KR20230125908A - 2d x-ray generating method from 3d computed tomography data and image processing apparatus - Google Patents

Abstract

A method of generating a virtual X-ray image from 3D CT data includes receiving three-dimensional (3D) CT data by an image processing device, and receiving a first constant for determining a specific direction and beam intensity by the image processing device. and the image processing device generating a virtual X-ray image based on the amount of change in a beam attenuated in each of a plurality of slices according to the specific direction in the 3D CT data and the intensity of the beam determined by the first constant. Include steps.

Description

Method and image processing device for generating virtual X-ray images from 3D CT data

A technique described below relates to a technique of generating an X-ray image in a specific direction from 3D CT data. In particular, the technique described below is a technique for generating an X-ray image of a specific dose.

CT (computed tomography) is one of the representative medical images. A medical staff may analyze and diagnose a region to be imaged using 3D CT (computed tomography) data. Furthermore, a medical staff needs to check a 2D image of a specific direction (viewpoint) in a 3D CT. Conventional techniques generate a 2D image by adding or averaging pixels in a specific direction while projecting 3D stack CT data in a specific direction.

US registered patent US 8,755,584

In X-rays, when high-speed electrons collide with atoms, the speed of high-speed electrons decreases, and the contrast of the actual projected image changes according to the intensity of the source dose. Conventional techniques simply use 3D stack CT data to Since a method of calculating pixel values in a certain direction is used, all characteristics of an X-ray image cannot be contained in terms of radiation physics. In addition, the conventional technique did not consider intensity information of beams by using a method of simply summing CT data values.

The technology to be described below aims to provide a virtual X-ray image that well represents X-ray characteristics in terms of radiation physics by mathematically reflecting the decrease in radiation intensity when radiation collides with fine particles. In addition, the technology to be described below is intended to provide a virtual X-ray image according to a desired dose according to the needs of the user.

A method of generating a virtual X-ray image from 3D CT data includes receiving three-dimensional (3D) CT data by an image processing device, and receiving a first constant for determining a specific direction and beam intensity by the image processing device. and the image processing device generating a virtual X-ray image based on the amount of change in a beam attenuated in each of a plurality of slices according to the specific direction in the 3D CT data and the intensity of the beam determined by the first constant. Include steps.

An image processing device that generates a virtual X-ray image from 3D CT data is an input device that receives three-dimensional (3D) CT data, a specific direction and a first constant that determines the beam intensity, and a virtual X-ray image from 3D CT data A storage device for storing a program for generating an image and a virtual X-ray based on the amount of change in a beam attenuated in each of a plurality of slices according to the specific direction in the 3D CT data and the intensity of the beam determined by the first constant It includes an arithmetic device that generates an image.

In another aspect, a learning method of a learning model that converts low-dose X-rays into high-dose X-rays is a low-dose virtual X-ray image using the method of generating a virtual X-ray image from the 3D CT data described above by an image processing device. and generating a learning data set composed of high-dose virtual X-ray images, and the image processing device learning a learning model for converting a low-dose X-ray image into a high-dose X-ray image using the learning data set. Include steps.

The technology described below can generate a 2D image that well represents X-ray characteristics in terms of radiation physics from 3D stacked CT data. Furthermore, the technique described below may generate a virtual X-ray image having a desired specific dose by reflecting the intensity value of the imaging beam. Medical staff can perform accurate diagnosis and treatment by using the image of the dose suitable for the diagnosis site and disease characteristics. Furthermore, an artificial intelligence model may be constructed by using low-dose X-ray images and high-dose X-ray images generated by a technique described below as learning data.

1 is an example of a system for generating a virtual X-ray image from 3D CT data.
2 is an example of a process of generating a virtual X-ray image from 3D CT data.
3 is a result of generating a virtual X-ray image by changing the beam intensity.
4 is an example of a process of learning an X-ray reconstruction model using a virtual X-ray image.
5 is an image conversion result of a reconstruction model for converting a low-dose X-ray image into a high-dose X-ray image.
6 is an example of an image processing device generating a virtual X-ray image from 3D CT data.

Since the technology to be described below can have various changes and various embodiments, specific embodiments will be illustrated in the drawings and described in detail. However, this is not intended to limit the technology described below to specific embodiments, and it should be understood to include all modifications, equivalents, or substitutes included in the spirit and scope of the technology described below.

