Water-fat content inverse correction method and system

Abstract

The embodiment of the invention provides a water-fat inverse correction method and system. The method comprises the following steps: acquiring scanning data of a target object; generating an initial first image and an initial second image based on the scanning data; judging whether water-fat inverse artifacts exist in the initial first image and the initial second image or not; if yes, performing at least one round of iterative optimization based on the initial first image and the initial second image to obtain a target first image and a target second image; wherein each round of iteration comprises the steps of adding a first mark in the first image and adding a second mark in the second image; and performing image optimization based on the first mark and the second mark.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a method and system for correcting water-fat separation. Background Art

[0002] Water Fat Imaging is one of the commonly used fat suppression sequences in clinical magnetic resonance imaging. In water fat imaging, water and fat are stimulated and imaged at the same time. The reconstruction algorithm separates the water signal and the fat signal based on the chemical shift of water and fat, and reconstructs the water map and fat map respectively. Water fat imaging can be applied to all parts of the body and challenging scenarios such as uneven main magnetic field, and has important application value in clinical practice. However, due to the natural ambiguity of water and fat in water fat imaging, water fat separation has always been one of the most common artifacts faced by water fat imaging.

[0003] The embodiments of this specification provide a water-fat separation correction method and system to obtain a high-quality water-fat image. Summary of the invention

[0004] One or more embodiments of the present specification provide a water-fat separation correction method, which includes: acquiring scan data of a target object; generating an initial first image and an initial second image based on the scan data; determining whether there are water-fat separation artifacts in the initial first image and the initial second image; if so, performing at least one round of iterative optimization based on the initial first image and the initial second image to obtain a target first image and a target second image; wherein each round of iteration includes: adding a first mark in the first image and adding a second mark in the second image; and performing image optimization based on the first mark and the second mark.

[0005] In some embodiments, the first tag and the second tag are added by receiving user input.

[0006] In some embodiments, the image optimization based on the first mark and the second mark includes: creating an optimization queue based on the image elements corresponding to the first mark or the second mark; obtaining the neighborhood elements of the top element of the optimization queue multiple times according to the first-in-first-out rule to determine multiple neighborhood element sets; performing optimization calculations on each top element and the elements in its corresponding neighborhood element set, and determining the target phase of the top element based on the results of the optimization calculations.

[0007] In some embodiments, the optimization calculation of each of the top elements and the elements in its corresponding neighborhood element set includes: obtaining the phase of the image element corresponding to the first mark or the second mark; determining the candidate phase of each top element and the phase of the calculated element in its corresponding neighborhood element set; and performing optimization calculation based on the candidate phase corresponding to each of the top elements and the phase of the calculated element in its corresponding neighborhood element set.

[0008] In some embodiments, the optimization calculation based on the candidate phase corresponding to each of the top elements and the phases of the elements in its corresponding neighborhood element set includes: for each top element, calculating the difference between the candidate phase corresponding to the top element and the average / Gaussian weighted average of the phases of the calculated elements in its corresponding neighborhood element set; assigning a target candidate phase to the top element so that it has a minimum difference with the average / Gaussian weighted average of the calculated elements in the neighborhood element set; adding the assigned target candidate phase to the element phase set corresponding to the top element; and re-adding the top element to the priority queue.

[0009] In some embodiments, the method also includes: before performing the optimization calculation, determining whether the number of calculations of the top element reaches a preset number; if so, using the target candidate phase corresponding to the majority of the element phase set corresponding to the top element as the target phase of the top element; if not, performing the optimization calculation based on the candidate phase corresponding to each of the top elements and the phases of the elements in its corresponding neighborhood element set.

[0010] In some embodiments, the performing image optimization based on the first mark and the second mark includes: performing the image optimization by using a preset image optimization algorithm.

[0011] In some embodiments, the preset image optimization algorithm includes a water-fat separation algorithm based on region growing or a water-fat separation algorithm based on deep learning.

[0012] One or more embodiments of the present specification provide a water-fat separation correction system, which includes: a scanning data acquisition module for acquiring scanning data of a target object; an image generation module for generating an initial first image and an initial second image based on the scanning data; a judgment module for judging whether there are water-fat separation artifacts in the initial first image and the initial second image; an iterative optimization module for performing at least one round of iterative optimization based on the initial first image and the initial second image to obtain a target image when there are water-fat separation artifacts in the initial first image and the initial second image; wherein each round of iteration includes: adding a first mark in the first image and adding a second mark in the second image; and performing image optimization based on the first mark and the second mark.

[0013] One or more embodiments of the present specification provide a water-fat separation correction device, including a processor, wherein the processor is used to execute the water-fat separation correction method.

[0014] One or more embodiments of the present specification provide 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 the above-mentioned water-fat separation correction method.

