A method for segmenting renal magnetic resonance images

Abstract

This invention relates to the field of magnetic resonance image processing technology, specifically to a method for segmenting renal magnetic resonance images, comprising the following steps: acquiring bilateral renal magnetic resonance scan data of a child with hydronephrosis using T2WI and DWI sequences from a plain magnetic resonance imaging (MRI) scan; establishing a renal region image segmentation model framework by combining a U-net network and a Mask R-CNN network structure; and segmenting the renal region image of the bilateral renal magnetic resonance scan data using the renal region image segmentation model framework to determine the hydronephrotic kidney region and the contralateral kidney region. This invention achieves adaptive enhancement of details in the hydronephrotic kidney within an artificial intelligence algorithm. It enhances details in the hydronephrotic kidney when segmentation performance can be improved, and avoids enhancing details in the hydronephrotic kidney when segmentation performance cannot be improved, reducing ineffective image enhancement computational overhead. Adaptive enhancement achieves a balance between segmentation accuracy and efficiency, resulting in highly efficient and accurate renal region image segmentation.

Description

Technical Field

[0001] This invention relates to the field of magnetic resonance image processing technology, and specifically to a method for segmenting renal magnetic resonance images. Background Technology

[0002] Hydronephrosis is kidney damage caused by obstruction of urine flow from the kidney to the bladder. It is a common urological disease in children, with the most frequent cause being ureteropelvic junction obstruction (UPJO) due to intrinsic abnormalities in the proximal ureter. Hydronephrosis is not a static but dynamic process, and the optimal timing for treatment remains controversial, especially in pediatric patients. Assessment of split kidney function is a crucial factor in surgical treatment; previous reports have suggested that children with poor preoperative split kidney function have a poorer prognosis. Therefore, it is necessary to identify a non-invasive and accurate imaging biomarker for assessing split kidney function in children with hydronephrosis.

[0003] Assessing renal function using radionuclide renal scintigraphy is considered a reference standard and may aid in treatment decisions. However, radionuclide renal scintigraphy suffers from poor image quality and low soft tissue resolution. Furthermore, due to the use of radioactive tracers and ionizing radiation, it cannot be used for frequent follow-ups. Studies have reported that renal morphological volume assessed using contrast-enhanced CT (CE-CT), non-contrast-enhanced CT (NCE-CT), and contrast-enhanced MRI (CE-MRI) has a moderate to excellent correlation with renal function. However, CT examinations involve ionizing radiation, and gadolinium-containing contrast agents may increase the risk of renal systemic fibrosis. Non-contrast-enhanced MRU (NCE-MRU) is widely used in the assessment of pediatric patients with hydronephrosis due to its high resolution, lack of radiation exposure, absence of contrast agents, and short scan time. Magnetic resonance urography (MRU) has been widely and routinely used for imaging pediatric patients with hydronephrosis, enabling renal volume measurement without the use of contrast agents while ensuring high soft tissue resolution and contrast. In our research paper, "Non-contrast enhanced magnetic resonance urography forming evaluating split kidney function in pediatric patients with hydronephrosis: comparison with renal scintigraphy," we demonstrated a high correlation between split kidney volume and split kidney function, avoiding the use of radionuclides and ionizing radiation. Therefore, split kidney function can be assessed through magnetic resonance imaging analysis.

[0004] However, while existing technologies can determine kidney morphological features, diffusion-weighted features, and other characteristics through magnetic resonance image analysis to assess renal function in cases of hydronephrosis, the lack of efficient and accurate kidney region image segmentation methods and insufficient capture of details in the affected kidney in magnetic resonance image analysis affect the accuracy and efficiency of kidney feature extraction, ultimately impacting the efficiency and accuracy of renal function assessment. Summary of the Invention

[0005] The purpose of this invention is to provide a method for segmenting renal magnetic resonance images, in order to solve the technical problem that the lack of efficient and accurate renal region image segmentation methods in the existing magnetic resonance image analysis leads to the impact on the accuracy and efficiency of renal feature extraction, and ultimately affects the efficiency and accuracy of renal function assessment.

[0006] To solve the above-mentioned technical problems, the present invention specifically provides the following technical solution:

[0007] A method for segmenting renal magnetic resonance images includes the following steps:

[0008] Bilateral renal MRI data of children with hydronephrosis were obtained using T2WI sequences and diffusion-weighted imaging (DWI) sequences.

[0009] A kidney region image segmentation model framework was established by combining the U-net network and Mask R-CNN network structures.

