Method, device, equipment and storage medium for distinguishing calcification point types

Abstract

The present invention relates to the field of medical image processing technology, and discloses a method, device, equipment and storage medium for distinguishing types of calcification points, the method comprising: obtaining target signals of multiple angles of breast tissue, the target signals comprising absorption signals, scattering signals and / or refraction signals; the target signals of multiple angles are obtained based on target rays, the target rays are rays after source rays of multiple different angles act on the breast tissue; according to the target signals, three-dimensional CT reconstruction of the breast tissue is performed to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue; according to the three-dimensional CT values ​​of the calcification points in the breast tissue, the types of the calcification points are determined. The present invention can easily distinguish the types of calcification points in breast tissue, and the method is simple and the distinction accuracy is high.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a method, device, equipment and storage medium for distinguishing calcification point types. Background Art

[0002] Breast calcifications are dot-like high-density shadows that appear on breast images during breast imaging examinations. The formation of this high-density shadow is caused by the increase in calcium content in local tissue cells. There are two main types of calcifications: Type I calcifications, whose main components are calcium oxalate dihydrate (CaC2O4·2H2O); and Type II calcifications, whose main calcium phosphate components are hydroxyapatite (Ca5(PO4)3(OH)). These two types of calcifications are associated with benign breast lesions and malignant breast lesions, respectively, so it is necessary to accurately distinguish the types of calcifications. However, in related technologies, it is difficult to quickly and accurately distinguish the types of calcifications on breast tissue. Summary of the invention

[0003] In view of this, the present invention provides a method, device, equipment and storage medium for distinguishing the types of calcification points, so as to solve the problem that it is difficult to quickly and accurately distinguish the types of calcification points on breast tissue.

[0004] In a first aspect, the present invention provides a method for distinguishing calcification point types, the method comprising:

[0005] Acquire target signals of breast tissue at multiple angles, the target signals including absorption signals, scattering signals and / or refraction signals; the target signals at multiple angles are acquired based on target rays, the target rays are rays after source rays at multiple different angles act on the breast tissue; the absorption signal is the remaining ray signal after the source ray passes through the breast tissue and is partially absorbed by the breast tissue, the scattering signal is the ray signal of the source ray that deviates from the original propagation direction and propagates in a dispersed manner after passing through the breast tissue, and the refraction signal is the ray signal of the source ray that changes its propagation direction at the boundary of different tissue structures during the process of the source ray passing through the breast tissue;

[0006] Performing three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue;

[0007] The type of the calcification point is determined according to the three-dimensional CT value of the calcification point in the breast tissue.

[0008] In an optional implementation, performing the three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain the three-dimensional CT values ​​of different tissue structures in the breast tissue includes:

[0009] Construct three-dimensional volume space;

[0010] For the target signal at each angle, back-project the target signal onto a corresponding voxel in the three-dimensional volume space along the reverse direction of the ray path;

[0011] The back-projection results of the target signal at each angle in the three-dimensional volume space are accumulated to obtain the three-dimensional CT values ​​of different tissue structures in the breast tissue.

[0012] In an optional implementation, the target signal at each angle is back-projected onto a corresponding voxel in the three-dimensional volume space along the reverse direction of the ray path, including:

[0013] Generating input information of a back-projection model for the target signal at each angle respectively;

[0014] Input the input information into the pre-trained back-projection model; wherein the back-projection model is constructed based on the UNet model, the back-projection model includes an encoding module and a decoding module, the encoding module includes a three-dimensional convolution block, a residual block and a downsampling block, the decoding module includes a three-dimensional convolution block, a residual block and an upsampling block, and the three-dimensional convolution blocks in the encoding module and the decoding module are EfficientNet;

[0015] According to the output result of the back-projection model, a target voxel of the target signal back-projected on the three-dimensional volume space and a value of the target voxel are determined.