Terms such as first, second, A, B, etc. may be used to describe various elements, but the elements are not limited by the above terms, and are merely used to distinguish one element from another. used only as For example, without departing from the scope of the technology described below, a first element may be referred to as a second element, and similarly, the second element may be referred to as a first element. The terms and/or include any combination of a plurality of related recited items or any of a plurality of related recited items.

In terms used in this specification, singular expressions should be understood to include plural expressions unless clearly interpreted differently in context, and terms such as “comprising” refer to the described features, numbers, steps, operations, and components. , parts or combinations thereof, but it should be understood that it does not exclude the possibility of the presence or addition of one or more other features or numbers, step-action components, parts or combinations thereof.

Prior to a detailed description of the drawings, it is to be clarified that the classification of components in the present specification is merely a classification for each main function in charge of each component. That is, two or more components to be described below may be combined into one component, or one component may be divided into two or more for each more subdivided function. In addition, each component to be described below may additionally perform some or all of the functions of other components in addition to its main function, and some of the main functions of each component may be performed by other components. Of course, it may be dedicated and performed by .

In addition, in performing a method or method of operation, each process constituting the method may occur in a different order from the specified order unless a specific order is clearly described in context. That is, each process may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

The technique described below is a technique of generating a constant 2D image from 3D stacked CT data.

Hereinafter, a virtual X-ray image or X-ray image means a virtual 2D X-ray image generated from 3D stacked CT data.

Hereinafter, it will be described that the image processing device generates a virtual X-ray image from 3D stack CT data. The image processing device may be implemented as a variety of devices capable of constant data processing. For example, the image processing device may be implemented as a PC, a server on a network, a smart device, a chipset in which a dedicated program is embedded, and the like.

1 is an example of a system 100 for generating virtual X-ray images from 3D CT data. In FIG. 1 , an example in which the image processing device is a computer terminal 130 and a server 140 is shown.

The CT equipment 110 generates a CT image of a patient. 3D CT data generated by the CT equipment 110 may be stored in an Electronic Medical Record (EMR) 120 or a separate database.

In FIG. 1 , the user A may generate a virtual X-ray image from 3D CT data using the computer terminal 130 . The computer terminal 130 may receive 3D CT data from the CT equipment 110 or the EMR 120 through a wired or wireless network. In some cases, the computer terminal 130 may be a device physically connected to the CT equipment 110. The computer terminal 130 generates an X-ray image of a plane in a certain direction from 3D CT data. At this time, the computer terminal 130 may generate X-ray images of various doses. That is, the computer terminal 130 may generate at least one of a low-dose virtual X-ray image and a high-dose X-ray image. A process of generating a virtual X-ray image will be described later. User A can check the virtual X-ray image on the computer terminal 130 .

The server 140 may receive 3D CT data from the CT equipment 110 or the EMR 120 . The server 140 generates an X-ray image of a plane in a certain direction from 3D CT data. The server 140 may generate X-ray images of various doses. That is, the server 140 may generate at least one of a low-dose virtual X-ray image and a high-dose X-ray image. A process of generating a virtual X-ray image will be described later. The server 140 may transmit the virtual X-ray image to the terminal of user A. User A may check the virtual X-ray image.

The computer terminal 130 and/or the server 140 may store the virtual X-ray image in the EMR 120.

2 is an example of a process of generating a virtual X-ray image from 3D CT data. 2 shows a process of generating a virtual X-ray image in a specific direction from 3D CT data assuming that X-rays are irradiated in a z-axis direction in a 3-dimensional space.

CT equipment generates images by irradiating X-rays with a constant beam intensity. 2 assumes that the X-ray beam travels in the z-axis direction. I ₀ represents the intensity of the beam. 3D CT data is a stacked set of L 2D CT slices. T _i means the ith slide in the entire CT slide set T _all = {T ₁ ,T ₂ ,...,T _L }. Meanwhile, an arbitrary coordinate value in 3D CT data may be defined as (x, y, z).

A virtual X-ray image may be physically expressed as in Equation 1 below.

where Δz is the actual distance between z-axis CT slices. μ _z = μ _z (x, y) is an attenuation coefficient due to X-ray absorption at the patient's volume position corresponding to (x, y, z) coordinates on a three-dimensional CT.