[0015] In some embodiments of the present specification, when water-fat separation artifacts appear in the scanned image, by marking some water-fat separation pixels, the water-fat separation results in the entire image are optimized by using the water-fat separation algorithm and the marked water-fat pixels. This correction method is a progressive iterative optimization. If a single correction cannot completely eliminate the water-fat separation artifacts, the next marking and correction optimization can be performed based on the results of each iteration. Through multiple corrections, a better water-fat separation correction effect is obtained, effectively improving the image quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] 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:

[0017] Figure 1 is an exemplary schematic diagram of an application scenario of the water-fat separation correction system shown in some embodiments of this specification;

[0018] Figure 2 is an exemplary flow chart of a water-fat separation correction method according to some embodiments of the present specification;

[0019] Figure 3 is an exemplary flow chart of image optimization according to some embodiments of this specification;

[0020] Figure 4 is an exemplary flow chart for performing optimization calculations according to some embodiments of this specification;

[0021] Figure 5 is an exemplary flow chart of a method for determining a target phase of a top element according to some embodiments of the present specification;

[0022] Figure 6 is an exemplary module diagram of a water-fat separation correction system according to some embodiments of this specification;

[0023] Figure 7 is an exemplary schematic diagram of a water image according to some embodiments of the present specification;

[0024] Figure 8 is an exemplary schematic diagram of a fat image according to some embodiments of the present specification;

[0025] Fig. 9 It is an exemplary schematic diagram of another interactive water-lipid classification back-calibration algorithm according to some embodiments of this specification. DETAILED DESCRIPTION

[0026] 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.

[0027] 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.

[0028] 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.

[0029] 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.

[0030] In magnetic resonance imaging, the appearance of water-fat swap artifacts will seriously affect the quality of the image and thus affect the diagnosis. Water fat swap artifact is a relatively common artifact in magnetic resonance water-fat imaging. It refers to the misidentification caused by the swapping of water and fat signals due to factors such as local field deviation during the imaging process. This artifact usually appears in areas such as near the air-soft tissue interface or the junction of bone and soft tissue, where uneven local field intensity may cause the water and fat signals to swap. In this case, water signals appear in areas that should be displayed as fat on the image, and vice versa, making it easy for doctors to make misjudgments when making diagnoses. Currently, when water fat swap artifacts appear, users can only rescan or select other alternative sequences, and there is a lack of effective post-processing-based solutions.

[0031] Both traditional algorithms and deep learning algorithms are committed to improving the accuracy of water-fat separation to reduce the impact of water-fat separation artifacts. Water-fat separation algorithms are usually based on the assumption of phase smoothness. In the phase smoothness assumption, the phase change caused by factors such as B0 field inhomogeneity should be spatially slow. Therefore, the uncertainty in the water-fat separation algorithm can be solved by selecting a water-fat combination that is more consistent with the phase smoothness assumption. For example, for the water-fat separation algorithm based on region growing, in the acquisition of a dual echo sequence with the same and opposite phases, one echo is in the same direction as water and fat, and the other echo is in the opposite direction as water and fat. Through the signal model of the dual echo signal and water-fat imaging, two candidate amplitudes for each pixel can be solved (generally a solution with a larger amplitude and a solution with a smaller amplitude). If the true component of the pixel is mainly water, the amplitude of the pixel in the water map corresponds to the large solution, and the amplitude of the pixel in the fat map corresponds to the small solution; conversely, if the true component of the pixel is mainly fat, the amplitude of the pixel in the water map corresponds to the small solution, and the amplitude of the pixel in the fat map corresponds to the large solution.

[0032] The goal of the water-fat separation algorithm is to distinguish whether each pixel should be dominated by water or fat through the phase smoothness assumption. The water-fat phase difference corresponding to each pixel can be further solved through the large-small solution, that is, the phase difference when the pixel is dominated by water and the phase difference when the pixel is dominated by fat. Ideally, when the phase difference of each pixel is assigned to the correct phase difference, its phase map will tend to be smooth, which better meets the phase smoothness assumption.

[0033] In view of this, some embodiments of the present specification provide a water-fat separation correction method, which can quickly help users correct water-fat separation artifacts when water-fat separation artifacts appear, ensuring that users obtain high-quality diagnostic-level water-fat images in a timely manner.

[0034] Figure 1It is an exemplary schematic diagram of an application scenario of the water-lipid separation correction system shown in some embodiments of this specification.

[0035] like Figure 1 As shown, the application scenario 100 of the water-fat separation correction system may include an imaging device 110 , a processing device 120 , a terminal 130 , a storage device 140 and a network 150 .

[0036] The imaging device 110 is used to scan the target object to collect scanning data related to the target object. The scanning data is used to generate one or more medical images of the target object. In some embodiments, the imaging device 110 can be single-modal or multi-modal. In some embodiments, the single-modal device can be a magnetic resonance imaging device, and the multi-modal device can include a positron emission tomography-magnetic resonance imaging (PET-CT) scanner, a computer tomography-magnetic resonance imaging (CT-MRI) scanner, etc. The target object can be biological or non-biological. As an example only, the target object includes a patient, an artificial object (such as an artificial phantom), etc. For another example, the target object also includes a specific part, organ and / or tissue of a patient. In some embodiments, the magnetic resonance imaging device includes a magnet assembly and a coil assembly. The coil assembly includes a gradient coil assembly and a radio frequency (RF) coil assembly. In some embodiments, the coil assembly includes a plurality of coil channels, each of which can be used to obtain corresponding scanning data and used for imaging.

[0037] The processing device 120 is used to process data and / or information obtained from the imaging device 110, the terminal 130 and / or the storage device 140. For example, the processing device 120 can be used to obtain an initial first image and an initial second image; perform at least one round of iterative optimization based on the initial first image and the initial second image to obtain a target first image and a target second image; wherein each round of iteration includes: adding a first tag in the first image and adding a second tag in the second image; performing image optimization based on the first tag and the second tag. In some embodiments, the processing device 120 is a single server or a server group. The server group can be centralized or distributed. In some embodiments, the processing device 120 can be local or remote. For example, the processing device 120 can access information and / or data from the imaging device 110, the terminal 130 and / or the storage device 140 through the network 150. For another example, the processing device 120 can be directly connected to the imaging device 110, the terminal 130 and / or the storage device 140 to access information and / or data. In some embodiments, the processing device 120 is implemented on a cloud platform. For example, the cloud platform includes private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, cloud-to-cloud, multi-cloud, etc. or any combination thereof. In some embodiments, the processing device 120 is part of the imaging device 110. For example, the processing device 120 is integrated in the imaging device 110.

