SPECT bone imaging bladder interference elimination method and system

Abstract

The invention belongs to the field of medical image processing, and discloses an SPECT bone imaging bladder interference elimination method and system, and the method comprises the steps: obtaining an SPECT bone imaging image, segmenting a pelvic skeleton region in the SPECT bone imaging image through a pre-trained nnUNet network model, and generating a pelvic skeleton mask; judging whether bladder interference exists or not according to the radioactivity counting information in the external rectangular area of the pelvic skeleton mask; if bladder interference exists, generating a bladder mask in an external rectangular area of the pelvic skeleton mask according to the radioactivity counting information in combination with a foreground segmentation method and an edge optimization method; eliminating or replacing radioactivity counts in the bladder mask area to generate an SPECT bone imaging image after bladder interference is eliminated; outputting correction result information containing the SPECT bone imaging image after the bladder interference is eliminated; therefore, automatic and precise identification and elimination of bladder interference are realized.

Description

Technical Field

[0001] This application relates to the field of medical image processing, and more specifically, to a method and system for eliminating bladder interference in SPECT bone scintigraphy. Background Technology

[0002] In anteroposterior plane images of SPECT bone scintigraphy, the bladder in the central pelvic region often becomes one of the main interfering factors when interpreting the images due to the accumulation of highly radioactive urine. This strong radioactive signal, like a "bright light," can completely obscure small or moderate abnormal radioactive concentrations in adjacent bones, such as the sacrum, sacroiliac joint, pubis, ischium, and femoral head and neck—high-incidence sites for bone metastases from various malignant tumors, including prostate and breast cancer. This significantly increases the risk of missing early metastatic lesions and seriously affects diagnostic accuracy.

[0003] Furthermore, the intense radiation from the bladder can produce a "sparkling" effect on image reconstruction and display, potentially creating artifacts of reduced radiation distribution in the surrounding bone areas. These artifacts may be misinterpreted as osteolytic destruction, ischemic necrosis, or other lesions, leading to unnecessary false positives and potentially causing patients to undergo additional examinations or experience unnecessary anxiety.

[0004] Currently, the most common clinical method for eliminating bladder interference involves doctors or nuclear medicine technicians manually identifying and removing the interfering area of ​​the bladder. During manual removal, a regular elliptical or circular area is typically selected. However, this manual method has several problems, including strong subjectivity, inaccurate determination of bladder boundaries, a tendency to over- or under-remove areas, and the inability to repeat the operation. This approach makes it difficult to accurately determine bladder boundaries, and excessive removal can lead to cold areas in surrounding tissues, which may be misdiagnosed as bone destruction, thus affecting the objectivity and accuracy of the diagnosis.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] The purpose of this application is to provide a method and system for eliminating bladder interference in SPECT bone scintigraphy, which aims to solve the interference caused by the accumulation of highly radioactive urine in the bladder on image interpretation in SPECT bone scintigraphy images, as well as the problems of strong subjectivity, inaccurate boundaries, and non-repeatable operation in existing manual subtraction methods.

[0007] In a first aspect, this application provides a method for eliminating bladder interference in SPECT bone scintigraphy images, used to eliminate bladder interference in SPECT bone scintigraphy images, wherein the SPECT bone scintigraphy images include anterior plane images and / or posterior plane images, and the method includes the following steps: A1. Obtain SPECT bone scintigraphy images; A2. Using a pre-trained nnUNet network model, segment the pelvic bone region in the SPECT bone imaging image to generate a pelvic bone mask; A3. Based on the radioactivity count information within the outer rectangular region of the pelvic bone mask, determine whether bladder interference exists; A4. If bladder interference exists, a bladder mask is generated in the outer rectangular region of the pelvic bone mask based on the radioactivity count information, combined with the foreground segmentation method and the edge optimization method. A5. Eliminate or replace the radioactive count in the bladder mask area to generate a SPECT bone scintigraphy image after eliminating bladder interference; A6. Output information containing the correction results of the SPECT bone scintigraphy image after eliminating bladder interference.

[0008] Secondly, this application provides a SPECT bone scintigraphy bladder interference cancellation system for canceling bladder interference in SPECT bone scintigraphy images, wherein the SPECT bone scintigraphy images include anterior and / or posterior plane images, and the system includes: Image acquisition module, used to acquire SPECT bone scintigraphy images; The segmentation module is used to segment the pelvic bone region in the SPECT bone imaging image using a pre-trained nnUNet network model and generate a pelvic bone mask. The interference detection module is used to determine whether there is bladder interference based on the radioactivity count information in the outer rectangular area of ​​the pelvic bone mask; A bladder mask generation module is used to generate a bladder mask in the outer rectangular region of the pelvic bone mask when bladder interference is present, based on the radioactivity counting information and in combination with a foreground segmentation method and an edge optimization method. The interference cancellation module is used to eliminate or replace the radioactive count in the bladder mask area and generate a SPECT bone scintigraphy image after eliminating bladder interference. The output module is used to output correction result information containing the SPECT bone scintigraphy image after eliminating bladder interference.

[0009] Beneficial Effects: This application provides a method and system for eliminating bladder interference in SPECT bone scintigraphy. By introducing a pre-trained nnUNet network model, the pelvic bone region is accurately segmented, laying the foundation for subsequent bladder interference identification and elimination. This method can intelligently determine the presence of bladder interference based on the radioactivity count information within the rectangular region circumscribed by the pelvic bone mask, avoiding the subjectivity of traditional manual identification. After confirming the presence of bladder interference, a precise bladder mask is generated by combining foreground segmentation and edge optimization methods, effectively solving the problems of inaccurate bladder boundaries and the tendency for over- or under-delineation during manual drawing. Finally, by eliminating or replacing the radioactivity count within the bladder mask region, a SPECT bone scintigraphy image with bladder interference eliminated is generated, significantly improving image quality. This allows for clear display of small or moderate abnormal radioactive concentrations in adjacent bones, effectively reducing the risk of missed diagnoses of early metastatic lesions and improving diagnostic accuracy. In addition, this method can also output correction result information containing correction result information, providing clinicians with objective and repeatable diagnostic basis, overcoming the shortcomings of manual operation in the existing technology that is not repeatable, and is of great significance to improving the clinical application value of SPECT bone imaging. Attached Figure Description

[0010] Figure 1 A flowchart of a SPECT bone scintigraphy method for eliminating bladder interference provided in this application.

[0011] Figure 2 This is a schematic diagram of a SPECT bone scintigraphy bladder interference elimination system provided in this application.

