JP6129125B2 - Radiation image analysis apparatus and method, and program - Google Patents

Description

  The present invention relates to an image analysis apparatus, method, and program for analyzing a radiographic image obtained by photographing a subject.

  Conventionally, when a radiographic image of a subject is captured by radiation that has passed through the subject, the greater the thickness of the subject, the greater the effect such as a decrease in radiation transmittance inside the subject, and the quality of the acquired radiographic image Is known to fluctuate. For this reason, the thickness of the subject is roughly estimated based on various information such as imaging conditions and radiographic signal value, histogram width of the radiographic signal value, and the length of the subject in the predetermined direction. There has been proposed a technique for changing image processing conditions such as gradation processing on a captured radiographic image and imaging conditions applied to radiographic image capturing in accordance with the thickness of the subject.

  For example, Patent Documents 1 and 2 correspond to the relationship between body thickness and pixel value by measuring the pixel value of an image obtained by radiographing a simulated subject having a known thickness in advance under a known imaging condition. And a rough body thickness distribution is estimated according to the pixel value of the subject image based on the correspondence table, and the subject image is scattered according to the body thickness distribution of the subject image. A technique is disclosed in which a line component is estimated and a post-processing image obtained by subtracting the scattered radiation component from the subject image is obtained.

Japanese Patent Laid-Open No. 02-244881 Japanese Patent Laid-Open No. 06-014911

  Here, in order to obtain a detailed body thickness that reflects the internal structure of the subject such as the lung field inside the subject, the subject is obtained from the pixel value of the subject image obtained by actually photographing the subject. It is preferable to calculate the thickness.

  However, in the subject image, the primary ray component (primary ray component) that has passed through the subject and directly irradiated to the radiation detector, and the scattered ray component (scattered ray component) scattered in the subject are scattered. It is included. For this reason, when the method of estimating the body thickness of the subject based on the pixel value of the subject image is applied as in Patent Documents 1 and 2, due to the influence of the scattered radiation component included in the subject image. It is difficult to accurately estimate the body thickness distribution of the subject.

  Therefore, the applicant of the present invention uses a fact that the scattered radiation component generated as the body thickness increases in Japanese Patent Application No. 2013-229951 and the like, and estimates the body thickness distribution based on the scattered radiation component of the subject image. Proposed. As a result of further research, in the actual imaging environment, not only scattered radiation components generated from the subject but also scattered radiation components generated from elements provided in the imaging environment other than the subject are included in the subject image. I paid attention to the possibility. For example, an imaging environment in which an imaging table for imaging a subject in a supine position is placed between a radiation source and a radiation detector, and the top plate of the imaging table is covered with a material that generates scattered radiation such as acrylic resin If this is the case, the radiation emitted from the radiation source passes through the subject and the imaging table and is detected by the radiation detector. A scattered radiation component made of a material that generates scattered radiation such as resin is included in the subject image. In addition, if the imaging room wall contains a material that generates scattered radiation, and the radiation detector is placed in the vicinity of the imaging room wall, it is generated from the imaging room wall during radiography of the subject. There is a possibility that the scattered radiation component is included in the subject image.

  In order to estimate the body thickness distribution more accurately based on the scattered radiation component in the subject image, the scattered radiation component included in the subject image is generated from elements provided in the imaging environment other than the subject. It is preferable to reduce the influence of the scattered radiation component and estimate the body thickness distribution based on the scattered radiation component generated from the subject.

  The present invention has been made in view of the above circumstances, and estimates a body thickness distribution of a subject by analyzing a subject image that is a radiographic image obtained by photographing the subject in a predetermined photographing environment. Therefore, an object of the present invention is to provide information for correcting an error in the body thickness distribution of a subject image due to the influence of scattered radiation generated by an imaging environment.

  The radiological image analysis apparatus according to the present invention includes an image acquisition unit that acquires a reference image, which is a radiographic image captured without placing a subject in a predetermined imaging environment, and a scattered radiation component included in the acquired reference image. The body thickness distribution of the virtual subject in the reference image is estimated based on the scattered radiation component included in the reference image, assuming that the virtual subject is a virtual subject representing the imaging environment of the reference image. Distribution of body thickness correction amount for correcting body thickness distribution error of subject image, which is a radiographic image obtained by placing a desired subject in an imaging environment, and estimating body thickness distribution of estimated virtual subject And a correction distribution acquisition unit that acquires the correction distribution of the reference image.

  The radiological image analysis method according to the present invention includes an image acquisition step of acquiring a reference image that is a radiographic image captured without placing a subject in a predetermined imaging environment, and a scattered radiation component included in the acquired reference image. The body thickness distribution of the virtual subject in the reference image is estimated based on the scattered radiation component included in the reference image, assuming that the virtual subject is a virtual subject representing the imaging environment of the reference image. Distribution of body thickness correction amount for correcting body thickness distribution error of subject image, which is a radiographic image obtained by placing a desired subject in an imaging environment, and estimating body thickness distribution of estimated virtual subject And a correction distribution acquisition step of acquiring as a correction distribution of the reference image.

  The radiographic image processing method according to the present invention may be provided as a program for causing a computer to execute the method.

  The “imaging environment” is specified by all elements (excluding the subject) that are located at the imaging location and substantially affect the scattered radiation component included in the radiographic image and their relative arrangement. . For example, the imaging table is an element that constitutes the imaging environment. When radiation emitted from a radiation source passes through the radiation detector and is scattered by the wall or floor of the imaging room, the radiation that has passed through the radiation detector. The wall part or floor part reached by is an element constituting the photographing environment.

  Note that “a reference image that is a radiographic image captured without placing a subject in a predetermined imaging environment” and “a subject image that is a radiographic image captured by arranging a desired subject in the imaging environment” Means an image photographed in a photographing environment that can be considered to be substantially the same, although the subject is arranged or not. In addition, the imaging environments that can be regarded as substantially equivalent include not only the case where all the elements that substantially affect the radiographic image located at the imaging location match, but also among the elements located at the imaging location. This includes the case where elements that substantially affect the radiographic image can be regarded as being in common. For example, each object that affects the radiation detected by the radiation detector, such as the material and size of the imaging stand located at the imaging location, and the material and size of the wall portion of the imaging room irradiated with radiation that has passed through the radiation detector Are similar, and the relative distance and arrangement of the objects are similar.

  The “body thickness” means the total thickness of the subject region excluding the air region on the path of the irradiated radiation.

  Further, “based on the scattered radiation component included in the reference image on the assumption that the scattered radiation component included in the acquired reference image is generated by a virtual object that is a virtual object representing the imaging environment of the reference image. "Estimating the body thickness distribution of the virtual subject in the reference image" means that a certain subject (virtual subject) is placed in the reference image, and the radiation passes through the placed subject so that the scattered radiation component Therefore, in the subject image, the scattered radiation component is increased at a plurality of positions in the reference image in correspondence with the increase in the scattered radiation component as the body thickness of the subject on the radiation path increases. This means that the body thickness corresponding to the scattered radiation component is estimated so that the body thickness of the virtual subject increases as the number increases.

  In the radiological image analyzer according to the present invention, the body thickness distribution of the subject in the subject image is acquired based on the scattered radiation component included in the subject image, and the position of each position in the body thickness distribution of the subject is determined according to the correction distribution. It is preferable to further include a body thickness distribution acquisition unit that acquires a corrected body thickness distribution of the subject whose body thickness is corrected.

  In addition, when the radiological image analysis apparatus according to the present invention includes a body thickness distribution acquisition unit, the radiographic image analysis apparatus further includes an imaging condition acquisition unit that acquires imaging conditions of the subject image and the reference image, and the correction distribution acquisition unit is different from each other. The correction distribution is acquired for each of the plurality of reference images to which the imaging condition is applied, the correction distribution of the reference image corresponding to the imaging condition is associated with the imaging condition, and the body thickness distribution acquisition unit is configured to acquire the imaging condition of the subject image. The correction distribution of the reference image associated with the imaging conditions similar to the above is acquired, and the corrected body thickness distribution is acquired by correcting the body thickness at each position in the body thickness distribution of the subject image according to the acquired correction distribution. May be.

  The above "imaging conditions" include imaging dose, tube voltage, product of tube current and irradiation time, distance between radiation source and detection surface of radiation detector, source target and filter material, and radiation detection used for imaging. Including at least one of the material of the radiation shielding plate provided to the radiation detector as necessary to protect the electronic type, etc., as well as the type of device, air gap amount (distance from the subject to the radiation detector), etc. If it is.

  “Similar imaging conditions” includes not only the case where all imaging conditions are the same, but also the case where the elements of the imaging conditions that are factors that substantially affect the radiographic image can be regarded as almost common. . For example, if the imaging dose, tube voltage, product of tube current and irradiation time, and the distance between the radiation source and the detection surface of the radiation detector are within the allowable range, the imaging conditions are treated as being similar. it can.

  Further, when the radiological image analysis apparatus according to the present invention includes a body thickness distribution acquisition unit, the correction distribution acquisition unit acquires a correction distribution for a reference image to which a specific imaging condition is applied in an imaging environment, and the body thickness distribution The acquisition unit acquires a corrected body thickness distribution obtained by correcting the body thickness at each position in the body thickness distribution of the subject image to which the imaging conditions different from the specific imaging conditions are applied in the imaging environment according to the acquired correction distribution. May be.

  Further, when the radiological image analysis apparatus according to the present invention includes a body thickness distribution acquisition unit, the image acquisition unit acquires a plurality of reference images corresponding to the imaging environment, and the correction distribution acquisition unit acquires the plurality of acquired plurality of reference images. The correction distribution is acquired for each of the reference images, and the body thickness distribution acquisition unit determines the position of each position in the body thickness distribution of the subject according to the average correction distribution obtained by averaging the plurality of correction distributions corresponding to the plurality of acquired reference images. It is preferable to obtain a corrected body thickness distribution of the subject whose body thickness has been corrected.

  In addition, when the radiological image analysis apparatus according to the present invention includes a body thickness distribution acquisition unit, the body thickness distribution acquisition unit performs the body thickness distribution of the subject in the subject image based on the scattered radiation component included in the subject image. When it is determined that at least part of the acquired body thickness distribution may cause an error due to a scattered radiation component from the imaging environment, the body thickness distribution of the subject is determined according to the correction distribution. The corrected body thickness distribution of the subject may be acquired by correcting the body thickness at each position.

  Further, when the radiological image analysis apparatus of the present invention includes a body thickness distribution acquisition unit, the body thickness distribution acquisition unit acquires a first virtual model of a subject having a predetermined first body thickness distribution. Estimating a model acquisition unit and imaging conditions of the subject image, applying the imaging conditions of the subject image, and estimating a primary line image obtained by radiography of the virtual model of the subject from the virtual model of the subject Estimate the radiographic image obtained by radiography of the subject from the combined image of the primary ray image and the scattered radiation image obtained by radiography of the virtual model of the subject and the estimated scattered ray image estimated from the virtual model of the subject A first estimated image generating unit that generates an estimated image of a subject image that is a corrected image, and a first body thickness distribution of the virtual model is modified to reduce a difference between the estimated image of the subject image and the subject image. 1 And Tadashibu may be the first body thickness distribution of the virtual model of the modified analyte to and a first body thickness distribution determiner for determining a body thickness distribution of the subject.

  Further, when the radiological image analysis apparatus according to the present invention includes a body thickness distribution acquisition unit, the body thickness distribution acquisition unit includes a plurality of model images that are radiographic images obtained by imaging a plurality of models different from the subject, and A model information acquisition unit that acquires model information that associates the body thickness distribution of the model image acquired based on the scattered radiation component of the model image, and imaging target information that represents the characteristics of the subject of the subject image Acquiring and acquiring imaging target information representing model characteristics of a plurality of model images based on the plurality of model information, respectively, and specifying a model image having pre-imaging target information similar to the imaging target information of the subject image A new body thickness distribution determining unit that determines the body thickness distribution associated with the specified model image as the body thickness distribution of the subject image may be provided.

  As the “model”, any model that can virtually represent the subject can be used. For example, when the subject is a human body, the model is preferably a comparative human body.

  The above “photographing target information” is specified by any method as long as it represents the characteristics of the subject to be photographed, such as the skeleton shape and size of the subject to be photographed, the physical features such as muscle mass and fat mass. It may be a thing. For example, the histogram width of the signal value of the radiographic image (difference between the maximum value and the minimum value of the histogram of the signal value), the representative length of the specific part of the subject in the radiographic image (eg, the width of the abdomen, etc.), the height of the subject, Weight, sex, age (adult, child), etc. may be used. Further, the shooting target information may be one or plural. Further, the shooting target information may be input by user input, or may be shooting target information extracted from the subject image and the model image.

