KR101065838B1 - Sinogram inpainting apparatus and method - Google Patents

Abstract

The sinogram reconstruction device reconstructs the sinogram to generate an initial image corresponding to the sinogram and identifies a metal material, a metal material that identifies the metal object in the original image. A metal region generating unit for generating a metal projection region, which is a region corresponding to a metal material by forward projection, and detecting and detecting a peripheral region related to pixel values inside the metal projection region among the peripheral regions outside the metal projection region. A reconstructing unit which obtains pixel values inside the metal projection area from the surrounding peripheral area to generate a reconstructed sinogram, and reconstructs the reconstructed sinogram to generate a reconstructed image corresponding to the reconstructed sinogram, It includes an image generating unit for bonding the metal material.

Description

SINOGRAM INPAINTING APPARATUS AND METHOD}

The disclosed technique relates to a sinogram reconstruction apparatus.

Readability of CT images may be severely degraded by objects having a high density inside the subject. Objects with significantly higher densities than surrounding body tissues have relatively large attenuation coefficients, such as beam hardening, beam scatter, and starburst on the entire CT image as well as the surrounding area. Star-burst, streak artifacts can occur. In particular, in the case of the oral cavity, damaged teeth are treated with a high density of prosthesis such as metal, and thus, an effective metal artifact reduction (MAR) method is required to obtain a CT image with guaranteed readability.

Among embodiments, a sinogram reconstruction apparatus reconstructs a sinogram to generate an initial image corresponding to the sinogram, and to identify a metal object in the original image. A metal material identification unit, a metal area generation unit for forward projection of the metal material to generate a metal projection area corresponding to the metal material, and the metal projection area among the peripheral areas outside the metal projection area. A reconstruction unit for detecting a peripheral region related to an internal pixel value, obtaining a pixel value inside the metal projection region from the detected peripheral region, and generating a reconstructed sinogram, and reconstructing the reconstructed sinogram To generate a reconstructed image corresponding to the reconstructed sinogram, and to combine the identified metallic material with the reconstructed image. It includes the intelligent generation unit.

Among the embodiments, the sinogram reconstruction method reconstructs a sinogram to generate an initial image corresponding to the sinogram, and to identify a metal object in the initial image. Forward projecting the metal material to produce a metal projection area that is a region corresponding to the metal material, wherein the metal projection area is associated with a pixel value inside the metal projection area of a peripheral area outside the metal projection area. Detecting a peripheral region, obtaining a pixel value inside the metal projection region from the detected peripheral region, generating a restored sinogram, and reconstructing the restored sinogram to recover the restored sinogram Generating a reconstructed image corresponding to and combining the identified metallic material with the reconstructed image.

Description of the disclosed technology is only an embodiment for structural or functional description, the scope of the disclosed technology should not be construed as limited by the embodiments described in the text. That is, the embodiments may be variously modified and may have various forms, and thus the scope of the disclosed technology should be understood to include equivalents capable of realizing the technical idea.

On the other hand, the meaning of the terms described in the present application should be understood as follows.

Singular expressions should be understood to include plural expressions unless the context clearly indicates otherwise, and terms such as "include" or "have" refer to features, numbers, steps, operations, components, parts, or parts thereof described. It is to be understood that the combination is intended to be present, but not to exclude in advance the possibility of the presence or addition of one or more other features or numbers, steps, operations, components, parts or combinations thereof.

Each step may occur differently from the stated order unless the context clearly dictates the specific order. That is, each step may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. Terms defined in commonly used dictionaries should be interpreted to be consistent with meaning in the context of the relevant art and can not be construed as having ideal or overly formal meaning unless expressly defined in the present application.

1 is a block diagram illustrating a configuration of a sinogram restoration apparatus according to an embodiment of the disclosed technology, and FIG. 2 is a diagram illustrating an example in which the sinogram restoration apparatus of FIG. 1 restores a sinogram. .

Referring to FIG. 1, the sinogram reconstruction apparatus 100 may include a metal material identification unit 110, a metal region generation unit 120, a reconstruction unit 130, and an image generation unit 140.