[0010] Using a kidney region image segmentation model framework, kidney region image segmentation is performed on the bilateral kidney magnetic resonance scan data to determine the hydronephrosis-side kidney region and the contralateral kidney region.

[0011] As a preferred embodiment of the present invention, the method for constructing the kidney region image segmentation model framework includes:

[0012] A hydronephrosis enhancement module was constructed using a U-net network to enhance the image of the hydronephrosis kidney in bilateral renal MRI scan data.

[0013] The first hydronephrosis and contralateral kidney segmentation module was constructed using a Mask R-CNN network to segment the hydronephrosis and contralateral kidney regions in enhanced bilateral renal MRI data.

[0014] A second segmentation module for the hydronephrotic kidney and the contralateral kidney was constructed using a Mask R-CNN network to segment the regions of the hydronephrotic kidney and the contralateral kidney in bilateral renal MRI scan data.

[0015] The enhancement triggering module of the hydronephrosis enhancement module is adjusted based on the comparison results of the segmentation performance between the first hydronephrosis-side kidney and contralateral kidney segmentation module and the second hydronephrosis-side kidney and contralateral kidney segmentation module.

[0016] As a preferred embodiment of the present invention, the hydronephrosis-side kidney enhancement module uses a U-net network to segment the hydronephrosis-side kidney region in the bilateral renal magnetic resonance scanning data to obtain a hydronephrosis-side kidney mask image, and performs pixel overlay and fusion of the hydronephrosis-side kidney mask image and the bilateral renal magnetic resonance scanning data to obtain bilateral renal magnetic resonance scanning data with hydronephrosis-side kidney enhancement.

[0017] ;

[0018] In the formula, Enhanced bilateral renal MRI data for the hydronephrosis-affected kidney. This is MRI data of both kidneys. This is a mask image of the hydronephrotic kidney obtained by segmentation using a U-net network. U-net is the U-net network. To enhance the feedback regulation of the trigger module, the enhancement coefficient of the hydronephrosis-side kidney in the enhancement module is adjusted.

[0019] As a preferred embodiment of the present invention, the first hydronephrotic kidney and contralateral kidney segmentation module is used to segment the hydronephrotic kidney region and the contralateral kidney region in the enhanced bilateral renal magnetic resonance scan data of the hydronephrotic kidney using a Mask R-CNN network, so as to obtain the enhanced mask image of the hydronephrotic kidney and the mask image of the contralateral kidney.

[0020] The structural expression for the segmentation module of the first hydronephrotic kidney and the contralateral kidney is as follows:

[0021] ;

[0022] In the formula, This is a masked image of the contralateral kidney. The image shows a mask-enhanced image of the kidney on the hydronephrosis side, using the Mask R-CNN network.

[0023] As a preferred embodiment of the present invention, the second hydronephrotic kidney and contralateral kidney segmentation module is used to segment the hydronephrotic kidney region and the contralateral kidney region in bilateral renal magnetic resonance scan data using a Mask R-CNN network to obtain a mask image of the hydronephrotic kidney and a mask image of the contralateral kidney.

[0024] The structural expression for the segmentation module of the second hydronephrotic kidney and the contralateral kidney is as follows:

[0025] ;

[0026] In the formula, This is a masked image of the contralateral kidney. The image shows a masked image of the kidney on the side with hydronephrosis. Mask R-CNN is the Mask R-CNN network.

[0027] In a preferred embodiment of the present invention, the enhancement coefficient of the hydronephrosis-side kidney fed back by the enhancement triggering module consists of two parts: a start / stop parameter controlling the start and stop of the hydronephrosis-side kidney enhancement module, and an adaptive intensity parameter controlling the enhancement intensity of the hydronephrosis-side kidney enhancement module for the hydronephrosis-side kidney, wherein:

[0028] The enhancement coefficient of the hydronephrotic kidney is:

[0029] ;

[0030] In the formula, The enhancement coefficient of the hydronephrotic kidney. These are start / stop parameters. This is an adaptive intensity parameter.

[0031] As a preferred embodiment of the present invention, the start / stop parameters are:

[0032] ;

[0033] In the formula, GT is the actual masked image of the hydronephrotic kidney region. Enhanced image of the hydronephrotic kidney with masking. Image of the hydronephrotic kidney with a mask. For Kullback-Leibler divergence calculation, This is a pixel overlay operation formula.