[0016] In an optional implementation, the back-projection model further includes a spatial attention module, the spatial attention module obtains a spatial attention weight value based on the input information, the spatial attention weight value is used to multiply the input information to obtain the weighted input information, the encoding module performs feature extraction based on the weighted input information, and the decoding module performs back-projection prediction based on the features extracted by the encoding module;

[0017] The spatial attention module includes a maximum pooling layer, an average pooling layer and a shared fully connected layer.

[0018] In an optional embodiment, determining the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue includes:

[0019] In the case where the target signal includes the absorption signal, the scattering signal, and the refraction signal, if a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the absorption signal is within a first preset range, a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the scattering signal is within a second preset range, and a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the refraction signal is within a third preset range, then it is determined that the calcification point is a first calcification point type; and the first preset range, the second preset range, and the third preset range correspond to the first calcification point type;

[0020] If the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the absorption signal is within a fourth preset range, the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the scattering signal is within a fifth preset range, and the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the refraction signal is within a sixth preset range, then the calcification point is determined to be a second calcification point type; the fourth preset range, the fifth preset range, and the sixth preset range correspond to the second calcification point type.

[0021] In an optional embodiment, determining the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue includes:

[0022] In a case where the target signal includes at least two of the absorption signal, the scattering signal and the refraction signal, a three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the at least two target signals respectively;

[0023] Acquire a ratio of the three-dimensional CT values ​​of the calcification point obtained after three-dimensional reconstruction of two different target signals;

[0024] The type of the calcification point is determined according to the magnitude relationship between the ratio and a preset threshold.

[0025] In a second aspect, the present invention provides a device for distinguishing calcification point types, the device comprising:

[0026] an acquisition module, for acquiring target signals of breast tissue at multiple angles, the target signals including absorption signals, scattering signals and / or refraction signals; the target signals at multiple angles are acquired based on target rays, the target rays are rays after source rays at multiple different angles act on the breast tissue; the absorption signal is the remaining ray signal after the source ray passes through the breast tissue and is partially absorbed by the breast tissue, the scattering signal is the ray signal of the source ray that deviates from the original propagation direction and propagates in a dispersed manner after passing through the breast tissue, and the refraction signal is the ray signal of the source ray that changes its propagation direction at the boundary of different tissue structures during the process of the source ray passing through the breast tissue;

[0027] A reconstruction module, used for performing three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue;

[0028] The type distinguishing module is used to determine the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue.

[0029] In a third aspect, the present invention provides a three-dimensional CT device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for distinguishing calcification point types according to the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0030] In an optional embodiment, a source grating, a phase grating and an absorption grating are sequentially arranged on the ray optical path, the source grating is used to divide the source ray beam emitted by the ray source into multiple coherent sub-ray beams, the phase grating is used to generate a self-imaging pattern, and the absorption grating is used to generate moiré fringes.

[0031] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for distinguishing calcification point types according to the first aspect or any corresponding embodiment thereof.

[0032] In a fifth aspect, the present invention provides a computer program product, including computer instructions, where the computer instructions are used to enable a computer to execute the method for distinguishing calcification point types according to the first aspect or any corresponding embodiment thereof.

[0033] The method, device, equipment and storage medium for distinguishing the types of calcification points provided by the embodiments of the present invention use three-dimensional information to distinguish the types of calcification points in breast tissue. The method first realizes three-dimensional CT reconstruction of breast tissue (based on absorption signals, scattering signals and / or refraction signals). The reconstructed CT value can truly reflect the absorption, scattering and / or refraction of X-rays by different tissue structures. Then, type I calcification points and type II calcification points can be distinguished based on the effect of different types of calcification points on signals of different mechanisms of X-rays, which is not only fast and convenient, but also accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the related technologies, the drawings required for use in the specific embodiments or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 FIG. 4 is a flow chart of a method for distinguishing calcification point types according to an embodiment of the present invention.

[0036] Figure 2 4 is a structural block diagram of a device for distinguishing calcification point types according to an embodiment of the present invention.

