Medical image processing method and device, image processing equipment and medium

Abstract

Embodiments of the present application disclose a medical image processing method and device, an image processing apparatus and a medium. The method comprises: determining an arterial input function of an arterial blood vessel for supplying blood to a target part and attenuation coefficient enhancement information of a single voxel corresponding to the arterial blood vessel according to a plurality of groups of dynamic enhancement CT images collected at different time points; inputting the arterial input function and the attenuation coefficient enhancement information into a preset estimation model to obtain a preset blood flow parameter of the target part, the preset blood flow parameter comprising a blood flow velocity and an average transit time, and the preset estimation model being constructed based on a box-type attenuation function. The technical problem that the prior art cannot accurately determine the blood flow parameter of the dynamic enhancement CT is solved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of image processing, and in particular to a medical image processing method and device, an image processing apparatus and a medium. BACKGROUND

[0002] Global burden of disease studies show that the overall risk of stroke in the Chinese population is 39.9%, which is the first cause of years of life lost due to disease. Stroke is divided into two types, namely ischemic stroke and hemorrhagic stroke. Ischemic stroke is caused by complete or partial obstruction of blood supply, and hemorrhagic stroke is caused by non-traumatic rupture of cerebral blood vessels leading to blood accumulation in the brain. Clinically, information for diagnosing stroke is usually obtained through brain blood flow parameters measured by dynamic contrast-enhanced CT (Computed Tomograph, CT for short) imaging, such as cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT) and time to peak (TTP).

[0003] Dynamic contrast-enhanced CT is an imaging method that detects the circulation of contrast agents in the brain along with blood flow after intravenous injection of iodine contrast agents. During acquisition, the head that has not been enhanced is first scanned, and the reconstructed image is defined as a phantom image. The brain projection data acquired at multiple consecutive time points is then reconstructed into a three-dimensional brain dynamic contrast-enhanced CT image corresponding to each time point. A series of dynamic contrast-enhanced images can be obtained by subtracting the phantom image from the image at each time point acquired later, which contains blood flow parameter information. According to the indicator dilution theory, the iodine contrast agent attenuation coefficient enhancement information of each voxel in the brain tissue can be expressed as the convolution of the blood flow propagation residual function and the arterial input function. If the enhancement information and the arterial input function can be measured by dynamic contrast-enhanced CT imaging, the blood flow propagation residual function can be derived according to the indicator dilution theory. The blood flow propagation residual function can be decomposed into a single voxel blood flow parameter. From the perspective of mathematical modeling, this is a process of solving an inverse problem. The traditional method for solving this inverse problem is based on singular value decomposition (SVD) and a series of improved methods. However, due to the limitation of radiation dose limits or actual scanning conditions, dynamic contrast-enhanced CT usually has lower image quality than normal CT, so the inverse problem to be solved is usually ill-posed, that is, the derived blood flow parameter may be unstable or inaccurate.

[0004] In summary, the prior art has the technical problem of being unable to accurately determine the blood flow parameters of dynamic contrast-enhanced CT. SUMMARY

[0005] The embodiment of the present application provides a medical image processing method, device, image processing equipment and medium, and solves the technical problem that the prior art cannot accurately determine blood flow parameters of dynamic enhancement CT.

[0006] In a first aspect, the embodiment of the present application provides a medical image processing method, which comprises the following steps:

[0007] According to a plurality of groups of dynamic enhancement CT images collected at different time points, an arterial input function of an arterial blood vessel for supplying blood to a target part and attenuation coefficient enhancement information of a single voxel corresponding to the arterial blood vessel are determined.

[0008] The arterial input function and the attenuation coefficient enhancement information are input into a preset estimation model to obtain preset blood flow parameters of the target part, the preset blood flow parameters comprising blood flow velocity and average transit time, and the preset estimation model being constructed based on a box attenuation function.

[0009] In a second aspect, the embodiment of the present application further provides a medical image processing device, which comprises the following modules:

[0010] The acquisition module is configured to determine, according to a plurality of groups of dynamic enhancement CT images collected at different time points, an arterial input function of an arterial blood vessel for supplying blood to a target part and attenuation coefficient enhancement information of a single voxel corresponding to the arterial blood vessel.