[0034] As a preferred embodiment of the present invention, the adaptive intensity parameter is set based on the hydronephrosis-side kidney segmentation performance in the hydronephrosis-side kidney enhancement module. for: ;

[0035] In the formula, GT is the actual masked image of the hydronephrotic kidney region. The image shows a mask of the hydronephrotic kidney obtained by segmentation using a U-net network.

[0036] As a preferred embodiment of the present invention, the adaptive intensity parameter is set based on the segmentation performance difference between the first hydronephrotic kidney and contralateral kidney segmentation module and the second hydronephrotic kidney and contralateral kidney segmentation module. for: ;

[0037] In the formula, GT is the actual masked image of the hydronephrotic kidney region. The image shows a mask of the hydronephrotic kidney obtained by segmentation using the U-net network. This is for Kullback-Leibler divergence calculation.

[0038] As a preferred embodiment of the present invention, the enhanced mask image of the hydronephrotic kidney and the mask image of the contralateral kidney output by the first hydronephrotic kidney and the contralateral kidney segmentation module are used as the output of the kidney region image segmentation model framework, and the output of the kidney region image segmentation model framework is used as the kidney region segmentation result of the bilateral kidney magnetic resonance scan data.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] This invention acquires MRI plain scan and diffusion-weighted imaging data of children with hydronephrosis, and uses artificial intelligence algorithms to perform kidney-specific image feature analysis on the image data. The artificial intelligence algorithm achieves adaptive enhancement of details of the hydronephrotic side of the kidney. It enhances the details of the hydronephrotic side of the kidney when it can improve segmentation performance, and does not need to enhance the details of the hydronephrotic side of the kidney when it cannot improve segmentation performance, thereby reducing the computational overhead of ineffective image enhancement. Adaptive enhancement can achieve a balance between segmentation accuracy and segmentation efficiency, and can achieve efficient and accurate kidney region image segmentation. Attached Figure Description

[0041] To more clearly illustrate the embodiments of the present invention or the technical solutions in 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 merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0042] Figure 1 This is a flowchart of a method for segmenting renal MRI parenchymal images provided in an embodiment of the present invention;

[0043] Figure 2 This is a schematic diagram of the kidney region image segmentation model framework provided in an embodiment of the present invention;

[0044] Figure 3 This is a segmentation result of the kidney region image provided in an embodiment of the present invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] like Figure 1 As shown, the present invention provides a method for segmenting renal magnetic resonance images, comprising the following steps:

[0047] Bilateral renal MRI data of children with hydronephrosis were obtained using T2WI sequences and diffusion-weighted imaging (DWI) sequences.

[0048] The specific imaging scanning parameters are as follows: T2WI scanning parameters: 3.0 T MR scanner, 8-channel cardiac coil, TR 5300ms, TE 66.55ms, slice thickness 3mm, FA 110°; DWI parameters: TR / TE (ms), 4000 / shortest TE; matrix 128×96, bandwidth, 250 kHz; slice thickness 3mm, b-value (s / mm2) 0, 800.

[0049] A kidney region image segmentation model framework was established by combining the U-net network and Mask R-CNN network structures.

[0050] Using a kidney region image segmentation model framework, kidney region image segmentation was performed on bilateral renal MRI scan data to determine the region of the hydronephrosis-related kidney and the contralateral kidney region, such as... Figure 3 As shown.

[0051] To obtain the morphological features and diffusion-weighted features of the hydronephrotic and contralateral kidneys, this invention first requires image segmentation of the hydronephrotic and contralateral (or normal) kidneys on bilateral renal magnetic resonance imaging (MRI) scans. This involves segmenting the hydronephrotic and contralateral (or normal) kidney regions on the MRI scan images. Then, based on the segmented hydronephrotic and contralateral kidney regions, the morphological features and diffusion-weighted features of the kidneys are extracted. Therefore, this invention constructs an image segmentation model framework for the hydronephrotic and contralateral kidney regions.

[0052] The construction methods for the kidney region image segmentation model framework include:

[0053] A hydronephrosis enhancement module was constructed using a U-net network to enhance the image of the hydronephrosis kidney in bilateral renal MRI scan data.

[0054] The first hydronephrosis and contralateral kidney segmentation module was constructed using a Mask R-CNN network to segment the hydronephrosis and contralateral kidney regions in enhanced bilateral renal MRI data.

[0055] A second segmentation module for the hydronephrotic kidney and the contralateral kidney was constructed using a Mask R-CNN network to segment the regions of the hydronephrotic kidney and the contralateral kidney in bilateral renal MRI scan data.