[0037] Figure 3 It is a schematic diagram of the hardware structure of a three-dimensional CT device according to an embodiment of the present invention.

[0038] Figure 4 Schematic diagram of the relative positions of optical elements of a three-dimensional CT device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0040] In the X-ray band, type I calcification points and type II calcification points have different absorption, refraction and scattering of X-rays. Therefore, the absorption, refraction and scattering information of calcification points can be used to distinguish between type I and type II calcification points. However, X-ray two-dimensional imaging technology is projection imaging, and the image is superimposed with the absorption information of all tissue structures on the X-ray projection path. Therefore, it is impossible to obtain information such as the absorption of calcification points from a single two-dimensional image. If you want to obtain information such as the absorption of rays at calcification points, you need to use two (or more) two-dimensional images with different mechanisms or different conditions to eliminate the superimposed projection effects of different tissues. For example, you can collect low-energy projection images and high-energy projection images respectively, and then make a ratio of the intensities of calcification points at the two energies. In this way, the projection superposition effect is suppressed. Therefore, by setting the threshold, you can distinguish the types of calcification points.

[0041] The embodiment of the present invention provides a method for distinguishing the types of calcification points in breast tissue by using three-dimensional information. The method first realizes three-dimensional CT reconstruction of breast tissue (based on absorption signals, scattering signals and / or refraction signals). The reconstructed CT value can truly reflect the absorption, scattering and / or refraction of X-rays by different tissue structures. Then, type I calcification points and type II calcification points can be distinguished according to the effect of different types of calcification points on signals of different mechanisms of X-rays, which is not only fast and convenient, but also accurate.

[0042] According to an embodiment of the present invention, an embodiment of a method for distinguishing calcification point types is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of executable computer instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0043] In this embodiment, a method for distinguishing calcification point types is provided, which can be used in various computer devices, including three-dimensional CT devices and other computer devices independent of the three-dimensional CT devices. Figure 1 FIG. 4 is a flow chart of a method for distinguishing calcification point types according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0044] Step S101, acquiring target signals of breast tissue at multiple angles, wherein the target signals include absorption signals, scattering signals and / or refraction signals; the target signals at multiple angles are acquired based on target rays, and the target rays are rays after source rays at multiple different angles act on the breast tissue; the absorption signal is the ray signal remaining after the source ray passes through the breast tissue and is partially absorbed by the breast tissue, the scattering signal is the ray signal of the source ray that deviates from the original propagation direction and propagates in a dispersed manner after passing through the breast tissue, and the refraction signal is the ray signal of the source ray that changes its propagation direction at the boundary of different tissue structures during the process of the source ray passing through the breast tissue.

[0045] Specifically, the above-mentioned rays can be X-rays or other rays. The source ray is a ray emitted from a ray source (specifically, an X-ray tube) to the breast tissue of the subject from multiple different angles (specifically, multiple angles determined at preset angle intervals within a range of 360 degrees). During the examination, the subject lies on the examination bed of the three-dimensional CT device and places the breast tissue in the ray field. During the ray source beam emission process, the rack carries the ray source and the detector (i.e., the ray detection device) and rotates one circle. During the rotation process, the ray source emits source rays at multiple rotation angles, and the detector receives the rays (including absorbed rays, scattered rays and / or refracted rays) acting on the breast tissue in real time to obtain corresponding ray signals (including absorbed signals, scattered signals and / or refracted signals). In addition, before the ray is emitted from the ray source and reaches the breast tissue, it will pass through a phase modulator, which can be a grating or a crystal, etc. This phase modulator will divide the ray wavefront into multiple coherent sub-waves and introduce a certain phase difference. When the radiation passes through breast tissue, the phase of the radiation changes due to the different refractive indices of the radiation in different tissue structures. These phase changes are detected and quantified by detectors. This is usually achieved through interferometry techniques, such as using another set of gratings or interferometers to capture the information of phase changes.