[0012] Labeling Explanation: 1. Image Acquisition Module; 2. Segmentation Module; 3. Interference Detection Module; 4. Bladder Mask Generation Module; 5. Interference Cancellation Module; 6. Output Module. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0014] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0015] Please refer to Figure 1 This application discloses a method for eliminating bladder interference in SPECT bone scintigraphy images, which includes anterior and / or posterior plane images. The method includes the following steps: A1. Obtain SPECT bone scintigraphy images; A2. Using a pre-trained nnUNet network model, segment the pelvic bone region in the SPECT bone imaging image and generate a pelvic bone mask; A3. Determine whether there is bladder interference based on the radioactivity count information within the circumscribed rectangular area of ​​the pelvic bone mask; A4. If bladder interference exists, a bladder mask is generated in the outer rectangular region of the pelvic bone mask based on the radioactivity count information, combined with the foreground segmentation method and the edge optimization method. A5. Eliminate or replace the radioactive count in the bladder mask area to generate a SPECT bone scintigraphy image after eliminating bladder interference; A6. Output includes correction results information for the SPECT bone scintigraphy image after eliminating bladder interference.

[0016] This application utilizes a pre-trained nnUNet network model for precise segmentation of the pelvic skeletal region. By combining radioactivity counting information, foreground segmentation methods, and edge optimization techniques, it achieves automated and accurate identification and elimination of bladder interference. This effectively overcomes the limitations of traditional manual methods and significantly improves the diagnostic accuracy and objectivity of SPECT bone scintigraphy images.

[0017] Specifically, the SPECT bone imaging images mentioned in this application refer to images of bone radioactivity distribution obtained through single-photon emission computed tomography (SPECT) technology. These images may include anterior and / or posterior plane images, primarily used to assess bone metabolic activity and lesions.

[0018] The nnUNet network model is a highly adaptive deep learning segmentation framework that demonstrates outstanding performance in medical image segmentation. This model can automatically adjust its network structure and hyperparameters based on training data, thereby achieving accurate segmentation of the pelvic skeletal region. The pelvic skeletal mask is a binary image of the pelvic skeletal region segmented by the nnUNet network model, primarily used to limit the scope of subsequent bladder interference detection and bladder mask generation.

[0019] Furthermore, radioactivity counting information refers to the radioactivity intensity values ​​of each pixel or voxel in a SPECT bone scintigraphy image, directly reflecting the degree of radioactive tracer accumulation in the tissue. Foreground segmentation methods aim to separate target regions (e.g., the bladder) from the background in an image, while edge optimization methods are used to refine the boundaries of the segmentation results, making them closer to the actual anatomical structure.

[0020] In the SPECT bone scintigraphy method for eliminating bladder interference, the first step is to acquire SPECT bone scintigraphy images. These images can be anterior-planar images, posterior-planar images, or a combination of both from the SPECT scanning device. Images can be acquired directly from the SPECT device interface or retrieved from a Picture Archiving and Communication System (PACS). For example, an automatic acquisition mechanism can be set up so that after new SPECT bone scintigraphy images are generated, the system can automatically identify and load these images for subsequent processing.

[0021] Next, a pre-trained nnUNet network model is used to segment the pelvic bone region in the acquired SPECT bone scan image to generate a pelvic bone mask. The nnUNet network model has learned a large number of features from bone images during training, thus accurately identifying and delineating the boundaries of the pelvic bones. In practical applications, the SPECT bone scan image can be used as input, and the nnUNet network model can be used for inference to obtain an accurate pelvic bone mask. This mask is a binary image where the pelvic bone region is labeled as the foreground and other regions are labeled as the background.

[0022] Subsequently, based on the radioactivity count information within the rectangular region circumscribed by the pelvic bone mask, it is determined whether bladder interference exists. This step is crucial for identifying bladder interference. For example, the total or average radioactivity count within the rectangular region circumscribed by the pelvic bone mask can be calculated and compared with a preset threshold. If the radioactivity count in this region is significantly higher than the threshold, bladder interference may be present. As a preferred implementation, the distribution characteristics of the radioactivity count within this region can be analyzed, such as whether there are localized high-radioactivity clusters, to aid in the determination.

[0023] If bladder interference is detected, a bladder mask is generated in the outer rectangular region of the pelvic bone mask based on the radioactivity count information, combined with the foreground segmentation method and the edge optimization method.

[0024] Specifically, foreground segmentation methods can employ various techniques. For example, threshold-based segmentation methods (such as the OTSU algorithm) can be used, setting a radioactivity count threshold to initially identify regions above this threshold as foreground regions. Alternatively, region growing segmentation methods can be employed, starting from one or more seed points and progressively expanding the region based on similarity criteria (such as radioactivity count similarity) until a stopping condition is met. Furthermore, clustering segmentation methods (such as K-means) or edge detection-based segmentation methods (such as the Canny operator combined with an active contour model) can also be used for foreground region identification.

[0025] After initially segmenting the bladder region, edge optimization methods are needed to refine the bladder mask to improve its accuracy. Edge optimization methods can employ morphological operations (such as dilation, erosion, opening, and closing) to smooth boundaries, remove small noise points, or connect broken areas. For example, a Gaussian smoothing filter can be used to preprocess the image to reduce the impact of noise on edge detection. Then, the Canny edge detection algorithm is applied to identify the precise boundaries of the bladder, and the Snake model is used to iteratively optimize these boundaries to better fit the true contour of the bladder.

[0026] After generating the bladder mask, the radioactivity counts within the bladder mask region are eliminated or replaced to generate a SPECT bone scintigraphy image free of bladder interference. Eliminating radioactivity counts can be achieved by directly setting all pixel or voxel values ​​within the bladder mask region to zero, thus completely removing the bladder's radioactive signal. As a preferred implementation, replacing radioactivity counts is more complex. For example, based on radioactivity count information from adjacent bone regions around the bladder mask, a replacement value consistent with the radioactivity distribution trend of the surrounding bone regions can be generated through interpolation, fitting, or gradient-based smoothing algorithms. Subsequently, the radioactivity counts in the bladder region are replaced with these generated values ​​to achieve a more natural transition.

[0027] Finally, the output includes correction results for the SPECT bone scintigraphy image after eliminating bladder interference. This correction information may include the processed SPECT bone scintigraphy image itself, or other auxiliary information such as the area of ​​the bladder mask, average radioactivity count, and a comparison report of the images before and after processing. This information can be output as digital image files, text reports, or structured data for physicians' diagnostic reference.