  In the radiological image analysis apparatus according to the present invention, the image acquisition unit acquires a plurality of radiographic images captured without arranging a subject in an imaging environment, and acquires an image obtained by averaging the plurality of radiographic images as a reference image. It is preferable.

  Further, in the radiological image analysis apparatus according to the present invention, the image acquisition unit generates a reduced image in which the resolution of the radiographic image captured without placing the subject in the imaging environment is reduced to a predetermined resolution, and the generated reduction An image may be acquired as a reference image.

  In the radiological image analysis apparatus according to the present invention, the correction distribution acquisition unit includes a second virtual model acquisition unit that acquires a virtual model of an imaging environment having a predetermined second body thickness distribution, and a reference image imaging condition. Estimating the acquired primary line image estimated from the virtual model of the shooting environment and the scattered radiation image obtained by radiography of the virtual model of the shooting environment from the virtual model of the shooting environment by acquiring and applying the shooting conditions of the reference image A second estimated image generating unit that generates an image obtained by combining the scattered radiation image as an estimated image of a reference image that is an image obtained by estimating a reference image; and correcting the second body thickness distribution of the virtual model of the imaging environment, A second correction unit that reduces a difference between the estimated image of the reference image and the reference image; and a second body thickness distribution of the virtual model of the corrected shooting environment is determined as a correction distribution corresponding to the shooting environment. It may be provided with a two-body thickness distribution determiner.

  Each of the above “estimated images” includes an estimated primary line image obtained by estimating the primary line image obtained by radiography of the virtual model from the virtual model, and an estimated scattered ray obtained by estimating the scattered ray image obtained by the radiography of the virtual model from the virtual model. Any image can be used as long as it can be substantially regarded as a combined image. For example, an estimated primary ray image is generated by applying a function for generating an estimated primary ray image to a virtual model, and an estimated scattered ray image is separately generated by applying a function for generating an estimated scattered ray image to a virtual model. The estimated image may be estimated by applying a function for generating the estimated image to the virtual model.

  Further, the “first body thickness distribution” and the “second body thickness distribution” are not limited as long as they indicate a predetermined body thickness distribution, and may have an arbitrary thickness distribution. For example, the “first body thickness distribution” or the “second body thickness distribution” can be a uniform distribution. In addition, when the subject is a predetermined part of the human body, the first body thickness distribution can be a body thickness distribution of the same part of the human body different from the subject. Further, the “first body thickness distribution” may be the same distribution as the “second body thickness distribution” or may be a different distribution.

  According to the present invention, a reference image, which is a radiographic image captured without placing a subject in a predetermined imaging environment, is acquired, and the scattered radiation component included in the acquired reference image determines the imaging environment of the reference image. Assuming that it is caused by a virtual subject that is a virtual subject to represent, the body thickness distribution of the virtual subject in the reference image is estimated based on the scattered radiation component included in the reference image, and the estimated virtual subject Correction distribution of a reference image, which is a distribution of body thickness correction amount for correcting an error in body thickness distribution of a subject image, which is a radiographic image obtained by arranging a desired subject in an imaging environment. As a result, it is possible to provide the acquired correction distribution as useful information for correcting an error in the body thickness distribution of the subject image due to the influence of the scattered radiation component caused by the imaging environment.

1 is a schematic block diagram showing a configuration of a radiographic image capturing system to which a radiological image analysis apparatus according to a first embodiment of the present invention is applied. The schematic block diagram which shows the structure of the correction distribution acquisition part and the body thickness distribution acquisition part which concern on the 1st Embodiment of this invention. The figure which shows an example of the correspondence table of body thickness distribution The figure for demonstrating an example of the production | generation method of an estimated image The figure for demonstrating another example of the production | generation method of an estimated image The flowchart which shows the flow of a process of the radiographic image analyzer which concerns on the 1st Embodiment of this invention. The flowchart which shows the flow of a process of the correction distribution acquisition part which concerns on the 1st Embodiment of this invention. The flowchart which shows the flow of a process of the body thickness distribution acquisition part which concerns on the 1st Embodiment of this invention. The schematic block diagram which shows the structure of the body thickness distribution acquisition part which concerns on the 3rd Embodiment of this invention. The flowchart which shows the flow of a process of the body thickness distribution acquisition part which concerns on the 3rd Embodiment of this invention. The flowchart which shows the flow of a process of the radiographic image analyzer which concerns on the 2nd Embodiment of this invention. The flowchart which shows the flow of a process of the radiographic image analyzer which concerns on the modification of the 2nd Embodiment of this invention. The figure for demonstrating the example which applies this invention to long imaging | photography. The figure for demonstrating the example applied to the division | segmentation photography of this invention

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a schematic block diagram showing a configuration of a radiographic image capturing system to which a radiographic image analyzing apparatus according to a first embodiment of the present invention is applied. As shown in FIG. 1, the radiographic image capturing system according to the present embodiment includes an image capturing device 10, a control device 20 that controls the system, and an image analysis device 30 (radiation image analysis device).

  The imaging apparatus 10 includes a radiation source 12 that irradiates a subject K with X-rays, and a radiation detector 14 that detects X-rays transmitted through the subject K and acquires a radiation image of the subject K. In the present embodiment, between the subject K and the radiation detector 14, a scattered radiation removal grid for removing scattered radiation scattered by the subject K out of X-rays transmitted through the subject K. (Grid) is not placed.

  The control device 20 controls the radiation source drive control unit 22 for driving and controlling the radiation source 12 in accordance with the set imaging conditions and the radiation detector 14, and the radiographic image of the subject (for example, the subject image Ik and the reference image Ir). And a detector control unit 24 that stores the information in the storage unit 32.

  The image analysis device 30 includes an input unit 38 that accepts various inputs from the operator to the image analysis device 30, a display unit 39, a central processing unit (CPU), a semiconductor memory, a communication interface, and a storage unit 32 such as a hard disk or an SSD. The image analysis program according to the present embodiment is installed in the image analysis apparatus 30. Then, by executing the image analysis program, the central processing unit and the memory of the image analysis device 30 cooperate to perform the imaging control unit 31 that controls the radiation source drive control unit 22 and the detector control unit 24, and the image acquisition unit. 33, an imaging condition acquisition unit 34, a correction distribution acquisition unit 35, a body thickness distribution acquisition unit 36, and a display control unit 37. The input unit 38 includes a keyboard, a mouse, a touch panel, and the like. The display unit 39 includes a CRT, a liquid crystal display, and the like, and displays a radiation image acquired by the imaging apparatus 10 and information necessary for various other desired processes.

The storage unit 32 stores the subject image Ik and the reference image Ir acquired by the imaging control unit 31 that controls the detector control unit 24 and the radiation source drive control unit 22 and the respective imaging conditions. Further, the storage unit 32 stores a correspondence table LUT in which density values (pixel values) and body thickness are associated with each other for each of a plurality of imaging conditions. The storage unit 32 also stores a virtual model M of the subject K having an initial body thickness distribution T ₀ ^k (x, y) and a virtual model Mr of an imaging environment having an initial body thickness distribution T ₀ ^r . Furthermore, the storage unit 32 stores the structures (anatomical structures such as lung fields, bones, and organs) included in the virtual model Mk for the subject image, the arrangement of the structures, and the radiation of the structures. Characteristic information indicating characteristics and the like is preset and stored based on the arrangement and composition of anatomical structures such as lung fields and bones in the thoracoabdominal region of the comparison subject. Note that the imaging environment virtual model Mr is a virtual subject representing the imaging environment of the reference image Ir. The storage unit 32 appropriately stores various parameters necessary for each process and generated images (such as an estimated primary ray image and an estimated scattered ray image). The body thickness in this specification means the total thickness of the subject area excluding the air area on the path of the irradiated radiation.

  The display control unit 37 causes the display unit 39 to appropriately display information necessary for the image analysis processing according to the present embodiment, information necessary for the imaging control processing of the imaging control unit 31, and the like.

  The image acquisition unit 33 acquires the reference image Ir and the subject image Ik from the detector control unit 24 or the storage unit 32. Note that the present invention is not limited to this embodiment, and the present invention can be applied to any type of subject.

  The subject image Ik is a radiographic image obtained by applying a predetermined imaging condition and placing a human body as the subject K in a predetermined imaging environment.

  The “imaging environment” is specified by all elements (excluding the subject) that are located at the imaging location and substantially affect the scattered radiation component included in the radiographic image and their relative arrangement. . For example, the imaging table is an element that constitutes the imaging environment. When radiation emitted from a radiation source passes through the radiation detector and is scattered by the wall or floor of the imaging room, the radiation that has passed through the radiation detector. The wall part or floor part reached by is an element constituting the photographing environment. As shown in FIG. 1, in the imaging room 17, a radiation source 12 and a radiation detector 14 arranged at positions facing each other, and an imaging table 16 for lying position photography that supports the back surface of the radiation detector 14 are arranged. ing. The radiation detector 14 includes a glass substrate 14a and a scintillator 14b that is a phosphor layer that converts radiation into light and necessary electronic components (not shown) from the position where the radiation source 12 is disposed toward the radiation irradiation direction. They are provided in this order. The glass substrate is provided with a signal output unit and a sensor unit (not shown). Light emitted from the scintillator is converted into electric charges by the sensor unit, and a radiation image is read. In this example, the imaging environment includes a radiation source 12, a radiation detector 14, an imaging table 16 including an acrylic resin top plate that is a material that generates scattered radiation, and a wall portion 17A that is irradiated with radiation. It is specified by its relative arrangement. In some cases, a radiation shield for protecting the electronic component is disposed between the scintillator 14b and the electronic component, and the presence / absence and material of the radiation shield are factors that affect the scattered radiation component. For example, the radiation shield may have a plate shape, and lead (Pb), copper (Cu), tungsten (W), or the like may be used as the material.

  The reference image Ir is a radiographic image captured without the subject K being placed in an imaging environment equivalent to the imaging environment of the subject image Ik. In the present embodiment, first, in a state where the subject is not arranged, the radiographic imaging is performed under a predetermined imaging condition for imaging the subject image Ik to obtain the reference image Ir, and then included in the imaging room 17. It is assumed that the subject image Ik is obtained by performing radiation imaging under the same imaging conditions as the reference image in a state where the subject is placed on the imaging stage 16 without changing the presence or arrangement of elements.

  Note that the reference image Ir and the subject image Ik mean images taken in a photographing environment that can be regarded as being substantially equivalent although the subject is placed or not. Imaging environments that can be regarded as being substantially equivalent include not only the case where all elements that substantially affect the radiographic image located at the imaging location match, but also radiographic images among the elements located at the imaging location. This includes the case where elements that substantially affect the system can be regarded as almost common. For example, the material and size of the imaging stand located at the imaging location, the material and size of the wall part of the imaging room irradiated with radiation that has passed through the radiation detector, and whether there is a radiation shield provided to the radiation detector as required There are cases where the objects that affect the radiation detected by the radiation detector, such as material and material, are similar, and the relative distances and arrangement of the objects are similar.

  Further, it is preferable that the reference image Ir and the subject image Ik are images captured by applying imaging conditions that can be regarded as substantially equivalent. The imaging conditions that can be regarded as substantially equivalent are not only the case where all the imaging conditions match, but also that the elements of the imaging conditions that cause the radiation image substantially have the same elements. Including cases that can be considered. Here, the imaging conditions are substantially equivalent when the imaging dose, the tube voltage, the product of the tube current and the irradiation time, and the distance SID between the radiation source and the detection surface of the radiation detector are within the allowable range. Treat as.

  The imaging condition acquisition unit 34 captures the imaging dose, the tube voltage, the product of the tube current and the irradiation time, and the distance between the radiation source and the detection surface of the radiation detector for the subject image Ik and the reference image Ir from the storage unit 32. Get as a condition. The imaging conditions acquired by the imaging condition acquisition unit 34 are appropriately used for processing by a correction distribution acquisition unit 35 described later, processing by a body thickness distribution acquisition unit 36, and other various desired image processing.

  Without being limited to this embodiment, “imaging conditions” include imaging dose, tube voltage, tube current and irradiation time, distance between radiation source and detection surface of radiation detector, source target and filter material, Including at least one of the types of radiation detectors used for radiography, the amount of air gap (distance from the subject to the radiation detector), the presence or absence of radiation shields provided to the radiation detector, and the materials as necessary If it is.

  The body thickness distribution acquisition unit 36 acquires the body thickness distribution Tk of the subject K of the subject image Ik based on the scattered radiation component included in the subject image Ik, and determines the subject K's distribution according to the correction distribution Tr described later. A corrected body thickness distribution Tk ′ of the subject K obtained by correcting the body thickness at each position in the body thickness distribution Tk is acquired.