[0046] Regarding the absorbed rays (corresponding to the absorption signal), after the rays emitted by the radiation source pass through the breast tissue, the differences in density, atomic number, etc. of different tissue structures will lead to differences in the absorption of the rays. By detecting (using a detector) the remaining rays emitted from the breast tissue, the absorption of the rays by the breast tissue can be reflected.

[0047] Step S102: performing three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue.

[0048] In the embodiment of the present invention, three-dimensional CT reconstruction needs to be performed separately for different target signals. That is to say, if the target signal includes an absorption signal, then three-dimensional CT reconstruction needs to be performed based on the absorption signal; if the target signal includes a scattering signal, then three-dimensional CT reconstruction needs to be performed based on the scattering signal; if the target signal also includes a refraction signal, then three-dimensional CT reconstruction needs to be performed based on the refraction signal.

[0049] In some specific implementations, step S102, i.e., performing three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue, includes:

[0050] Step S1021, constructing a three-dimensional volume space coordinate system.

[0051] Step S1022: for the target signal at each angle, back-project the target signal onto a corresponding line segment of the three-dimensional volume space coordinate system along the reverse direction of the ray path.

[0052] Specifically, the back projection may be a convolution operation, in which case it may be referred to as convolution back projection.

[0053] Before back-projection, the target signal can be pre-processed, which can specifically include denoising, smoothing, and inhomogeneity correction, etc., to improve the quality of three-dimensional CT reconstruction. In addition, the target signal can be resampled to a unified coordinate system as needed to facilitate subsequent back-projection and other processing. In order to avoid the problem of blurring of the reconstructed three-dimensional CT image caused by simple back-projection, the target signal can also be filtered before back-projection, for example, Laplace filtering or a filter function (such as Shepp-Logan filter) can be used to multiply the Fourier transformed target signal.

[0054] In some specific implementations, step S1022, that is, back-projecting the target signal at each angle onto a corresponding voxel in the three-dimensional volume space along the reverse direction of the ray path, specifically includes:

[0055] Step S10221, generating input information of a back-projection model for the target signal at each angle respectively;

[0056] Step S10222, inputting the input information into the pre-trained back-projection model; wherein the back-projection model is constructed based on the UNet model, the back-projection model includes an encoding module and a decoding module, the encoding module includes a three-dimensional convolution block, a residual block and a down-sampling block, the decoding module includes a three-dimensional convolution block, a residual block and an up-sampling block, and the three-dimensional convolution blocks in the encoding module and the decoding module are EfficientNet;

[0057] That is, when building the back-projection model based on the UNet model, the convolutional blocks in the UNet model are replaced with EfficientNet, which is an efficient convolutional neural network (CNN) model;

[0058] Step S10223: Determine the target voxel of the target signal back-projected in the three-dimensional volume space and the value of the target voxel according to the output result of the back-projection model.

[0059] Specifically, the back-projection model also includes a spatial attention module, the spatial attention module obtains a spatial attention weight value based on the input information, the spatial attention weight value is used to multiply the input information to obtain the weighted input information, the encoding module performs feature extraction based on the weighted input information, and the decoding module performs back-projection prediction based on the features extracted by the encoding module;

[0060] The spatial attention module includes a maximum pooling layer, an average pooling layer and a shared fully connected layer.

[0061] Step S1023, accumulating the back-projection results of the target signal at each angle in the three-dimensional volume space coordinate system to obtain the three-dimensional CT values ​​of different tissue structures in the breast tissue.

[0062] In the embodiment of the present invention, the back-projection results of the target signal at each angle in the three-dimensional volume space coordinate system are accumulated to form a three-dimensional accumulated voxel data set, which is the reconstructed three-dimensional CT image.

[0063] The reconstructed three-dimensional CT image may also be subjected to some post-processing to further improve the quality of the reconstructed three-dimensional CT image. The post-processing may include, for example, contrast enhancement, edge sharpening, and the like.

[0064] The embodiment of the present invention utilizes deep learning technology to perform back-projection of two-dimensional CT signals, which is not only highly efficient but also highly accurate and reliable.