[0028] The SPECT bone scintigraphy bladder interference elimination method of this application effectively solves the problems of low accuracy, strong subjectivity and non-repeatable operation of traditional manual methods in eliminating bladder interference in SPECT bone scintigraphy images through a systematic procedure.

[0029] Specifically, this method first acquires SPECT bone scintigraphy images automatically, ensuring the standardization of data input. Then, a pre-trained nnUNet network model is used to accurately segment the pelvic bone region, generating a pelvic bone mask. This step provides accurate anatomical localization for subsequent bladder interference assessment and mask generation, thus avoiding the subjectivity and inaccuracy of manual drawing.

[0030] When determining the presence of bladder interference, this application utilizes radioactivity counting information within the rectangular region circumscribed by the pelvic bone mask for objective evaluation, rather than relying on subjective human judgment, thereby improving the accuracy and consistency of the assessment. When bladder interference is detected, this method combines foreground segmentation and edge optimization methods to automatically generate a bladder mask within the defined pelvic bone region. This automated and refined mask generation process is significantly superior to the coarse and irregular traditional manual drawing, enabling more accurate identification of the bladder boundary and avoiding over- or under-masking.

[0031] In the process of eliminating bladder interference, this application provides two strategies: eliminating or replacing the radioactive count within the bladder mask area. This can either completely remove the interfering signal or achieve a more natural image transition through intelligent replacement, thereby generating high-quality SPECT bone scintigraphy images with bladder interference eliminated. Ultimately, the output image contains correction result information, providing clearer and more objective imaging evidence for clinical diagnosis.

[0032] Compared with existing technologies, the core innovation of this application lies in its automated, intelligent, and high-precision bladder interference elimination process. Traditional methods mainly rely on human experience and manual operation, which is inefficient and the results are greatly affected by subjective factors. This application introduces a deep learning model (nnUNet) for bone segmentation and combines various image processing techniques (such as foreground segmentation and edge optimization) to achieve full automation from interference identification to elimination, greatly improving processing efficiency and the objectivity of results. For example, in the generation of the bladder mask, this application uses foreground segmentation and edge optimization methods, which can generate a more accurate bladder boundary than manual delineation. This avoids the problem of over-subtraction leading to cold area artifacts in surrounding tissues, which is common in traditional methods, and effectively reduces the risk of misdiagnosis. In addition, this application also provides a flexible interference elimination strategy (elimination or replacement), which can select the most appropriate processing method according to actual needs, further improving image quality and diagnostic value. These technological advancements together make the method of this application have significant advantages in SPECT bone imaging image analysis, and can provide clinicians with more reliable diagnostic support.

[0033] In some implementations, step A3 includes: A301. Based on the radioactivity count information within the circumscribed rectangular region of the pelvic bone mask, obtain the maximum radioactivity count within the circumscribed rectangular region; A302. If the maximum radioactivity count within the circumscribed rectangular region is the maximum radioactivity count in the entire SPECT bone scintigraphy image, then bladder interference is determined to exist.

[0034] Specifically, in step A301, obtaining the maximum radioactivity count within the circumscribed rectangular region refers to scanning the radioactivity counts of all pixels within the circumscribed rectangular region of the pelvic bone mask and identifying the pixel with the highest value. This maximum radioactivity count can serve as an indicator of the strongest radioactivity within that region.

[0035] In step A302, the maximum radioactivity count within the circumscribed rectangular region is compared with the maximum radioactivity count across the entire SPECT bone scintigraphy image. If they are equal, it indicates that a point within the pelvic region (typically the location of the bladder) is the area with the highest radioactivity in the entire image. In this case, the bladder typically exhibits extremely high radioactivity uptake, strongly suggesting the presence of bladder interference. The aim is to provide a rapid, intuitive, and highly confident preliminary method for assessing bladder interference.

[0036] This application's solution effectively addresses the efficiency issues that traditional methods may face when identifying significant bladder interference by comparing the maximum radioactivity count within the circumscribed rectangular region of the pelvic bone mask with the maximum radioactivity count across the entire SPECT bone scintigraphy image. When significant radioactivity accumulation exists in the bladder region, its radioactivity count is often much higher than that of the surrounding bone tissue, sometimes even becoming the highest point in the entire image. This direct comparison allows for the rapid and accurate identification of such typical, high-intensity bladder interference, avoiding more complex analysis processes and thus improving the efficiency and accuracy of the assessment.

[0037] The above technical solution provides a simple, rapid, and reliable preliminary assessment mechanism for bladder interference. This method is particularly suitable for identifying bladder conditions with high levels of radioactive accumulation and significant interference, effectively avoiding the risk of misdiagnosis or missed diagnosis of bone lesions due to bladder interference. Furthermore, this assessment method has high specificity, reducing false positives and thus improving the overall accuracy and efficiency of SPECT bone scintigraphy image analysis.

[0038] Furthermore, step A3 also includes: A303. If the maximum radioactivity count within the outer rectangular region of the pelvic bone mask is not the maximum radioactivity count in the entire SPECT bone scan image, then execute: Based on the radioactivity count information within the outer rectangular region of the pelvic bone mask, connected regions with radioactivity counts exceeding a preset count threshold are identified within the outer rectangular region and designated as suspicious connected regions. Morphological feature analysis was performed on the suspicious connected regions to obtain the morphological feature analysis results; Spatial distribution analysis was performed on the suspected connected regions to obtain the spatial distribution analysis results; Based on the morphological feature analysis results and spatial distribution analysis results, determine whether there are any suspicious connected regions that conform to the typical morphological features and spatial distribution features of the bladder; If a suspicious connected region exists that matches the typical morphological and spatial distribution characteristics of the bladder, then bladder interference is determined to exist; otherwise, bladder interference is determined not to exist.

[0039] Specifically, when the maximum radioactivity count within the circumscribed rectangular region of the pelvic bone mask is not the maximum radioactivity count in the entire SPECT bone scintigraphy image, a more refined analysis is needed to avoid missing potential bladder interference. First, based on the radioactivity count information within the circumscribed rectangular region of the pelvic bone mask, connected regions within this region whose radioactivity count (either the maximum or average radioactivity count) exceeds a preset count threshold are identified, and these regions are designated as suspicious connected regions. The preset count threshold can be set based on clinical experience or statistical analysis, aiming to filter out regions with high radioactivity accumulation. A connected region refers to a region in the image where pixel values ​​are continuous (the identification method for connected regions can utilize existing technologies and is not limited here).

[0040] Furthermore, morphological feature analysis is performed on the identified suspicious connected regions to obtain morphological feature analysis results. Morphological feature analysis may include, but is not limited to, calculating and evaluating at least one of the geometric features of the region, such as area, perimeter, roundness, aspect ratio, and compactness. For example, the bladder typically exhibits a relatively regular elliptical or circular shape and has a certain size range.

