A liver parameter quantitative analysis method and system

Abstract

An embodiment of the present specification provides a method for quantitative analysis of liver parameters, the method comprising determining a first segmentation result through a first segmentation model based on a first image; determining a second segmentation result through a second segmentation model based on the first segmentation result and a second image; determining an image to be counted based on the second segmentation result; and determining liver parameters of at least one statistical area based on the image to be counted.

Description

Technical Field

[0001] The present specification relates to the medical field, and in particular to a method and system for quantitative analysis of liver parameters. Background Art

[0002] Quantitative analysis of liver parameters is a non-invasive quantitative imaging method for detecting fat content, iron content, etc. in liver tissue, which provides a means for clinical diagnosis, classification, treatment and efficacy evaluation of fatty liver and other diseases. However, there will be certain deviations when setting regions of interest and performing quantitative statistics on whole liver imaging data including intrahepatic vessels, lesions, volume effects, marginal areas and liver parenchyma.

[0003] Therefore, a better method for quantitative analysis of liver parameters is urgently needed. Summary of the invention

[0004] One of the embodiments of the present specification provides a method for quantitative analysis of liver parameters. The method for quantitative analysis of liver parameters includes: determining a first segmentation result based on a first image by using a first segmentation model; determining a second segmentation result based on the first segmentation result and a second image by using a second segmentation model; determining an image to be counted based on the second segmentation result; and determining a liver fat score of at least one statistical area based on the image to be counted.

[0005] One of the embodiments of the present specification provides a liver fat quantitative analysis system, which includes: a first segmentation module, used to determine a first segmentation result based on a first image through a first segmentation model; a second segmentation module, used to determine a second segmentation result based on the first segmentation result and the second image through a second segmentation model; a statistical image determination module, used to determine an image to be counted based on the second segmentation result; and a liver fat score determination module, used to determine the liver fat score of at least one statistical area based on the image to be counted.

[0006] One of the embodiments of the present specification provides a liver fat quantitative analysis device, comprising at least one processor and at least one memory; the at least one memory is used to store computer instructions; the at least one processor is used to execute at least part of the computer instructions to implement a liver parameter quantitative analysis method.

[0007] One of the embodiments of the present specification provides a computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes a method for quantitative analysis of liver parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0009] Figure 1 is a schematic diagram of an application scenario of a liver parameter quantitative analysis system according to some embodiments of this specification;

[0010] Figure 2 is a block diagram of a liver parameter quantitative analysis system according to some embodiments of the present specification;

[0011] Figure 3 is a schematic diagram of a method for quantitative analysis of liver parameters according to some embodiments of this specification;

[0012] Figure 4 is an exemplary flow chart of a method for quantitative analysis of liver parameters according to some embodiments of this specification;

[0013] Figure 5 is a schematic diagram of a region of interest according to some embodiments of this specification;

[0014] Figure 6 is another schematic diagram of a region of interest according to other embodiments of this specification;

[0015] Figure 7 is a schematic diagram of model training according to some embodiments of this specification;

[0016] Figure 8 is a schematic diagram of a liver segment identification model according to some embodiments of this specification;

[0017] Fig. 9 It is a schematic diagram of the falciform ligament identification model shown in some embodiments of this specification. DETAILED DESCRIPTION

[0018] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0019] It should be understood that the "system", "device", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0020] As shown in this specification and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0021] Flowcharts are used in this specification to illustrate the operations performed by the system according to the embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed precisely in order. Instead, the steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0022] There are two main existing technologies for quantifying liver parameters. One is to manually place a certain number of regions of interest on the magnetic resonance liver parameter image data and perform quantitative statistics. This method is time-consuming and may result in some human errors. The second is to combine deep learning technology to segment the whole liver on the magnetic resonance liver parameter image data and automatically place the regions of interest for quantitative statistics. Currently, neither of these two methods excludes intrahepatic blood vessels, lesions, volume effects, and edge areas, so the quantitative statistical results will be inaccurate.

[0023] The liver parameter quantitative analysis method and system in one or more embodiments of the present specification can be applied to various scenarios, such as physical examination, medical research, drug efficacy evaluation, etc.

[0024] Through the liver parameter quantitative analysis method and system, one or more functions can be achieved, for example, extracting effective liver parenchyma areas, automatically setting regions of interest in effective liver parenchyma areas, etc. The liver parameter quantitative analysis method and system can improve the accuracy of quantitative analysis, reduce or avoid dependence on doctors, and achieve one or more beneficial effects.

[0025] It should be understood that the application scenarios of the liver parameter quantitative analysis method and system of the present application are merely some examples or embodiments of the present application. For ordinary technicians in this field, without paying any creative work, the present application can also be applied to other similar scenarios based on these drawings.

[0026] Figure 1It is a schematic diagram of an application scenario of a liver parameter quantitative analysis system according to some embodiments of this specification.