[0065] Step S103: determining the type of the calcification point in the breast tissue according to the three-dimensional CT value of the calcification point.

[0066] In some specific implementations, step S103, i.e., determining the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue, includes:

[0067] In the case where the target signal includes the absorption signal, the scattering signal, and the refraction signal, if a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the absorption signal is within a first preset range, a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the scattering signal is within a second preset range, and a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the refraction signal is within a third preset range, then it is determined that the calcification point is the first calcification point type; and the first preset range, the second preset range, and the third preset range correspond to the first calcification point type;

[0068] If the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the absorption signal is within a fourth preset range, the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the scattering signal is within a fifth preset range, and the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the refraction signal is within a sixth preset range, then the calcification point is determined to be the second calcification point type; the fourth preset range, the fifth preset range, and the sixth preset range correspond to the second calcification point type.

[0069] Specifically, the first preset range, the second preset range and the third preset range need to be determined based on the optical characteristics (absorption, scattering and refraction) of the calcification points of the first calcification point type to the rays and the intensity of the source rays. The fourth preset range, the fifth preset range and the sixth preset range need to be determined based on the optical characteristics (absorption, scattering and refraction) of the calcification points of the second calcification point type to the rays and the intensity of the source rays.

[0070] This embodiment obtains different target signals (absorption signal, scattering signal and refraction signal) based on the different optical properties (absorption, scattering and refraction) of calcification points to rays, and then performs three-dimensional CT reconstruction based on the different target signals to obtain the quantitative values ​​of the absorption coefficient, scattering coefficient and refraction coefficient of each point in the breast tissue, and finally realizes the distinction of the types of calcification points based on the quantitative values ​​of the absorption coefficient, scattering coefficient and refraction coefficient of the calcification points.

[0071] In another optional specific implementation, step S103, i.e., determining the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue, includes:

[0072] Step S103a, when the target signal includes at least two of the absorption signal, the scattering signal and the refraction signal, performing three-dimensional CT reconstruction based on at least two of the target signals to obtain a three-dimensional CT value of the calcification point;

[0073] Step S103b, obtaining a ratio of the three-dimensional CT values ​​of the calcification point obtained after three-dimensional reconstruction of two different target signals;

[0074] Step S103c: determining the type of the calcification point according to the magnitude relationship between the ratio and a preset threshold.

[0075] Specifically, the preset threshold is used to determine the range of the ratio corresponding to different types of calcification points, and there may be multiple preset thresholds, which may be determined specifically according to the optical properties (absorption, scattering and refraction) of different types of calcification points to rays.

[0076] In the embodiment of the present invention, if the target signal includes only two of the above, then only one ratio may be obtained. If the target signal includes all three, then two ratios or three ratios may be obtained.

[0077] In summary, the embodiment of the present invention performs three-dimensional CT reconstruction based on the absorption ray signal, scattered ray signal and refracted ray signal after the ray acts on the breast tissue, and the reconstructed three-dimensional CT value can truly and accurately reflect the optical characteristics of the various structures of the breast tissue to the ray. The two different types of calcification points, type I and type II, have great differences in various signals of the ray, so it is easy to distinguish these two different types of calcification points by using the differences in the optical characteristics of the different types of calcification points.

[0078] In addition, the method for distinguishing the types of calcification points provided in the embodiment of the present invention does not need to obtain images of additional imaging mechanisms or imaging conditions to couple signals to achieve the distinction between different types of calcification points. Only one inspection operation is required to obtain the target signal of breast tissue (including absorption signal, scattering signal and / or refraction signal), and three-dimensional CT reconstruction is performed based on the target signal to achieve the distinction between different types of calcification points. The distinction method is simple and the inspection operation is also simple.