[0056] The enhancement trigger module of the hydronephrosis enhancement module, namely the compare module, is adjusted based on the comparison results of the segmentation performance between the first hydronephrosis-side kidney and contralateral kidney segmentation module and the second hydronephrosis-side kidney and contralateral kidney segmentation module.

[0057] U-Net is a convolutional neural network for image segmentation, shaped like a "U," primarily composed of an encoder (Contracting Path) and a decoder (Expanding Path). The encoder extracts features progressively through convolutional and pooling layers, while the decoder restores image resolution through upsampling and convolutional operations. U-Net's unique feature is its skip connections, which directly pass feature maps from the encoder to the decoder, helping to recover more precise details. This design allows U-Net to better preserve detail when processing image segmentation tasks, thereby improving segmentation accuracy and precision.

[0058] Mask R-CNN (Mask Region-based Convolutional Neural Network) is an object detection and instance segmentation model that extends Faster R-CNN. It can not only detect the location of objects in an image and identify their categories, but also generate pixel-level segmentation masks for each detected object, thereby achieving more refined object segmentation.

[0059] like Figure 2 As shown, this invention constructs a kidney region image segmentation model framework to segment the hydronephrosis side kidney region and the contralateral kidney region. It consists of a hydronephrosis side kidney enhancement module based on the U-net network, and two hydronephrosis side kidney and contralateral kidney segmentation modules based on the Mask R-CNN network. The hydronephrosis side kidney enhancement module is used to first segment the hydronephrosis side kidney region on the original bilateral kidney MRI scan data using the U-net network. The segmentation mask result of the hydronephrosis side kidney region is superimposed and fused onto the original bilateral kidney MRI scan data to form hydronephrosis side kidney enhanced bilateral kidney MRI scan data. It can enhance a specific region (hydronephrosis side kidney region) in the original bilateral kidney MRI scan data while keeping other regions unchanged, thereby highlighting the details of the hydronephrosis side kidney region.

[0060] Since this invention aims to assess the renal function of children with hydronephrosis, the hydronephrotic kidney is of greater concern in the assessment process. Therefore, the morphological and diffusion-weighted features of the hydronephrotic kidney play a dominant role, requiring higher accuracy in their extraction. Unilateral enhancement of the hydronephrotic kidney region is employed to meet the accuracy requirements for its morphological and diffusion-weighted features, while the morphological and diffusion-weighted features of the corresponding normal kidney play a supporting role. This allows for a lower requirement for the normal kidney, eliminating the need for enhancement processing. This approach satisfies the extraction needs of the renal morphological and diffusion-weighted features while reducing unnecessary computation.

[0061] Then, the enhanced details of the bilateral renal MRI scan data were re-segmented through the Mask R-CNN network structure (i.e., the segmentation module of the first hydronephrotic kidney and the contralateral kidney) to achieve higher precision in the segmentation of the hydronephrotic kidney region. The enhanced details were used to obtain more accurate segmentation results of the hydronephrotic kidney region.

[0062] The kidney region image segmentation model framework also includes a Mask R-CNN network structure (i.e., a segmentation module for the second hydronephrosis-side kidney and the contralateral kidney). This module is used to segment the hydronephrosis-side kidney region on the original bilateral renal MRI scan data. It essentially acts as a quantification component for enhancing the hydronephrosis-side kidney, monitoring the segmentation performance of the hydronephrosis-side kidney region obtained from the original bilateral renal MRI scan data and comparing it with the segmentation performance of the hydronephrosis-side kidney region obtained from bilateral renal MRI scan data with enhanced details. If the former's performance is lower than the latter's, it is fed back to the enhancement coefficient of the hydronephrosis-side kidney. Above, that is, start / stop parameters The activation of the enhancement module for the hydronephrosis-related kidney is triggered, enabling the module to enhance the image of the hydronephrosis-related kidney. In other words, the segmentation of the first hydronephrosis-related kidney and the contralateral kidney is based on the enhanced bilateral renal MRI scan data of the hydronephrosis-related kidney. More detailed information after enhancement can be extracted.