[0038] The terminal 130 includes a mobile device 131, a tablet computer 132, a laptop computer 133, etc. or any combination thereof. In some embodiments, the terminal 130 interacts with other components in the scene 100 through the network 150. For example, the terminal 130 can send one or more control instructions to the imaging device 110 through the network 150 to control the imaging device 110 to scan the target object according to the instructions. In some embodiments, the mobile device 131 includes a smart home device, a wearable device, a mobile device, a virtual reality device, an augmented reality device, etc. or any combination thereof.

[0039] In some embodiments, the terminal 130 is a part of the processing device 120. In some embodiments, the terminal 130 is integrated with the processing device 120 as an operation console of the imaging device 110. For example, a user / operator (e.g., a doctor or a nurse) can control the operation of the imaging device 110 through the operation console, for example, scanning a target object, acquiring an initial first image and an initial second image, etc.

[0040] The storage device 140 is used to store data (e.g., the first initial image, the second initial image, the target first image, the target second image, and the intermediate image in the image optimization process, etc.), instructions and / or any other information. In some embodiments, the storage device 140 stores data obtained from the imaging device 110, the processing device 120 and / or the terminal 130. For example, the storage device 140 stores multiple channel images corresponding to multiple coil channels obtained from the imaging device 110, multiple filtered channel images, channel merged images, etc. In some embodiments, the storage device 140 stores data and / or instructions that can be executed or used by the processing device 120 to perform the exemplary methods described in this specification.

[0041] In some embodiments, the storage device 140 may include one or a combination of mass storage, removable storage, volatile read-write memory, read-only memory (ROM), etc. In some embodiments, the storage device 140 may be implemented by the cloud platform described in this specification.

[0042] In some embodiments, the storage device 140 can be connected to the network 150 to communicate with one or more components in the scene 100 (e.g., the processing device 120, the terminal 130). One or more components in the scene 100 can read data or instructions in the storage device 140 through the network 150. In some embodiments, the storage device 140 can be part of the processing device 120, or it can be independent and directly or indirectly connected to the processing device 120.

[0043] The network 150 includes any suitable network capable of facilitating information and / or data exchange of the scene 100. In some embodiments, one or more components of the scene 100 (e.g., the imaging device 110, the processing device 120, the terminal 130, the storage device 140) can exchange information and / or data with one or more components of the scene 100 via the network 150. For example, the processing device 120 obtains the initial first image and the initial second image from the imaging device 110 via the network 150. In some embodiments, the network 150 may include one or more network access points. For example, the network 150 may include wired and / or wireless network access points.

[0044] It should be noted that the above description of the application scenario 100 is provided for illustrative purposes only and is not intended to limit the scope of the present embodiment. For those of ordinary skill in the art, various changes and modifications can be made under the guidance of the present embodiment. For example, the components and / or functions of the application scenario 100 can be changed or altered according to the specific implementation scheme. As an example only, some other components can be added to the application scenario 100, such as a power module that can supply power to one or more components of the application scenario 100, and other devices or modules.

[0045] Figure 2 FIG. 1 is an exemplary flow chart of a method for correcting water-fat separation according to some embodiments of the present specification. Figure 2 As shown, the process 200 includes the following steps. In some embodiments, the process 200 can be performed by a water-fat separation correction system or processing device.

[0046] Step 202: Obtain scan data of the target object.

[0047] The target object includes the entirety or part of a biological object and / or a non-biological object (eg, a phantom, etc.) involved in the scanning process. For example, the target object may be the abdomen, chest, or other parts of a patient.

[0048] Scanning data refers to the data signal obtained by exciting the subject and collecting signals in a magnetic field using a magnetic resonance imaging device.

[0049] In some embodiments, the processing device may obtain the scan data by scanning the target object using magnetic resonance imaging (MRI) imaging technology.

[0050] In some embodiments, the processing device may also obtain the scan data of the target object by reading from a storage device, a database, etc. For example, the scan data may be pre-scanned by using an imaging device to scan the target object and stored in a storage device.

[0051] Step 204: Generate an initial first image and an initial second image based on the scan data.

[0052] In some embodiments, the processing device can generate the initial first image and the initial second image by image reconstruction based on the scan data. This embodiment does not limit the specific reconstruction method, for example, image reconstruction can be performed by an image reconstruction algorithm.

[0053] The initial first image refers to one of the water-fat images obtained after scanning the target object by a magnetic resonance imaging device. For example, the initial first image is a water image obtained by scanning.

[0054] The initial second image refers to another type of water-fat image obtained after scanning the target object by a magnetic resonance imaging device. For example, the initial second image is a fat image obtained by scanning.

[0055] In magnetic resonance imaging, a water map is a fat-suppressed image obtained by separating the water signal. In the water map, the signal of water molecules will be highlighted, while the signal of tissues or substances mainly composed of fat will be minimized or suppressed. The water map can better display water-containing tissues such as tissue structures, blood vessels, and masses. A fat map is a fat image obtained by separating the fat signal. In the fat map, the signal of fat molecules will be highlighted, while the signal of tissues or substances mainly composed of water will be minimized or suppressed. The fat map can better display fat tissues, lipomas, fat masses and other fat-containing structures.