[0079] In this embodiment, a device for distinguishing the types of calcification points is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made are not repeated here. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0080] This embodiment provides a device for distinguishing the types of calcification points, such as Figure 2 As shown, including:

[0081] An acquisition module 201 is used to acquire target signals of breast tissue at multiple angles, wherein the target signals include absorption signals, scattering signals and / or refraction signals; the target signals at multiple angles are acquired based on target rays, wherein the target rays are rays of source rays at multiple different angles after acting on the breast tissue; the absorption signal is a ray signal remaining after the source ray passes through the breast tissue and is partially absorbed by the breast tissue, the scattering signal is a ray signal of the source ray that deviates from the original propagation direction and propagates in a dispersed manner after passing through the breast tissue, and the refraction signal is a ray signal of the source ray that changes its propagation direction at the boundary of different tissue structures during the process of the source ray passing through the breast tissue;

[0082] A reconstruction module 202 is used to perform three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue;

[0083] The type distinguishing module 203 is used to determine the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue.

[0084] In some optional implementations, the reconstruction module 202 includes:

[0085] A three-dimensional volume space construction unit, used for constructing a three-dimensional volume space;

[0086] A back-projection unit, configured to back-project the target signal at each angle onto a corresponding voxel in the three-dimensional volume space along a reverse direction of a ray path;

[0087] The accumulation unit is used to accumulate the back-projection results of the target signal at each angle in the three-dimensional volume space to obtain the three-dimensional CT values ​​of different tissue structures in the breast tissue.

[0088] In some optional embodiments, the back projection unit includes:

[0089] An information generating subunit, used for generating input information of a back-projection model for the target signal at each angle;

[0090] A model prediction subunit, used for inputting the input information into the pre-trained back-projection model; wherein the back-projection model is constructed based on the UNet model, the back-projection model includes an encoding module and a decoding module, the encoding module includes a three-dimensional convolution block, a residual block and a down-sampling block, the decoding module includes a three-dimensional convolution block, a residual block and an up-sampling block, and the three-dimensional convolution blocks in the encoding module and the decoding module are EfficientNet;

[0091] A determination subunit is used to determine the target voxel of the back-projection of the target signal in the three-dimensional volume space and the value of the target voxel according to the output result of the back-projection model.

[0092] In some optional implementations, the back-projection model further includes a spatial attention module, the spatial attention module obtains a spatial attention weight value based on the input information, the spatial attention weight value is used to multiply the input information to obtain the weighted input information, the encoding module performs feature extraction based on the weighted input information, and the decoding module performs back-projection prediction based on the features extracted by the encoding module;

[0093] The spatial attention module includes a maximum pooling layer, an average pooling layer and a shared fully connected layer.

[0094] In some optional embodiments, the type distinguishing module 203 is specifically used to, when the target signal includes the absorption signal, the scattering signal and the refraction signal, determine that the calcification point is a first calcification point type if the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the absorption signal is within a first preset range, the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the scattering signal is within a second preset range, and the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the refraction signal is within a third preset range; the first preset range, the second preset range and the third preset range correspond to the first calcification point type; if the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the absorption signal is within a fourth preset range, the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the scattering signal is within a fifth preset range, and the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the refraction signal is within a sixth preset range, determine that the calcification point is a second calcification point type; the fourth preset range, the fifth preset range and the sixth preset range correspond to the second calcification point type.

[0095] In some other optional embodiments, the type distinction module 203 is specifically used to, when the target signal includes at least two of the absorption signal, the scattering signal and the refraction signal, obtain the three-dimensional CT value of the calcification point based on at least two of the target signals after three-dimensional CT reconstruction; obtain the ratio of the three-dimensional CT values ​​of the calcification point obtained after three-dimensional reconstruction of two different target signals; and determine the type of the calcification point based on the relationship between the ratio and a preset threshold.

[0096] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0097] The calcification point type distinguishing device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0098] The embodiment of the present invention also provides a three-dimensional CT device having the above Figure 2 The device for distinguishing the types of calcification points shown.