[0063] When the former's performance is detected to be higher than or equal to the latter's, it indicates that the enhancement of the hydronephrotic kidney will not have a positive effect on improving segmentation performance, and will also be reflected in the enhancement coefficient of the hydronephrotic kidney. Above, that is, The enhancement module for the hydronephrotic kidney is turned off, and image enhancement of the hydronephrotic kidney is not performed. Ultimately, the segmentation of the first hydronephrotic kidney and the contralateral kidney is based on bilateral renal MRI scan data. The above is based on the original bilateral renal MRI scan data. At this point, the region enhancement operation is a redundant operation and does not need to be performed. It achieves the desired accuracy while improving efficiency. Therefore, by controlling the start and stop operation of the enhancement module of the hydronephrotic kidney through start and stop parameters, the efficiency and accuracy of image segmentation can be balanced.

[0064] Meanwhile, the enhancement coefficient of the hydronephrotic kidney It also integrates adaptive intensity parameters. Adaptive intensity parameters There are two setting methods. The first is to initially separate the kidney from the hydronephrosis side. Related, The lower the value, the better the initial segmentation performance of the hydronephrotic kidney. The higher the reliability of the initial segmentation result of the hydronephrotic kidney superimposed on the original bilateral renal MRI data, the better the detail enhancement effect will be. Adaptive intensity parameters under certain conditions Integration is not of practical significance. It only has practical significance under certain circumstances, thus setting it up , ,when The lower, The higher the value, the higher the reliability of the superposition. Ultimately, the segmentation of the first hydronephrotic kidney and the contralateral kidney is based on the enhanced bilateral renal MRI scan data of the hydronephrotic kidney. This involves adding unnecessary augmentation operations, sacrificing efficiency for improved accuracy.

[0065] The second type is the difference in segmentation performance between the segmentation modules for the first hydronephrotic kidney and the contralateral kidney and the second hydronephrotic kidney and the contralateral kidney, due to the adaptive intensity parameters. Only in It only has practical significance under certain circumstances, that is, in Under the given conditions, this is practically significant, indicating that the segmentation module for the first hydronephrotic kidney and the contralateral kidney requires enhancement operations from the enhancement module for the hydronephrotic kidney. The greater the dependence on enhancement operations, the greater the difference in segmentation performance (equivalent to the difference in segmentation results between the first and second hydronephrotic kidney segmentation modules). The larger the value of the segmentation module (the more dependent it is on enhancement operations), the more it aims to match the segmentation performance of the second hydronephrotic kidney and the contralateral kidney. Therefore, the reliability of the initial segmentation results of the hydronephrotic kidney superimposed on the original bilateral renal MRI data should also be higher, and the superimposition degree should be set accordingly. , ,when The higher, The higher the value, the higher the reliability of the superposition. Ultimately, the segmentation of the first hydronephrotic kidney and the contralateral kidney is based on the enhanced bilateral renal MRI scan data of the hydronephrotic kidney. This involves adding unnecessary augmentation operations, sacrificing efficiency for improved accuracy.

[0066] In summary, this invention can achieve adaptive enhancement of details in the hydronephrotic kidney. It enhances details in the hydronephrotic kidney when segmentation performance can be improved, and does not need to enhance details in the hydronephrotic kidney when segmentation performance cannot be improved. This reduces the computational overhead of ineffective image enhancement. The adaptive enhancement can achieve a balance between segmentation accuracy and segmentation efficiency, thus achieving efficient and accurate kidney region image segmentation.

[0067] The hydronephrosis-side kidney enhancement module uses the U-net network to segment the hydronephrosis-side kidney region in the bilateral renal magnetic resonance scan data to obtain the hydronephrosis-side kidney mask image, and performs pixel overlay and fusion of the hydronephrosis-side kidney mask image and the bilateral renal magnetic resonance scan data to obtain the bilateral renal magnetic resonance scan data with hydronephrosis-side kidney enhancement.

[0068] ;

[0069] In the formula, Enhanced bilateral renal MRI data for the hydronephrosis-affected kidney. This is MRI data of both kidneys. This is a mask image of the hydronephrotic kidney obtained by segmentation using a U-net network. U-net is the U-net network. To enhance the feedback regulation of the trigger module, the enhancement coefficient of the hydronephrosis-side kidney in the enhancement module is adjusted.

[0070] The first segmentation module for the hydronephrotic kidney and the contralateral kidney is used to segment the regions of the hydronephrotic kidney and the contralateral kidney in the enhanced bilateral renal magnetic resonance scan data of the hydronephrotic kidney using the Mask R-CNN network, so as to obtain the mask-enhanced image of the hydronephrotic kidney and the mask image of the contralateral kidney.

[0071] The structural expression for the segmentation module of the first hydronephrotic kidney and the contralateral kidney is as follows:

[0072] ;

[0073] In the formula, This is a masked image of the contralateral kidney. The image shows a mask-enhanced image of the kidney on the hydronephrosis side, using the Mask R-CNN network.