[0011] The blood flow parameter determination module is configured to input the arterial input function and the attenuation coefficient enhancement information into a preset estimation model to obtain preset blood flow parameters of the target part, the preset blood flow parameters comprising blood flow velocity and average transit time, and the preset estimation model being constructed based on a box attenuation function.

[0012] In a third aspect, the embodiment of the present application further provides an image processing equipment, which comprises the following modules:

[0013] One or more processors;

[0014] A storage device configured to store one or more programs;

[0015] When the one or more programs are executed by the one or more processors, the one or more processors implement the medical image processing method according to any embodiment.

[0016] In a fourth aspect, the embodiment of the present application further provides a storage medium containing computer executable instructions, which are used to execute the medical image processing method according to any embodiment when executed by a computer processor.

[0017] The technical scheme of the medical image processing method provided in the embodiment of the present application comprises: determining an arterial input function of an arterial blood vessel for supplying blood to a target part and attenuation coefficient enhancement information of a single voxel corresponding to the arterial blood vessel according to a plurality of groups of dynamic enhancement CT images collected at different time points; inputting the arterial input function and the attenuation coefficient enhancement information into a preset estimation model to obtain a preset blood flow parameter of the target part, the preset blood flow parameter comprising a blood flow velocity and an average transit time, and the preset estimation model being constructed based on a box-shaped attenuation function. Since the preset estimation model constructed based on the box-shaped attenuation function is more in line with actual conditions compared with an ideal physiological model required when singular value decomposition is performed, the preset blood flow parameter determined based on the preset estimation model is more accurate. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0019] Figure 1 is a flowchart of the medical image processing method provided in the first embodiment of the present application;

[0020] Figure 2 is a flowchart of the medical image processing method provided in the second embodiment of the present application;

[0021] Figure 3 is a structural block diagram of the medical image processing device provided in the third embodiment of the present application;

[0022] Figure 4 is another structural block diagram of the medical image processing device provided in the third embodiment of the present application;

[0023] Figure 5 is a structural block diagram of the image processing device provided in the fourth embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely by embodiments with reference to the drawings. Obviously, the described embodiments are some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort are within the protection scope of the present application.

[0025] Embodiment one

[0026] Figure 1is a flowchart of the medical image processing method provided by Embodiment One of the present application. The technical solution of this embodiment is applicable to the case where a preset estimation model constructed based on a box-shaped attenuation function is used to determine the preset blood flow parameter corresponding to a dynamic enhancement CT image. The method can be executed by the medical image processing device provided by the present application, which can be realized in the form of software and / or hardware and applied in the processor of an image processing device. The method specifically includes the following steps:

[0027] S101, according to a plurality of groups of dynamic enhancement CT images collected at different time points, determine the arterial input function of the arterial blood vessel supplying blood to the target site and the attenuation coefficient enhancement information of the single voxel corresponding to the arterial blood vessel.

[0028] Wherein, the target site is the site to be analyzed for blood flow velocity parameter, preferably but not limited to brain, target tumor area and other sites, and the present embodiment takes the brain as an example to illustrate the technical solution.

[0029] It can be understood that if the target site is the brain, then the dynamic enhancement CT image is a brain dynamic enhancement CT image, and the arterial blood vessel supplying blood to the brain tissue is the carotid artery.

[0030] Wherein, the different time points refer to the image acquisition time points after the contrast agent enters the brain tissue. The timing and number of image acquisition time points can be set according to the needs of clinicians. In this embodiment, the reconstructed image corresponding to the scan data collected at each time point is taken as a group of dynamic enhancement CT images.

[0031] Wherein, the format of the dynamic enhancement CT image is DICOM format.