[0027] like Figure 1 As shown, the application scenario 100 of the liver parameter quantitative analysis system may include an imaging device 110 , a network 120 , a terminal 130 , a processing device 140 , and a storage device 150 .

[0028] The imaging device 110 can be used to scan the target object to obtain scanning data and form an image. The imaging device 110 can be a medical imaging device, for example, a CT (Computed Tomography) imaging device, a PET (Positron Emission Computed Tomography) imaging device, an MRI (Magnetic Resonance Imaging) device, a PET-CT imaging device, a PET-MRI device, etc. In some embodiments, the imaging device 110 can include a gantry 111, a detector 112, a scanning area 113, and a scanning bed 114. The target object can be placed on the scanning bed 114 to be scanned. The gantry 111 can support the detector 112. In some embodiments, the detector 112 can include one or more detector units. The detector unit can include a single-row detector and / or a multi-row detector. The detector unit can include a scintillation detector (for example, a cesium iodide detector) and other detectors. In some embodiments, the gantry 111 can rotate, for example, in a CT imaging device, the gantry 111 can rotate clockwise or counterclockwise around the gantry rotation axis. In some embodiments, the imaging device 110 may further include a radiation scanning source, which may rotate together with the gantry 111. The radiation scanning source may emit a radiation beam (eg, X-ray) toward a target object, and the radiation beam attenuates after passing through the target object and is detected by the detector 112, thereby generating an image signal.

[0029] The processing device 140 can process data and / or information obtained from the imaging device 110, the terminal 130 and / or the storage device 150. For example, the processing device 140 can process the image data detected and generated by the detector 112 to obtain an image. In some embodiments, the processing device 140 can be a single server or a server group. The server group can be centralized or distributed. In some embodiments, the processing device 140 can be local or remote. For example, the processing device 140 can access information and / or data from the imaging device 110, the terminal 130 and / or the storage device 150 through the network 120. For another example, the processing device 140 can directly connect to the imaging device 110, the terminal 130 and / or the storage device 150 to access information and / or data. In some embodiments, the processing device 140 can be implemented on a cloud platform. For example, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, a cross-cloud, a multi-cloud, etc., or a combination of several of them.

[0030] The terminal 130 may include a mobile device 131, a tablet computer 132, a laptop computer 133, etc. or any combination thereof. In some embodiments, the terminal 130 may interact with other components in the application scenario 100 of the liver fat quantitative analysis system through a network. For example, the terminal 130 may send one or more control instructions to the imaging device 110 to control the imaging device 110 to scan the target object according to the instructions. In some embodiments, the terminal 130 may be part of the processing device 140. In some embodiments, the terminal 130 may be integrated with the processing device 140 as an operating console of the imaging device 110. For example, a user / operator (e.g., a doctor) of the application scenario 100 of the liver fat quantitative analysis system may control the operation of the device imaging device 110 through the operating console, such as scanning the target object, etc.

[0031] The storage device 150 may store data (e.g., scan data of a target object, a first image, a second image, etc.), instructions, and / or any other information. In some embodiments, the storage device 150 may store data obtained from the imaging device 110, the terminal 130, and / or the processing device 140. For example, the storage device 150 may store the first image and the second image of the target object obtained from the imaging device 110. In some embodiments, the storage device 150 may store data and / or instructions that the processing device 140 may execute or use to perform the exemplary method described in this specification. In some embodiments, the storage device 150 may include a large-capacity memory, a removable memory, a volatile read-write memory, a read-only memory (ROM), etc., one or a combination of several thereof. In some embodiments, the storage device 150 may be implemented by the cloud platform described in this specification. For example, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, a cross-cloud, a multi-cloud, etc., one or a combination of several thereof.

[0032] In some embodiments, the storage device 150 may be connected to the network 120 to communicate with one or more components (e.g., the processing device 140, the terminal 130, etc.) in the application scenario 100 of the liver parameter quantitative analysis system. One or more components in the application scenario 100 of the liver parameter quantitative analysis system may read data or instructions in the storage device 150 through the network 120. In some embodiments, the storage device 150 may be part of the processing device 140, or may be independent and directly or indirectly connected to the processing device.

[0033] The network 120 may include any suitable network capable of facilitating information and / or data exchange of the application scenario 100 of the liver parameter quantitative analysis system. In some embodiments, one or more components of the application scenario 100 of the liver parameter quantitative analysis system (e.g., the imaging device 110, the terminal 130, the processing device 140, the storage device 150, etc.) may exchange information and / or data with one or more components of the application scenario 100 of the liver parameter quantitative analysis system through the network 120. For example, the processing device 140 may obtain scan data from the imaging device 110 through the network 120. In some embodiments, the network 120 may include one or more network access points. For example, the network 120 may include a wired and / or wireless network access point, such as a base station and / or an Internet exchange point, through which one or more components of the application scenario 100 of the liver fat quantitative analysis system may be connected to the network 120 to exchange data and / or information.