The metal material identification unit 110 reconstructs the sinogram 210 to generate an initial image 212, and identifies a metal object 214 in the initial image 212. do. For example, the metal material identification unit 110 may generate the initial image 212 by filtering the backgram (FBP) through the sinogram 210. In addition, the metal material identification unit 110 may identify the metal material 214 in the initial image 212 according to a threshold. Here, the pixels having a high value may correspond to the metal material 214. In addition, the threshold value may be set in advance, or may be set to a predetermined upper percentage of the highest value among the pixels constituting the original image 212.

The metal region generator 120 forward-projects the metal material 214 to generate a region 220 corresponding to the metal material in the sinogram. Hereinafter, a region corresponding to the metal material will be referred to as a metal projection region 220.

The restoration unit 130 detects a peripheral area related to the pixel value inside the metal projection area 220 among the peripheral areas outside the metal projection area 220, and detects the peripheral area inside the metal projection area 220 from the detected peripheral area. The reconstructed sinogram 230 is generated by acquiring pixel values. Here, the reconstruction unit 130 may generate the reconstructed sinogram 230 by detecting the peripheral area of all the pixels in the metal projection area 220.

The image generating unit 140 generates the reconstructed image 240 by filtering the restored sinogram 230 and post-projection projection (FBP), and combines the metallic material 242 to the reconstructed image 240, thereby improving MAR. Generated image 244.

3 is a block diagram illustrating an example of a restoration unit of FIG. 1. Referring to FIG. 3, the reconstructor 130 includes an associated peripheral region detector 310 and an interpolator 320.

The associated peripheral area detector 310 detects a peripheral area related to pixel values inside the metal projection area among the peripheral areas outside the metal projection area. Hereinafter, a peripheral area related to pixel values inside the metal projection area among the peripheral areas outside the metal projection area will be referred to as a related peripheral area. For example, the associated peripheral region may be pixels including attenuation coefficient information about an object region through which light reaching the restoration target pixel passes. Here, the pixel to be restored may correspond to each pixel in the metal projection area. Here, the relevant peripheral area may be pixels surrounded by two sinusoids passing through the pixel to be restored.

The interpolation unit 320 generates a restored sinogram by interpolating the inside of the metal projection area from the relevant peripheral area. For example, the interpolator 320 may generate a restored sinogram by performing linear interpolation.

FIG. 4 is a diagram for describing an example of detecting, by an associated peripheral region detector of FIG. 3, a related peripheral region related to a restoration target pixel positioned in the center of a detector.

Referring to FIG. 4, the detector 410 obtains projection images from various angles, and generates a sinogram by arranging the obtained projection images two-dimensionally according to each angle. For example, the detector 410 may generate a sinogram by obtaining projection images while rotating in units of 10 degrees. Here, the reconstruction value of the reconstruction target pixel P of the sinogram is determined by an attenuation coefficient of the object region through which the ray reaching the pixel P passes. If the influence of incident light to be scattered by the medium is small, the value of pixel P is only affected by region A 420. Therefore, when the detector 410 rotates, the surrounding pixels including the attenuation coefficient information of the area A influence the determination of the reconstruction value of the pixel P. For example, as illustrated in FIG. 4B, when the detector 410 is rotated 10 degrees, the pixels that affect the reconstruction value of the reconstruction target pixel P are three pixels on the right side of the sinogram. In addition, as shown in FIG. 4C, when the detector 410 rotates 20 degrees, the pixels that affect the reconstruction value of the reconstruction target pixel P are five pixels next to the second from the right on the sinogram. Thus, when the detector 410 rotates by 0, 10, 20, 340, 350 degrees, the relevant peripheral region detector 310 detects the related peripheral region as shown in FIG. 4E on the sinogram. Can be detected.

FIG. 5 is a diagram for describing an example of detecting, by an associated peripheral region detector of FIG. 1, a related peripheral region related to a restoration target pixel positioned at an edge of a detector.