[0032] Wherein, the arterial input function (AIF) describes the time behavior characteristics of the contrast agent passing through the imaging voxel. It can be understood that in fact, the input concentration time curve of each voxel is not exactly the same, so the intensity change information of the contrast agent corresponding to each voxel is also different, so only the attenuation coefficient enhancement information of the contrast agent of a single voxel of the blood supplying blood vessel can be estimated. The estimation method of the attenuation coefficient enhancement information of the single voxel in this embodiment is not limited, and the attenuation coefficient enhancement information of the single voxel can be determined using the prior art.

[0033] The method for determining the arterial input function comprises: determining the pixel mean value in the carotid artery region of the plurality of sets of brain dynamic enhancement CT images collected at different time points, taking the relative enhancement corresponding to the pixel mean value as the first relative enhancement, and determining the pixel mean value in the jugular vein region of the plurality of sets of brain dynamic enhancement CT images, and taking the relative enhancement corresponding to the pixel mean value as the second relative enhancement; constructing the arterial input function curve corresponding to the first relative enhancement and the venous output function curve corresponding to the second relative enhancement with the collection time as the horizontal axis and the relative enhancement as the vertical axis; correcting the arterial input function curve according to the curve area corresponding to the venous output function curve to update the arterial input function curve; and determining the arterial input function according to the updated arterial input function curve. Since the blood outflow of the venous blood vessels is the same as the blood inflow of the arterial blood vessels, the curve area corresponding to the arterial input function should be the same as the curve area corresponding to the venous output function, so the arterial input function curve can be corrected according to the curve area corresponding to the venous output function curve, thereby solving the problem of inaccurate arterial input function caused by the low pixel value of the arterial region on the dynamic enhancement CT image, and improving the accuracy of the arterial input function.

[0034] In some embodiments, the method for determining the carotid artery region comprises: summing the corresponding frames of the plurality of sets of dynamic enhancement CT images collected at different time points to obtain a first image, then segmenting the first image using a first preset threshold to obtain a carotid artery template, and using the carotid artery template to determine the carotid artery region of the dynamic enhancement CT image collected at each time point; and correspondingly, segmenting the first image using a second preset threshold to obtain a jugular vein template, and using the jugular vein template to determine the jugular vein region of the dynamic enhancement CT image collected at each time point.

[0035] It can be understood that the dynamic enhancement CT image is a CT image after preprocessing. The preprocessing method includes but is not limited to filter denoising processing and motion correction processing. The motion correction processing includes intra-frame motion correction and inter-frame motion correction. In this embodiment, the plurality of sets of dynamic enhancement CT images after filter denoising processing are preferably subjected to intra-frame motion correction processing first to update the dynamic enhancement CT images; then a reference frame image is selected from the plurality of sets of updated dynamic enhancement CT images, and rigid registration is performed on other frame images according to the reference frame image to update the plurality of sets of dynamic enhancement CT images again. It can be understood that the dynamic enhancement CT images after intra-frame motion correction and inter-frame motion correction have no motion artifacts or only have a small amount of motion artifacts, and the situation that the blood input function and the single voxel attenuation coefficient enhancement information are low in accuracy due to motion artifacts can be avoided.

[0036] S102, input the arterial input function and the attenuation coefficient enhancement information into a preset estimation model to obtain preset blood flow parameters of the target site, the preset blood flow parameters including blood flow velocity and mean transit time, and the preset estimation model being constructed based on a box attenuation function.

[0037] After the arterial input function and the attenuation coefficient enhancement information of the single voxel are obtained, the arterial input function and the attenuation coefficient enhancement information are input into a preset estimation model constructed based on a box attenuation function to obtain preset blood flow parameters. The preset blood flow parameters include blood flow velocity and mean transit time. The blood flow velocity is the linear velocity of the blood flow in the brain in the blood vessel, that is, the distance of a particle in the blood vessel in a unit of time; the mean transit time is the average time of the contrast agent staying in the tissue.