[0099] See also Figure 3 , Figure 3 is a schematic diagram of the structure of a three-dimensional CT device provided by an optional embodiment of the present invention, such as Figure 3 As shown, the three-dimensional CT device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses for communication, and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the three-dimensional CT device, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Figure 3 A processor 10 is taken as an example.

[0100] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0101] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0102] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required by at least one function; the data storage area may store data created according to the use of the three-dimensional CT device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the three-dimensional CT device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0103] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0104] The three-dimensional CT device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 can be connected via a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0105] The input device 30 can receive input digital or character information and generate key signal input related to the user settings and function control of the three-dimensional CT device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator rod, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device may be a touch screen.

[0106] The three-dimensional CT device also includes a communication interface, which is used for the three-dimensional CT device to communicate with other devices or a communication network.

[0107] Also, see Figure 4 A source grating G0, a phase grating G1 and an absorption grating G2 are sequentially arranged on the ray optical path of the three-dimensional CT device. The source grating G0 is used to divide the source ray beam emitted by the ray source into multiple coherent sub-ray beams, the phase grating G1 is used to generate a self-imaging pattern, and the absorption grating G2 is used to generate moiré fringes.

[0108] In this embodiment, the absorption grating G2 rotates around the optical axis, and differs from the self-imaging pattern generated by the phase grating G1 by a small angle, so as to form moiré fringes on the detector that are much larger than the grating period, so that they can be directly resolved by the detector. The source grating G0 solves the problem of insufficient coherence of conventional ray sources.

[0109] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0110] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0111] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for distinguishing calcification point types, characterized in that: The method comprises: Acquire target signals of breast tissue at multiple angles, the target signals including absorption signals, scattering signals and / or refraction signals; the target signals at multiple angles are acquired based on target rays, the target rays are rays after source rays at multiple different angles act on the breast tissue; the absorption signal is the remaining ray signal after the source ray passes through the breast tissue and is partially absorbed by the breast tissue, the scattering signal is the ray signal of the source ray that deviates from the original propagation direction and propagates in a dispersed manner after passing through the breast tissue, and the refraction signal is the ray signal of the source ray that changes its propagation direction at the boundary of different tissue structures during the process of the source ray passing through the breast tissue; Performing three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue; Determining the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue; Wherein, performing the three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain the three-dimensional CT values ​​of different tissue structures in the breast tissue includes: Constructing a three-dimensional volume space; for the target signal at each angle, back-projecting the target signal onto a corresponding voxel in the three-dimensional volume space along the reverse direction of the ray path; accumulating the back-projection results of the target signal at each angle in the three-dimensional volume space to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue; For the target signal at each angle, back-projecting the target signal onto a corresponding voxel in the three-dimensional volume space along the reverse direction of the ray path includes: Generating input information of a back-projection model for the target signal at each angle respectively; Input the input information into the pre-trained back-projection model; wherein the back-projection model is constructed based on the UNet model, the back-projection model includes an encoding module and a decoding module, the encoding module includes a three-dimensional convolution block, a residual block and a downsampling block, the decoding module includes a three-dimensional convolution block, a residual block and an upsampling block, and the three-dimensional convolution blocks in the encoding module and the decoding module are EfficientNet; Determining, according to an output result of the back-projection model, a target voxel of the back-projection of the target signal on the three-dimensional volume space and a value of the target voxel; Determining the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue includes: In the case where the target signal includes the absorption signal, the scattering signal, and the refraction signal, if a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the absorption signal is within a first preset range, a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the scattering signal is within a second preset range, and a three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the refraction signal is within a third preset range, then it is determined that the calcification point is a first calcification point type; and the first preset range, the second preset range, and the third preset range correspond to the first calcification point type; If the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the absorption signal is within a fourth preset range, the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the scattering signal is within a fifth preset range, and the three-dimensional CT value of the calcification point obtained by performing three-dimensional CT reconstruction based on the refraction signal is within a sixth preset range, then the calcification point is determined to be a second calcification point type; the fourth preset range, the fifth preset range, and the sixth preset range correspond to the second calcification point type.