In Equation 1, since the beam intensity I ₀ is a proportional constant, it can be removed as shown in Equation 2 below in the image normalization process. Therefore, Equation 1 can be re-expressed as Equation 2 below.

On the other hand, for any z coordinate value (z = 1, 2, ..., L), it is established as in Equation 3 below according to the X-Ray mass attenuation theory.

where log is the natural log. And I ₀ is the intensity of the projected original beam when taking X-rays. ΔT _i means the absolute change amount of the beam attenuated from the CT slice T _i where the projection beam encounters the i-th. It has the same two-dimensional dimension as the ith CT slice (ΔT ₀ is assumed to be 0).

By substituting Equation 3 into Equation 2, Equation 4 below can be derived.

In Equation 4, ΔT _i (for all i ∈ {1,2,3,...,L}) can be approximated as shown in Equation 5 below.

In Equation 5, β is an arbitrary constant, and T _i means a 2D pixel value of the ith CT slice. Beam attenuation is obtained by applying attenuation to each of the slices by assuming 3D CT data to be the actual volume of the patient.

In addition, I ₀ in Equation 4 can be approximated as in Equation 6 below using an arbitrary constant value of α as follows.

In Equation 6, T _all = {T ₁ , T ₂ ,..., T _L } means the entire 3D CT data of the patient. I ₀ can be expressed by obtaining the maximum and minimum values of all pixel values of this 3D CT and multiplying the difference between these values by α.

When values approximated in Equations 5 and 6 are applied to Equation 4, the virtual X-ray image can finally be expressed as Equation 7 below.

Equation 7 is affected by the constants α and β. As a result, virtual X-ray images of various doses can be created by adjusting I ₀ using α and β.

The researcher created virtual X-ray images according to various beam intensities. In the experiment, the researcher fixed β = 1000 and created virtual X-ray images with different doses by changing the value of α. 3 is a result of generating a virtual X-ray image by changing the beam intensity. 3(A) to 3(D) are virtual X-ray images generated by applying different values of α. Fig. 3(A) is α=110, Fig. 3(B) is α=150, Fig. 3(C) is α=200, Fig. 3(D) is α=1000. That is, FIG. 3(A) corresponds to the lowest dose image and FIG. 3(D) corresponds to the highest dose image. Referring to FIG. 3 , it can be seen that the image becomes clearer as I ₀ , which functions as a dose, increases.

4 is an example of a process 200 of learning an X-ray reconstruction model using a virtual X-ray image. It will be described that the learning process is also performed by the image processing device. However, it may be explained that the learning process of the learning model is performed by a learning device, a computer device, or a model generating device that builds the learning model.

As described above, the image processing device may generate X-ray images of various doses from 3D CT data. Meanwhile, an application for reconstructing a low-dose X-ray image into a high-dose X-ray image using a learning model is being studied. Most of these image reconstruction models require learning data, and the quality of the reconstructed image is increased so that it can be learned with a large amount of learning data.

The image processing device generates a pair of low-dose X-ray images and high-dose X-ray images using the 3D CT data stored in a DB (database) in the above-described manner (210). The training data includes a pair of {low-dose X-ray image, high-dose X-ray image} generated using the same 3D CT data. The image processing device may generate a large amount of learning data set using the same 3D CT data. The training data set may include images generated from the same 3D CT data. Furthermore, the learning data set may include images generated using 3D CT data generated at different times or for a plurality of patients.

The image processing device learns the X-ray reconstruction model using the prepared learning data set (220). The X-ray reconstruction model can be implemented with various learning models. Typically, a dose reconstruction model may be constructed using a generative model.

The researcher prepares learning data by generating low-dose X-ray images and high-dose X-ray images from 3D CT data using the method described above, and converts low-dose X-ray images into high-dose X-ray images using the learning data. We built an X-ray reconstruction model that converts to . The researcher used a Generative Adversarial Network (GAN). 5 is an image conversion result of a reconstruction model for converting a low-dose X-ray image into a high-dose X-ray image. 5(A) is an input image (low-dose image), FIG. 5(B) is a generated image (high-dose image) output from a reconstruction model, and FIG. 5(C) is an actual high-dose image. To verify the performance of the model, the researcher obtained the image quality measurement PSNR (Peak Signal-to-Noise Ratio). As a result of verification, the average PSNR was 39.07, showing high performance close to the original.