[0041] Simultaneously, spatial distribution analysis is performed on the identified suspicious connected regions to obtain spatial distribution analysis results. Spatial distribution analysis may include, but is not limited to, assessing at least one of the following: the location of the region's centroid, its relative position to the pelvic bones, and symmetry. For example, the bladder is typically located in a specific anatomical position within the pelvic cavity and may exhibit a certain degree of symmetry in anteroposterior images.

[0042] Finally, based on the morphological and spatial distribution analysis results, it is determined whether there are any suspicious connected regions that conform to the typical morphological and spatial distribution characteristics of the bladder. For example, by comparing the calculated morphological feature values ​​with a preset reference range based on typical bladder morphology, it can be determined whether the suspicious connected region morphologically matches the bladder; by comparing the calculated spatial distribution characteristics with a preset reference range or pattern based on typical bladder anatomical location, it can be determined whether the suspicious connected region spatially matches the bladder. If such a region exists, bladder interference is determined to exist; otherwise, bladder interference is determined not to exist. This comprehensive judgment mechanism can effectively distinguish the bladder from other non-specific radioactive accumulations, improving the accuracy of the judgment.

[0043] This application's solution addresses the potential for missed detections when relying solely on maximum radioactivity counts to identify bladder interference by introducing morphological feature analysis and spatial distribution analysis. When bladder interference is not the highest radioactivity count across the entire SPECT bone scintigraphy image, identifying suspicious connected regions with radioactivity counts exceeding a preset threshold effectively separates potential bladder areas from the background. Morphological feature analysis of these suspicious connected regions allows for preliminary screening using the inherent geometric characteristics of the bladder, such as its shape and size. For example, the bladder typically has a relatively regular shape and a specific size range, which helps distinguish it from irregular lesions or artifacts. Simultaneously, spatial distribution analysis utilizes the bladder's specific anatomical location to further confirm whether suspicious areas are situated in typical bladder anatomical positions. For instance, the bladder is typically located in the lower central region of the pelvic cavity and has a specific relative positional relationship with the pelvic bones. By comprehensively considering morphological and spatial distribution features, this application's solution can more comprehensively and accurately identify areas with radioactivity intensity that are not the highest across the entire image but still exhibit bladder interference, thus avoiding misjudgments or missed detections caused by relying on a single indicator.

[0044] Using the above technical solution, even if the radioactivity count of bladder interference is not the highest radioactivity count in the entire SPECT bone scintigraphy image, potential bladder interference areas can be effectively identified through comprehensive analysis of the morphological characteristics and spatial distribution of suspicious connected regions. This significantly improves the accuracy and sensitivity of bladder interference assessment, especially when bladder radioactivity accumulation is not prominent or other high-radiation lesions are present. It effectively avoids missed diagnoses, thereby ensuring that subsequent interference elimination steps can be performed more comprehensively, ultimately obtaining more accurate SPECT bone scintigraphy image correction results.

[0045] In some implementations, step A4 includes: A401. Using the foreground segmentation method, segment all foreground connected regions within the bounding rectangular region of the pelvic bone mask; A402. Based on the radioactivity counting information, obtain the characteristic radioactivity count values ​​of each foreground connected region; the characteristic radioactivity count values ​​are the average or maximum radioactivity count values ​​within the foreground connected region. A403. Based on the characteristic radioactivity count value, sort each foreground connected region in descending order, and take the top N foreground connected regions as bladder candidate regions, where N is a preset positive integer value. A404. Obtain the area of ​​each bladder candidate region, and determine the final bladder region from each bladder candidate region based on the area; A405. Use edge optimization methods to perform edge optimization on the final bladder region to generate a bladder mask.

[0046] Foreground segmentation methods aim to identify regions with high radioactivity counts in an image, typically corresponding to high-uptake sites such as bones or the bladder. Specifically, foreground segmentation can employ various techniques. For example, thresholding segmentation identifies pixels above a radioactivity count threshold as foreground; region growing segmentation starts with one or more seed points and expands to surrounding pixels based on predefined similarity criteria (such as radioactivity count similarity) to form connected regions; clustering segmentation groups pixels according to their radioactivity count characteristics to form different foreground regions; and edge detection-based segmentation determines the contours of foreground regions by identifying boundaries of radioactivity count variations in the image. These methods can effectively separate all potential high-radioactivity connected regions from the circumscribed rectangular region of the pelvic bone mask.

[0047] Furthermore, after obtaining each foreground connected region, its radioactivity count features need to be quantified. The characteristic radioactivity count value can be understood as a numerical value representing the overall radioactivity level of that connected region. Specifically, it can be the average radioactivity count value of all pixels within the segmented foreground connected region, or the maximum radioactivity count value within that connected region. The purpose is to provide a quantitative basis for subsequent bladder region screening, in order to distinguish the radioactivity intensity of different connected regions.

[0048] Based on this, to prioritize regions with higher radioactivity counts, each foreground connected region is sorted in descending order according to its characteristic radioactivity count value. After sorting, the top N foreground connected regions are selected as candidate bladder regions, where N is a preset positive integer value. This step aims to initially screen out the high-radioactivity regions most likely representing the bladder, reducing the computational burden of subsequent processing and focusing on the most relevant regions.

[0049] Subsequently, the area of ​​each candidate bladder region is obtained. Based on these areas, the final bladder region is determined from multiple candidate regions. This step leverages the characteristic that the bladder typically has a specific size range in SPECT bone scintigraphy images, further improving the accuracy of bladder identification through area filtering.

[0050] Finally, to ensure the generated bladder mask has accurate boundaries, edge optimization methods are used to optimize the edges of the final bladder region. Edge optimization methods can smooth, correct, or refine the contour of the initially determined bladder region to eliminate jagged edges or irregular shapes that may occur during segmentation, thereby generating a more accurate and natural bladder mask.

[0051] This application's solution effectively improves the accuracy of bladder identification and the precision of mask generation by refining the bladder mask generation process into a series of logical steps. First, foreground segmentation comprehensively identifies all potential high-radioactivity regions within the circumscribed rectangular area of ​​the pelvic bone mask, providing a foundation for bladder identification. Then, by calculating the characteristic radioactivity count values ​​of each foreground connected region and sorting them in descending order, regions with the highest radioactive uptake intensity can be preferentially selected; these regions typically correspond to the radioactive characteristics of the bladder. Furthermore, by analyzing and filtering the area of ​​candidate bladder regions, false positive regions that do not conform to typical bladder dimensions can be eliminated, thus locating the bladder more accurately. Finally, edge optimization is used to refine the determined bladder region, ensuring that the generated bladder mask has smooth and accurate boundaries, avoiding errors caused by coarse segmentation, and providing high-quality input for subsequent bladder interference elimination.