2. The method according to claim 1, characterized in that The back-projection model also includes a spatial attention module, the spatial attention module obtains a spatial attention weight value based on the input information, the spatial attention weight value is used to multiply the input information to obtain the weighted input information, the encoding module performs feature extraction based on the weighted input information, and the decoding module performs back-projection prediction based on the features extracted by the encoding module; The spatial attention module includes a maximum pooling layer, an average pooling layer and a shared fully connected layer.

3. A device for distinguishing calcification point types, characterized in that: The device comprises: an acquisition module, for acquiring target signals of breast tissue at multiple angles, the target signals including absorption signals, scattering signals and / or refraction signals; the target signals at multiple angles are acquired based on target rays, the target rays are rays after source rays at multiple different angles act on the breast tissue; the absorption signal is the remaining ray signal after the source ray passes through the breast tissue and is partially absorbed by the breast tissue, the scattering signal is the ray signal of the source ray that deviates from the original propagation direction and propagates in a dispersed manner after passing through the breast tissue, and the refraction signal is the ray signal of the source ray that changes its propagation direction at the boundary of different tissue structures during the process of the source ray passing through the breast tissue; A reconstruction module is used to perform three-dimensional CT reconstruction of the breast tissue according to the target signal to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue; the reconstruction module specifically includes: a three-dimensional volume space construction unit, which is used to construct a three-dimensional volume space; a back-projection unit, which is used to back-project the target signal at each angle to a corresponding voxel in the three-dimensional volume space along the reverse direction of the ray path; and an accumulation unit, which is used to accumulate the back-projection results of the target signal at each angle in the three-dimensional volume space to obtain three-dimensional CT values ​​of different tissue structures in the breast tissue; The back-projection unit specifically includes: an information generating subunit, which is used to generate input information of the back-projection model for the target signal at each angle; A model prediction subunit, used for inputting the input information into the pre-trained back-projection model; wherein the back-projection model is constructed based on the UNet model, the back-projection model includes an encoding module and a decoding module, the encoding module includes a three-dimensional convolution block, a residual block and a down-sampling block, the decoding module includes a three-dimensional convolution block, a residual block and an up-sampling block, and the three-dimensional convolution blocks in the encoding module and the decoding module are EfficientNet; a determination subunit, configured to determine a target voxel for back-projection of the target signal on the three-dimensional volume space and a value of the target voxel according to an output result of the back-projection model; A type distinguishing module, used to determine the type of the calcification point according to the three-dimensional CT value of the calcification point in the breast tissue; The type distinguishing module is specifically used for determining, when the target signal includes the absorption signal, the scattering signal and the refraction signal, that the calcification point is a first calcification point type if the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the absorption signal is within a first preset range, the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the scattering signal is within a second preset range, and the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the refraction signal is within a third preset range; the first preset range, the second preset range and the third preset range correspond to the first calcification point type; if the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the absorption signal is within a fourth preset range, the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the scattering signal is within a fifth preset range, and the three-dimensional CT value of the calcification point obtained by three-dimensional CT reconstruction based on the refraction signal is within a sixth preset range, the calcification point is determined to be a second calcification point type; the fourth preset range, the fifth preset range and the sixth preset range correspond to the second calcification point type.

4. A three-dimensional CT device, characterized in that: include: A memory and a processor, wherein the memory and the processor are connected to each other for communication, the memory stores computer instructions, and the processor executes the method for distinguishing the types of calcification points according to any one of claims 1 to 2 by executing the computer instructions.

5. The three-dimensional CT device according to claim 4, characterized in that: A source grating, a phase grating and an absorption grating are arranged in sequence on the ray optical path. The source grating is used to divide the source ray beam emitted by the ray source into multiple coherent sub-ray beams, the phase grating is used to generate a self-imaging pattern, and the absorption grating is used to generate moiré fringes.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for distinguishing calcification point types according to any one of claims 1 to 2.

Images