[0034] It should be noted that the application scenario 100 of the liver parameter quantitative analysis system is provided only for the purpose of illustration and is not intended to limit the scope of the present application. For those of ordinary skill in the art, various modifications or variations can be made according to the description of the present application. For example, the application scenario 100 of the liver parameter quantitative analysis system can also include a database. For another example, the application scenario 100 of the liver parameter quantitative analysis system can implement similar or different functions on other devices. However, these changes and modifications will not deviate from the scope of the present application.

[0035] Figure 2 is a block diagram of a liver parameter quantitative analysis system according to some embodiments of the present specification.

[0036] In some embodiments, the liver parameter quantitative analysis system 200 may include a first segmentation module 210, a second segmentation module 220, a statistical image determination module 230, a liver parameter determination module 240, and the like.

[0037] The first segmentation module 210 is used to determine a first segmentation result based on the first image by using a first segmentation model.

[0038] The second segmentation module 220 is used to determine a second segmentation result by using a second segmentation model based on the first segmentation result and the second image.

[0039] The statistical image determination module 230 is used to determine the image to be counted based on the second segmentation result.

[0040] In some embodiments, the statistical image determination module 230 can be used to determine the liver segment area through the liver segment recognition model based on the third image (the third image comes from the first image and / or the second image), and at least one statistical area includes the liver segment area.

[0041] In some embodiments, the statistical image determination module 230 may be configured to determine the falciform ligament of the liver by using a falciform ligament recognition model based on at least a portion of the first image and / or the second image.

[0042] In some embodiments, the statistical image determination module 230 can be used to divide the liver area into a left liver part and a right liver part according to the falciform ligament; select a slice, and determine a first number of left liver parts and / or a second number of right liver parts according to the area ratio of the left liver part to the right liver part; set a first number of regions of interest in the left liver part, and / or set a second number of regions of interest in the right liver part, and the at least one statistical region includes at least one region of interest.

[0043] In some embodiments, the statistical image determination module 230 may be configured to exclude blood vessels from the second segmentation result based on a blood vessel recognition model.

[0044] In some embodiments, the statistical image determination module 230 may be configured to automatically set at least one region of interest in at least one statistical region based on the first image and the size of the region of interest.

[0045] The liver parameter determination module 240 is used to determine the liver parameters of at least one statistical region based on the image to be statistically analyzed.

[0046] It should be noted that the above description of the candidate item display and determination system and its modules is only for convenience of description and does not limit the present specification to the scope of the embodiments. It is understandable that, after understanding the principle of the system, those skilled in the art may arbitrarily combine the modules or form a subsystem to connect with other modules without deviating from the principle. In some embodiments, Figure 2 The first segmentation module 210, the second segmentation module 220, the statistical image determination module 230 and the liver parameter determination module 240 disclosed in the embodiment can be different modules in a system, or one module can realize the functions of two or more modules mentioned above. For example, each module can share a storage module, or each module can have its own storage module. Such variations are all within the protection scope of the present application.

[0047] Figure 3 It is a schematic diagram of a method for quantitative analysis of liver parameters according to some embodiments of the present specification.

[0048] like Figure 3 As shown, the process 300 includes the following steps. In some embodiments, the process 300 may be executed by the processing device 140 .

[0049] Step 310 : Determine a first segmentation result based on the first image by using a first segmentation model. In some embodiments, step 310 may be performed by the first segmentation module 210 .

[0050] The first image refers to an image of the liver region. The first image may be a magnetic resonance (MR) image, a B-ultrasound image, a computed tomography (CT) image, a positron emission tomography (PET) image, a magnetic resonance spectroscopy (MRS) image, etc. In some embodiments, the first image may be acquired and / or reconstructed by the imaging device 110. In some embodiments, the first image may include a fat fraction (FF) map of the liver region, a water map, a fat map, a water-fat in-phase (IP) map, a water-fat out-phase (OP) map, a transverse relaxation rate (R2*) map, etc.

[0051] Preferably, the first image is a three-dimensional image. The image value (e.g., grayscale value) of each voxel in the first image can reflect a certain attribute of the corresponding position. Taking the fat fraction map as an example, the value of each voxel in the three-dimensional fat fraction map can reflect the fat signal intensity, fat signal saturation or fat content of the corresponding position; the value of each pixel in the two-dimensional fat fraction map can reflect the fat signal intensity, fat signal saturation or fat content of the corresponding position.

[0052] The first segmentation refers to a technical means of distinguishing the whole liver area from adjacent organs, adjacent tissues, etc. Through the first segmentation, it can be determined whether each specific voxel in the image belongs to the whole liver area.

[0053] The first segmentation model refers to a machine learning model that implements the first segmentation. The first segmentation model can be VGG, FCN, UNet, VNet, DeepLab, etc. and any combination thereof.