[0056] like Figure 7 and Figure 8 As shown, Figure 7 is an exemplary schematic diagram of a water image according to some embodiments of the present specification, Figure 8 is an exemplary schematic diagram of a fat image according to some embodiments of the present specification, Figure 7 and Figure 8 The location indicated by the arrow in the middle image represents the artifact area in the image.

[0057] Step 206 , determining whether there are water-fat separation artifacts in the initial first image and the initial second image.

[0058] In some embodiments, the processing device can determine whether there is a water-fat separation artifact by an image recognition algorithm, for example, the initial first image and the initial second image are used as algorithm inputs, and the algorithm outputs whether there is a water-fat separation artifact in the image and marks the artifact area. In some embodiments, the image recognition algorithm used includes a deep learning model, for example, the initial first image and the initial second image are input into the deep learning model to determine whether there is a water-fat separation artifact.

[0059] In some embodiments, the processing device may also output the initial first image and the initial second image through a display device, and determine whether there is a water-fat separation artifact by receiving user input. For example, the processing device may display the initial first image and the initial second image through a display device, and the user may view the displayed images and outline the water-fat separation artifact area in the initial first image and the initial second image through an interactive method. This embodiment does not limit the user interaction method.

[0060] Step 208: Perform at least one round of iterative optimization based on the initial first image and the initial second image to obtain a target first image and a target second image.

[0061] The target first image refers to an image obtained after the image quality of the initial first image is optimized. Image quality optimization includes removing or reducing image artifacts in the image, such as water-fat separation artifacts, etc. For example, the target first image is a water image after the water-fat separation artifact correction is performed on the initial first image.

[0062] The target second image refers to an image obtained after the image quality of the initial second image is optimized. For example, the target second image is a fat image after the water-fat anti-artifact correction is performed on the initial second image.

[0063] Iterative optimization refers to the process of continuously improving and optimizing the method or algorithm for water-fat separation artifact correction in images through continuous iterations. In the scenario of water-fat separation artifact correction, iterative optimization includes subsequent optimization based on each round of iterative optimization, and continuously adjusting the optimization algorithm / method according to actual conditions. Through multiple iterations, the effect of water-fat separation artifact correction can be gradually improved to obtain higher quality images.

[0064] In some embodiments, the processing device can evaluate the results of iterative optimization, for example, whether the image quality of the water image and the fat image after the current round of iterative optimization meets preset requirements (for example, whether they can be used for medical diagnosis). If the requirements are not met, the next round of iterative optimization can be performed based on the results of the current round of iterative optimization. If the requirements are met, the optimization can be stopped and the water image and fat image obtained in the current round of iteration can be used as the target first image and the target second image.

[0065] In some embodiments, each round of iteration in the iterative optimization includes operations as shown in steps 2082-2084.

[0066] Step 2082, adding a first mark in the first image and adding a second mark in the second image.

[0067] The first mark refers to the pixel corresponding to the water-fat separation artifact area marked in the corresponding first image when the water-fat separation artifact occurs in the water-fat imaging. For example, when the initial first image is a water image, the first mark is the pixel corresponding to the fat marked in the water image.

[0068] The second mark refers to the pixel corresponding to the water-fat separation artifact marked in the corresponding second image when the water-fat separation artifact occurs in the water-fat imaging. For example, when the initial second image is a fat image, the second mark is the pixel corresponding to the water-fat separation artifact marked in the fat image.

[0069] It should be noted that, in the first round of iteration, the first image is the initial first image, and the second image is the initial second image. In subsequent rounds of iteration, the first image is the result of the optimization in the previous round of iteration (for example, the first image after optimizing the initial first image). Similarly, the second image is also the result of the optimization in the previous round of iteration (for example, the first image after optimizing the initial second image).

[0070] In some embodiments, the user may manually add the first mark and the second mark. For example, the processing device adds the first mark and the second mark by receiving user input. The user input method includes external device input, for example, using a mouse to directly mark pixels in the image, and using the marking result as the first mark and the second mark. The user input method may also be other methods, which are not limited in this specification.

[0071] In some embodiments, the first mark and the second mark can also be automatically added by the system. For example, the processing device can process the first image and the second image through an image recognition algorithm or a neural network model based on deep learning, identify the water-fat separation anti-artifact area therein, and mark the pixel points corresponding to the water-fat separation anti-artifact area as the first mark and the second mark.

[0072] The first mark and the second mark can be a pixel point corresponding to the water-fat separation artifact area, or a certain area, which is not limited in this specification. For example, the location and size of the water-fat separation artifact vary. When a large area of ​​water-fat separation artifact appears, such as a large area of ​​a certain tissue, and the area is dominated by water signals or fat signals, a polygon can be used to outline the area, and the water-fat phase can be assigned to the area in advance, providing prior knowledge for the water-fat separation algorithm used in subsequent image optimization.

[0073] Step 2084: optimize the image based on the first mark and the second mark.

[0074] In some embodiments, the processing device can perform image optimization using a preset image optimization algorithm, including a water-fat separation algorithm based on region growing and a water-fat separation algorithm based on deep learning.