6 is an example of the image processing device 300 generating a virtual X-ray image from 3D CT data. The image processing device 300 corresponds to the above image processing devices (130 and 140 in FIG. 1). The image processing device 300 may be physically implemented in various forms. For example, the image processing device 300 may have a form of a computer device such as a PC, a network server, and a data processing dedicated chipset.

The image processing device 300 may include a storage device 310, a memory 320, an arithmetic device 330, an interface device 340, a communication device 350, and an output device 360.

The storage device 310 may store 3D CT data generated by CT equipment.

The storage device 310 may store codes or programs for generating virtual X-ray images from 3D CT data.

The storage device 310 may store virtual X-ray image(s) generated from 3D CT data.

In addition, the storage device 310 may store a reconstruction model to be learned using the learning data described in FIG. 4 . Furthermore, the storage device 310 may store a reconstruction model learned using the learning data.

The memory 320 may store data and information generated during a process of generating a virtual X-ray image by the image processing device 300 or/or learning a reconstruction model.

The interface device 340 is a device that receives certain commands and data from the outside. The interface device 340 may receive 3D CT data from a physically connected input device or an external storage device. The interface device 340 may receive specific direction information desired to be generated using 3D CT data. The interface device 340 may receive a constant β related to attenuation of the CT slice and/or a constant α determining the beam intensity. The interface device 340 may transmit the generated virtual X-ray image to an external object.

The communication device 350 refers to a component that receives and transmits certain information through a wired or wireless network. The communication device 350 may receive 3D CT data from an external object. The communication device 350 may receive specific direction information desired to be generated using 3D CT data. The communication device 350 may receive a constant β related to attenuation of the CT slice and/or a constant α determining the beam intensity. Alternatively, the communication device 350 may transmit the generated virtual X-ray image to an external object such as a user terminal.

Since the interface device 340 and the communication device 350 are configured to send and receive certain data from a user or other physical object, they can also be collectively referred to as input/output devices. When limited to information or data input functions, the interface device 340 and the communication device 350 may also be referred to as input devices.

The output device 360 is a device that outputs certain information. The output device 360 may output an interface required for data processing, a virtual X-ray image, a reconstructed X-ray image, and the like.

The computing device 330 may generate a virtual X-ray image from 3D CT data.

The virtual X-ray image may be expressed by Equation 4 as described above. As shown in Equation 4, the virtual X-ray image may be expressed as the amount of change of the beam attenuated in the CT slice(s) and the intensity of the X-ray beam together with the 3D CT data. Meanwhile, the absolute amount of change ΔT _i of the beam attenuated in the CT slice T _i can be set to a constant β as in Equation 5, and the intensity of the beam I ₀ is set to a constant α as shown in Equation 6. It can be.

The arithmetic device 330 may generate virtual X-ray images having various doses by adjusting α. That is, the arithmetic device 330 may generate a low-dose X-ray image and a high-dose X-ray image by applying different values of α to the same 3D CT data. The arithmetic device 330 sets a constant attenuation with a constant β for each slice of 3D CT data to implement an environment like an actual X-ray imaging. Through this, the arithmetic device 330 may generate a virtual X-ray image, such as an X-ray image of a specific direction, from 3D CT data. The arithmetic device 330 may generate a virtual X-ray image based on the specific directions, β and α input by the user.

The arithmetic device 330 may generate a low-dose virtual X-ray image and a high-dose virtual X-ray image from the same 3D CT data by repeating the process of generating the virtual X-ray image while changing α.

In addition, the arithmetic device 330 may generate a large amount of learning data using 3D CT data as described in FIG. 4 . The training data includes a pair of low-dose X-ray images and high-dose X-ray images generated from the same 3D CT data. The computing device 330 may learn an X-ray reconstruction model using the learning data. Furthermore, the arithmetic device 330 may convert an input low-dose X-ray image into a high-dose X-ray image by using the learned reconstruction model.

The arithmetic device 330 may be a device such as a processor, an AP, or a chip in which a program is embedded that processes data and performs certain arithmetic operations.

In addition, the above-described CT image processing method or virtual X-ray image generation method may be implemented as a program (or application) including an executable algorithm that may be executed on a computer. The program may be stored and provided in a temporary or non-transitory computer readable medium.