[0038] The construction method of the preset estimation model includes:

[0039] The change curve of the attenuation coefficient enhancement information of the single voxel with time can be expressed as:

[0040]

[0041] wherein μ is the blood flow parameter to be estimated, h(μ, t) is the scaled residual function of the blood flow, C a (t) is the arterial input function (AIF), and ε(t) is noise. The scaled residual function of the blood flow can be further expressed as a function of the blood flow velocity (CBF), the tissue density ρ and the residual function r(t), as follows:

[0042] h(μ, t) = CBF· ρ· r(t) (2)

[0043] wherein it is set that μ includes the blood flow velocity (CBF) and the mean transit time (MTT). r(t) is constructed as a function based on the box attenuation function, and then according to the above formula, the following can be obtained:

[0044]

[0045] wherein T0=0.632×MTT. It can be understood that the convolution in formula (1) can be expanded as a sampling of the integral of the arterial input function, as follows:

[0046]

[0047] It can be appreciated that formula (3) is brought into formula (4) to obtain h (μ, t) by deconvolution, so as to obtain CBF and MTT. After CBF and MTT are obtained, the product of the two can be calculated to obtain blood volume (CBV). In addition, the abscissa corresponding to the maximum value of Y (t) in formula (4) can be determined, and the time corresponding to the abscissa is the time to peak (TTP). Thus, all clinically commonly used blood flow parameters can be obtained.

[0048] The technical scheme of the medical image processing method provided in the embodiment of the application comprises: determining an arterial input function of an arterial blood vessel for supplying blood to a target part and attenuation coefficient enhancement information of a single voxel corresponding to the arterial blood vessel according to a plurality of groups of dynamic enhancement CT images collected at different time points; inputting the arterial input function and the attenuation coefficient enhancement information into a preset estimation model to obtain preset blood flow parameters of the target part, the preset blood flow parameters comprising blood flow velocity and mean transit time, and the preset estimation model being constructed based on a box-shaped attenuation function. Since the preset estimation model constructed based on the box-shaped attenuation function is more in line with actual conditions than an ideal physiological model required when singular value decomposition is performed, the preset blood flow parameters determined based on the preset estimation model are more accurate.

[0049] Embodiment two

[0050] Figure 2 is a flowchart of the medical image processing method provided in the embodiment two of the application. The embodiment of the application adds a step of estimating a preset blood flow parameter distribution on the basis of the above-mentioned embodiment.

[0051] Correspondingly, the method of the embodiment comprises:

[0052] S201, determining an arterial input function of an arterial blood vessel for supplying blood to a target part and attenuation coefficient enhancement information of a single voxel corresponding to the arterial blood vessel according to a plurality of groups of dynamic enhancement CT images collected at different time points.

[0053] S202, inputting the arterial input function and the attenuation coefficient enhancement information into a preset estimation model to obtain preset blood flow parameters of the target part, the preset blood flow parameters comprising blood flow velocity and mean transit time, and the preset estimation model being constructed based on a box-shaped attenuation function.

[0054] S203, determining an updating function of a mean value and a variance of the preset blood flow parameters determined by the preset estimation model by a Bayesian machine learning method.

[0055] The embodiment of the application further deduces and estimates the preset blood flow parameters by the Bayesian machine learning method on the basis of the above-mentioned embodiment, so as to determine a distribution of the preset blood flow parameters.

[0056] Since Y(t) corresponds to N discrete sampling time points in the scan, the noise ε(t) can be modeled as a zero-mean Gaussian distribution at each acquisition time point, and each image frame is independent in time, formula (4) can be represented by the following discrete expression:

[0057]

[0058] where i is the index of the image frame, ω i is proportional to the weight of the detection event (i.e. the acquisition time length of the brain three-dimensional enhanced CT image indexed by i). According to formula (4) to construct the data mismatch term, then the problem of derivation becomes the problem of finding the parameter μ that maximizes the log-likelihood function, as follows:

[0059]

[0060] Since ω i = 1 in the case of the same acquisition time length of each three-dimensional dynamic enhanced CT image, the above optimization problem is equivalent to a nonlinear least squares fitting problem.