  The body thickness distribution acquisition unit 36 corrects the body thickness distribution Tk ′ of the subject K by correcting the body thickness at each position in the body thickness distribution Tk according to the correction distribution described later in order to acquire the corrected body thickness distribution Tk. Any method capable of obtaining the above may be applied. In the present embodiment, the body thickness distribution acquisition unit 36 determines whether one or both of the correction distribution for each pixel of the reference image and the body thickness distribution for each pixel of the subject image have mutually corresponding pixel positions. The corrected body thickness distribution after correction is obtained by appropriately scaling up and down so as to match and subtracting the correction distribution from the body thickness distribution.

  In addition, the body thickness distribution acquisition unit 36, based on the scattered radiation component included in the subject image Ik, causes the scattered radiation component to increase as the body thickness on the radiation path of the subject increases. Any method capable of acquiring the body thickness distribution Tk of the subject K can be employed. In addition, since this embodiment is an example in which the body thickness distribution acquisition unit 36 also has the function of the correction distribution acquisition unit 35, the detailed function of the body thickness distribution acquisition unit 36 according to this embodiment is described with reference to the correction distribution acquisition unit. This will be described later together with the description of 35.

  Here, as shown in Japanese Patent Application No. 2013-229941, the inventors of the present application are researching a method for estimating the body thickness distribution of the subject K based on the scattered radiation component of the subject image Ik. . In these research processes, it was noted that the subject image Ik includes both a scattered ray component generated from the subject K and a scattered ray component generated from the imaging environment itself. Furthermore, when estimating the body thickness distribution of the subject K based on the scattered radiation component of the subject image Ik, the body thickness distribution of the subject K should be estimated based on the scattered radiation component generated from the subject K. In some cases, in reality, the body thickness distribution of the subject K is estimated based on the scattered radiation component including the scattered radiation component generated from the subject K and the scattered radiation component generated from the imaging environment itself, and thus the scattering generated from the imaging environment. The present inventors have found a problem that an error may occur in the body thickness distribution Tk estimated based on the scattered ray component of the subject image Ik due to the line component.

  For example, in the imaging environment shown in FIG. 1, an imaging table 16 for imaging a subject in a supine position is disposed between a radiation source and a radiation detector, and the top plate of the imaging table 16 is scattered by acrylic resin or the like. In an imaging environment configured (or covered) with a material that generates a line, the radiation X emitted from the radiation source 12 passes through the subject K and the top of the imaging table 16 to generate scattered light. Since the scattered light is detected by the radiation detector 14, the subject image Ik depends on a material that generates scattered rays such as a scattered ray component generated by the subject K and an acrylic resin top plate of the imaging table 16. The scattered radiation component is included in the subject image Ik. Further, when the wall of the imaging room 17 includes a material that generates scattered radiation, and the radiation detector 14 is disposed in the vicinity of the wall of the imaging room 17, the imaging room is used for radiography of the subject K. There is a possibility that the subject image Ik includes the scattered radiation component generated from the wall portion 17A irradiated with the radiation of the 17 walls.

  As a result of intensive studies, the present inventors have made use of the scattered radiation component of the reference image Ir photographed without arranging the subject K in the photographing environment equivalent to the subject image Ik, and the scattered radiation component generated by the photographing environment itself. It was found that the distribution of errors in body thickness distribution can be calculated.

  That is, since the reference image is a radiographic image captured without arranging the subject, the scattered radiation component included in the reference image is considered to be a scattered radiation component generated from one or more elements constituting the imaging environment. Therefore, the reference image Ir is regarded as an image in which a virtual subject that is a virtual subject representing the imaging environment is arranged, and the scattered radiation component included in the reference image Ir is a scattered radiation component generated by the virtual subject. Assuming that the body thickness distribution of the subject is estimated based on the scattered radiation component of the subject image, the body thickness distribution of the virtual subject can be estimated based on the scattered radiation component of the reference image Ir. . Then, since the thickness of the virtual subject should not actually exist in the reference image Ir photographed in a state where the subject K is not arranged, the estimated body thickness distribution is generated from the photographing environment itself. It represents the distribution of errors in body thickness that occur based on scattered radiation components. Therefore, the body thickness distribution of the virtual subject is estimated based on the scattered radiation component of the reference image Ir, and the estimated body thickness distribution is used to correct the body thickness distribution Tk of the subject image Ik. By obtaining a correction distribution that is a correction amount distribution, it is possible to provide useful information for correcting an error in body thickness that occurs based on a scattered radiation component generated from the imaging environment itself.

  Therefore, the correction distribution acquisition unit 35 regards that the scattered radiation component included in the acquired reference image Ir is generated by a virtual subject that is a virtual subject representing the imaging environment of the reference image Ir, Based on the scattered radiation component included in the reference image Ir, the body thickness distribution of the virtual subject of the reference image Ir is estimated, and the estimated body thickness distribution of the virtual subject is imaged by arranging a desired subject in the imaging environment. It is acquired as a correction distribution of a reference image, which is a distribution of body thickness correction amounts for correcting an error in the body thickness distribution of the subject image Ik, which is a radiographic image.

  Further, “based on the scattered radiation component included in the reference image on the assumption that the scattered radiation component included in the acquired reference image is generated by a virtual object that is a virtual object representing the imaging environment of the reference image. "Estimating the body thickness distribution of the virtual subject in the reference image" means that a certain subject (virtual subject) is placed in the reference image, and the radiation passes through the placed subject so that the scattered radiation component Therefore, in the subject image, the scattered radiation component is increased at a plurality of positions in the reference image in correspondence with the increase in the scattered radiation component as the body thickness of the subject on the radiation path increases. This means that the body thickness corresponding to the scattered radiation component is estimated so that the body thickness of the virtual subject increases as the number increases.

  In the subject image, the correction distribution acquisition unit 35 responds to the fact that the scattered radiation component increases as the body thickness of the subject on the radiation path increases, and a plurality of reference images are generated according to the scattered radiation component of the reference image. Any method capable of calculating the body thickness distribution of the subject included in the reference image can be applied so that the body thickness of the virtual subject increases as the scattered radiation component increases at the position. For example, the correction distribution acquisition unit 35 may calculate the body thickness distribution of the virtual subject (imaging environment) from the reference image by the same method as the body thickness distribution acquisition unit 36, and by a method different from the body thickness distribution acquisition unit 36, The body thickness distribution of the virtual subject (imaging environment) may be calculated from the reference image.

  In the present embodiment, the body thickness distribution acquisition unit 36 also has the function of the correction distribution acquisition unit 35, and the correction distribution acquisition unit 35 uses the same method as the body thickness distribution acquisition process performed by the body thickness distribution acquisition unit 36 to generate a virtual image from the reference image. An example in which a body thickness distribution of a subject (imaging environment) is calculated and acquired as a correction distribution will be described. That is, in the process in which the body thickness distribution acquisition unit 36 acquires the body thickness distribution and the correction distribution acquisition process in the correction distribution acquisition unit 35, the body thickness distribution acquisition unit 36 acquires the subject image Ik and the subject in the process of acquiring the body thickness distribution. The point using the virtual model Mk of the specimen K is that the reference image Ir and the virtual model Mr of the imaging environment (the virtual model of the virtual subject representing the imaging environment) are used in the correction distribution acquisition process by the correction distribution acquisition unit 35. Except for this, common processing is executed.

  FIG. 2 is a block diagram showing a schematic configuration of the correction distribution acquisition unit 35 and the body thickness distribution acquisition unit 36 of the present invention. The body thickness distribution acquisition unit 36 in the first embodiment includes a virtual model acquisition unit 41, an estimated image generation unit 42, a correction unit 43, and a body thickness distribution determination unit 44. Since each element and each function of the body thickness distribution acquisition unit 36 are the same as each element and each function of the correction distribution acquisition unit 35, the common part of the body thickness distribution acquisition unit 36 and the correction distribution acquisition unit 35 is the body thickness distribution. Detailed description of either the acquisition unit 36 or the correction distribution acquisition unit 35 may be omitted.

When the body thickness distribution acquisition unit 36 performs the process of acquiring the body thickness distribution of the subject image Ik, the virtual model acquisition unit 41 determines the virtual model Mk of the subject K having the initial body thickness distribution T ₀ ^k. get. Further, the virtual model acquisition unit 41 has an initial body thickness distribution T ₀ ^r when the body thickness distribution acquisition unit 36 functions as the correction distribution acquisition unit 35 and performs a process of acquiring the correction distribution of the imaging environment. A virtual model Mr of the shooting environment is acquired.

Here, the virtual model may be data in which body thickness according to a predetermined initial body thickness distribution is associated with each position on the xy plane. The virtual model Mr of the imaging environment is data that virtually represents the subject K in which the body thickness according to the initial body thickness distribution T ₀ ^k (x, y) is associated with each position on the xy plane. Similarly, the virtual model Mr of the imaging environment virtually represents the virtual subject of the imaging environment in which the body thickness according to the initial body thickness distribution T ₀ ^r (x, y) is associated with each position on the xy plane. It is data to represent. Note that the initial body thickness distributions of the virtual model Mk for the subject image and the virtual model Mr of the imaging environment are corrected by the correction unit 43 described later in the course of the correction distribution acquisition process or the body thickness distribution acquisition process. Therefore, the distribution may be an arbitrary distribution, both may have different distributions, or both may have similar or the same distribution. Further, the initial body thickness distribution of the reference image Ir and the subject image Ik may be generated and acquired during the process of acquiring the initial body thickness distribution of each virtual model Mr, Mk, and each virtual model Mr, Mk Prior to the acquisition process, it may be set in advance.

In the present embodiment, the virtual model Mr of the shooting environment has a uniform distribution having a predetermined value. Further, the initial body thickness distribution T ₀ ^k for acquiring the body thickness distribution of the subject K is generated and acquired by the virtual model acquisition unit 41.

The virtual model acquisition unit 41 acquires the imaging conditions of the subject image Ik, and acquires a table in which the pixel values corresponding to the imaging conditions of the subject K are associated with the body thickness from the storage unit 32. FIG. 3 shows an example of a table in which pixel values are associated with body thicknesses. Then, the virtual model acquisition unit 41 specifies the body thickness corresponding to the pixel value of each pixel of the subject image Ik based on the table shown in FIG. 3, so that the initial body thickness distribution T ₀ of the subject image Ik. ^{Get k} . The above processing is expressed by the following formula (1). In Expression (1), Ik (x, y) represents the pixel value of each pixel in the subject image, and T ₀ (x, y) represents the initial body thickness distribution at each pixel position.

  When the virtual model Mr of the imaging environment is used as the virtual model, the estimated image generation unit 42 estimates the primary line image Ip obtained by estimating the primary line image obtained by radiation imaging of the virtual model Mr of the imaging environment and the imaging environment. An image obtained by synthesizing the estimated scattered radiation image Is obtained by estimating the scattered radiation image obtained by radiography of the virtual model Mr is generated as an estimated image Im of the reference image Ir that is an image obtained by estimating the reference image Ir.

  Further, when the virtual model Mk of the subject K is used as the virtual model, the estimated image generation unit 42 estimates the primary line image Ip obtained by estimating the primary line image obtained by radiography of the virtual model Mk and the virtual model Mk. An image obtained by synthesizing the estimated scattered radiation image Is obtained by estimating the scattered radiation image obtained by the radiation imaging is generated as an estimated image Im of the subject image Ik that is an image obtained by estimating the subject image Ik.

Hereinafter, an example in which the estimated image generation unit 42 generates an estimated image will be described. In the following description, when the virtual model is a virtual model Mr of the subject K uses the initial body thickness distribution T ₀ ^k of virtual models Mk of the subject K as the initial material thickness distribution T _0, the object image Ik Is used. As for when the virtual model is a virtual model Mr of the shooting environment, the initial body thickness with an initial body thickness distribution T ₀ ^r of the virtual model Mr photographing environment as a distribution T _0, the reference image instead of the subject image Ik Ir is used.

4 and 5 are diagrams for explaining a method of generating the estimated image Im. As shown in FIG. 4, the estimated image generation unit 42 obtains an estimated primary line image Ip obtained when the virtual model M is imaged under the same imaging conditions as the subject image Ik by the following equation (2). ), And using the generated estimated primary line image Ip, an estimated scattered radiation image Is is generated according to Expression (3). Then, the estimated image generation unit 42 generates the estimated image Im by synthesizing the estimated primary line image Ip and the estimated scattered radiation image Is as shown in Expression (4). When the estimated primary ray image Ip and the estimated scattered ray image Is are generated for the first time, the initial body thickness distribution T ₀ (x, y) is used in the estimation equations (2) and (3) (formulas). (N = 1 in (2) and (3)).