[0052] By employing the aforementioned technical solution, more accurate identification and finer boundary delineation of the bladder region can be achieved when generating a bladder mask in SPECT bone scintigraphy images. This solution, through multi-stage screening and optimization, effectively reduces the risk of misidentifying non-bladder regions as bladder and improves the geometric accuracy of the bladder mask. This provides high-quality input for subsequent bladder interference elimination steps, resulting in SPECT bone scintigraphy images with higher diagnostic value and reliability after bladder interference elimination, avoiding the loss or misjudgment of bone lesion information due to inaccurate bladder masking.

[0053] Preferably, step A404 may include: Obtain the area of ​​each candidate bladder region; Obtain the patient's body shape parameters or overall size information of SPECT bone scan images; Calculate the expected area range of the bladder based on body shape parameters or overall size information; The area of ​​each bladder candidate region is compared with the expected area range to determine the final bladder region from each bladder candidate region.

[0054] Specifically, after obtaining the area of ​​each bladder candidate region, additional reference information needs to be introduced to improve the accuracy of bladder identification. Among these, the patient's body shape parameters can be understood as indicators reflecting the patient's physiological characteristics, such as at least one of the patient's height, weight, body surface area, age, and gender. These parameters are correlated with the physiological dimensions of human organs and can be used to estimate the appropriate bladder size. Simultaneously, the overall size information of the SPECT bone scan image can refer to the image's pixel resolution, physical dimensions (e.g., pixel pitch obtained through DICOM metadata), or the overall size of the pelvic bone mask in the image, aiming to provide a measure of the actual physical space represented by the image.

[0055] Based on the acquired patient body shape parameters or the overall size information of the SPECT bone scan image, the expected area range of the bladder can be calculated. This expected area range is a pre-defined reasonable interval established based on statistical analysis of a large amount of clinical data or medical experience, used to limit the area that a normal bladder may occupy in the image. For example, a regression model can be constructed, taking the patient's body shape parameters or the overall size information of the SPECT bone scan image as input, and outputting a predicted mean and standard deviation of the bladder area, thereby determining a reasonable area range.

[0056] Subsequently, the area of ​​each candidate bladder region is compared with the calculated expected area range. This comparison process effectively filters out candidate regions whose area falls within a reasonable physiological range, thereby eliminating non-bladder regions that are too large or too small due to morphological abnormalities, pathological changes, or artifacts. Thus, the final bladder region is determined from the candidate bladder regions that meet the expected area range.

[0057] Through the above technical solution, this application can significantly improve the accuracy and robustness of bladder region identification in SPECT bone scintigraphy images. By considering individual patient differences and the actual physical size of the image, the determined expected bladder area range is more targeted, making the final bladder region determination more reliable. This not only helps reduce interference elimination errors caused by inaccurate bladder identification but also effectively avoids the false elimination of normal bone regions, thereby generating more accurate SPECT bone scintigraphy images with greater clinical diagnostic value.

[0058] Furthermore, the step of comparing the area of ​​each candidate bladder region with the expected area range to determine the final bladder region from the candidate bladder regions may include: If only one bladder candidate region falls within the expected area range, then the bladder candidate region whose area falls within the expected area range shall be taken as the final bladder region. If the areas of multiple bladder candidate regions fall within the expected area range, the radioactivity count distribution characteristics and spatial location characteristics of the corresponding bladder candidate regions are obtained, and the final bladder region is determined from the corresponding bladder candidate regions based on the radioactivity count distribution characteristics and spatial location characteristics. If no bladder candidate region falls within the expected area range, the bladder candidate region with the largest area will be used as the final bladder region.

[0059] Specifically, the steps described above, which compare the area of ​​each candidate bladder region with the expected area range to determine the final bladder region from each candidate bladder region, can adopt different strategies based on different comparison results.

[0060] Specifically, when only one candidate bladder region falls within the expected area range, that candidate region is directly identified as the final bladder region. This indicates that the candidate region's size closely matches the expected bladder size, thus possessing a high degree of reliability.

[0061] Furthermore, when multiple candidate bladder regions fall within the expected area range, more refined discrimination criteria are needed to avoid misjudgment. In this case, the radioactivity count distribution characteristics and spatial location characteristics of the corresponding candidate bladder regions will be obtained. Radioactivity count distribution characteristics may include, but are not limited to, the average radioactivity count within the region, the maximum radioactivity count, the standard deviation of the radioactivity count, or the characteristics of the radioactivity count histogram. These characteristics reflect the degree and uniformity of radioactive tracer aggregation within the region. Spatial location characteristics may include, but are not limited to, the centroid coordinates of the region, its relative distance to specific anatomical landmarks within the pelvic skeletal region (such as the pubic symphysis, sacrum, etc.), or the shape characteristics of the region. These characteristics help determine whether the region is located in the typical anatomical position of the bladder. Based on these characteristics, machine learning models or expert rule systems can be used to further filter from multiple candidate regions that meet the area criteria to select the region that best matches the typical characteristics of the bladder as the final bladder region.

[0062] Furthermore, if no candidate bladder region falls within the expected area range, it may indicate that the expected area range is set too strictly, or that there are anomalies in the image. In this case, to ensure the identification of potential bladder regions, the candidate bladder region with the largest area will be used as the final bladder region. This strategy is based on the experience that the bladder is typically a large radioactive accumulation organ in the pelvic region, and its possibility as a bladder is prioritized even if its area slightly exceeds the expected range.

[0063] This application's solution effectively addresses potential ambiguity in comparing the area of ​​candidate bladder regions with the expected area range by introducing a multi-level decision-making logic. When only one candidate region meets the area requirement, the solution directly and efficiently determines the final bladder region, avoiding unnecessary complex calculations. When multiple candidate regions fall within the expected range, further analysis of radioactivity count distribution characteristics and spatial location features allows for more refined differentiation of these candidate regions using richer image information and anatomical prior knowledge, thereby improving the accuracy and specificity of bladder identification. For example, the bladder typically exhibits high radioactivity accumulation and is located in a specific position within the pelvis; these features help identify the true bladder from multiple regions of similar size. When the areas of all candidate regions do not meet the expected range, the solution selects the candidate region with the largest area as the final bladder region, ensuring a reasonable identification result even in extreme cases. This avoids the omission of bladder regions due to strict area limitations, thus enhancing the robustness of the method.