[0061] Equation (6) can be extended to a full Bayesian inference problem, which can provide the overall probability distribution of the preset blood flow parameters instead of only the maximum likelihood estimate, thus directly reflecting the uncertainty information of each parameter measurement. According to Bayes' theorem, the posterior distribution of the estimated parameter μ with respect to the attenuation coefficient enhancement information Y is:

[0062]

[0063] where P(Y|μ) is the likelihood function, P(μ) is the prior probability of the estimated parameter μ, and P(Y) is the prior probability of the attenuation coefficient enhancement information Y. It can be understood that the fundamental difference between this formula and formula (6) is that μ no longer represents the maximum likelihood estimate but the overall probability distribution. Estimating μ within the Bayesian framework becomes an inference problem, which is essentially to maximize P(μ|Y) to obtain μ according to the given attenuation coefficient enhancement information (Y). Since it is difficult to directly calculate the posterior distribution, the variational Bayesian method is used for approximation. The basic idea of this variational Bayesian method is to find an analytical distribution q(μ|Y) that is close to the posterior distribution P(μ|Y), so that the relative entropy of the two will be minimized. The log evidence of the distribution model of formula (7) can be written as:

[0064]

[0065] where P(Y,μ) is the joint distribution, E *is the expectation of q(μ|Y). ELBO is the lower bound of the evidence, and KL is the divergence of the approximate distribution q(μ|Y) and P(Y, μ). Since the relative entropy is always positive, ELBO provides a lower bound of the log-likelihood function. Thus we can find the μ that makes the analytic distribution closer to the true posterior distribution by maximizing ELBO, as follows:

[0066]

[0067] The relative entropy in the bracket of equation (9) is also always positive, and the optimal solution of μ can be obtained by equating the numerator and the denominator.

[0068] For the numerator, it is necessary to ensure that it is tractable. Using variational Bayes to decompose q(μ|Y), it is assumed that the true posterior distribution of each parameter is a multi-parameter Gaussian. Thus the approximate posterior distribution can be expressed as:

[0069]

[0070] where k = 1, μ1 corresponds to blood flow velocity, σ1 is the variance of the blood flow velocity distribution, and m1 is the mean of the blood flow velocity distribution; k = 2, μ2 corresponds to blood flow velocity, σ2 is the variance of the blood flow velocity distribution, and m2 is the mean of the blood flow velocity distribution; μ is a matrix containing the blood flow velocity and the mean transit time to be estimated, σ is a matrix including the variance σ1 of the blood flow velocity and the variance σ2 of the mean transit time, and m is a matrix containing the mean m1 of the blood flow velocity and the mean of the mean transit time m2.

[0071] For the denominator, the likelihood function of equation (6) is inserted, and a conjugate distribution, i.e., a multi-variate Gaussian distribution, is selected as the prior distribution, with the mean m0 and the variance σ0. Thus the denominator can be expressed as:

[0072]

[0073] In order to ensure the tractability of the denominator and to be applicable to equation (4), g(μ) can be expressed by the first-order Taylor expansion of the posterior distribution, as follows:

[0074]

[0075] where J is the Jacobian matrix. After applying this linear transformation, equation (11) becomes:

[0076]

[0077] By equating equation (10) and equation (13), the update functions of the mean and the variance of μ are obtained, as follows:

[0078]

[0079] It is understandable that σ0 and m0 are the initial values ​​of σ and m, respectively.

[0080] S204. Determine the distribution of the preset blood flow parameters based on the update function of the variance and mean of the preset blood flow parameters.

[0081] Since μ includes blood flow velocity and mean flow time, once the update functions for the mean and variance of μ are determined, the update functions for the mean and variance of blood flow velocity (CBF) and mean flow time (MTT) can be determined. Furthermore, the distribution of blood flow velocity and the distribution of mean flow time can be determined based on the update functions for the variance and mean of blood flow velocity. The variance of blood flow velocity contains uncertainty information about blood flow velocity, and the variance of mean flow time contains uncertainty information about mean flow time.

[0082] After determining the distribution of blood flow velocity and the distribution of mean flow time, this embodiment preferably further calculates the ratio of the variance to the mean of blood flow velocity to obtain a first coefficient of variation representing the uncertainty of blood flow velocity, and calculates the ratio of the variance to the mean of mean flow time to obtain a second coefficient of variation representing the uncertainty of mean flow time. It is understood that the larger the coefficient of variation, the greater the uncertainty of the corresponding parameter, and consequently, the higher the unreliability of the dynamic contrast-enhanced CT image scanning.