[0054] In some embodiments, the input of the first segmentation model is one or more first images, and the output of the first segmentation model is three-dimensional data that segments the whole liver area from the non-liver area, that is, the first segmentation result.

[0055] In some embodiments, each element value of the first segmentation result represents the probability that the voxel at the corresponding position belongs to the whole liver region. If the element value is large, for example, the element value is 1 or close to 1, the possibility that the voxel at the corresponding position belongs to the whole liver region is relatively large. If the element value is small, for example, the element value is 0 or close to 0, the possibility that the voxel at the corresponding position belongs to the whole liver region is relatively small.

[0056] For the training method of the first segmentation model, see Figure 7 and its description.

[0057] By implementing the first segmentation to achieve preliminary segmentation of the first image and eliminating the non-liver area, interference from the non-liver area can be avoided, the accuracy of the subsequent segmentation process can be improved, and the workload of sample labeling in the second segmentation model training process can be reduced.

[0058] Step 320 : Determine a second segmentation result by using a second segmentation model based on the first segmentation result and the second image. In some embodiments, step 320 may be performed by the second segmentation module 220 .

[0059] The second image refers to an image of the liver region. The second image may be a magnetic resonance (MR) image, a B-ultrasound image, a computed tomography (CT) image, a positron emission tomography (PET) image, a magnetic resonance spectroscopy (MRS) image, etc. In some embodiments, the second image may be acquired and / or reconstructed by the imaging device 110. In some embodiments, the second image may include a fat fraction (FF) map of the liver region, a water map, a fat map, a water-fat in-phase (IP) map, a water-fat out-phase (OP) map, a transverse relaxation rate (R2*) map, etc.

[0060] Preferably, the second image is a three-dimensional image. The second image may be the same as or different from the first image. For example, in some embodiments, the first image and the second image are both fat fraction maps of the liver region. For another example, in some embodiments, the first image is a fat fraction map of the liver region, and the second image is a water phase map, a lipid phase map, a water-lipid isophase map, a water-lipid antiphase map, a transverse relaxation rate map, etc. of the liver region of the same target object (e.g., a physical examination guest, a patient, a volunteer).

[0061] The second segmentation refers to a further segmentation of the whole liver area based on the above-mentioned first segmentation result. Through the second segmentation, invalid areas in the whole liver area, such as lesion areas, volume effect areas, etc., can be excluded. In some embodiments, the lesion area includes a cyst area, a tumor area, a sclerosis area, an intrahepatic bile duct stone area, a fibrosis area, and an inflammation area. The volume effect area refers to an area composed of voxels that do not completely belong to the whole liver area in the image. It can be understood that the voxels in the first image or the second image correspond to the volume of the corresponding position of the target object, and the volume may belong to the liver in part and the other part to the adjacent organs or tissues. For example, the volume corresponding to the voxels at the junction of the liver and the diaphragm in the first image or the second image may belong to the liver in part, and the other part to the diaphragm.

[0062] The second segmentation model refers to a machine learning model for implementing the second segmentation. The second segmentation model can be VGG, FCN, UNet, VNet, DeepLab, etc. or any combination thereof.

[0063] In some embodiments, the input of the second segmentation model may include the first segmentation result and at least one second image.

[0064] In some embodiments, at least one second image may be segmented based on the first segmentation result, and the segmented at least one second image may be used as an input to the second segmentation model. The segmentation process may adjust the voxel values ​​of voxels that do not belong to the whole liver region in the at least one second image, for example, setting all the voxel values ​​to 0.

[0065] In some embodiments, the output of the second segmentation model may be an image after excluding invalid areas (e.g., lesion areas and volume effect areas). The invalid areas may be excluded by adjusting the grayscale value of the voxels. For example, the voxel values ​​of the invalid areas in the fat fraction map are all set to 255.

[0066] For the training method of the second segmentation model, see Figure 7 and its description.

[0067] Step 330: determine the image to be counted based on the second segmentation result.

[0068] In some embodiments, the statistical image determination module 230 may slice the second segmentation result to obtain an image to be counted. The slice may be a cross-sectional slice, a coronal slice, a sagittal slice, etc. For example, a plurality of two-dimensional cross-sectional slices of the three-dimensional second segmentation result may be obtained, and at least a portion of the plurality of two-dimensional cross-sectional slices may be used as the image to be counted.

[0069] In some embodiments, the statistical image determination module 230 may divide the second segmentation result into liver segment regions, and use at least one region of the liver segment region as the image to be statistically analyzed. For liver segments and methods of dividing into liver segment regions, see the following description.

[0070] In some embodiments, before determining the image to be counted based on the second segmentation result, the statistical image determination module 230 may perform operations such as excluding a blood vessel region and an edge region in the second segmentation result.