Here, (x, y) is the coordinates of the pixel position of the subject image Ik, Ip (x, y) is the estimated primary line image at the pixel position (x, y), and Is (x, y) is the pixel position (x , Y), the estimated scattered radiation image at I, (x, y) is the dose at the pixel position (x, y), Im (x, y) is the estimated image at the pixel position (x, y), and μ is the line attenuation of the subject. The coefficient, K _s (x, y, Tn (x ′, y ′), θ _{x ′, y ′} ) represents a point spread function according to the object thickness at the pixel position (x, y). A convolution kernel. Note that the dose Io (x, y) is a dose at which radiation is detected by the detector when it is assumed that no subject exists, and the distance (SID) between the radiation source 12 and the detection surface of the radiation detector 14, Varies according to tube voltage and radiography dose. Further, θ _{x ′ and y ′} represent parameters specified by shooting conditions and characteristic information of the virtual model M.

Note that the estimated image Im may be an image estimated to be obtained when the virtual model M (Mr, Mk) is radiographed, and is substantially the same as an image obtained by combining the estimated primary ray image Ip and the estimated scattered ray image Is. Anything can be considered. For example, as shown in FIG. 5, the estimated image Im is generated by convolving and integrating the kernel combining the primary ray component and the scattered ray component using the following equation (5) instead of the equations (2) to (4). May be. Here, K _{p + s} (x, y, T _n−1 (x ′, y ′), θ _{x ′, y ′} ) is a kernel representing a point diffusion function that combines the primary ray component and the scattered ray component. Any model function may be used as long as an estimated image obtained by synthesizing the estimated primary ray image and the estimated scattered ray image from the image obtained by radiography can be generated.

Note that K _{p + s} (x, y, T _n−1 (x ′, y ′), θ _{x ′, y ′} ) can be obtained experimentally according to the imaging conditions and the like.

In the present embodiment, the kernel K _s (x, y, Tn (x ′, y ′), θ _{x ′, y ′} ), K _{p + s} (x, y, T _n−1 ) is based on the shooting conditions at the time of shooting. (X ′, y ′), θ _{x ′, y ′} ) may be calculated, but various imaging conditions and kernels K _s (x, y, Tn (x ′, y ′), θ _{x ′, y ′)} ), K _{p + s} (x, y, T _n−1 (x ′, y ′), θ _{x ′, y ′} ) are stored in the storage unit 32, and irradiation field information at the time of imaging is stored. Based on the subject information and shooting conditions, this table is referred to and kernels K _s (x, y, Tn (x ′, y ′), θ _{x ′, y ′} ), K _{p + s} (x, y, T _{n) −1} (x ′, y ′), θx _{′, y ′} ) is obtained.

When the estimated image Im is an estimated image of the reference image Ir, the correcting unit 43 determines the difference between the estimated image Im of the reference image Ir and the reference image Ir based on the estimated image Im of the reference image Ir and the reference image Ir. The initial body thickness distribution T ₀ ^r (T _n−1 ^r ) of the virtual model Mr of the imaging environment is corrected so that becomes smaller.

In addition, when the estimated image Im is an estimated image of the subject image Ik, the correcting unit 43 estimates the estimated image Im of the subject image Ik based on the estimated image Im of the subject image Ik and the subject image Ik. The initial body thickness distribution T ₀ k (T _n−1 ^k ) of the virtual model Mk of the subject K is corrected so that the difference between the subject image Ik and the subject image Ik becomes small.

Further, the correction unit 43, in order to perform the correction processing of body thickness distribution T _n-1, the body thickness distribution T _n-1 as the difference of the estimated image Im and the subject image Ik (reference image Ir) is reduced Any method that can acquire the correction value of each position is applicable. In the present embodiment, the correction unit 43 varies the body thickness distribution T _n-1 of the virtual model Mk (Mr) for each partial region of one or more pixels of the virtual model Mk (Mr), and the estimated image Im A process of calculating the body thickness of the partial region that reduces the difference from the sample image Ik (reference image Ir) is performed. Then, the body thickness distribution of the virtual model is corrected based on the calculated body thickness of each partial region.

The function of the correction unit 43 will be described below in the case where the generated estimated image Im is an estimated image of the subject image Ik and the virtual model M is the virtual model Mk of the subject K. In the description of the processing of the correction unit 43, when the generated estimated image Im is an estimated image of the subject image Ik and the virtual model is the virtual model Mk of the subject K, the following body thickness distribution T _n The body thickness distribution T _n−1 ^k of the subject image Ik is used as ₋₁ . Further, in the description of the processing of the correcting unit 43, when the estimated image Im is estimated image of the reference image Ir is body thickness distribution T body thickness of the virtual model Mr of _n-1 as the imaging environment distribution T _{n over 1} ^r And the reference image Ir is used instead of the subject image Ik.

The correction part 43 shall calculate | require the correction value of the body thickness of body thickness distribution _Tn-1 using the steepest descent method. Using the following formulas (6) and (7), among the pixels of the virtual model M, only the body thickness of one specific coordinate is changed in T _n-1 (x, y), and the error function f _error is By repeatedly calculating dT _n-1 (x, y) based on the primary partial differentiation (gradient), the output value of the error function f _error can be minimized. Then, the body thickness of one specific coordinate when the output value of the error function f _error is minimized is determined as the correction value of the body thickness of the specific coordinate. Similarly, the body thickness distribution of each pixel is corrected by obtaining a correction value of the body thickness for each of the other pixels, and the corrected body thickness distribution _Tn is acquired.

However, in Formula (6), (alpha) is an update coefficient which is a parameter showing the update speed of body thickness. As an example of a method for calculating the differential value portion of K _{p + s} shown in Expression (7), for example, when an extremely small value dt is added to T _n−1 (x, y), a change in value is calculated by Expression (8). Thus, the value of K _{p + s} in equation (7) can be obtained.

The body thickness distribution determination unit 44 uses the body thickness distribution T _n-1 (n is a natural number) of the virtual model Mk of the subject K modified based on the estimated image Im of the subject image Ik as the body thickness of the subject image Ik. The distribution Tk is determined.

Further, the body thickness distribution determination unit 44 uses the body thickness distribution T _n-1 (n is a natural number) of the virtual model Mr of the imaging environment modified based on the estimated image Im of the reference image Ir as a correction distribution of the subject image Ik. Determine to Tr.

Further, the body thickness distribution determination unit 44 determines the corrected body thickness distribution T _n-1 as the body thickness distribution Tk of the subject image Ik when a predetermined end condition is satisfied. Further, when the condition set as the end condition is satisfied, the body thickness distribution determination unit 44 determines the corrected body thickness distribution T _n-1 as the body thickness distribution of the imaging environment, and the determined body thickness distribution is determined. The correction distribution Tr is determined.

The method for determining the end condition of the body thickness distribution determination unit 44 will be described using the process of determining the body thickness distribution of the subject image Ik as an example. In the following processing, in the case of processing for determining the body thickness distribution of the imaging environment, the reference image Ir is used instead of the subject image Ik, and the body thickness distribution (first body thickness distribution) of the virtual model Mk of the subject K is used. Instead, the body thickness distribution (second body thickness distribution) of the virtual model Mr of the imaging environment may be used. As shown in the equations (9) and (10), the body thickness distribution determination unit 44 defines an error value V _error indicating the difference between the subject image Ik and the estimated image Im, and sets an error value as an end condition. It is determined whether or not V _error is less than or equal to a threshold value. Further, as shown in Expression (10), the sum of squares of the pixel values of the difference image Id obtained by subtracting the estimated image Im from the subject image Ik is defined as an error function f _error . Note that any determination method capable of determining that the difference between the subject image Ik and the estimated image Im has become sufficiently small to be acceptable can be applied as the end condition.
Further, the error function f _error is not limited to the above example, and the error function f _error can be defined by any method that represents the difference between the subject image Ik (or the reference image Ir) and the estimated image Im. For example, as shown in the following equation (11), the sum of absolute values of the pixel values of the difference image Id obtained by subtracting the estimated image Im from the subject image Ik (or the reference image Ir) may be used as the error function f _error .

In addition, in Formula (1)-(11), the same code | symbol is attached | subjected to the same element and description is abbreviate | omitted. Any optimization method that minimizes the error value V _error indicating the difference between the subject image Ik and the estimated image Im can be applied. For example, a simplex method, a steepest descent method, or a conjugate gradient method can be used. When the estimated image Im is an estimated image of the reference image Ir, the reference image Ir is used instead of the subject image Ik in the equations (7), (9), (10), and (11).

  6 is a flowchart showing the flow of processing of the image analysis apparatus 30, FIG. 7 is a flowchart showing the flow of correction distribution determination processing, and FIG. 8 is a flowchart of body thickness distribution determination processing. The processing flow of the image analysis device 30 will be described with reference to FIGS.

  First, as illustrated in FIG. 6, the image acquisition unit 33 acquires a reference image Ir (S01). Then, the correction distribution acquisition unit 35 acquires the correction distribution Tr based on the scattered radiation component of the reference image Ir (S02).

The correction distribution acquisition process in S02 will be described in detail with reference to FIG. The virtual model acquisition unit 41 acquires the virtual model Mr of the imaging environment having the initial body thickness distribution T ₀ ^r from the storage unit 32 (S11).

  Next, the estimated image generation unit 42 estimates an estimated primary line image Ip obtained by estimating a primary line image obtained by radiography of the imaging environment virtual model Mr and a scattered radiation image obtained by radiography of the imaging environment virtual model Mr. An image obtained by combining the estimated scattered radiation image Is is generated as an estimated image Im of the reference image Ir that is an image obtained by estimating the reference image Ir (S12).

  Subsequently, the body thickness distribution determination unit 44 determines whether the difference between the estimated image Im of the reference image Ir and the reference image Ir satisfies the end condition using the conditional expressions (9) and (10) (S13). ).

When the error value V _error does not satisfy the termination condition (S13: No), the body thickness distribution determination unit 44 uses the conditional expressions (6) and (7) to calculate the body thickness distribution T _n-1 (n = 1). In the case of (2), a correction process for correcting the initial body thickness distribution T ₀ ^r ) is performed (S14).

When the corrected body thickness distribution _Tn is acquired, the body thickness distribution determination unit 44 updates the value of n by incrementing it by 1 (assuming n = n + 1), and the virtual model acquisition unit 41 corrects the corrected body thickness. Distribution _Tn is acquired (S11). Then, the obtained body thickness distribution T _n, estimating image generating unit 42 and the body thickness distribution determiner 44 performs similarly to each of the above processing of S11 to S13. The reference to the error value _{V error} indicating a difference between the estimated image Im of the image Ir and the reference image Ir is satisfies the termination condition, in the same manner as described above, correction processing of the body thickness distribution _{T n} and (S14), was modified body an acquisition process of the virtual model Mr which has a thickness distribution _{T n} (S11), the body thickness distribution _{T n} generation process of a new estimation image Im with the (S12), the newly generated estimation image Im and the reference image Ir The process of determining whether the difference between the two and the end condition is satisfied (S13) is repeated.

On the other hand, when the body thickness distribution determination unit 44 determines that the error value V _error satisfies the termination condition (S13: Yes), the body used for the error value V _error when the termination condition is satisfied. The thickness distribution Tn is determined as the body thickness distribution of the reference image Ir, the body thickness distribution is acquired as the correction distribution Tr, and the correction distribution acquisition process is terminated (S15).

  Returning to FIG. 6, the subsequent process will be described. The image acquisition unit 33 acquires the subject image Ik (S03). Next, the body thickness distribution acquisition unit 36 estimates and acquires the body thickness distribution Tk based on the scattered radiation component of the subject image Ik (S04).

The correction distribution acquisition process in S04 will be described in detail with reference to FIG. The virtual model acquisition unit 41 acquires the virtual model Mk of the subject K having the initial body thickness distribution T ₀ ^k from the storage unit 32 (S21).

  Next, the estimated image generation unit 42 estimates the primary line image Ip obtained by radiography of the virtual model Mr in the imaging environment of the subject K and the scattering obtained by radiography of the virtual model Mr of the subject K. An image obtained by synthesizing the estimated scattered radiation image Is obtained by estimating the line image is generated as an estimated image Im that is an image obtained by estimating the subject image Ik (S22).

  Subsequently, the body thickness distribution determination unit 44 determines whether or not the difference between the subject image Ik and the estimated image Im satisfies the end condition using the conditional expressions (9) and (10) (S23).

When the error value V _error does not satisfy the termination condition (S23: No), the body thickness distribution determination unit 44 uses the conditional expressions (6) and (7) to calculate the body thickness distribution T _n-1 (n = 1). In the case of ( ₂ ), a correction process for correcting the initial body thickness distribution T ₀ ) is performed (S24).