[0083] It is understood that the blood flow velocity distribution estimated in this embodiment only represents the blood flow velocity distribution of capillaries within brain tissue. Clinical diagnosis requires reference to the blood flow velocity distribution outside brain tissue. Therefore, after obtaining the blood flow velocity distribution within brain tissue, this embodiment further determines the blood flow velocity distribution outside brain tissue through hematocrit correction. Specifically, the mean of the blood flow velocity is multiplied by a preset coefficient, and the variance of the blood flow velocity is multiplied by the square of this preset coefficient to obtain the blood flow velocity distribution outside brain tissue. The preset coefficient can be selected as 0.733.

[0084] In one embodiment, after obtaining the blood volume, the blood volume outside the brain tissue is determined by hematocrit correction. Specifically, the blood volume is multiplied by a preset coefficient to obtain the blood volume outside the brain tissue. The preset coefficient can be selected as 0.733.

[0085] The technical solution of the medical image processing method provided in this embodiment of the invention can determine preset blood flow parameters by using a preset estimation model, or estimate the preset estimation model based on variational Bayes to determine the distribution of preset blood flow parameters. The distribution of preset blood flow parameters can be used to determine the reliability of preset blood flow parameters determined by dynamic contrast-enhanced CT, the accuracy of scanning protocol settings, or to compare the advantages and disadvantages of different image processing methods.

[0086] Example 3

[0087] Figure 3 This is a structural block diagram of a medical image processing apparatus provided in an embodiment of the present invention. The apparatus is used to execute the medical image processing method provided in any of the above embodiments, and the apparatus may be implemented in software or hardware. The apparatus includes:

[0088] The acquisition module 11 is used to determine the arterial input function of the artery used to supply blood to the target site and the attenuation coefficient enhancement information of the single voxel corresponding to the artery based on multiple sets of dynamically enhanced CT images acquired at different time points.

[0089] The blood flow parameter determination module 12 is used to input the arterial input function and attenuation coefficient enhancement information into the preset estimation model to obtain the preset blood flow parameters of the target site. The preset blood flow parameters include blood flow velocity and average flow time. The preset estimation model is constructed based on the box attenuation function.

[0090] like Figure 4 As shown, the device also includes an input module 10, which is used to: determine the average pixel value in the carotid artery region of multiple sets of dynamic contrast-enhanced CT images of the brain acquired at different time points, and take the relative enhancement corresponding to the average pixel value as the first relative enhancement; determine the average pixel value in the jugular vein region of the multiple sets of dynamic contrast-enhanced CT images of the brain, and take the relative enhancement corresponding to the average pixel value as the second relative enhancement; construct an arterial input function curve corresponding to the first relative enhancement and a venous output function curve corresponding to the second relative enhancement with the acquisition time as the horizontal axis and the relative enhancement as the vertical axis; correct the arterial input function curve according to the area of ​​the curve corresponding to the venous output function curve to update the arterial input function curve; and determine the arterial input function according to the updated arterial input function curve.

[0091] like Figure 4 As shown, the device also includes an estimation module 13, which is used to determine the update function of the mean and variance of the preset blood flow parameters determined by the preset estimation model using a Bayesian machine learning method; and to determine the distribution of the preset blood flow parameters based on the update function of the variance and mean of the preset blood flow parameters. The distribution of these preset blood flow parameters can be used to determine the reliability of the preset blood flow parameters determined by dynamic contrast-enhanced CT, the accuracy of the scanning protocol settings, or to compare the advantages and disadvantages of different image processing methods.

[0092] Optionally, the estimation module 13 is further configured to determine a first coefficient of variation representing the uncertainty of blood flow velocity based on the variance and mean of blood flow velocity; and to determine a second coefficient of variation representing the uncertainty of average flow time based on the variance and mean of average flow time.