[0071] The vascular region may include the right hepatic vein, middle hepatic vein, left hepatic vein, umbilical vein, inferior vena cava, hepatic artery, portal vein, hepatic duct, common bile duct and their branches at all levels.

[0072] In some embodiments, the statistical image determination module 230 may exclude the vascular region in the second segmentation result through a vascular recognition model. The vascular recognition model refers to a machine learning model for realizing intrahepatic vascular recognition. The vascular recognition model may be VGG, FCN, UNet, VNet, DeepLab, etc. or any combination thereof. In some embodiments, the input of the vascular recognition model may be the second segmentation result, and the output of the vascular recognition model may be the vascular region annotated in the second segmentation result, for example, all voxel values ​​of the vascular region identified in the second segmentation result are set to 255.

[0073] In some embodiments, the statistical image determination module 230 may exclude the vessel region in the second segmentation result by multiscale vessel enhancement filtering.

[0074] In some embodiments, the statistical image determination module 230 may exclude the edge region in the second segmentation result in a variety of ways. For example, by using an eroded edge image method. For another example, the edge region in the second segmentation result may be excluded by removing peripheral voxels. For example, the value of the voxel in the outermost layer (or the outermost layers) in the second segmentation result is set to 255.

[0075] Step 340: determine liver parameters of at least one statistical region based on the image to be statistically analyzed.

[0076] Liver parameters refer to parameters used to evaluate the degree of liver lesions. Liver parameters may include liver fat fraction, iron content parameters, perfusion parameters (eg, blood flow, blood flow velocity, etc.), diffusion parameters (eg, diffusion coefficient, kurtosis coefficient), etc.

[0077] In some embodiments, the liver parameter determination module 240 may determine the liver parameters in a variety of ways.

[0078] For example, in some embodiments, taking liver fat score as an example, the liver parameter determination module 240 can statistically calculate at least one statistical area separately to obtain at least one liver fat score, and summarize the statistics of at least one liver fat score (for example, taking the arithmetic mean, weighted mean, etc.), and use the summary result as the liver fat score.

[0079] For another example, in some embodiments, again taking the liver fat fraction as an example, the liver parameter determination module 240 can directly calculate the liver fat fraction based on the fat fraction map. As mentioned above, the value of each pixel (or voxel) in the fat fraction map can reflect the fat signal intensity, fat signal saturation or fat content at the corresponding position. The liver parameter determination module 240 can calculate the average value of each pixel (or voxel) in the fat fraction map after the second segmentation, and use the average value as the liver fat fraction.

[0080] In some embodiments, the statistical image determination module 230 may determine at least one statistical region by various methods.

[0081] In some embodiments, the statistical image determination module 230 may determine a liver segment region through a liver segment recognition model based on the third image, and determine at least one statistical region based on the liver segment region.

[0082] In some embodiments, the third image may be the first image and / or the second image. In some embodiments, the third image may be the processed first image and / or the second image. For example, the third image may be the first image and / or the second image after sharpening, and the sharpening may increase the contrast between the vascular walls of the portal vein, the hepatic artery, the hepatic vein, the hepatic bile duct and its bifurcations and the surrounding voxels, which is conducive to more accurate liver segmentation.

[0083] Liver segments can reflect the structural characteristics of the liver's internal duct system, and each segment in the liver can be regarded as a functional anatomical unit of the liver. The liver segment area includes: caudate lobe (segment I), upper segment of left lateral lobe (segment II), lower segment of left lateral lobe (segment III), left internal lobe (segment IV), upper segment of right anterior lobe (segment V), upper segment of right posterior lobe (segment VI), lower segment of right posterior lobe (segment VII), and lower segment of right anterior lobe (segment VIII).

[0084] The liver segment recognition model refers to a machine learning model for realizing liver segment recognition. The liver segment recognition model can be VGG, FCN, UNet, VNet, DeepLab, etc. or any combination thereof. In some embodiments, Figure 8 As shown, the input of the liver segment recognition model is the third image, and the output of the liver segment recognition model is the segmented liver segment region, for example, the contour of each segment of the circled liver segment in the third image.

[0085] For the training method of the liver segment recognition model, see Figure 7 and its description.

[0086] In some embodiments, the statistical image determination module 230 may set at least one statistical region in the liver segment region, for example, evenly distribute the statistical regions in the liver segment region.

[0087] In some embodiments, the statistical image determination module 230 may determine at least one statistical region by other means. For example, the whole liver region is divided into two parts based on the falciform ligament of the liver, and at least one statistical region is set in each of the two parts. Figure 4 and its description.

[0088] Figure 4 is an exemplary flow chart of a method for quantitative analysis of liver parameters according to some embodiments of the present specification.

[0089] Step 410: determine the falciform ligament of the liver using a falciform ligament recognition model based on at least a portion of the first image and / or the second image.

[0090] The falciform ligament is a longitudinal ligament on the diaphragmatic surface of the liver, and its side view resembles a sickle.