When the corrected body thickness distribution Tn is acquired, the body thickness distribution determination unit 44 increases the value of n by one and updates it (n = n + 1), and the virtual model acquisition unit 41 corrects the corrected body thickness distribution. Tn is acquired (S21). Then, the obtained body thickness distribution T _n, estimating image generating unit 42 and the body thickness distribution determiner 44 performs similarly to each of the above processing of S21 to S23. Then, until the error value V _error indicating the difference between the subject image Ik and the estimated image Im satisfies the termination condition, the correction processing (S24) of the body thickness distribution Tn and the corrected body thickness distribution Tn are performed as described above. The difference between the acquisition process (S21) of the virtual model Mk having, the generation process (S22) of the new estimated image Im using the body thickness distribution Tn, and the difference between the newly generated estimated image Im and the subject image Ik ends. The process of determining whether the condition is satisfied (S23) is repeated.

On the other hand, when the body thickness distribution determining unit 44 determines that the error value V _error satisfies the termination condition (S23: Yes), the body used for the error value V _error when the termination condition is satisfied. The thickness distribution Tn is determined as the body thickness distribution Tk of the subject image Ik, and the body thickness distribution acquisition process is terminated (S25).

  Then, the body thickness distribution acquisition unit 36 aligns the corresponding positions of the subject image Ik and the reference image Ir, and corresponds to each other based on the body thickness distribution of the subject image Ik and the correction distribution of the reference image Ir. At the position, the body thickness correction amount value of the correction distribution Tr is subtracted from the body thickness value of the body thickness distribution Tk to determine the corrected body thickness distribution Tk ′ of the subject image Ik, and image analysis processing in the present embodiment Is finished (S05). In the image analysis process, the processes of S01 and S02, and the processes of S03 and S04 may be executed in a different order or may be executed in parallel.

  According to the present embodiment, the reference image Ir, which is a radiographic image captured without placing a subject in a predetermined imaging environment, is acquired, and the scattered radiation component included in the acquired reference image is captured of the reference image. The body thickness distribution of the virtual subject of the reference image Ir is estimated based on the scattered radiation component included in the reference image Ir, assuming that the virtual subject is a virtual subject representing the environment. The body thickness distribution of the virtual subject is a body thickness correction amount distribution for correcting an error in the body thickness distribution Tk of the subject image Ik, which is a radiographic image taken by arranging a desired subject in the imaging environment. By acquiring the acquired correction distribution Tr as a correction distribution Tr of a certain reference image, it is useful for correcting the error of the body thickness distribution Tn of the subject image Ik due to the influence of the scattered radiation component generated by the imaging environment. It can be provided as a broadcast.

  Further, the body thickness distribution acquisition unit 36 acquires the body thickness distribution Tk of the subject image Ik based on the scattered radiation component of the subject image Ik, and the correction distribution Tr acquired based on the scattered radiation component of the reference image Ir. Accordingly, in order to obtain the corrected body thickness distribution Tk ′ of the subject K in which the body thickness at each position in the body thickness distribution Tk is corrected, the body thickness distribution error caused by the imaging environment is preferably corrected and corrected appropriately. The corrected body thickness distribution Tk ′ of the subject K can be obtained. Further, the correction distribution acquisition unit 35 acquires the body thickness distribution of the imaging environment based on the scattered radiation component from the reference image Ir by the same method as the body thickness distribution acquisition unit 36, and corrects the acquired body thickness distribution of the imaging environment. Since it acquires as distribution, the said effect can be heightened more.

  The correction distribution acquisition unit 35 also acquires a virtual model acquisition unit 41 (second virtual model acquisition unit) that acquires a virtual model of a photographing environment having a predetermined body thickness distribution (second body thickness distribution), and a reference image. The imaging conditions of the imaging environment are obtained by applying the imaging conditions of the reference image and applying the imaging conditions of the reference image to the estimated primary line image estimated from the imaging environment virtual model and the radiation imaging of the imaging environment virtual model. An estimated image generation unit 42 (second estimated image generation unit) that generates, as an estimated image of a reference image that is an image obtained by estimating a reference image, an image obtained by combining the estimated scattered radiation image estimated from A correction unit 43 (second correction unit) that corrects the body thickness distribution of the virtual model of the shooting environment so that the difference between the reference image and the reference image is small, and the body thickness distribution of the virtual model of the corrected shooting environment corresponds to the shooting environment Do Since the body thickness distribution determining unit 44 (second body thickness distribution determining unit) that determines the positive distribution is provided, the body thickness distribution of the virtual subject representing the imaging environment based on the scattered radiation component from the reference image Ir is accurate. By using the body thickness distribution of the virtual subject as a correction distribution, it is possible to accurately correct an error in the body thickness distribution caused by the scattered radiation component due to the imaging environment.

  Further, as in the above-described embodiment, the body thickness distribution acquisition unit 36 is a virtual in which the body thickness distribution acquisition unit acquires a virtual model of the subject K having a predetermined body thickness distribution (first body thickness distribution). A model acquisition unit 41 (first virtual model acquisition unit) obtains imaging conditions for the subject image Ik, applies the imaging conditions for the subject image Ik, and is obtained by radiography of the virtual model of the subject K. An estimated primary line image obtained by estimating a line image from a virtual model of the subject K and a scattered radiation image obtained by radiography of the virtual model of the subject K are combined with an estimated scattered ray image estimated from the virtual model of the subject K. An estimated image generation unit 42 (first estimated image generation unit) that generates an image as an estimated image Im of a subject image Ik that is an image obtained by estimating a radiographic image obtained by radiography of the subject K; and a subject image Ik Estimated image of The correction unit 43 (first correction unit) that corrects the body thickness distribution of the virtual model so that the difference between m and the subject image Ik is small, and the corrected body thickness distribution Tn of the virtual model of the subject K is the subject. When the body thickness distribution determining unit 44 (first body thickness distribution determining unit) that determines the body thickness distribution Tk of K is provided, an accurate body thickness distribution based on the scattered radiation component is acquired from the subject image Ik. be able to.

Further, as shown in this embodiment, the virtual model acquisition unit 41 further acquires the virtual model Mk (or Mr) having a modified body thickness distribution T _n, estimating image generating unit 42 has been modified body the thickness further generates an estimated image Im from a virtual model Mk having a distribution T _n (or Mr), correcting section 43 further difference is reduced for the generated estimation image Im and the object image Ik (or the reference image Ir) when further modify the body thickness distribution T _n of the virtual model M as on the basis of the virtual model with a modified body thickness distribution T _n, by repeating the modification of the body thickness distribution T, the estimated image Im is The body thickness distribution T can be corrected more accurately so as to be approximated by the subject image Ik (or the reference image Ir). Therefore, the correction distribution Tr of the reference image Ir can be determined more accurately by using the body thickness distribution T _{n + 1} of the corrected virtual model Mr of the shooting environment as the correction distribution Tr of the shooting environment. Moreover, the body thickness distribution Tk of the subject image Ik can be determined more accurately by setting the corrected body thickness distribution Tn _{+ 1} of the virtual model M of the subject K as the body thickness distribution Tk of the subject K.

Further, according to the present embodiment, the virtual model M is obtained when the body thickness distribution acquisition unit 36 is sufficiently small that the difference between the estimated image Im and the subject image Ik (or the reference image Ir) is acceptable. the body thickness distribution T _n is determined as the body thickness distribution Tk of the subject K (or corrected distribution Tr of the reference image Ir). For this reason, the body thickness distribution is corrected repeatedly so that the estimated image Im and the subject image Ik (or reference image Ir) approximate to each other, and the subject image (or reference image Ir) is very accurately corrected. The body thickness distribution can be determined. Further, the body thickness distribution acquisition unit 36 determines whether or not the difference between the estimated image Im and the subject image Ik (or the reference image Ir) is equal to or less than a threshold value, whereby the estimated image Im and the subject image Ik ( Alternatively, it is preferably determined whether or not the difference from the reference image Ir) is sufficiently small to an acceptable level, and the estimated image Im and the subject image Ik (or the reference image Ir) are approximated to a body thickness distribution. As described above, the body thickness distribution of the subject image can be determined very accurately by correcting the body thickness distribution repeatedly. Note that the threshold value of the end condition may be different or may be the same when using the virtual model of the subject K and when using the virtual model of the imaging environment.

Further, as shown in the present embodiment, the correction unit 43 varies the body thickness of one partial region of the body thickness distribution T _n−1 of the virtual model M for each partial region of one or more pixels of the virtual model M. Then, the body thickness of the one part when the difference between the estimated image Im and the subject image Ik is minimized is calculated, and the body thickness distribution of the virtual model M is corrected by the calculated body thickness of each part. case, the correction value of the body thickness of each pixel can be accurately calculated, it is possible to obtain a suitably modified body thickness distribution T _n.

  In addition, the correction distribution acquisition unit 35 acquires the body thickness distribution according to the scattered radiation component of the reference image Ir by a method common to the body thickness distribution acquisition unit 36, and uses the acquired body thickness distribution as the correction distribution Tr. By correcting the body thickness distribution Tk using the correction distribution Tr, the influence of the error of the body thickness distribution due to the scattered radiation component due to the imaging environment is suitably reduced, and the corrected body that is the body thickness distribution of the subject K with high accuracy. A thickness distribution Tk ′ can be obtained. In addition, the correction distribution acquisition unit 35 and the body thickness distribution acquisition unit 36 share the components and functions of the virtual model acquisition unit 41, the estimated image generation unit 42, the correction unit 43, and the body thickness distribution determination unit 44. The configuration of the image processing apparatus can be further simplified. However, for desired matters required for the image processing apparatus, the correction distribution acquisition unit 35 and the body thickness distribution acquisition unit 36 are respectively a virtual model acquisition unit, an estimated image generation unit, a correction unit, and a body thickness determination unit. The body thickness estimation processing based on the scattered radiation component may be performed by different methods.

  In the above-described embodiment, the estimated image generation unit 42 may apply various methods that can generate the estimated primary line image Ip and the estimated scattered radiation image Is.

In the above-described embodiment, the imaging condition acquisition unit 34 acquires the imaging conditions of the subject image Ik and the reference image Ir, and the estimated image generation unit 42 uses the acquired imaging conditions to generate the estimated image Im. A parameter that varies depending on the shooting condition (for example, θ _{x ′, y ′} in the above formulas (3) and (5)) is selected, and the estimated image Im is generated using the selected parameters (S22, S32). It is preferable to carry out.

In the above embodiment, the correction unit 43 acquires the imaging conditions of the subject image Ik (or the reference image Ir), and varies according to the imaging conditions used for generating the estimated image Im according to the acquired imaging conditions. (For example, θ _{x ′, y ′} in the above formulas (7) and (8)) are selected, and the body thickness distribution correction process (S14 and S24) of the estimated image Im is performed using the selected parameters. It is preferable. In this case, the estimated image Im can be generated by appropriately setting parameters that vary depending on the imaging condition in accordance with the imaging condition of the subject image Ik (or the reference image Ir). Accurately estimated and generated. Therefore, as a result, the body thickness distribution of the subject K or (the body thickness distribution of the imaging environment) can be determined more accurately.

  In the image analysis device 30, the image acquisition unit 33 acquires a plurality of reference images in one shooting environment, and the correction distribution acquisition unit 35 acquires correction distributions for the acquired plurality of reference images, respectively. The thickness distribution acquisition unit 36 corrects the body thickness at each position in the body thickness distribution Tk of the subject K according to the average correction distribution obtained by averaging the plurality of correction distributions Tr corresponding to the acquired plurality of reference images. The corrected body thickness distribution Tk ′ of the specimen K may be acquired. In this case, for correcting the body thickness distribution Tk, instead of the correction distribution acquired from one reference image, an average correction distribution obtained by averaging a plurality of correction distributions Tr acquired from a plurality of reference images is used. In addition, it is possible to use the correction distribution that reflects the error of the body thickness due to the imaging environment more accurately by suppressing the influence of the scattered radiation component variation for each imaging, and more preferably corrects the body thickness distribution Tk of the subject K. can do.

  Further, in the image analysis apparatus 30, the image acquisition unit 33 acquires a plurality of radiographic images captured without arranging the subject in the imaging environment, and acquires an image obtained by averaging the plurality of radiographic images as a reference image. preferable. In this case, an average image of a plurality of radiographic images taken without placing a subject in an equivalent imaging environment is used as a reference image, thereby suppressing the influence of dispersion of scattered radiation components in each reference image. can do. As a result, it is possible to acquire a correction distribution reflecting an error in the body thickness due to the imaging environment as a result, and to more appropriately correct the body thickness distribution Tk of the subject K.