A non-transitory readable medium is not a medium that stores data for a short moment, such as a register, cache, or memory, but a medium that stores data semi-permanently and can be read by a device. Specifically, the various applications or programs described above are CD, DVD, hard disk, Blu-ray disk, USB, memory card, ROM (read-only memory), PROM (programmable read only memory), EPROM (Erasable PROM, EPROM) Alternatively, it may be stored and provided in a non-transitory readable medium such as EEPROM (Electrically EPROM) or flash memory.

Temporary readable media include static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), and enhanced SDRAM (Enhanced SDRAM). SDRAM, ESDRAM), Synchronous DRAM (Synclink DRAM, SLDRAM) and Direct Rambus RAM (DRRAM).

This embodiment and the drawings accompanying this specification clearly represent only a part of the technical idea included in the foregoing technology, and those skilled in the art can easily understand it within the scope of the technical idea included in the specification and drawings of the above technology. It will be obvious that all variations and specific examples that can be inferred are included in the scope of the above-described technology.

Claims (11)

Receiving three-dimensional (3D) CT data by an image processing device;
receiving, by the image processing device, a first constant for determining a specific direction and beam intensity; and
Generating, by the image processing device, a virtual X-ray image based on the amount of change in a beam attenuated in each of a plurality of slices according to the specific direction in the 3D CT data and the intensity of the beam determined by the first constant A method for generating virtual X-ray images from 3D CT data containing According to claim 1,
The image processing device determines the amount of change in the beam attenuated in each of the plurality of slices by applying a second constant for determining the amount of change in the beam attenuated in the slice. How to generate a virtual X-ray image from 3D CT data. According to claim 1,
The method of generating a virtual X-ray image from 3D CT data in which the image processing device generates the virtual X-ray image expressed by the following formula.

(Image is a virtual X-ray image, ΔT _i is the absolute amount of change in the beam attenuated from the CT slice T _i where the projected beam encounters the i-th, L is the total number of slices, z is the specific direction, I ₀ is the X- is the intensity of the ray projection beam) According to claim 3,
The method of generating a virtual X-ray image from 3D CT data in which I ₀ is expressed by the following formula.

(Where α is the first constant, T _all = {T ₁ , T ₂ ,..., T _L } is a set of all CT slices) According to claim 1,
The image processing device generates a low-dose virtual X-ray image and a high-dose virtual X-ray image by repeating a process of generating a virtual X-ray image while changing the value of the first constant while using the 3D CT data A method of generating a virtual X-ray image from 3D CT data further comprising the step of doing. generating a learning data set composed of a low-dose virtual X-ray image and a high-dose virtual X-ray image by using the method of generating a virtual X-ray image from the 3D CT data according to claim 5 by an image processing device; and
Learning to convert low-dose X-rays into high-dose X-rays comprising the step of the image processing device learning a learning model for converting a low-dose X-ray image into a high-dose X-ray image using the learning data set How the model is trained. An input device that receives three-dimensional (3D) CT data, a specific direction, and a first constant for determining beam intensity;
a storage device for storing a program for generating a virtual X-ray image from 3D CT data; and
3D CT including an arithmetic device generating a virtual X-ray image based on the beam intensity determined by the amount of change in the beam attenuated in each of the plurality of slices according to the specific direction in the 3D CT data and the first constant An image processing device that creates a virtual X-ray image from data. According to claim 7,
The processing device generates a virtual X-ray image from 3D CT data that determines the amount of change in the beam attenuated in each of the plurality of slices by applying a second constant for determining the amount of change in the beam attenuated in the slice. . According to claim 7,
The processing device generates a virtual X-ray image from 3D CT data generating the virtual X-ray image expressed by the following formula.

(Image is a virtual X-ray image, ΔT _i is the absolute amount of change in the beam attenuated from the CT slice T _i where the projected beam encounters the i-th, L is the total number of slices, z is the specific direction, I ₀ is the X- is the intensity of the ray projection beam) According to claim 9,
The I ₀ is an image processing device that generates a virtual X-ray image from 3D CT data expressed by the following formula.

(Where α is the first constant, T _all = {T ₁ , T ₂ ,..., T _L } is a set of all CT slices) According to claim 7,
The arithmetic device generates a low-dose virtual X-ray image and a high-dose virtual X-ray image by repeating a process of generating a virtual X-ray image while changing a value of the first constant while using the 3D CT data. An image processing device that generates virtual X-ray images from 3D CT data.

Images