[0091] The falciform ligament recognition model refers to a machine learning model for realizing falciform ligament recognition. The falciform ligament recognition model can be VGG, FCN, UNet, VNet, DeepLab, etc. or any combination thereof. In some embodiments, Fig. 9 As shown, the input of the falciform ligament recognition model is at least a part of the first image and / or the second image, and the output of the falciform ligament recognition model is the marked falciform ligament in at least a part of the first image and / or the second image, for example, the outline of the circled falciform ligament, setting the grayscale value of all voxels of the falciform ligament to 255, etc.

[0092] For the training method of the falciform ligament recognition model, see Figure 7 and its description.

[0093] Step 420 , dividing the liver region into a left liver portion and a right liver portion according to the falciform ligament.

[0094] In some embodiments, the statistical image determination module 230 may divide the liver region on the left side of the falciform ligament in the second segmentation result into a left liver portion, and divide the liver region on the right side of the falciform ligament into a right liver portion.

[0095] Step 430, selecting a slice with the largest liver area, and determining a first number of left liver parts and / or a second number of right liver parts according to an area ratio of the left liver part to the right liver part.

[0096] In some embodiments, the statistical image determination module 230 may select a slice from the second segmentation result. For example, the slice with the largest liver area in the human body cross-sectional slice of the second segmentation result is selected. For another example, the middle slice of the second segmentation result is selected.

[0097] In some embodiments, the statistical image determination module 230 may calculate the area ratio of the left liver portion to the right liver portion in the slice, for example, the area ratio is 1:3.

[0098] In some embodiments, the statistical image determination module 230 can determine the first number of left liver parts and the second number of right liver parts according to the total number of regions of interest and the above-mentioned area ratio. The total number of regions of interest can be obtained by user definition, based on experience setting, according to system parameter setting (for example, resolution, liver area, area of ​​the region of interest), etc. For example, the user sets the total number of regions of interest to 20 and the left and right liver area ratio to 1:3, then the first number is 5 and the second number is 15.

[0099] Step 440: setting a first number of regions of interest in the left liver part, and / or setting a second number of regions of interest in the right liver part, wherein the at least one statistical region includes at least one region of interest.

[0100] In some embodiments, the statistical image determination module 230 may divide the left liver part and / or the right liver part into several sub-regions by various methods, for example, by dividing by area, by angle, by experience, and the like.

[0101] In some embodiments, the statistical image determination module 230 may divide the left liver part in the slice into a first number of sub-regions by area, and divide the right liver part into a second number of sub-regions by area. For example, the left liver part is divided into 5 sub-regions by area, and the right liver part is divided into 15 sub-regions by area.

[0102] In some embodiments, the statistical image determination module 230 may divide the left liver portion in the slice into a first number of sub-regions according to an angle, and divide the right liver portion into a second number of sub-regions according to an angle. Figure 5 As shown, the right liver part is divided into five sub-regions (sub-region 521, sub-region 523, sub-region 525, sub-region 527, and sub-region 529) in equal angles.

[0103] The shape of the region of interest may be a sector, a hexagon, a rectangle, etc. The size of the region of interest may be determined based on factors such as the total liver area, the left liver area, the right liver area, the total number of regions of interest, the shape of the left liver, the shape of the right liver, etc.

[0104] In some embodiments, the statistical image determination module 230 can automatically set at least one region of interest in at least one statistical region based on the first image and the size of the region of interest. For example, the statistical image determination module 230 can set the region of interest at the center of each sub-region. For another example, the region of interest in the sub-region with the smallest variance of image values ​​(e.g., grayscale values) among all feasible regions of interest is used as the region of interest in the sub-region.

[0105] In some embodiments, the statistical image determination module 230 can set multiple regions of interest in multiple slices, for example, multiple regions of interest are set in the slice with the largest liver area in the human cross-sectional slice of the second segmentation result and in the human cross-sectional slice at the golden section height of the liver in the second segmentation result.

[0106] In some embodiments, the statistical image determination module 230 can calculate liver parameters, such as liver fat scores, for each region of interest, and display them to the user (e.g., a doctor). For example, each region of interest is highlighted in the second segmentation result, and the corresponding liver fat score is displayed on each region of interest.

[0107] Since liver tissue is large and fat and iron are unevenly distributed in the liver, setting up multiple regions of interest and measuring liver parameters (such as fat fraction, iron content, etc.) in multiple regions of interest respectively is conducive to accurate analysis of various parts of the liver.

[0108] In some embodiments, the statistical image determination module 230 may set multiple regions of interest in other ways, for example, setting a region of interest only in the right liver, or only in the left liver.

[0109] In some embodiments, the statistical image determination module 230 may use at least one region of interest as at least one statistical region.