[0075] The water-fat separation algorithm based on deep learning is a method for segmenting water and fat tissue using a deep neural network. In some embodiments, the water-fat separation algorithm based on deep learning includes a convolutional neural network (CNN) or its variants, such as U-Net, Fully Convolutional Network (FCN), etc. The processing device can use a large amount of labeled data to learn the characteristic representation of water and fat tissue during the training process, and use it to predict the segmentation of new images. When the original water and fat images (for example, the initial first image and the initial second image) are input, a segmentation prediction image can be obtained after forward propagation of the network, in which each pixel is marked as belonging to the water or fat category. The network achieves accurate segmentation by learning the spatial distribution, structural differences, phase differences, amplitude differences and other characteristics of water and fat tissue.

[0076] The water-fat separation algorithm based on region growing is a pixel-level image segmentation method used to separate water and fat tissue in medical images. The algorithm is based on the principle of similarity and connectivity between pixels and achieves segmentation by gradually growing and merging pixels. In the water-fat separation algorithm based on region growing, a seed pixel is selected as the starting point, and then the neighboring pixels are judged whether they belong to the same region as the seed pixel according to certain criteria (such as pixel gray value, phase continuity, spatial position, etc.). If the conditions are met, it is marked as part of the same region or different regions and added to the seed pixel set; if the conditions are not met, it can be put back into the queue and wait for re-judgment in the subsequent growth process. The process is repeated continuously until all pixels that meet the conditions are marked as the same region. The region growing algorithm can achieve a relatively accurate segmentation of the boundaries of water-fat tissue by considering the spatial relationship and local features between pixels.

[0077] For example, the processing device can use the first marker and the second marker as seed points to assign corresponding water-fat phase differences to water points and fat points, respectively. In the iterative optimization process based on region growth, the assigned water points and fat points as seed points will be used as the center to spread and grow to the periphery. The judgment of growth is based on the condition of local phase smoothness. For example, when a water seed point is used as the center point to spread to the surrounding neighborhood, the water phase difference and fat phase difference of the neighborhood points are compared with the L1 distance of the phase difference of the water point. L1 distance is a commonly used distance measurement method, also known as Manhattan distance or city block distance. It is used to measure the difference or similarity between two vectors. In the water-fat separation back-correction algorithm, L1 distance can be used to compare the difference between water pixels and fat pixels. By calculating the L1 distance between the two and classifying and screening pixels according to the degree of difference, the purpose of water-fat separation back-correction is achieved. The principle of assigning the phase of the neighborhood points is to make the phase of the local area smoother, that is, to minimize the L1 distance from the central water point, and to assign the correct water-fat phase to the neighborhood. When the neighborhood of the seed point is assigned an appropriate phase, they will continue to spread to the surrounding neighborhood as new seed points. When the foreground of the entire phase map is involved in the calculation and assigned a phase, the phase map will be a smooth phase map. Based on the smoothed phase difference after solution, the amplitude map of water and the amplitude map of lipid can be solved by the least squares method.

[0078] For more information about the water-fat separation algorithm based on region growing, see Figure 3 and Figure 4 The relevant description will not be repeated here.

[0079] In some implementations of this specification, when a water fat swap artifact appears in the scanned image, the water fat swap result in the entire image is optimized by marking some of the water fat swap pixels and using the marked water fat pixels through the water fat separation algorithm. This correction method is a progressive iterative optimization. If a single correction cannot completely eliminate the water fat swap artifact, the next marking and correction optimization can be performed according to the results of each iteration. For example, when the water fat swap artifact cannot be effectively corrected in the first iteration, or the position of the water fat marker point is incorrectly marked, the water fat marker can be changed or added according to the first optimization result, and a second correction can be performed. Since the identification of water fat swap artifacts requires a certain amount of experience accumulation, in actual work, multiple water fat swap corrections are likely to occur. In the embodiments provided in this specification, the current round of iteration can use the previous correction result as a basis to adjust the position, quantity, method, etc. of the water fat marker point, and perform multiple corrections, thereby obtaining a better water fat swap correction effect and effectively improving the image quality.

[0080] Figure 3is an exemplary flow chart of image optimization according to some embodiments of this specification. Figure 3 As shown, the process 300 includes the following steps. In some embodiments, the process 300 is performed by a water-lipid separation correction system or processing device.

[0081] Step 302: Create an optimization queue based on the image element corresponding to the first mark or the second mark.

[0082] An image element refers to a pixel in an image. For example, the image element corresponding to the first mark and the second mark refers to a pixel corresponding to the water-fat anti-artifact area in the image. In some embodiments, the corresponding image element can be a pixel or a region composed of multiple pixels.

[0083] The optimized queue is a First-In-First-Out (FIFO) queue created based on image elements. A FIFO queue is a data structure that operates and accesses elements in the order they are entered. In a FIFO queue, the first element added is processed or accessed first, while the last element added remains in a waiting state in the queue until the previous element is processed. A FIFO queue can be implemented using various data structures, such as arrays or linked lists.

[0084] In some embodiments, the processing device obtains the optimized queue by extracting image elements corresponding to the first mark and / or the second mark from the first image or the second image, and filling the image elements into the created queue.

[0085] In some embodiments, for the top element of the optimization queue, the processing device performs corresponding optimization processing by executing the operations illustrated in steps 304 and 306 .

[0086] Step 304 , obtaining neighboring elements of the top element of the optimization queue multiple times according to a first-in-first-out rule, and determining multiple sets of neighboring elements.

[0087] The first-in-first-out rule is a queue rule that specifies the order in which elements are processed in the queue. According to the first-in-first-out rule, the first element added to the queue will be processed first, and the last element added to the queue will be processed last.