[0093] The technical solution of the medical image processing device provided in this invention involves an acquisition module that determines the arterial input function of the artery supplying blood to the target site and the attenuation coefficient enhancement information of the individual voxels corresponding to the artery based on multiple sets of dynamically enhanced CT images acquired at different time points. A blood flow parameter determination module inputs the arterial input function and attenuation coefficient enhancement information into a preset estimation model to obtain preset blood flow parameters for the target site. These preset blood flow parameters include blood flow velocity and average flow time. The preset estimation model is constructed based on a box-shaped attenuation function. Because the preset estimation model based on the box-shaped attenuation function is more consistent with reality than the ideal physiological model required for singular value decomposition, the accuracy of the preset blood flow parameters determined based on this preset estimation model is higher.

[0094] The medical image processing device provided in the embodiments of the present invention can execute the medical image processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0095] Example 4

[0096] Figure 5 This is a schematic diagram of the image processing device provided in Embodiment 4 of the present invention, as shown below. Figure 5 As shown, the device includes a processor 201, a memory 202, an input device 203, and an output device 204; the number of processors 201 in the device can be one or more. Figure 5 Taking a processor 201 as an example; the processor 201, memory 202, input device 203, and output device 204 in the device can be connected via a bus or other means. Figure 5 Taking the example of a connection between China and Israel via a bus.

[0097] The memory 202, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the medical image processing method in this embodiment of the invention (e.g., acquisition module 11 and blood flow parameter determination module 12). The processor 201 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 202, thereby implementing the aforementioned medical image processing method.

[0098] The memory 202 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function; the data storage area may store data created based on terminal usage. Furthermore, the memory 202 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 202 may further include memory remotely located relative to the processor 201, which can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0099] The input device 203 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the device.

[0100] The output device 204 may include a display device such as a display screen, for example, the display screen of a user terminal, for at least outputting preset blood flow parameters and / or the distribution of preset blood flow parameters.

[0101] Example 5

[0102] This invention also provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform a medical image processing method, the method comprising:

[0103] Based on multiple sets of dynamically enhanced CT images acquired at different time points, the arterial input function of the artery used to supply blood to the target site and the attenuation coefficient enhancement information of the individual voxel corresponding to the artery are determined.

[0104] The arterial input function and the attenuation coefficient enhancement information are input into a preset estimation model to obtain preset blood flow parameters for the target site. The preset blood flow parameters include blood flow velocity and average flow time. The preset estimation model is constructed based on a box-shaped attenuation function.

[0105] Of course, the computer-executable instructions provided in the embodiments of the present invention are not limited to the method operations described above, but can also perform related operations in the medical image processing method provided in any embodiment of the present invention.

[0106] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the medical image processing methods described in the various embodiments of the present invention.

[0107] It is worth noting that in the embodiments of the above-mentioned medical image processing device, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of the present invention.

[0108] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A medical image processing method, characterized in that, include: Based on multiple sets of dynamically enhanced CT images acquired at different time points, the arterial input function of the artery used to supply blood to the target site and the attenuation coefficient enhancement information of the individual voxel corresponding to the artery are determined. The arterial input function and the attenuation coefficient enhancement information are input into a preset estimation model to obtain preset blood flow parameters for the target site. The preset blood flow parameters include blood flow velocity and average flow time. The preset estimation model is constructed based on a box-shaped attenuation function. Wherein, the target site is the brain, the artery is the carotid artery, and the corresponding method for determining the arterial input function includes: The average pixel value in the carotid artery region of multiple sets of dynamic contrast-enhanced CT images of the brain acquired at different time points is determined, and the relative enhancement corresponding to the average pixel value is taken as the first relative enhancement. The average pixel value in the jugular vein region of the multiple sets of dynamic contrast-enhanced CT images of the brain is determined, and the relative enhancement corresponding to the average pixel value is taken as the second relative enhancement. With acquisition time as the horizontal axis and relative enhancement as the vertical axis, an arterial input function curve corresponding to the first relative enhancement and a venous output function curve corresponding to the second relative enhancement are constructed. The area of ​​the curve corresponding to the arterial input function is the same as the area of ​​the curve corresponding to the venous output function. The arterial input function curve is corrected based on the area of ​​the curve corresponding to the venous output function curve, so as to update the arterial input function curve; Determine the arterial input function based on the updated arterial input function curve; The method for constructing the preset estimation model includes: The curve showing the change of the attenuation coefficient enhancement information of a single voxel over time is as follows: （1）； Where µ is the blood flow parameter to be estimated. For the scaling residual function of blood flow, For arterial input function, For noise; The scaling residual function of the blood flow is: （2）； µ includes blood flow velocity and mean flow time. For blood flow velocity, For tissue density, It is a residual function; Will Constructed as a function based on a box-shaped decay function, the scaling residual function of the blood flow is: （3）； Where MTT is the average flow time. t represents the acquisition time of the dynamically enhanced CT image; The convolution unrolling of the curve showing the attenuation coefficient enhancement information of a single voxel over time is a sampling of the integral of the arterial input function, as follows: （4）。 2. The method according to claim 1, characterized in that, Also includes: The product of the blood flow velocity and the average flow time is calculated as the blood volume.