[0074] The second segmentation module for the hydronephrotic kidney and the contralateral kidney is used to segment the regions of the hydronephrotic kidney and the contralateral kidney in the bilateral renal magnetic resonance scan data using the Mask R-CNN network, so as to obtain the mask image of the hydronephrotic kidney and the mask image of the contralateral kidney.

[0075] The structural expression for the segmentation module of the second hydronephrotic kidney and the contralateral kidney is as follows:

[0076] ;

[0077] In the formula, This is a masked image of the contralateral kidney. The image shows a masked image of the kidney on the side with hydronephrosis. Mask R-CNN is the Mask R-CNN network.

[0078] The enhancement coefficient of the hydronephrosis-side kidney fed back by the enhancement trigger module consists of two parts: start / stop parameters that control the activation and deactivation of the enhancement module for the hydronephrosis-side kidney, and adaptive intensity parameters that control the enhancement intensity of the hydronephrosis-side kidney enhancement module for the hydronephrosis-side kidney.

[0079] The enhancement coefficient of the hydronephrotic kidney is:

[0080] ;

[0081] In the formula, The enhancement coefficient of the hydronephrotic kidney. These are start / stop parameters. This is an adaptive intensity parameter.

[0082] The start / stop parameters are:

[0083] ;

[0084] In the formula, GT is the actual masked image of the hydronephrotic kidney region. Enhanced image of the hydronephrotic kidney with masking. Image of the hydronephrotic kidney with a mask. For Kullback-Leibler divergence calculation, This is a pixel overlay operation formula.

[0085] The second segmentation module for the hydronephrosis-side kidney and the contralateral kidney is used to segment the hydronephrosis-side kidney region on the original bilateral renal MRI scan data. It essentially acts as a quantification component for the enhancement effect of the hydronephrosis-side kidney, monitoring the segmentation performance of the hydronephrosis-side kidney region obtained from the original bilateral renal MRI scan data and comparing it with the segmentation performance of the hydronephrosis-side kidney region obtained from bilateral renal MRI scan data with enhanced details. When the former's performance is lower than the latter, it is fed back to the enhancement coefficient of the hydronephrosis-side kidney. Above, that is, start / stop parameters The activation of the enhancement module for the hydronephrosis-related kidney is triggered, enabling the module to enhance the image of the hydronephrosis-related kidney. In other words, the segmentation of the first hydronephrosis-related kidney and the contralateral kidney is based on the enhanced bilateral renal MRI scan data of the hydronephrosis-related kidney. More detailed information after enhancement can be extracted.

[0086] When the former's performance is detected to be higher than or equal to the latter's, it indicates that the enhancement of the hydronephrotic kidney will not have a positive effect on improving segmentation performance, and will also be reflected in the enhancement coefficient of the hydronephrotic kidney. Above, that is, The enhancement module for the hydronephrotic kidney is turned off, and image enhancement of the hydronephrotic kidney is not performed. Ultimately, the segmentation of the first hydronephrotic kidney and the contralateral kidney is based on bilateral renal MRI scan data. The above is based on the original bilateral renal MRI scan data. At this point, the region enhancement operation is a redundant operation and does not need to be performed. It achieves the desired accuracy while improving efficiency. Therefore, by controlling the start and stop operation of the enhancement module of the hydronephrotic kidney through start and stop parameters, the efficiency and accuracy of image segmentation can be balanced.

[0087] Adaptive intensity parameters are set based on the segmentation performance of the hydronephrosis-side kidney in the enhanced module. for: ;

[0088] In the formula, GT is the actual masked image of the hydronephrotic kidney region. The image shows a mask of the hydronephrotic kidney obtained by segmentation using a U-net network.

[0089] Enhancement coefficient of the hydronephrotic kidney It also integrates adaptive intensity parameters. Adaptive intensity parameters There are two setting methods. The first is to initially separate the kidney from the hydronephrosis side. Related, The lower the value, the better the initial segmentation performance of the hydronephrotic kidney. The higher the reliability of the initial segmentation result of the hydronephrotic kidney superimposed on the original bilateral renal MRI data, the better the detail enhancement effect will be. Adaptive intensity parameters under certain conditions Integration is not of practical significance. It only has practical significance under certain circumstances, thus setting it up , ,when The lower, The higher the value, the higher the reliability of the superposition. Ultimately, the segmentation of the first hydronephrotic kidney and the contralateral kidney is based on the enhanced bilateral renal MRI scan data of the hydronephrotic kidney. This involves adding unnecessary augmentation operations, sacrificing efficiency for improved accuracy.