[0064] Through the above technical solution, this application can significantly improve the accuracy and robustness of bladder region identification in SPECT bone scintigraphy images. Specifically, this solution provides a clear and reasonable bladder region determination strategy in different scenarios (single matching, multiple matching, and no matching) through a hierarchical decision-making mechanism. Especially when the areas of multiple bladder candidate regions fall within the expected range, the introduction of radioactive count distribution features and spatial location features for secondary screening can effectively distinguish regions with similar morphology but different properties, avoiding misjudgment and ensuring the accurate identification of the final bladder region. In addition, in the special case where no candidate region area falls within the expected range, by selecting the candidate region with the largest area, the risk of bladder regions being missed is effectively reduced, further improving the practicality and reliability of the method and providing a more accurate foundation for subsequent bladder interference elimination.

[0065] In some possible implementations, step A5 includes: The radioactivity count in the bladder mask area was set to 0 to obtain a SPECT bone scintigraphy image after eliminating bladder interference.

[0066] Specifically, setting the radioactivity count to 0 within the bladder mask area means that after acquiring the bladder mask, the radioactivity count values ​​of the image pixels covered by the mask are directly modified to zero. This operation aims to completely eliminate the strong signal caused by the radioactive tracer in the urine within the bladder area, thereby eliminating its obscuring or interference with the surrounding bone structures. This results in a SPECT bone scintigraphy image after eliminating bladder interference, where no radioactive signal is displayed in the bladder area, allowing for a clearer presentation of the radioactivity distribution in adjacent bones.

[0067] The above-described technical solution enables rapid and complete elimination of bladder interference in SPECT bone scintigraphy images. This method is simple to operate and computationally efficient, effectively preventing high radioactivity signals from the bladder region from interfering with the diagnosis of surrounding bone structures, particularly the pelvic bones. As a result, physicians can more clearly observe the radioactive uptake of bones near the bladder, improving the detection rate and diagnostic accuracy of bone lesions, especially early or minor lesions, thereby enhancing the clinical application value of SPECT bone scintigraphy images.

[0068] In some other possible implementations, step A5 includes: A501. Obtain the boundary information of the bladder mask region; A502. Based on the boundary information, obtain the radioactivity count information of the adjacent bone region around the bladder mask region; A503. Calculate the radioactivity count gradient information of the adjacent bone region based on the radioactivity count information of the adjacent bone region; A504. Based on radioactivity count gradient information, generate alternative radioactivity counts for the bladder mask region; A505. Replace the radioactivity counts in the bladder mask area with replacement radioactivity counts to generate a SPECT bone scintigraphy image after eliminating bladder interference.

[0069] Specifically, in step A501, obtaining the boundary information of the bladder mask region refers to identifying and extracting the edge pixels or voxel set of the generated bladder mask using image processing algorithms. This boundary information can be a list of pixel or voxel coordinates or a binary image in which the boundary pixels are labeled. Its purpose is to clearly define the precise boundary between the bladder region and surrounding tissues, providing a basis for subsequent analysis of neighboring regions.

[0070] In step A502, based on the boundary information, the radioactivity count information of the adjacent bone region surrounding the bladder mask area is obtained. This can be understood as defining a neighboring region surrounding the bladder mask by extending outwards by a certain distance (which can be preset according to actual needs). This neighboring region is typically set to include the bone structures around the bladder, such as the pelvic bones. Then, the radioactivity count values ​​of all pixels or voxels within this neighboring region are extracted. The purpose is to collect the actual radioactivity distribution data around the bladder interference area as a basis for inferring the radioactivity count inside the bladder.

[0071] In practical applications, step A503 involves calculating the radioactivity count gradient information of adjacent bone regions based on their radioactivity count information. Specifically, this means applying gradient operators (such as the Sobel operator, Prewitt operator, or Laplacian operator) to the radioactivity count data of adjacent bone regions to calculate the spatial rate of change of radioactivity counts. Gradient information can reflect the trend of radioactivity counts diffusing or attenuating outwards from the bladder boundary. Its purpose is to capture local variation patterns in radioactivity counts, providing a basis for smoothly filling the bladder region.

[0072] Further, in step A504, generating replacement radioactivity counts for the bladder mask region based on radioactivity count gradient information refers to using interpolation algorithms (such as linear interpolation, bilinear interpolation, cubic spline interpolation, or diffusion-based image inpainting techniques) to predict and generate radioactivity counts within the bladder mask region based on the radioactivity counts and gradient information of neighboring regions. These replacement counts aim to simulate the radioactivity distribution that the region should have when the bladder is absent, ensuring a smooth transition with the radioactivity counts of the surrounding bone region. The goal is to eliminate bladder interference while maintaining the natural continuity of the image and avoiding the introduction of new artifacts.

[0073] Therefore, in step A505, the radioactivity count within the bladder mask area is replaced with a replacement radioactivity count to generate a SPECT bone scintigraphy image free of bladder interference. Specifically, this involves covering the original, interfered radioactivity count data within the bladder mask with the replacement radioactivity count generated in step A504. The resulting image will no longer contain the strong radioactivity signal from the bladder, but will instead show a radioactivity distribution that blends naturally with the surrounding bone region. The aim is to provide a clearer, more diagnostically valuable SPECT bone scintigraphy image, facilitating accurate assessment of lesions in the pelvic region by physicians.

[0074] This application's solution acquires refined boundary information of the bladder mask region and, based on this, extracts radioactivity count information and its gradient information from adjacent bone regions, thereby accurately capturing the true radioactivity distribution pattern of the bones surrounding the bladder. It is precisely this in-depth analysis of these local radioactivity distribution characteristics that enables the system to generate replacement radioactivity counts that smoothly transition with the surrounding region based on gradient information. This replacement mechanism avoids the image discontinuities and information loss that may result from simple zeroing. By simulating the radioactivity distribution when the bladder is absent, it effectively fills in the bladder region, thus eliminating interference while maximizing the preservation of the image's diagnostic value and visual naturalness.

[0075] Through the above technical solution, this application can avoid the loss of local image information and visual artifacts caused by simply zeroing out the bladder interference. By replacing the bladder mask area with the radioactivity count and gradient information of the adjacent bones, an image that smoothly connects with the radioactivity distribution of the surrounding bone area can be generated, thereby effectively improving the overall quality and diagnostic accuracy of SPECT bone scintigraphy images. This replacement method not only eliminates the strong radioactivity interference from the bladder, but also allows doctors to more clearly and accurately assess the true radioactivity uptake of the bones around the bladder, which has significant clinical significance, especially for the diagnosis of lesions in the pelvic region, avoiding misdiagnosis or missed diagnosis due to improper interference elimination.