3. The method according to claim 1, characterized in that, Also includes: The update function for the mean and variance of the preset blood flow parameters determined by the preset estimation model is determined by Bayesian machine learning. The distribution of the preset blood flow parameters is determined based on the update function of the variance and mean of the preset blood flow parameters.

4. The method according to claim 3, characterized in that, Also includes: A first coefficient of variation is determined based on the variance and mean of the blood flow velocity to represent the uncertainty of the blood flow velocity; A second coefficient of variation is determined based on the variance and mean of the average flow time to represent the uncertainty of the average flow time.

5. A medical image processing device, characterized in that, include: The acquisition module is used to determine the arterial input function of the artery used to supply blood to the target site and the attenuation coefficient enhancement information of the single voxel corresponding to the artery based on multiple sets of dynamically enhanced CT images acquired at different time points. The blood flow parameter determination module is used to input the arterial input function and the attenuation coefficient enhancement information into a preset estimation model to obtain preset blood flow parameters of the target site. The preset blood flow parameters include blood flow velocity and average flow time. The preset estimation model is constructed based on a box attenuation function. The input module is used to determine the average pixel value in the carotid artery region of multiple sets of dynamic contrast-enhanced CT images of the brain acquired at different time points, and to use the relative enhancement corresponding to the average pixel value as the first relative enhancement. It also determines the average pixel value in the jugular vein region of the same multiple sets of dynamic contrast-enhanced CT images of the brain, and to use the relative enhancement corresponding to the average pixel value as the second relative enhancement. Using acquisition time as the horizontal axis and relative enhancement as the vertical axis, it constructs an arterial input function curve corresponding to the first relative enhancement and a venous output function curve corresponding to the second relative enhancement. The area of ​​the curve corresponding to the arterial input function is the same as the area of ​​the curve corresponding to the venous output function. The arterial input function curve is corrected based on the area of ​​the curve corresponding to the venous output function curve to update the arterial input function curve. Finally, the arterial input function is determined based on the updated arterial input function curve. The method for constructing the preset estimation model includes: The curve showing the change of the attenuation coefficient enhancement information of a single voxel over time is as follows: （1）； Where µ is the blood flow parameter to be estimated. For the scaling residual function of blood flow, For arterial input function, For noise; The scaling residual function of the blood flow is: （2）； µ includes blood flow velocity and mean flow time. For blood flow velocity, For tissue density, It is a residual function; Will Constructed as a function based on a box-shaped decay function, the scaling residual function of the blood flow is: （3）； Where MTT is the average flow time. t represents the acquisition time of the dynamically enhanced CT image; The convolution unrolling of the curve showing the attenuation coefficient enhancement information of a single voxel over time is a sampling of the integral of the arterial input function, as follows: （4）。 6. An image processing device, characterized in that, The image processing device includes: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the medical image processing method as described in any one of claims 1-4.

7. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the medical image processing method as described in any one of claims 1-4.

Images