[0090] An adaptive intensity parameter is set based on the segmentation performance difference between the first segmentation module for the hydronephrotic kidney and the contralateral kidney and the second segmentation module for the hydronephrotic kidney and the contralateral kidney. for: ;

[0091] In the formula, GT is the actual masked image of the hydronephrotic kidney region. The image shows a mask of the hydronephrotic kidney obtained by segmentation using the U-net network. This is for Kullback-Leibler divergence calculation.

[0092] The second type is the difference in segmentation performance between the segmentation modules for the first hydronephrotic kidney and the contralateral kidney and the second hydronephrotic kidney and the contralateral kidney, due to the adaptive intensity parameters. Only in It only has practical significance under certain circumstances, that is, in Under the given conditions, this is practically significant, indicating that the segmentation module for the first hydronephrotic kidney and the contralateral kidney requires enhancement operations from the enhancement module for the hydronephrotic kidney. The greater the dependence on enhancement operations, the greater the difference in segmentation performance (equivalent to the difference in segmentation results between the first and second hydronephrotic kidney segmentation modules). The larger the value of the segmentation module (the more dependent it is on enhancement operations), the more it aims to match the segmentation performance of the second hydronephrotic kidney and the contralateral kidney. Therefore, the reliability of the initial segmentation results of the hydronephrotic kidney superimposed on the original bilateral renal MRI data should also be higher, and the superimposition degree should be set accordingly. , ,when The higher, The higher the value, the higher the reliability of the superposition. Ultimately, the segmentation of the first hydronephrotic kidney and the contralateral kidney is based on the enhanced bilateral renal MRI scan data of the hydronephrotic kidney. This involves adding unnecessary augmentation operations, sacrificing efficiency for improved accuracy.

[0093] The enhanced mask image of the hydronephrotic kidney and the mask image of the contralateral kidney, output by the segmentation module of the first hydronephrotic kidney and the contralateral kidney, are used as the output of the kidney region image segmentation model framework, and the output of the kidney region image segmentation model framework is used as the kidney region segmentation result of the bilateral kidney magnetic resonance scan data.

[0094] This invention acquires MRI plain scan and diffusion-weighted imaging data of children with hydronephrosis, and uses artificial intelligence algorithms to perform kidney-specific image feature analysis on the image data. The artificial intelligence algorithm achieves adaptive enhancement of details of the hydronephrotic side of the kidney. It enhances the details of the hydronephrotic side of the kidney when it can improve segmentation performance, and does not need to enhance the details of the hydronephrotic side of the kidney when it cannot improve segmentation performance, thereby reducing the computational overhead of ineffective image enhancement. Adaptive enhancement can achieve a balance between segmentation accuracy and segmentation efficiency, and can achieve efficient and accurate kidney region image segmentation.

[0095] The above embodiments are merely exemplary embodiments of this application and are not intended to limit this application. The scope of protection of this application is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to this application within its substance and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of this application.

Claims

1. A method for segmenting renal magnetic resonance images, characterized in that, Includes the following steps: Bilateral renal MRI data of children with hydronephrosis were obtained using T2WI sequences and diffusion-weighted imaging (DWI) sequences. A kidney region image segmentation model framework was established by combining the U-net network and Mask R-CNN network structures. Using a kidney region image segmentation model framework, the kidney region image is segmented from the bilateral renal magnetic resonance scan data to determine the hydronephrosis-side kidney region and the contralateral kidney region. The method for constructing the kidney region image segmentation model framework includes: A hydronephrosis enhancement module was constructed using a U-net network to enhance the image of the hydronephrosis kidney in bilateral renal MRI scan data. The first hydronephrosis and contralateral kidney segmentation module was constructed using a Mask R-CNN network to segment the hydronephrosis and contralateral kidney regions in enhanced bilateral renal MRI data. A second segmentation module for the hydronephrotic kidney and the contralateral kidney was constructed using a Mask R-CNN network to segment the regions of the hydronephrotic kidney and the contralateral kidney in bilateral renal MRI scan data. The enhancement triggering module of the hydronephrosis enhancement module is adjusted based on the comparison results of the segmentation performance between the first hydronephrosis-side kidney and contralateral kidney segmentation module and the second hydronephrosis-side kidney and contralateral kidney segmentation module.