  Further, in the image analysis device 30, the correction distribution acquisition unit 35 acquires the correction distribution Tr for each of the plurality of reference images Ir to which different shooting conditions are applied, and associates the correction distribution corresponding to the shooting condition with each other, The body thickness distribution acquisition unit 36 acquires a correction distribution associated with imaging conditions similar to the imaging conditions of the subject image Ik, and each of the body thickness distributions Tk of the subject image Ik according to the acquired correction distribution Tr. It is preferable to obtain a corrected body thickness distribution Tk ′ obtained by correcting the body thickness at each position. For example, for each of the first to m-th shooting rooms, a reference image shot under different first to n-th shooting conditions is acquired, and a correction distribution is acquired for each reference image. It is conceivable to prepare a correspondence table in which photographing conditions are associated with correction distributions. For example, the first to nth imaging conditions can represent the product of tube voltage, tube current, and irradiation time. Then, at the time of photographing the subject image, a correction distribution corresponding to the photographing room of the subject image and the photographing condition of the subject image (or the photographing condition similar to the photographing condition of the subject image) is acquired based on the association table. do it. In this case, since the imaging conditions of the subject image Ik and the reference image Ir are similar, the body based on the imaging conditions is more than the case where the imaging conditions of the subject image Ik and the imaging conditions of the reference image Ir are different from each other. The body thickness distribution can be corrected using the correction distribution in which the variation with respect to the thickness correction distribution is suppressed, and the corrected body thickness distribution representing the body thickness distribution of the subject K can be acquired more accurately. It should be noted that for the similar determination of the shooting conditions, a condition in which one or more elements of the shooting conditions are common or approximate can be adopted as long as it is possible to determine that the shooting conditions are sufficiently similar.

  The correction distribution acquisition unit 35 acquires the correction distribution Tr for the reference image Ir to which specific shooting conditions are applied in the shooting environment, and the body thickness distribution acquisition unit 36 determines the reference image according to the acquired correction distribution Tr. You may acquire the correction | amendment body thickness distribution which each corrected the body thickness of each position in the body thickness distribution of the subject image to which the imaging conditions different from specific imaging conditions were applied in the imaging environment equivalent to an imaging environment. When there is little variation in the correction distribution due to the imaging conditions, the subject image to which the imaging conditions different from the specific imaging conditions are applied using the correction distribution acquired for the reference image to which the specific imaging conditions are applied By correcting the body thickness distribution, the body thickness distribution of the subject image can be corrected easily and suitably.

  Also, in a large-scale medical institution having a plurality of radiation imaging systems, there may be cases where the elements and arrangement of the imaging environment such as the imaging table, imaging apparatus, imaging room wall and floor are common in the plurality of imaging rooms. possible. In such a case, a correction distribution based on a reference image captured without placing a subject under a predetermined shooting condition in a shooting environment of a specific shooting room is acquired, and the acquired correction distribution is replaced with another equivalent distribution. It may be used for correcting the body thickness distribution of the subject image in the imaging room having the imaging environment. In addition, if the elements and arrangements that are the main factors of the generation of scattered radiation caused by the shooting environment are substantially common among the elements of the shooting environment, even if the types and arrangements of other elements are different, It can be considered that the environment is substantially common. For this reason, even in a plurality of different shooting rooms, among the elements of the shooting environment, when one or more main elements contributing to the generation of scattered radiation caused by the shooting environment and the arrangement thereof are substantially common , Acquiring a correction distribution based on a reference image shot without placing a subject under a predetermined shooting condition in a shooting environment of one shooting room, and using the acquired correction distribution in a shooting room having another equivalent shooting environment It can be used to correct the body thickness distribution of the subject image.

  Further, for example, it is conceivable to perform the process of acquiring the correction distribution Tr of the present invention when a predetermined photographing environment is installed. In this case, when a new shooting environment is installed, the reference image Ir is shot and acquired in the shooting environment, the correction distribution Tr is acquired based on the acquired reference image Ir, and then the same shooting environment (shooting of the reference image) is acquired. The acquired correction distribution Tr is used to correct the body thickness distribution of each subject image with respect to a plurality of subject images taken in a plurality of different subjects photographed in an imaging environment equivalent to the environment). Can do. Even in a shooting environment in which the correction distribution is once acquired, it is preferable to update the correction distribution Tr by appropriately performing a process of acquiring the correction distribution Tr after a desired period. In this case, it is possible to suitably obtain and use a correction distribution corresponding to the imaging environment by reducing the influence of a change over time such as deterioration of each element such as a radiation source arranged in the imaging environment.

  As a second embodiment which is a modification of the present embodiment, the body thickness distribution acquisition unit 36 calculates the body thickness distribution of the subject in the subject image Ik based on the scattered radiation component included in the subject image Ik. When it is determined that there is a possibility that an error due to the scattered radiation component from the imaging environment has occurred in at least a part of the acquired body thickness distribution, the body thickness distribution Tk of the subject according to the correction distribution Tr The corrected body thickness distribution Tk ′ of the subject K obtained by correcting the body thickness at each position in FIG.

  For example, FIG. 11 shows a flowchart showing a processing flow of the image analysis apparatus 30 according to the second embodiment. In FIG. 11, the processes of S41, S42, S43, S44, and S46 are the same as the processes of S01, S02, S03, S04, and S05 in FIG. As shown in FIG. 11, the image analysis apparatus 30 acquires a reference image (S41), a correction distribution acquisition process based on the reference image (S42), and a subject image, as in FIG. The process (S43) and the process (S44) for acquiring the body thickness distribution based on the subject image are sequentially executed. Next, when the body thickness distribution acquisition unit 36 determines whether there is a possibility that an error in the body thickness distribution has occurred, and determines that there is a possibility that an error in the body thickness distribution has occurred ( S45, Y), a process of obtaining a corrected body thickness distribution by correcting the body thickness distribution (S46). On the other hand, when it is determined that there is no possibility of an error in the body thickness distribution (S45, N), the process of correcting the body thickness distribution and acquiring the corrected body thickness distribution (S46) is omitted. In this case, since the body thickness distribution can be corrected only when necessary, the calculation load can be reduced.

  FIG. 12 is a flowchart showing a processing flow of the image analysis apparatus 30 according to the modification of the second embodiment. As shown in FIG. 12, a process for acquiring a reference image (a process corresponding to S01 in FIG. 6) and a process for acquiring a correction distribution based on the reference image (a process corresponding to S02 in FIG. 6) are performed on the subject image. May be executed after the process of acquiring the body thickness distribution based on the process (the process corresponding to S04 in FIG. 6). In FIG. 12, the processing of S51, S52, S54, S55, and S46 is the same as the processing of S03, S04, S01, S02, and S05, and detailed description thereof is omitted. As illustrated in FIG. 12, the image analysis apparatus 30 executes a process (S51) of acquiring a subject image, then executes a process of acquiring a body thickness distribution (S52), and then performs a body thickness distribution acquisition unit 36. Is used to determine whether there is a possibility that an error in the body thickness distribution has occurred, and if it is determined that an error in the body thickness distribution has occurred (S53, Y), a reference image is acquired. The processing (S54) for performing correction, the processing for acquiring the correction distribution based on the reference image (S55), and the processing for correcting the body thickness distribution and acquiring the correction body thickness distribution (S56) can be executed in this order. . When it is determined that there is no possibility of an error in the body thickness distribution (Y in S54), a reference image acquisition process (S54) and a process of acquiring a correction distribution based on the reference image (S55) The process of correcting the body thickness distribution and acquiring the corrected body thickness distribution (S56) can be omitted.

  In the processing for determining whether an error has occurred in the body thickness distribution in FIGS. 11 and 12 (S45, S53), the body thickness distribution acquisition unit 36 calculates the body thickness distribution calculated for each pixel from the subject image Ik. By setting each pixel corresponding to the body thickness distribution as a target pixel, the body thickness of the target pixel is a value obtained by averaging each body thickness of 3 × 3 pixels including the target pixel and pixels adjacent to the target pixel, Smooth the body thickness distribution. Then, for each pixel corresponding to each other, the body thickness distribution after smoothing is subtracted from the body thickness distribution before smoothing. Then, when the absolute value of the difference value in the difference value distribution obtained by subtraction exceeds a predetermined reference value, it is determined that there is a possibility that an error has occurred in the acquired body thickness distribution. Further, for example, in order to prevent a change in body thickness distribution at the edge of the subject region from being erroneously recognized as an error in the body thickness distribution caused by the imaging environment, a region where the subject of the subject image Ik exists by a well-known method. And the region excluding the edge of the subject region and the vicinity of the edge is set as a target region for determining the body thickness distribution, and the above determination processing may be executed only for the body thickness distribution included in the target region. The above reference value is calculated based on the standard shape of the subject as a reference value that is larger than the maximum value of the change in body thickness between two neighboring positions. When the absolute value of the change in body thickness between two nearby positions exceeds a reference value, it can be set to determine that the body thickness distribution may include an error due to the imaging environment. The two neighboring positions may be the positions of the pixels adjacent to each other, or may be positions that are separated from each other by several to several tens of pixels as long as they are sufficiently close to each other.

  In addition, the above-described method of “determining that at least a part of the body thickness distribution may have an error due to a scattered radiation component due to the imaging environment” means that at least a part of the body thickness distribution of the subject is Anything can be determined that the scattered radiation component from the light source can reach the radiation detector and is calculated to be larger than the actual body thickness distribution of the subject and may cause an error in the body thickness distribution. It may be a method. In addition, for example, the body thickness obtained by adding a predetermined thickness to the maximum body thickness assumed based on the shape of a standard subject is acquired as a reference value, and the maximum body thickness included in the body thickness distribution is used as a reference. When the value is larger than the value, it may be set so as to determine that the body thickness distribution may include an error due to the imaging environment.

  In the radiological image analysis method according to the present invention, it is preferable to shorten the processing time. For this purpose, the image acquisition unit 33 generates a reduced image in which the resolution of the radiographic image captured without placing the subject in the imaging environment is reduced to a predetermined resolution, and acquires the generated reduced image as a reference image. It is preferable to do. For example, as a modification of the first embodiment, in the process (S01) of acquiring the reference image in FIG. 6, the image acquisition unit 33 is a radiation imaged without arranging the subject K in a predetermined imaging environment. A reduced image in which the resolution of the image is reduced to a predetermined resolution is generated, and the generated reduced image is acquired as a reference image. Then, in the process (S02) in which the correction distribution acquisition unit 35 acquires the correction distribution of the reference image Ir, the correction distribution Tr of the shooting environment is acquired by performing the processes of S11 to S15 in FIG. . Thereafter, the correction distribution Tr may be appropriately interpolated and enlarged as necessary to match the resolution of the subject image Ik. In this case, it is possible to reduce the processing load of the process of acquiring the correction distribution of the reference image (for example, S02 in FIG. 6) and increase the processing time.

  Similarly, the image acquisition unit 33 generates a reduced image in which the resolution of a radiographic image captured by arranging the subject is reduced to a predetermined resolution, and acquires the generated reduced image as the subject image Ik. It is preferable to do. For example, as a modification of the first embodiment, in the process (S03) of acquiring the subject image in FIG. 6, the image acquisition unit 33 has taken the radiation imaged by arranging the subject in a predetermined imaging environment. A reduced image in which the resolution of the image is reduced to a predetermined resolution is generated, and the generated reduced image is acquired as a subject image Ik (reduced subject image). Then, in the process in which the body thickness distribution acquisition unit 36 acquires the body thickness distribution Tk of the reduced subject image Ik (S04 in FIG. 6), the processes in S21 to S25 in FIG. 8 are performed on the reduced subject image Ik. Thus, the body thickness distribution Tk of the reduced subject image is acquired. Then, the body thickness distribution Tk may be appropriately interpolated and enlarged as necessary. In this case, the processing load of the process for acquiring the body thickness distribution of the subject image (for example, S04 in FIG. 6) can be reduced, and the processing time can be increased. It should be noted that the reduced reference image and the reduced subject image preferably have the smallest possible resolution within a range that maintains the resolution necessary for processing.

  Another modification of the body thickness acquisition process of the body thickness distribution acquisition unit 36 will be described as a third embodiment. FIG. 9 is a schematic block diagram of the body thickness distribution acquisition unit 36 in the third embodiment, and FIG. 10 is a flowchart showing the processing of the body thickness distribution acquisition unit 36 in the third embodiment. Note that the image analysis apparatus 30 in the second and third embodiments has the same configuration and functions as those in the first embodiment except for the body thickness distribution acquisition unit 36, and therefore other than the body thickness distribution acquisition unit 36. A description of the configuration and functions is omitted.