[0110] Setting the statistical region in the second segmentation result can avoid the statistical region from overlapping with an invalid region (eg, a lesion region, a volume effect region, etc.), which is beneficial to improving the accuracy of quantitative analysis of liver fat.

[0111] It should be noted that the above description of process 400 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 400 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0112] Figure 5 is a schematic diagram of a region of interest according to some embodiments of the present specification.

[0113] like Figure 5 As shown, taking the case where the total number of regions of interest is 5 as an example, the liver parameter determination module 240 determines 5 sub-regions (sub-region 521, sub-region 523, sub-region 525, sub-region 527, and sub-region 529) in the slice 510 (e.g., the slice with the largest liver area). Then, the liver parameter determination module 240 sets a diamond-shaped region of interest (region of interest 531, region of interest 533, region of interest 535, region of interest 537, and region of interest 539) in each sub-region.

[0114] Figure 6 is another schematic diagram of a region of interest according to other embodiments of the present specification.

[0115] like Figure 6As shown, again taking the case where the total number of regions of interest is 5 as an example, the liver parameter determination module 240 determines 4 sub-regions (sub-region 631, sub-region 633, sub-region 635, sub-region 637) in slice 610 (e.g., the slice with the largest liver area), and determines 1 sub-region (sub-region 641) in slice 620 (e.g., the slice at the golden section height). Next, the liver parameter determination module 240 sets a diamond-shaped region of interest (region of interest 651, region of interest 653, region of interest 655, region of interest 657) in each sub-region of slice 610, and sets a diamond-shaped region of interest (region of interest 661) in sub-region 641 of slice 620.

[0116] Figure 7 It is a schematic diagram of model training according to some embodiments of this specification.

[0117] like Figure 7 As shown, in some embodiments, the initial model 710 can be trained based on a large number of training samples with labels to update the parameters of the initial model to obtain the trained model 720. The initial model 710 can include an initial first segmentation model, an initial second segmentation model, an initial liver segment recognition model, an initial falciform ligament recognition model, and an initial blood vessel recognition model, and the trained model 720 can include a first segmentation model, a second segmentation model, a liver segment recognition model, a falciform ligament recognition model, and a blood vessel recognition model.

[0118] In some embodiments, a training set may be randomly selected from the training samples, and data augmentation may be performed on the training set; the initial model 710 may be iteratively trained several times on the training set to obtain a trained model 720 .

[0119] In some embodiments, the iterative training method may include: calculating the gradient of the loss function, and iteratively updating the parameters of the initial model 710 by the gradient descent method. The loss function may include a cross entropy loss function, a mean loss function, an exponential loss function, a logarithmic loss function, a square loss function, etc. The gradient descent method may include a standard gradient descent method and a stochastic gradient descent method, etc.

[0120] In some embodiments, a variety of learning rate decay strategies can be used in iterative training, such as piecewise decay, reverse time decay, exponential decay, and adaptive decay.

[0121] In some embodiments, when the loss function converges or is less than a preset threshold, the iterative training can be terminated. In some embodiments, when the number of iterations reaches a preset number of iterations, the iterative training can be terminated.

[0122] In some embodiments, the training sample of the first segmentation model may be an image of a liver region, for example, a fat fraction map of the liver region. The sample identifier may be a gold standard annotation of the liver.

[0123] In some embodiments, the training samples of the second segmentation model may include the first segmentation result and an image of the liver region. The sample identification may be a gold standard annotation that excludes lesions and volume effect regions.

[0124] In some embodiments, the training sample of the second segmentation model may be an image of a liver region that has been segmented based on the first segmentation result. The sample identifier may be a gold standard annotation that excludes lesions and volume effect regions.

[0125] In some embodiments, the training samples of the liver segment recognition model may include images of the liver region or processed images of the liver region. The sample identification may be a gold standard annotation of the liver segment region.

[0126] In some embodiments, the training sample of the falciform ligament recognition model may be image data of the liver region, and the sample identification may be a gold standard annotation of the falciform ligament of the liver.

[0127] In some embodiments, the training sample of the falciform ligament recognition model may be the second segmentation result of the image data of the liver region, and the sample identifier may be the gold standard label of the falciform ligament of the liver in the second segmentation result.

[0128] In some embodiments, the training sample of the blood vessel recognition model may be an image of the liver region or a second segmentation result. For example, the training sample may be the transverse relaxation rate (R2*) of the liver region. The sample identifier may be a gold standard label of the intrahepatic blood vessels.

[0129] In some embodiments, the first segmentation model, the second segmentation model, the liver segment recognition model, the falciform ligament recognition model, and the blood vessel recognition model can be trained separately or jointly. For example, the first segmentation model and the second segmentation model can be jointly trained. For another example, the first segmentation model, the second segmentation model, and the liver segment recognition model can be jointly trained. For another example, the first segmentation model, the second segmentation model, and the blood vessel recognition model can be jointly trained.