[0076] In some preferred embodiments, a specific example is illustrated below. Assume that during the processing of a patient's SPECT bone scan image, a bladder mask has been successfully generated via step A4. To eliminate bladder interference and maintain image continuity, the system first accurately identifies the boundary pixels of the bladder mask. Next, a ring-shaped neighborhood is defined, centered on these boundary pixels and extending outwards by 5 pixels, and the radioactivity counts of all pixels within this neighborhood are extracted. Subsequently, the Sobel operator is used to calculate the gradient information of the radioactivity counts within this neighborhood to understand how the radioactivity counts change from the bladder edge outwards. Based on this gradient information, the system can employ a bilinear interpolation algorithm to smoothly extrapolate the replacement radioactivity count for each pixel inside the bladder mask from the radioactivity counts of the neighborhood. For example, if the radioactivity count in the bone region above the bladder is high, while the radioactivity count in the bone region below is low, the interpolation algorithm will generate a replacement count distribution that smoothly transitions from high to low. Finally, the original radioactivity counts in the bladder mask area are replaced with these generated alternative radioactivity counts, resulting in a SPECT bone scintigraphy image that eliminates bladder interference while maintaining the natural continuity of the image, allowing for clearer observation and diagnosis of bone lesions in the pelvic region.

[0077] Preferably, step A6 may include: A601. Add the SPECT bone scintigraphy image after eliminating bladder interference to the correction result information; A602. Obtain area information and average radioactivity count information of the bladder mask region; A603. Compare the area information with a preset area threshold, and compare the average radioactivity count information with a preset radioactivity count threshold; A604. If the area information is greater than the preset area threshold and the average radioactivity count information is greater than the preset radioactivity count threshold, a prompt message suggesting an additional abdominal SPECT / CT examination will be generated and added to the correction result information. A605. Output correction result information.

[0078] Specifically, the corrected results information can be understood as a comprehensive report or data structure that not only contains the processed SPECT bone scintigraphy images but also carries other relevant diagnostic or suggestive information. Including the SPECT bone scintigraphy images after eliminating bladder interference in the corrected results information aims to ensure that the corrected core diagnostic images are presented completely.

[0079] The area information of the bladder mask region refers to the number of pixels or voxels covered when the bladder mask is generated, reflecting the extent of bladder interference. The average radioactivity count information refers to the average radioactivity count of all pixels or voxels within the bladder mask region, reflecting the intensity of bladder interference. Obtaining this information aims to quantify the degree of bladder interference, providing an objective basis for subsequent judgments.

[0080] In practical applications, preset area thresholds and preset radioactivity count thresholds are reference values ​​determined based on clinical experience, statistical data, or expert consensus, used to assess the clinical significance of bladder interference. Comparing the area information of the bladder mask region with the preset area thresholds, and comparing the average radioactivity count information with the preset radioactivity count thresholds, aims to determine whether the bladder interference reaches a level requiring special attention.

[0081] Furthermore, if both the area information and the average radioactivity count information exceed a preset area threshold, it indicates that the radioactivity accumulation in the bladder region is not only extensive but also high in intensity. This may suggest abnormal radioactive tracer retention or metabolism, such as urinary retention, bladder lesions, or abnormal accumulation of tracers in non-bladder tissues. In this case, a suggestion to perform an abdominal SPECT / CT scan is generated and added to the corrected results information. This aims to provide clinicians with further diagnostic advice, guiding them to consider more detailed imaging examinations to clarify the underlying cause. Abdominal SPECT / CT scans can provide more precise anatomical localization and functional information, which is helpful for differential diagnosis. Finally, the corrected results information containing all relevant information is output, ensuring that clinicians receive a comprehensive and instructive diagnostic report.

[0082] This application's solution achieves intelligent assessment of the degree of bladder interference by quantitatively analyzing the area and average radioactivity count of the bladder mask region and comparing them with preset clinical thresholds. This quantitative assessment enables the system to identify bladder interference that may have clinical significance and indicate potential pathological conditions. When the range and intensity of bladder interference exceed the preset thresholds, the system proactively generates a prompt to perform an abdominal SPECT / CT examination. This mechanism effectively compensates for the insufficient information provided by simply outputting corrected images, elevating simple image correction to a comprehensive output with preliminary diagnostic assistance and clinical decision guidance value.

[0083] Through the aforementioned technical solution, this application can provide more comprehensive and intelligent SPECT bone scintigraphy correction results. This solution not only eliminates bladder interference but also achieves early identification and alerting of potential abnormalities by quantitatively analyzing the area and radioactivity count of the bladder mask region and combining it with preset thresholds. This significantly improves the efficiency and diagnostic accuracy of clinicians in interpreting SPECT bone scintigraphy images, avoids missing important clinical information due to bladder interference, and provides more precise guidance for subsequent diagnosis and treatment, thereby enhancing the overall intelligence and clinical value of the diagnostic process.

[0084] refer to Figure 2 This application provides a SPECT bone scintigraphy bladder interference cancellation system for canceling bladder interference in SPECT bone scintigraphy images, the SPECT bone scintigraphy images including anterior and / or posterior plane images, the system comprising: Image acquisition module 1 is used to acquire SPECT bone scintigraphy images (for details, please refer to step A1 above). Segmentation module 2 is used to segment the pelvic bone region in the SPECT bone imaging image using a pre-trained nnUNet network model and generate a pelvic bone mask (for details, please refer to step A2 above). Interference detection module 3 is used to determine whether there is bladder interference based on the radioactivity count information in the outer rectangular area of ​​the pelvic bone mask (for details, please refer to step A3 above). Bladder mask generation module 4 is used to generate a bladder mask in the outer rectangular region of the pelvic bone mask when there is bladder interference, based on radioactivity counting information and combining foreground segmentation and edge optimization methods (for details, please refer to step A4 above). Interference cancellation module 5 is used to eliminate or replace the radioactive count in the bladder mask area and generate a SPECT bone scintigraphy image after eliminating bladder interference (for details, please refer to step A5 above). Output module 6 is used to output correction result information containing SPECT bone scintigraphy images after eliminating bladder interference (for details, please refer to step A6 above).