2. The kidney magnetic resonance image segmentation method according to claim 1, characterized in that: The hydronephrosis-side kidney enhancement module uses the U-net network to segment the hydronephrosis-side kidney region in the bilateral renal magnetic resonance scan data to obtain a hydronephrosis-side kidney mask image, and performs pixel overlay and fusion of the hydronephrosis-side kidney mask image and the bilateral renal magnetic resonance scan data to obtain bilateral renal magnetic resonance scan data with hydronephrosis-side kidney enhancement. ； Where, Enhanced bilateral renal MRI data for the hydronephrosis-affected kidney. This is MRI data of both kidneys. This is a mask image of the hydronephrotic kidney obtained by segmentation using a U-net network. U-net is the U-net network. To enhance the feedback regulation of the trigger module, the enhancement coefficient of the hydronephrosis-side kidney in the enhancement module is adjusted.

3. The kidney magnetic resonance image segmentation method according to claim 2, characterized in that: The first hydronephrotic kidney and contralateral kidney segmentation module is used to segment the hydronephrotic kidney region and the contralateral kidney region in the enhanced bilateral renal magnetic resonance scan data of the hydronephrotic kidney using the Mask R-CNN network, so as to obtain the enhanced mask image of the hydronephrotic kidney and the mask image of the contralateral kidney. The structural expression for the segmentation module of the first hydronephrotic kidney and the contralateral kidney is as follows: ； Where, This is a masked image of the contralateral kidney. The image shows a mask-enhanced image of the kidney on the hydronephrosis side, using the Mask R-CNN network.

4. The kidney magnetic resonance image segmentation method according to claim 3, characterized in that: The second hydronephrotic kidney and contralateral kidney segmentation module is used to segment the hydronephrotic kidney region and the contralateral kidney region in bilateral renal magnetic resonance scan data using the Mask R-CNN network, to obtain the hydronephrotic kidney mask image and the contralateral kidney mask image. The structural expression for the segmentation module of the second hydronephrotic kidney and the contralateral kidney is as follows: ； Where, This is a masked image of the contralateral kidney. The image shows a masked image of the kidney on the side with hydronephrosis. Mask R-CNN is the Mask R-CNN network.

5. The kidney magnetic resonance image segmentation method according to claim 4, characterized in that: The enhancement coefficient of the hydronephrosis-side kidney fed back by the enhancement trigger module consists of two parts: a start / stop parameter that controls the start and stop of the enhancement module for the hydronephrosis-side kidney, and an adaptive intensity parameter that controls the enhancement intensity of the hydronephrosis-side kidney to the hydronephrosis-side kidney, wherein: The enhancement coefficient of the hydronephrotic kidney is: ； In the formula, The enhancement coefficient of the hydronephrotic kidney. These are start / stop parameters. This is an adaptive intensity parameter.

6. The kidney magnetic resonance image segmentation method according to claim 5, characterized in that: The start / stop parameters are: ； In the formula, GT is the actual masked image of the hydronephrotic kidney region. Enhanced image of the hydronephrotic kidney with masking. Image of the hydronephrotic kidney with a mask. For Kullback-Leibler divergence calculation, This is a pixel overlay operation formula.

7. The kidney magnetic resonance image segmentation method according to claim 6, characterized in that: The adaptive intensity parameters are set based on the hydronephrosis-side kidney segmentation performance in the hydronephrosis-side kidney enhancement module. for: ; In the formula, GT is the actual masked image of the hydronephrotic kidney region. The image shows a mask of the hydronephrotic kidney obtained by segmentation using a U-net network.

8. The kidney magnetic resonance image segmentation method according to claim 6, characterized in that: The adaptive intensity parameter is set based on the segmentation performance difference between the first segmentation module for the hydronephrotic kidney and the contralateral kidney and the second segmentation module for the hydronephrotic kidney and the contralateral kidney. for: ; In the formula, GT is the actual masked image of the hydronephrotic kidney region. The image shows a mask of the hydronephrotic kidney obtained by segmentation using the U-net network. This is for Kullback-Leibler divergence calculation.

9. The kidney magnetic resonance image segmentation method according to claim 1, characterized in that: The enhanced mask image of the hydronephrotic kidney and the mask image of the contralateral kidney, output by the segmentation module of the first hydronephrotic kidney and the contralateral kidney, are used as the output of the kidney region image segmentation model framework, and the output of the kidney region image segmentation model framework is used as the kidney region segmentation result of the bilateral kidney magnetic resonance scan data.

Images