  As shown in FIG. 9, the body thickness distribution acquisition unit 36 according to the third embodiment includes a plurality of model images that are radiographic images obtained by imaging a plurality of models different from the subject. A model information acquisition unit 51 that acquires model information obtained by associating the body thickness distribution of the model image acquired based on the scattered radiation component of the model image, and positions corresponding to each other between the subject image and the model image. The alignment unit 52 to be aligned and the imaging target information representing the characteristics of the subject of the subject image are acquired, and the imaging target information indicating the characteristics of the models of the plurality of model images is acquired based on the plurality of model information, respectively. A model image having pre-shooting target information similar to the shooting target information of the subject image is identified, and the body thickness distribution associated with the identified model image is determined as the body thickness distribution of the subject image. And a new body thickness distribution determiner 53. Hereinafter, the new body thickness distribution determining unit 53 is referred to as a body thickness distribution determining unit 53.

  As the “model”, any model that can virtually represent the subject can be used. For example, when the subject is a human body, the model is preferably a comparative human body. In addition, when there are a plurality of model images having similar imaging target information, among the model images having similar imaging target information, the subject image Ik is associated with a model image having the same imaging region and imaging condition. It is preferable to determine the obtained body thickness distribution as the body thickness distribution of the subject image.

  In addition, “photographing target information” is specified by any method as long as it represents the characteristics of the subject to be photographed, such as the shape of the skeleton of the subject to be photographed, the size of the skeleton, and the physical features such as muscle mass and fat mass. It may be. For example, the histogram width of the signal value of the radiographic image (difference between the maximum value and the minimum value of the histogram of the signal value), the representative length (for example, the width of the abdomen) of the subject in the radiographic image, the height of the subject, Weight, sex, age (adult, child), etc. may be used. Further, the shooting target information may be one or plural. Further, the shooting target information may be input by user input, or may be shooting target information extracted from the subject image and the model image.

  FIG. 10 is a flowchart of the process of the body thickness distribution acquisition unit 36 (a process corresponding to the process of S04 in FIG. 6) in the third embodiment. In the third embodiment, prior to the body thickness distribution acquisition process, the storage unit 32 stores the same model Gi (1 <i <n) as the subject image Ik, which is a human body different from the subject K. In the imaging environment, a plurality of model images Igi (1 <i <n), which are radiographic images obtained by radiographing portions overlapping with the subject image Ik under the same imaging conditions, and the model image Igi Model information Cgi (1 <i <n) that associates the body thickness distribution Tgi (1 <i <n) of the model Gi that is the subject of (1 <i <n) is created and stored in advance.

  The calculation of the body thickness distribution Tgi (x, y) (1 <i <n) of the model Gi is acquired based on the scattered radiation component of the model image Igi. In the third embodiment, as shown in the example of FIG. 3, a correspondence table that associates the pixel value of the model image with the body thickness of the subject is prepared for each combination of the imaging conditions and the characteristic information of the imaging target region. And remember. Then, an association table corresponding to the imaging condition of the model image and the region to be imaged is acquired, and by specifying the body thickness corresponding to the pixel value at each position of the model image based on the association table, the model image Get body thickness distribution.

  The processing of the body thickness distribution acquisition unit 36 in the third embodiment will be described below using FIG. First, the model information acquisition unit 51 acquires a plurality of model information (S31). Then, the alignment unit 52 performs processing such as relative movement, rotation, and enlargement / reduction on the subject image Ik and the model image Igi (1 <i <n) included in the model information, so that positions corresponding to each other are obtained. Alignment is performed (S32). Next, the body thickness distribution determination unit 53 acquires the histogram width of the subject image Ik, which is the photographing target information of the subject image Ik (S33), and similarly photographs from the model image Igi (1 <i <n). The histogram width of the subject image Ik, which is target information, is acquired (S34). Then, the body thickness distribution determination unit 53 compares the histogram widths of the signal values of the subject image Ik and the plurality of model images Igi, and specifies the model image Igo having the closest histogram width to the histogram width of the subject image Ik. (S35). In a radiographic image, as the subject's physique increases, the spread of the signal value tends to increase. Therefore, by using the histogram width of the signal value as the imaging target information, the subject K and the physique of the subject image Ik are used. It is possible to specify a model image Igo having a model Go similar in the subject.

  Then, the body thickness distribution determination unit 53 determines the body thickness distribution Tgo associated with the identified model image Igo as the body thickness distribution Tk of the subject image Ik, and the body thickness distribution acquisition analysis process of the third embodiment Is finished (S36).

  According to the third embodiment, the body thickness of a model image having imaging target information similar to the imaging target information of the subject K in the body thickness distribution of the model image captured in the same imaging environment as the subject image Ik. Since the distribution can be determined as the body thickness distribution of the subject image Ik, the body thickness distribution of the subject K determined with high accuracy can be acquired as the body thickness distribution. As a result, the corrected body thickness distribution obtained by correcting the body thickness distribution with the correction distribution can be made more accurate.

  In each of the above embodiments, the image acquisition unit 33 acquires the subject image Ik photographed without using the grid. However, the present invention is not limited to this, and the image acquisition unit 33 uses the grid. You may acquire the image which performed the process which removes the striped pattern resulting from a grid with respect to the radiographic image which image | photographed the subject K by radiography as the subject image Ik. The processing for removing the stripe pattern caused by the grid may be performed by various methods capable of removing the stripe pattern caused by the grid. For example, a method described in JP 2012-203504 A may be used. You can refer to it.

  In each of the above-described embodiments, the normal radiography in which the radiation transmitted through a predetermined imaging region of the subject is converted into a radiographic image using one radiation detector has been described as an example, but various imaging methods for the subject are used. Long imaging (for example, refer to Japanese Patent Application Laid-Open No. 2011-224337) for obtaining a radiographic image of a long imaging region (for example, the entire body of the subject) that does not fit in one image, and radiation There is divided imaging in which a radiographic image is captured for each divided area by using one divided area obtained by dividing the imaging area of the detector into a plurality of areas (see, for example, JP-A-2002-263089). The radiological image analysis method according to the present invention can be suitably applied to a radiographic image related to long imaging or divided imaging.

  In long imaging, a radiation detector formed by connecting a plurality of radiation detectors so as to ensure a long imaging region and combining them in a longitudinal shape may be used. FIG. 13 shows an example of a radiation detector for long imaging. As shown in FIG. 13, the radiation detectors 14A, 14B, and 14C having a substantially rectangular flat plate shape are connected to each other at an end portion by a connecting portion (not shown) to form one long imaging radiation detector. Each of the radiation detectors 14A, 14B, and 14C is disposed so that the end portions thereof overlap each other so that the radiation detector is disposed without interruption in the imaging region. Here, the radiation detectors 14A, 14B, and 14C are ISS (Irradiation Side Sampling) radiation detectors. The ISS type radiation detectors 14A, 14B, and 14C are glass substrates 14Aa, 14Ba, and 14Ca, and a scintillator 14Ab that is a phosphor layer that converts radiation into light along the radiation irradiation direction (see the dashed line arrow in the figure). , 14Bb, 14Cb are provided in this order. The glass substrate is provided with a signal output unit and a sensor unit (not shown). Light emitted from the scintillator is converted into electric charges by the sensor unit, and a radiation image is read. Further, the glass substrates 14Aa, 14Ba, and 14Ca have a substantially rectangular flat plate shape, and the scintillators 14Ab, 14Bb, and 14Cb have a substantially rectangular flat plate shape that is slightly smaller than the glass substrates 14Aa, 14Ba, and 14Ca.

  In such a long-radiation radiation detector, since the radiation detectors are located in the vicinity of each other particularly in a portion where the radiation detectors are overlapped, a glass substrate or a scintillator of one radiation detector, etc. The scattered radiation generated in the above reaches the other radiation detectors, and it is easy to cause an error in the calculation of the body thickness distribution based on the scattered radiation. The lower part of FIG. 13 shows correction distributions TrA, TrB, and TrC for each pixel of the reference images of the radiation detectors 14A, 14B, and 14C, respectively, in the longitudinal direction of the long imaging radiation detector. As shown in the correction distribution TrA, the radiation detector 14A has an end that is particularly close to the radiation detector 14B when the scattered radiation generated by the radiation detector 14B (particularly the end 14Bc) disposed in the vicinity arrives. The scattered dose is large in the portion 14Ac, and the error in the body thickness distribution increases as shown by Sa. Similarly, as shown in the correction distribution TrB, the radiation detector 14B reaches the scattered radiation generated by the radiation detector 14A (particularly the end portion 14Ac) and the radiation detector 14C (particularly the end portion 14Cc) disposed in the vicinity. As a result, the scattered dose is large especially at the end 14Bc in the vicinity of the radiation detector 14A, and the error in the body thickness distribution increases as shown by Sb1, and particularly at the end 14Bd located in the vicinity of the radiation detector 14C. The scattered dose is large, and the error of the body thickness distribution becomes large as shown by Sb2. Similarly, as shown in the correction distribution TrC, the radiation detector 14C has a special effect on the radiation detector 14B due to the arrival of scattered radiation generated by the radiation detector 14B (particularly the end 14Bd) disposed in the vicinity. The scattered dose is large at the end portion 14Cc located in the vicinity of, and the error in the body thickness distribution increases as shown by Sc.

  In a radiation imaging system using a radiation detector for long imaging composed of a plurality of radiation detectors as described above, when the radiation passes through the radiation detector depending on the material of the glass substrate or the material of the scintillator The resulting scattered dose may also be large. For this reason, it is very useful to acquire the correction distribution of the reference image by the method of the present invention and correct the error of the body thickness distribution by the acquired correction distribution. In this case, the body thickness distribution of the subject can be accurately calculated by suitably suppressing an error in the body thickness distribution due to the scattered radiation generated by the radiation detector disposed in the vicinity. Further, when a plurality of radiation detectors are combined and used as one long imaging radiation detector, the correction distribution of the reference image (see TrA, TrB, and Trc in FIG. 13) and the subject for each radiation detector. The body thickness distribution of the image may be calculated, and the correction distribution may be acquired by subtracting the correction distribution from the corresponding body thickness distribution for each individual radiation detector. Further, known image processing is performed so as to combine the radiographic images obtained by the respective radiation detectors to generate a reference image of the entire photographing region of the long imaging, and the correction distribution of the reference image (see TrABC in FIG. 13). ), A subject image of the entire imaging region of the long image is generated, the body thickness distribution of the subject image is calculated, and the correction distribution is obtained by subtracting the correction distribution from the body thickness distribution. Good. In the correction distribution TrABC of the long-distance radiation detector shown in FIG. 13, the overlapping area of the radiation detectors 14A and 14B is a correction distribution related to the overlapping area of the radiation detector 14A used for the reference image. The overlapping region of 14A and 14C is a correction distribution related to the overlapping region of the radiation detector 14C used for the reference image.

  Further, even in divided imaging in which radiation imaging is performed at the time of different imaging for each divided area by limiting the irradiation range according to the divided area, the correction distribution of the reference image according to the present invention is acquired, and the subject is obtained by the acquired correction distribution. It is preferable to correct an error in the body thickness distribution of the image. FIG. 14 shows an example of a divided area of the radiation detector 14 at the time of divided imaging. In the example shown in FIG. 14, five divided regions 141, 142, 143, 144, and 145 having a predetermined size that can image the region of interest of the subject are set for the radiation detector 14. In divided imaging, for example, a radiation shielding object (such as a radiation shielding plate) having an opening having a size corresponding to the divided area is used in order to limit the irradiation range according to the divided area. Then, by arranging the radiation shielding plate between the subject and the radiation detector so that the divided region and the opening face each other and performing radiation imaging, each divided region 141, Radiographic images representing the desired imaging target regions are acquired at 142, 143, 144, and 145, respectively. Note that the radiation detector may be read after each divided imaging to acquire the radiation image of each divided area every time. After the divided imaging for all the divided areas is completed, the radiation detector is read collectively. Thus, the radiation images of all the divided areas may be acquired together.

  In the divided shooting, the correction distribution of the reference image is acquired for each divided region, and the body thickness distribution of the divided image shot in each divided region is obtained using the correction distribution at the position corresponding to each divided region. Correction may be performed. In addition, when it can be considered that the correction distributions in the respective divided regions are substantially equal, a correction distribution corresponding to one divided region may be acquired and used as the correction distribution of all the divided regions. In this case, the calculation load of the process for obtaining the correction distribution is reduced, which is suitable for speeding up the process.

  In addition, in the IVR (Interventional Radiology) technology for performing examination / treatment while observing the target site with a moving image etc. using an image diagnostic device, a radiation shield having an opening, etc. In some cases, radiography is performed a plurality of times by limiting the range in which radiation is irradiated to an imaging region of a predetermined size including the target region of the subject. In such a case as well, the correction distribution may be acquired in one imaging region and used in common as the correction distribution of the imaging region for subsequent radiography. Further, even when radiography is performed a plurality of times while shifting the imaging area of a predetermined size little by little according to the progress of the treatment of the subject, the correction distributions in the shifted imaging areas can be regarded as almost equal. In this case, a correction distribution may be acquired in one shooting area and used as the correction distribution of the shooting area at each shifted position. Also in this case, it is suitable for reducing the calculation load of the process for obtaining the correction distribution and for speeding up the process.