[0130] In some embodiments, the training samples for the joint training of the first segmentation model and the second segmentation model may be images of the liver region. The sample identifier may be a gold standard annotation of the whole liver region excluding lesions and volume effect regions.

[0131] In some embodiments, the training samples for the joint training of the first segmentation model, the second segmentation model, and the liver segment recognition model may be images of the liver region, and the sample identifiers may be gold standard annotations in the liver segment that exclude lesions and volume effect regions.

[0132] In some embodiments, the training samples for the joint training of the first segmentation model, the second segmentation model, and the vascular recognition model may be images of the liver region. The sample identification may be a gold standard annotation of the entire liver region excluding lesions, volume effect regions, and vascular regions.

[0133] It should be noted that the above description of process 700 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 700 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0134] The beneficial effects that may be brought about by some embodiments of the present application include but are not limited to: (1) being able to exclude lesion areas, volume effect areas, vascular areas, and edge areas, making the quantitative analysis of liver parameters more accurate; (2) being able to intuitively display the quantitative liver parameters of each area of ​​interest in the liver segment area and / or cross-sectional slices; (3) combining the use of multiple image segmentation methods to make the image segmentation of the liver area more accurate. It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or any other possible beneficial effects.

[0135] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements and corrections to this specification. Such modifications, improvements and corrections are suggested in this specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

[0136] At the same time, this specification uses specific words to describe the embodiments of this specification. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures or characteristics in one or more embodiments of this specification can be appropriately combined.

[0137] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this specification, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0138] Similarly, it should be noted that in order to simplify the description disclosed in this specification and thus help understand one or more embodiments of the invention, in the above description of the embodiments of this specification, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the subject matter of this specification are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0139] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0140] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, documents, etc., cited in this specification are hereby incorporated by reference in their entirety. Except for application history documents that are inconsistent with or conflicting with the contents of this specification, documents that limit the broadest scope of the claims of this specification (currently or later attached to this specification) are also excluded. It should be noted that if the descriptions, definitions, and / or use of terms in the materials attached to this specification are inconsistent or conflicting with the contents described in this specification, the descriptions, definitions, and / or use of terms in this specification shall prevail.

[0141] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. A method for quantitative analysis of liver parameters, the method comprising: Based on the first image, determining a first segmentation result by using a first segmentation model; Based on the first segmentation result and the second image, determining a second segmentation result by a second segmentation model; Determine an image to be counted based on the second segmentation result; Determine at least one statistical area based on the image to be counted; Based on the first image and the size of the region of interest, at least one region of interest is automatically set in the at least one statistical region, and liver parameters of the region of interest are calculated, where the liver parameters are parameters used to evaluate the degree of liver lesions.

2. The method according to claim 1, further comprising: Based on a third image, a liver segment region is determined by a liver segment recognition model, wherein the third image is from the first image and / or the second image, and the at least one statistical region includes a liver segment region.

3. The method according to claim 1, further comprising: Based on at least a portion of the first image and / or the second image, the falciform ligament of the liver is determined by a falciform ligament recognition model.

4. The method according to claim 3, further comprising: Dividing the liver region into a left liver portion and a right liver portion according to the falciform ligament; Selecting a slice, and determining a first number of left liver parts and / or a second number of right liver parts according to an area ratio of the left liver part to the right liver part; A first number of regions of interest are set in the left liver part, and / or a second number of regions of interest are set in the right liver part, and the at least one statistical region includes at least one region of interest.

5. The method according to claim 1, wherein determining the image to be counted based on the second segmentation result comprises: The blood vessels in the second segmentation result are excluded based on the blood vessel recognition model.

6. A liver parameter quantitative analysis system, characterized in that: The system comprises: A first segmentation module, configured to determine a first segmentation result based on the first image by using a first segmentation model; A second segmentation module, configured to determine a second segmentation result by a second segmentation model based on the first segmentation result and the second image; A statistical image determination module, configured to determine an image to be counted based on the second segmentation result; and A liver parameter determination module is used to determine at least one statistical area based on the image to be counted, automatically set at least one region of interest in the at least one statistical area based on the first image and the size of the region of interest, and calculate the liver parameters of the region of interest, wherein the liver parameters are parameters used to evaluate the degree of liver lesions.

7. The system according to claim 6, characterized in that: The statistical image determination module is further used to determine the liver segment area based on a third image by a liver segment recognition model, wherein the third image comes from the first image and / or the second image, and the at least one statistical area includes the liver segment area.

8. A liver parameter quantitative analysis device, characterized in that: The apparatus comprises at least one processor and at least one memory; The at least one memory is used to store computer instructions; The at least one processor is configured to execute at least part of the computer instructions to implement the method according to any one of claims 1 to 5.

9. A computer storable medium, characterized in that: The storage medium stores computer instructions, and when at least part of the computer instructions are executed by a processor, the method described in any one of claims 1 to 5 is implemented.

Images