[0085] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for eliminating bladder interference in SPECT bone imaging, for eliminating bladder interference in SPECT bone imaging images including anterior planar images and / or posterior planar images, characterized in that, The method comprises the following steps: A1. Obtain a SPECT bone imaging image; A2. Segment the pelvic bone region in the SPECT bone imaging image using a pre-trained nnUNet network model to generate a pelvic bone mask; A3. Determine whether there is bladder interference according to the radioactivity count information in the circumscribed rectangular region of the pelvic bone mask; A4. If there is bladder interference, generate a bladder mask in the circumscribed rectangular region of the pelvic bone mask according to the radioactivity count information, combined with a foreground segmentation method and an edge optimization method; A5. Eliminate or replace the radioactivity count in the bladder mask region to generate a SPECT bone imaging image after eliminating the bladder interference; A6. Output the correction result information containing the SPECT bone imaging image after eliminating the bladder interference.

2. The method for eliminating bladder interference in SPECT bone imaging according to claim 1, characterized in that, Step A3 comprises: A301. Obtain the maximum radioactivity count in the circumscribed rectangular region of the pelvic bone mask according to the radioactivity count information in the circumscribed rectangular region; A302. If the maximum radioactivity count in the circumscribed rectangular region is the maximum radioactivity count in the full image of the SPECT bone imaging image, it is determined that there is bladder interference.

3. The method for eliminating bladder interference in SPECT bone imaging according to claim 2, wherein, Step A3 further comprises: A303. If the maximum radioactivity count in the circumscribed rectangular region of the pelvic bone mask is not the maximum radioactivity count in the full image of the SPECT bone imaging image, then: Identify the connected region with radioactivity count higher than the pre-designed threshold value in the circumscribed rectangular region of the pelvic bone mask as a suspicious connected region according to the radioactivity count information in the circumscribed rectangular region; Perform morphological feature analysis on the suspicious connected region to obtain a morphological feature analysis result; Perform spatial distribution analysis on the suspicious connected region to obtain a spatial distribution analysis result; Determine whether there is a suspicious connected region that meets the typical morphological features and spatial distribution characteristics of the bladder according to the morphological feature analysis result and the spatial distribution analysis result; If there is a suspicious connected region that meets the typical morphological features and spatial distribution characteristics of the bladder, it is determined that there is bladder interference, otherwise, it is determined that there is no bladder interference.

4. The method for eliminating bladder interference in SPECT bone imaging according to claim 1, wherein, Step A4 comprises: A401. Segment all foreground connected domains in the circumscribed rectangular region of the pelvic bone mask using a foreground segmentation method; A402. Obtain the feature radioactivity count value of each foreground connected domain according to the radioactivity count information; the feature radioactivity count value is the average radioactivity count value or the maximum radioactivity count value in the foreground connected domain; A403. Sort each foreground connected domain in descending order according to the feature radioactivity count value, and take the top N foreground connected domains as bladder candidate regions, N being a pre-set positive integer value; A404. Obtain the area of each bladder candidate region, and determine the final bladder region from each bladder candidate region according to the area; A405. Perform edge optimization on the final bladder region using an edge optimization method to generate the bladder mask.

5. The method for eliminating bladder interference in SPECT bone imaging according to claim 4, wherein, Step A404 comprises: Obtain the area of each bladder candidate region; obtaining a body parameter of a patient or overall size information of the SPECT bone imaging image; calculating an expected area range of the bladder according to the body parameter or the overall size information; comparing the area of each of the bladder candidate regions with the expected area range to determine a final bladder region from the bladder candidate regions.

6. The method for eliminating bladder interference in SPECT bone imaging according to claim 5, wherein, The step of comparing the area of each of the bladder candidate regions with the expected area range to determine a final bladder region from the bladder candidate regions comprises: if only one of the bladder candidate regions has an area falling within the expected area range, taking the bladder candidate region with the area falling within the expected area range as the final bladder region; if multiple of the bladder candidate regions have areas falling within the expected area range, obtaining a radioactivity count distribution feature and a spatial position feature of the corresponding bladder candidate regions, and determining the final bladder region from the corresponding bladder candidate regions according to the radioactivity count distribution feature and the spatial position feature; if none of the bladder candidate regions has an area falling within the expected area range, taking the bladder candidate region with the largest area as the final bladder region.

7. The method for eliminating bladder interference in SPECT bone imaging according to claim 1, wherein, Step A5 comprises: setting radioactivity counts in the bladder mask region to 0 to obtain a SPECT bone imaging image with bladder interference eliminated.

8. The method for eliminating bladder interference in SPECT bone imaging according to claim 1, wherein, Step A5 comprises: A501. obtaining boundary information of the bladder mask region; A502. obtaining radioactivity count information of adjacent bone regions around the bladder mask region according to the boundary information; A503. calculating radioactivity count gradient information of the adjacent bone regions according to the radioactivity count information of the adjacent bone regions; A504. generating replacement radioactivity counts of the bladder mask region based on the radioactivity count gradient information; A505. replacing radioactivity counts in the bladder mask region with the replacement radioactivity counts to generate a SPECT bone imaging image with bladder interference eliminated.

9. The method for eliminating bladder interference in SPECT bone imaging according to claim 1, wherein, Step A6 comprises: A601. adding the SPECT bone imaging image with bladder interference eliminated to correction result information; A602. obtaining area information and average radioactivity count information of the bladder mask region; A603. comparing the area information with a preset area threshold value, and comparing the average radioactivity count information with a preset radioactivity count threshold value; A604. if the area information is greater than the preset area threshold value and the average radioactivity count information is greater than the preset radioactivity count threshold value, generating prompt information suggesting to perform an abdominal SPECT / CT examination, and adding the prompt information to the correction result information; A605. outputting the correction result information.

10. A SPECT bone imaging bladder interference elimination system for performing bladder interference elimination on a SPECT bone imaging image, the SPECT bone imaging image comprising an anterior planar image and / or a posterior planar image, characterized in that, The system comprises: an image acquisition module configured to acquire a SPECT bone imaging image; a segmentation module configured to segment a pelvic bone region in the SPECT bone imaging image using a pre-trained nnUNet network model to generate a pelvic bone mask; an interference judgment module configured to judge whether there is bladder interference according to radioactivity count information in a circumscribed rectangular region of the pelvic bone mask; a bladder mask generation module configured to generate a bladder mask in a bounding rectangle region of the pelvic skeleton mask according to the radioactivity count information in the presence of bladder interference, by combining a foreground segmentation method and an edge optimization method; an interference elimination module configured to eliminate or replace the radioactivity count in the bladder mask region to generate a SPECT bone imaging image after eliminating the bladder interference; an output module configured to output correction result information containing the SPECT bone imaging image after eliminating the bladder interference.

Images