[0088] Neighborhood elements refer to pixels adjacent to the currently selected pixel. For example, adjacent can be elements without other elements in between, or elements with a spacing of no more than a preset number (e.g., 1 or 2 elements) in between. Neighborhood element set refers to a set consisting of neighborhood elements.

[0089] In some embodiments, the processing device can obtain the neighboring elements of the top element by region growing. Region growing refers to selecting a seed pixel or region as a starting point based on the principle of similarity between pixels, and judging whether to merge adjacent pixels or regions into the growing region according to certain criteria. These criteria include pixel gray value, color, texture characteristics, spatial position, etc.

[0090] Each of the multiple neighborhood element sets corresponds to a top element, and the elements in the neighborhood element sets of each top element may be different.

[0091] Step 306: Optimize each of the top elements and the elements in its corresponding neighborhood element set, and determine the target phase of the top element based on the result of the optimization calculation.

[0092] Optimization calculation refers to the process of calculating the difference between the candidate phase of the top element and the average / Gaussian weighted average of the phases of the calculated elements in the neighborhood element set, and assigning a target phase to the top element based on the calculated result.

[0093] The target phase is the phase that is ultimately assigned to the top element. Phase refers to the phase information of each pixel in the image data. In magnetic resonance imaging, phase refers to the phase angle of the signal, which contains information about the change in the magnetic field.

[0094] In some embodiments, the optimization calculation method includes calculating the difference between the candidate phase of the top element and the average value / Gaussian weighted average value of the phases of the elements in its corresponding neighborhood element set, and determining the target phase by multiple calculations (each element will be used as the top element multiple times, which can avoid the local minimum problem) of the distance between the candidate phase of the top element and the Gaussian weighted average value of its neighborhood phases.

[0095] For detailed information on determining the target phase, see Figure 4 and Figure 5 Description.

[0096] Figure 4 is an exemplary flow chart of performing optimization calculations according to some embodiments of this specification. Figure 4 As shown, the process 400 includes the following steps. In some embodiments, the process 400 is performed by a water-fat separation correction system or processing device.

[0097] Step 402: Acquire the phase of the image element corresponding to the first mark or the second mark.

[0098] The phase includes the phase information of the pixel corresponding to the marked area in the image, which can be used to determine which type of image area the pixel should be classified into, for example, water area or fat area.

[0099] In some embodiments, the phases of the elements corresponding to the first mark and the second mark may be obtained by calculating phase correlation or using other phase information processing techniques, or by manual allocation, which is not limited in this embodiment.

[0100] Step 404, determining the candidate phase of each top element and the phases of the calculated elements in its corresponding neighborhood element set.

[0101] In some embodiments, the priority queue is created based on the image element corresponding to the first mark or the second mark, so the candidate phase of the top element of the queue can be determined based on the phase of the image element corresponding to the first mark or the second mark. For example, based on the size of the phase value of the mark, a plurality of candidate phases close to the size of the phase of the mark are allocated to the corresponding element.

[0102] The calculated element refers to the neighboring element whose phase has been determined. Since the optimization calculation method is iterative calculation, the top element is calculated as the center in each iteration, and its neighboring elements have been the top elements in other rounds of iterations. Therefore, the phase of the calculated element is the phase of the element that has been the top element in other rounds of iterations and has determined the phase.

[0103] Step 406, performing optimization calculation based on the candidate phase corresponding to each of the top elements and the phases of the calculated elements in the corresponding neighborhood element set.

[0104] In some embodiments, each element in the first marking area and the second marking area will be pushed into the priority queue for calculation multiple times, that is, each element will be calculated as the top element of the queue multiple times. In order to avoid the problem of local minimum, each element is set with a calculation access upper limit, that is, a certain number of calculation times is set for each element.

[0105] Therefore, in some embodiments, before performing optimization calculations, the processing device will determine whether the number of calculations of the top element has reached a preset number of times. When the number of calculations of the top element has not reached the preset number of times, optimization calculations will be performed based on the candidate phases corresponding to each of the top elements and the phases of the elements in its corresponding neighborhood element set; when the number of calculations of the top element has reached the preset number of times, the target candidate phase corresponding to the majority in the element phase set will be used as the target phase of the top element.

[0106] Mode is a statistical term used to describe the value that appears most often in a set of data. In a data set, if there are multiple values ​​that appear the same number of times and are all maximum values, then the data set can have multiple modes.

[0107] For the target candidate phase in the element phase set, when there is only one mode, its corresponding target candidate phase can be directly used as the target phase of the top element. When there are multiple modes, you can arbitrarily select the target candidate phase corresponding to one mode as the target phase of the top element. You can also increase the number of optimization calculations for the top element and re-determine the mode after increasing the number of target candidate phases in the element phase set until the number of modes is 1, and then use the target candidate phase corresponding to the mode as the target phase. For a description of the target candidate phase, see Figure 5 Description in .

[0108] In some embodiments, an access upper limit is also set for the elements in the neighborhood element set. When performing optimization calculations, the processing device will determine whether the elements in the neighborhood element set have reached the calculation upper limit. If so, the phase of the element will not be calculated when calculating the neighborhood average / Gaussian weighted average.

[0109] For details on the optimization calculation process, please refer to the following Figure 5 Related description.

[0110] Figure 5 FIG. 1 is an exemplary flow chart of a method for determining a target phase of a top element according to some embodiments of the present specification. Figure 5 As shown, the process 500 includes the following steps. In some embodiments, the process 500 can be performed by a water-lipid separation correction system or processing device.

[0111] In some embodiments, for each top element of the priority queue, the optimization calculation may include the following operations.