  Note that the corrected body thickness distribution (corrected body thickness distribution) obtained by the present invention is various for determining the image processing conditions corresponding to the body thickness of the subject with respect to the subject image. Can be used for processing. For example, the corrected body thickness distribution obtained by the present invention is applied to gradation processing such as density and contrast, noise suppression processing, dynamic range adjustment processing, frequency enhancement processing, etc. on a subject image that is a still image or a moving image. It is possible to use it. Further, the corrected body thickness distribution obtained by the present invention can be used for any processing for determining an imaging condition corresponding to the body thickness for the subject image. When each image processing condition or imaging condition is determined using the corrected body thickness distribution obtained by the present invention, the determined image processing is performed by applying an accurate body thickness distribution to the subject image. The effect of improving the image quality due to conditions or shooting conditions can be enhanced.

  Further, for example, an energy subtraction technique or a time-lapse subtraction technique (T-sub technique) for acquiring a radiographic image by a difference between two radiographic images acquired by imaging a high energy energy and a low energy radiation with different tube voltages. ), In accordance with the corrected body thickness distribution of the subject obtained by the present invention, when the low energy image is subtracted from the high energy image, the weighting of the high energy image at the position where the body thickness is large is increased. You may perform the process which determines a weighting coefficient. In this case, by applying an accurate body thickness distribution to the subject image, the influence of the occurrence of the beam hardening phenomenon in which the radiation quality varies depending on the thickness of the subject is reduced. The image quality of the post-image can be preferably improved.

  Further, in the technical field of acquiring a plurality of radiographic images by capturing a plurality of radiographic images by imaging the same subject a plurality of times, such as a radioscopic image application technique and a tomosynthesis technique, the method of the present invention It is preferable to acquire the corrected body thickness distribution of the subject and determine the imaging condition of the subject to be subsequently imaged based on the acquired body thickness distribution. Since a subject image that is subsequently captured can be captured under appropriate imaging conditions according to the body thickness, the subsequently captured subject image can be an image with an image quality suitable for diagnosis. Further, for example, when generating and displaying a tomographic image at a slice position according to a user designation from a three-dimensional image obtained by the tomosynthesis technique, the range of the slice position is more determined based on the corrected body thickness distribution of the subject. An index indicating a slice interval or the like such as a slider bar that can be accurately specified and can select a slice position can be displayed. Also, in the technical field of acquiring the plurality of radiographic images by capturing the same subject multiple times from the same imaging position, such as the application technique of the radiographic image, the reference image is first acquired and corrected distribution by the method of the present invention It is preferable to acquire the corrected body thickness distribution by correcting the body thickness distribution of each subject image subsequently captured based on the acquired correction distribution.

  Each of the above embodiments is merely an example, and all of the above description should not be used to limit the technical scope of the present invention. The aspect of the present invention is not limited to the above-described individual examples (first to third embodiments, other modifications and application examples), and any combination of elements of the individual examples is not limited to the present invention. In addition, various modifications that can be conceived by those skilled in the art are also included. That is, various additions, modifications, and partial deletions can be made without departing from the concept and spirit of the present invention derived from the contents defined in the claims and equivalents thereof.

  In addition, the system configuration, the hardware configuration, the processing flow, the module configuration, the user interface, the specific processing content, etc. in the above embodiment have been variously modified without departing from the spirit of the present invention. It is included in the technical scope of the present invention. For example, some or all of the components of the image analysis apparatus may be configured by a single workstation, and may be configured by one or more workstations, servers, and storage devices connected via a network. It may be configured.

  In the above-described embodiment, the body thickness distribution determination process and the correction distribution are performed using the radiographic image acquired by the imaging apparatus 10 that captures a radiographic image (subject image or reference image) of the subject using the radiation detector 14. Although the determination process is performed, the radiographic image information of the subject is accumulated and recorded on a stimulable phosphor sheet as a radiation detector disclosed in JP-A-8-266529, JP-A-9-24039, and the like. Of course, the present invention can be applied even when a radiation image obtained by photoelectrically reading from a phosphor sheet is used.

DESCRIPTION OF SYMBOLS 10 Imaging device 20 Control apparatus 30 Image analyzer (radiation image analyzer)
31 Image acquisition unit 32 Storage unit 33 Image acquisition unit 34 Imaging condition acquisition unit 35 Correction distribution acquisition unit 36 Body thickness distribution acquisition unit 41 Virtual model acquisition unit (first virtual model acquisition unit, second virtual model acquisition unit)
42 Estimated image generator (first estimated image generator, second estimated image generator)
43 correction parts (first correction part, second correction part)
44 Body thickness distribution determination unit (first body thickness distribution determination unit, second body thickness distribution determination unit)
51 Model Information Acquisition Unit 52 Alignment Unit 53 Body Thickness Distribution Determination Unit (New Body Thickness Distribution Determination Unit)
K Subject Ik Subject image Ir Reference image Im Model image Mk Virtual model of subject image Mr Virtual model of reference image Tk Body thickness distribution of subject image Tr Body thickness distribution of reference image Tn Body thickness distribution of model image Tp Primary Line transmittance Ts Scattered line transmittance

Claims (13)

An image acquisition unit that acquires a reference image that is a radiographic image captured without placing a subject in a predetermined imaging environment;
The scattered radiation component included in the acquired reference image is regarded as being generated by a virtual subject that is a virtual subject representing the imaging environment of the reference image, and the scattered radiation component included in the reference image Based on the estimated body thickness distribution of the virtual subject based on the reference image, the estimated body thickness distribution of the virtual subject is a radiographic image obtained by placing a desired subject in the imaging environment. A radiological image analysis apparatus comprising: a correction distribution acquisition unit that acquires a correction distribution of a reference image that is a distribution of a body thickness correction amount for correcting an error in a body thickness distribution of a specimen image.   Obtaining the body thickness distribution of the subject of the subject image based on the scattered radiation component included in the subject image, and calculating the body thickness at each position in the body thickness distribution of the subject according to the correction distribution, respectively The radiological image analysis apparatus according to claim 1, further comprising a body thickness distribution acquisition unit that acquires the corrected body thickness distribution of the corrected subject. An imaging condition acquisition unit that acquires imaging conditions of the subject image and the reference image,
The correction distribution acquisition unit acquires the correction distribution for each of the plurality of reference images to which different shooting conditions are applied, and associates the shooting conditions with the correction distribution of the reference image corresponding to the shooting conditions. ,
The body thickness distribution acquisition unit acquires a correction distribution of the reference image associated with an imaging condition similar to the imaging condition of the subject image, and the body of the subject image according to the acquired correction distribution The radiological image analysis apparatus according to claim 2, wherein the corrected body thickness distribution obtained by correcting the body thickness at each position in the thickness distribution is acquired. Wherein the correction profile acquisition unit acquires the correction distribution for the reference image specified shooting condition is applied in the shooting environment,
The body thickness distribution acquisition unit calculates a body thickness at each position in the body thickness distribution of the subject image to which an imaging condition different from the specific imaging condition is applied in the imaging environment according to the acquired correction distribution. The radiological image analysis apparatus according to claim 2, wherein each corrected body thickness distribution is acquired. The image acquisition unit acquires a plurality of the reference images corresponding to the shooting environment,
The correction distribution acquisition unit acquires the correction distribution for each of the acquired reference images,
The body thickness distribution acquisition unit corrected the body thickness at each position in the body thickness distribution of the subject according to an average correction distribution obtained by averaging a plurality of correction distributions corresponding to the acquired reference images. radiographic image analyzer according to claim 2 Symbol placement to obtain a corrected body thickness distribution of the object.   The body thickness distribution acquisition unit acquires the body thickness distribution of the subject based on a scattered radiation component included in the subject image, and at least a part of the acquired body thickness distribution is scattered radiation from the imaging environment. When it is determined that an error due to a component may occur, the corrected body thickness distribution of the subject obtained by correcting the body thickness at each position in the body thickness distribution of the subject according to the correction distribution The radiographic image analyzer according to any one of claims 2 to 5, which is acquired. The body thickness distribution acquisition unit,
A first virtual model acquisition unit that acquires a virtual model of the subject having a predetermined first body thickness distribution;
Estimated primary obtained by acquiring imaging conditions of the subject image, applying the imaging conditions of the subject image, and estimating a primary line image obtained by radiography of the virtual model of the subject from the virtual model of the subject A radiographic image obtained by radiography of the subject, which is an image obtained by synthesizing a ray image and a scattered radiation image obtained by radiography of the virtual model of the subject with an estimated scattered ray image estimated from the virtual model of the subject A first estimated image generation unit that generates an estimated image of the subject image that is an estimated image;
A first correction unit that corrects the first body thickness distribution of the virtual model to reduce a difference between the estimated image of the reference image and the reference image;
7. The apparatus according to claim 2, further comprising: a first body thickness distribution determining unit that determines the first body thickness distribution of the corrected virtual model of the subject as the body thickness distribution of the subject. Radiation image analyzer. The body thickness distribution acquisition unit,
A model in which a plurality of model images, each of which is a radiographic image obtained by photographing a plurality of models different from the subject, and the body thickness distribution of the model image acquired based on the scattered radiation component of the model image A model information acquisition unit for acquiring information,
Obtaining imaging target information representing the characteristics of the subject of the subject image, acquiring imaging target information representing the characteristics of the model of the plurality of model images based on the plurality of model information, respectively, The model image having pre-shooting target information similar to the shooting target information of the subject image is specified, and the body thickness distribution associated with the specified model image is determined as the body thickness distribution of the subject image. The radiological image analysis apparatus according to claim 2, further comprising a new body thickness distribution determination unit.   The said image acquisition part acquires the some radiographic image image | photographed without arrange | positioning the test object in the said imaging environment, The image which averaged the said some radiographic image is acquired as said reference image. The radiological image analysis apparatus of Claim 1.   The image acquisition unit generates a reduced image obtained by reducing the resolution of a radiographic image captured without placing the subject in the imaging environment to a predetermined resolution, and acquires the generated reduced image as the reference image. The radiographic image analyzer of any one of Claim 1 to 9. The correction distribution acquisition unit
A second virtual model acquisition unit that acquires a virtual model of the imaging environment having a predetermined second body thickness distribution;
Obtaining the imaging conditions of the reference image, applying the imaging conditions of the reference image, the estimated primary line image estimated from the virtual model of the imaging environment and the scattered radiation image obtained by radiography of the virtual model of the imaging environment A second estimated image generation unit that generates an image obtained by combining the estimated scattered radiation image estimated from the virtual model of the imaging environment as an estimated image of the reference image that is an image obtained by estimating the reference image;
A second correction unit that corrects the second body thickness distribution of the virtual model of the shooting environment to reduce a difference between the estimated image of the reference image and the reference image;
11. The apparatus according to claim 1, further comprising: a second body thickness distribution determining unit that determines the second body thickness distribution of the modified virtual model of the imaging environment as a correction distribution corresponding to the imaging environment. The radiation image analysis apparatus described. A radiological image analysis method executed by a radiological image analyzer,
An image acquisition step of acquiring a reference image, which is a radiographic image captured without placing a subject in a predetermined imaging environment;
The scattered radiation component included in the acquired reference image is regarded as being generated by a virtual subject that is a virtual subject representing the imaging environment of the reference image, and the scattered radiation component included in the reference image Based on the estimated body thickness distribution of the virtual subject based on the reference image, the estimated body thickness distribution of the virtual subject is a radiographic image obtained by placing a desired subject in the imaging environment. A radiographic image analysis method comprising: a correction distribution acquisition step of acquiring as a correction distribution of the reference image, which is a distribution of a body thickness correction amount for correcting an error in a body thickness distribution of a specimen image. On the computer,
An image acquisition step of acquiring a reference image, which is a radiographic image captured without placing a subject in a predetermined imaging environment;
The scattered radiation component included in the acquired reference image is regarded as being generated by a virtual subject that is a virtual subject representing the imaging environment of the reference image, and the scattered radiation component included in the reference image Based on the estimated body thickness distribution of the virtual subject based on the reference image, the estimated body thickness distribution of the virtual subject is a radiographic image obtained by placing a desired subject in the imaging environment. An image analysis program for executing a correction distribution acquisition step of acquiring a correction distribution of a reference image, which is a distribution of a body thickness correction amount for correcting an error in a body thickness distribution of a specimen image.