the target acquisition view is any one of the acquisition views; The deformation ratio determination unit is used to determine the relative deformation ratio of each pixel in each tumor region based on the distance between each pixel in each tumor region and the centroid of the corresponding tumor region. The displacement determination unit is used to determine the relative displacement of each pixel in each tumor region based on the distance between each pixel in each tumor region and the centroid of the reference region. The deformation degree determination unit is used to determine the respiratory deformation degree of each pixel in each tumor region under the target acquisition view by using the relative deformation ratio and the relative displacement of each pixel in each tumor region.

5. The automatic segmentation and three-dimensional reconstruction system for tumor images according to claim 4, characterized in that, The deformation ratio determination unit is specifically used for: Based on the first distance between the j-th pixel and the centroid of the corresponding tumor region in the tumor image sequence under the target acquisition view within the target respiratory cycle, the first average distance between the j-th pixel and the centroid of the tumor region is determined, where j is a positive integer; Using the first distance of the j-th pixel in the second tumor region and the corresponding first average distance, the relative deformation ratio of the j-th pixel in the second tumor region under the target acquisition view is determined, and the second tumor region is the first tumor region in the initial tumor image.

6. The automatic segmentation and three-dimensional reconstruction system for tumor images according to claim 4, characterized in that, The displacement determination unit is specifically used for: Based on the second distance between the j-th pixel in each first tumor region and the centroid of the reference region in the tumor image sequence under the target acquisition view within the target respiratory cycle, the second average distance between the j-th pixel and the centroid of the reference region is determined, where j is a positive integer; Using the second distance of the j-th pixel in the second tumor region and the corresponding second average distance, the relative displacement of the j-th pixel in the second tumor region under the target acquisition view is determined, and the second tumor region is the first tumor region in the initial tumor image.

7. A three-dimensional reconstruction system for automatic segmentation of tumor images according to any one of claims 1-6, characterized in that, The image correction module is specifically used for: Based on the respiratory deformation degree of each pixel in the tumor region in the initial tumor image under each acquisition perspective, the deformation correction amount of each pixel in the tumor region in the initial tumor image under each acquisition perspective is determined respectively; Based on the deformation correction amount of each pixel in the tumor region in the initial tumor image under each acquisition view, the position of each pixel in the tumor region in the initial tumor image under each acquisition view is corrected to obtain the corrected tumor image under each acquisition view.

8. A three-dimensional reconstruction system for automatic segmentation of tumor images according to any one of claims 1-6, characterized in that, The 3D model construction module is specifically used for: The modified tumor images are sorted according to the image acquisition order to determine the spatial location of each modified tumor image; Each of the modified tumor images is subjected to consistency processing to obtain the corresponding consistent tumor images; The consistent tumor images are stacked according to their corresponding spatial locations to obtain an image stacking result; Based on the image stacking results, the three-dimensional model of the tumor is constructed.

9. A three-dimensional reconstruction system for automatic segmentation of tumor images according to any one of claims 1-6, characterized in that, The tumor image acquisition module is specifically used for: An image acquisition signal is sent to a tumor scanner so that the tumor scanner can perform a tumor scan on the target patient. Receive the raw signal data of the target patient collected by the tumor scanner; The original signal data is converted into a two-dimensional image format using a target reconstruction algorithm to obtain the initial tumor images from multiple acquisition perspectives.

Images