[0112] Step 502, calculating the difference between the candidate phase corresponding to the top element and the average value / Gaussian weighted average value of the phases of the calculated elements in the corresponding neighborhood element set.

[0113] Gaussian weighted average is a commonly used filtering method. The main idea is to use a Gaussian kernel to perform a convolution operation with each point in the neighborhood, and then sum the convolution results to obtain the output image. In Gaussian weighted average, when averaging the image, pixels at different positions are assigned different weights.

[0114] The processing device can calculate the Gaussian weighted average value in various ways, and this embodiment does not limit the specific calculation method.

[0115] Step 504: assign a target candidate phase to the top element of the queue so that the difference between the average value / Gaussian weighted average value of the phases of the calculated elements in the neighborhood element set is minimal.

[0116] The candidate phase refers to the phase that may be assigned to the water-fat type corresponding to the top element of the team. For example, when the doctor manually adds a mark, there are two marking tools, corresponding to the allocation of fat signals and water signals. When the water signal marking tool is used, the circled area is assigned to the candidate phase of water, and when the fat signal marking tool is used, the circled area is assigned to the candidate phase of fat.

[0117] The target candidate phase refers to the candidate phase among multiple candidate phases of the top element that is closest to the average value / weighted average value of the phases of the neighboring elements.

[0118] In some embodiments, the target candidate phase assigned to the top element of the queue may be the same as the phase assigned when the tag was added.

[0119] Step 506: Add the assigned target candidate phase to the element phase set corresponding to the top element of the team.

[0120] The smallest phase difference between two elements can be considered as the smallest difference between the corresponding two elements, and the smoothness of their phase is stronger.

[0121] The element phase set refers to a set of multiple target candidate phases assigned to the top element of the team. The multiple target candidate phases in the element phase set can be used to determine the target phase of the top element of the team.

[0122] Step 508: Add the top element of the queue back to the priority queue.

[0123] After adding the top element back into the priority queue, it can enter the subsequent iterative calculation according to the queue order.

[0124] It should be noted that the description of each process above 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 the process under the guidance of this specification. However, these modifications and changes are still within the scope of this specification. For example, adding a storage step, etc.

[0125] Based on the above inventive concept, the application workflow of the water-lipid content correction method disclosed in this specification is introduced below.

[0126] The workflow of the water-fat separation correction method disclosed in this specification can be placed on the image reconstruction page of the imaging device. The process can be summarized as: scanning the WFI protocol, reconstructing the water map and fat map, on the image browsing page, the doctor finds the water-fat separation artifact, uses the annotation tool to mark the separation artifact, starts the water-fat separation artifact correction algorithm, and finally obtains the corrected water map and fat map.

[0127] The input data of this process is the acquired complex image, because the calculation of the correction algorithm depends on complex calculation. Therefore, the correction algorithm can be embedded in the system's image reconstruction process to obtain the acquired complex image. After the doctor determines that the artifact appears, he starts to use the marking tool to mark it. The marking tool acts on the water map and fat map. The correction algorithm needs to obtain the position of the area marked by the doctor as the input of the correction algorithm. Specifically, the input of the correction algorithm is divided into two parts. The first part is the acquired original complex image. For example, the dual echo water-fat separation algorithm has two complex images. The same phase and anti-phase complex images are used as the input of the correction algorithm. The second part is the position information of the marked anti-phase artifact, which can be in the form of a marking point, and the input of the correction algorithm is the coordinates of the marking point. Alternatively, it can also be in the form of a marked area, and the input is a regional mask, which is divided into a water map mask and a fat map mask.

[0128] After the correction algorithm is finished, the water map and fat map are output to the user. The user will re-read whether there are still water-fat separation artifacts. If the water-fat separation artifacts are still found, the user can adjust the marking features such as the marking position, marking area, marking density, etc., and run the correction algorithm again to obtain better water-fat separation artifact correction effects through iteration.

[0129] It should be noted that the examples of water-fat separation anti-artifact correction algorithms are only for illustrative purposes, and a variety of water-fat separation anti-artifact correction algorithms can be used as image optimization algorithms in the technical solutions disclosed in this specification, such as the RIPE algorithm based on the traditional physical model, TRW-S, and water-fat separation algorithms based on deep learning. The above water-fat separation algorithms can all be combined with water-fat seed points. The water-fat seed points, as the prior knowledge of water-fat distribution of the above algorithms, can significantly improve the effect of water-fat separation output.

[0130] For example, see Fig. 9 , Fig. 9 It is an exemplary schematic diagram of another interactive water-lipid classification back-calibration algorithm according to some embodiments of this specification.

[0131] S1. The algorithm input is an initial first image and an initial second image, for example, a water image and a fat image.

[0132] S2. The user manually marks the water-fat separation area in the initial first image and the initial second image, and assigns the corresponding water phase and fat phase.

[0133] S3, phase initialization. For example, with P u and P v To represent the allocated water phase and lipid phase, the phase is initialized as

[0134] S4. Smoothing and filtering the allocated water phase and lipid phase. In some embodiments, the filtering method includes median filtering, mean filtering, Gaussian weighted mean filtering, etc.

[0135] S5, the pixel points in the fixedly marked water-fat separation area are not moved, and other phase foreground elements are updated. For example, each foreground element is assigned a phase P closest to the smoothing filter output L1. u / P v .

[0136] S6. Record the number of updated elements.

[0137] S7, determine whether the number of updated pixels is zero. If yes, proceed to S8, if no, proceed to step S4 and repeat the above process.

[0138] S8. Output the calibration result.

[0139] Figure 6 is an exemplary module diagram of a water-fat separation correction system according to some embodiments of this specification. Figure 5 As shown, the system 600 may include a scanning data acquisition module 610 , an image generation module 620 , a determination module 630 and an iterative optimization module 640 .

[0140] The scan data acquisition module 610 is used to acquire scan data of the target object.

[0141] The image generation module 620 is used to generate an initial first image and an initial second image based on the scan data.

[0142] The determination module 630 is used to determine whether there are water-fat separation artifacts in the initial first image and the initial second image.

[0143] The iterative optimization module 640 is used to perform at least one round of iterative optimization based on the initial first image and the initial second image to obtain a target image when there are water-fat separation artifacts in the initial first image and the initial second image; wherein each round of iteration includes: adding a first mark in the first image and adding a second mark in the second image; and performing image optimization based on the first mark and the second mark.

[0144] about Figure 6 For further description of each system shown, please refer to the relevant flow chart section, for example, Figure 2-Figure 5 Related description.

[0145] It should be understood that Figure 6The system and its modules shown can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art will understand that the above methods and systems can be implemented using computer executable instructions and / or included in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the system and its modules of this specification. Not only can the hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc. be implemented, it can also be implemented with software such as executed by various types of processors, and it can also be implemented by a combination of the above hardware circuits and software (for example, firmware).

[0146] It should be noted that the above description of the water-fat separation correction system and its modules is only for convenience of description and cannot limit the present specification to the scope of the embodiments. It is understandable that for those skilled in the art, after understanding the principle of the system, it is possible to arbitrarily combine the modules or form a subsystem to connect with other modules without deviating from this principle. In some embodiments, Figure 6 The scanning data acquisition module 610, image generation module 620, judgment module 630 and iterative optimization module 640 disclosed in the specification can be different modules in a system, or a 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 this specification.

[0147] 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.

[0148] 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.

[0149] 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.

[0150] 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.

[0151] 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.

[0152] 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.

[0153] 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 correcting water-fat separation, the method comprising: Obtaining scan data of a target object; generating an initial first image and an initial second image based on the scan data; Determining whether there are water-fat separation artifacts in the initial first image and the initial second image; If yes, at least one round of iterative optimization is performed based on the initial first image and the initial second image to obtain a target first image and a target second image; wherein each round of iteration includes: adding a first marker in the first image and adding a second marker in the second image; Image optimization is performed based on the first mark and the second mark. 2 . The method according to claim 1 , wherein the first mark and the second mark are added by receiving user input.

3. The method according to claim 1, wherein the performing image optimization based on the first mark and the second mark comprises: Creating an optimization queue based on the image element corresponding to the first mark or the second mark; Obtaining neighboring elements of the top element of the optimization queue multiple times according to a first-in-first-out rule to determine multiple sets of neighboring elements; An optimization calculation is performed on each of the top elements and the elements in its corresponding neighborhood element set, and a target phase of the top element is determined based on the result of the optimization calculation.

4. The method according to claim 3, wherein the optimizing calculation of each of the top elements and the elements in the corresponding neighborhood element set comprises: Acquire the phase of the image element corresponding to the first mark or the second mark; Determine the candidate phase of each top element and the phase of the calculated elements in its corresponding neighborhood element set; wherein the calculated elements are optimized based on the candidate phase corresponding to each top element and the phase of the calculated elements in its corresponding neighborhood element set.

5. The method according to claim 4, wherein the optimization calculation based on the candidate phase corresponding to each of the top elements and the phases of the elements in the corresponding neighborhood element set comprises: For each top element, Calculate the difference between the candidate phase corresponding to the top element of the team and the average value / Gaussian weighted average value of the phases of the calculated elements in the corresponding neighborhood element set; Assigning a target candidate phase to the top element of the team so that the phase has a minimum difference with the average / Gaussian weighted average of the phases of the calculated elements in the neighborhood element set; Add the assigned target candidate phase to the element phase set corresponding to the top element of the team; Add the top element of the queue back to the priority queue.

6. The method according to claim 5, further comprising: Before performing the optimization calculation, determine whether the number of calculations of the top element of the queue has reached a preset number; If yes, then the target candidate phase corresponding to the majority of the element phase set corresponding to the top element of the team is used as the target phase of the top element of the team; If not, an optimization calculation is performed based on the candidate phase corresponding to each of the top elements of the team and the phases of the elements in its corresponding neighborhood element set.

7. The method according to claim 1, wherein the performing image optimization based on the first mark and the second mark comprises: The image optimization is performed using a preset image optimization algorithm.

8. According to the method of claim 7, the preset image optimization algorithm includes a water-fat separation algorithm based on region growing or a water-fat separation algorithm based on deep learning.

9. A water-fat separation correction system, the system comprising: A scanning data acquisition module, used to acquire scanning data of a target object; An image generation module, configured to generate an initial first image and an initial second image based on the scan data; A judging module, used for judging whether there are water-fat separation artifacts in the initial first image and the initial second image; The iterative optimization module is used to perform at least one round of iterative optimization based on the initial first image and the initial second image to obtain a target image when there are water-fat separation artifacts in the initial first image and the initial second image; wherein each round of iteration includes: adding a first marker in the first image and adding a second marker in the second image; Image optimization is performed based on the first mark and the second mark.

10. A water-fat separation correction device, comprising a processor, wherein the processor is used to execute the water-fat separation correction method according to any one of claims 1 to 8.

Images