When the detector 510 is rotated according to 10, 20, 340, 350 degrees, as shown in (b) to (e) of FIG. 5, the associated peripheral region detector 310 is shown in (e) of FIG. The relevant peripheral area as shown can be detected. Here, the relevant peripheral region shown in (e) of FIG. 5 is smaller in size than the related peripheral region shown in (e) of FIG. When the object region is assumed to be a circle, the smaller the size of the region A is, the smaller the number of pixels corresponding to the region A is in the sinogram as the restoration target pixel is positioned at the edge.

6 and 7 are diagrams for describing an algorithm in which the related peripheral region detector of FIG. 3 detects an associated peripheral region.

In FIG. 6A, an area of the object through which light reaching the restoration target pixel passes is A, and two points where the edge 620 and the area A of the object meet are e1 and e2. Here, as the sensor 610 rotates, the trajectories of el and e2 on the sinogram correspond to the curve surrounding the relevant peripheral region. Referring to FIG. 6B, if one of two points where the edge 620 and the area A of the object meet is e, the coordinates (x, y) of e may be represented by θ as shown in Equation 1 below. have.

θ = arcsin (y / r)

Where r is the radius of the object and θ is the angle of point e.

The distance d of the pixel to be restored from the center on the detector 610 according to the rotation angle β of the detector 610 is expressed by Equation 2 below.

d = rsin (θ-β)

Therefore, the distance d corresponds to a sinusoid whose magnitude is r and the period is 2π, and the trajectories of two points e1 and e2 can be expressed as shown in FIG. 7.

Referring to FIG. 7, the related peripheral region detector 310 detects the region surrounded by two sinusoids of the trajectories e1 and e2 as the related peripheral regions 710 and 720. In FIG. 7A, the size of the associated peripheral region 710 is small because the pixel to be restored is located at the edge. In FIG. 7B, the size of the associated peripheral region 720 is reduced because the pixel to be restored is located in the center. Big.

8 is a block diagram illustrating another example of the restoration unit of FIG. 1. Referring to FIG. 8, the restoration unit 130 includes a related peripheral region detector 810 and a weight restoration unit 820.

The associated peripheral region detector 810 detects an associated peripheral region, which is a peripheral region related to pixel values inside the metal projection region, among the peripheral regions outside the metal projection region. Here, the relevant peripheral area detector 810 may detect related peripheral areas having different sizes according to positions on the sinogram of the pixel to be restored.

The weight reconstruction unit 820 applies Gaussian weights to the periphery of the relevant periphery, and obtains pixel values inside the metal projection area from the periphery to which the weight is applied to generate the reconstructed sinogram.

The weight reconstruction unit 820 may determine the reconstruction value of the reconstruction target pixel P by performing the following equation (3).

Where I (p) is the reconstruction value of the reconstruction target pixel, I (p, q) is the periphery pixel value of the reconstruction target pixel, R (p, q) is the value of the periphery of the associated periphery region (RNA), and W (p , q) is a Gaussian value, N is a binary mask to obtain only weights when there is a neighboring pixel value in the relevant surrounding area, 1 when the surrounding pixel value exists in the relevant surrounding area (RNA), and the surrounding pixel 0 if no value exists.

FIG. 9 is a diagram for describing an example of obtaining a binary mask 970 by obtaining a weight 950 of neighboring pixels for a related peripheral area of the weight recovery unit of FIG. 8.

Referring to FIG. 9, the weight restoring unit 820 obtains a relevant region Gaussian mask 930 by multiplying a value 910 of a related peripheral region (RNA) by a Gaussian weight (W) 920, and obtains a related region Gaussian mask ( 930 may be multiplied by the value 940 of the neighboring pixels to obtain a weighted value 950 of the neighboring pixels for the relevant peripheral area. In addition, the weight restorer 820 may multiply the relevant region Gaussian mask 930 by the binary mask 960 to obtain a weight value 970 when the value of the neighboring pixel exists.

The weight restoring unit 820 generates a Gaussian mask by applying Gaussian weights to the surrounding area, and multiplies the values of the pixels in the surrounding area by the Gaussian mask to calculate the values of the pixels in the surrounding area where the weight is reflected. As the pixels of the region are farther from the pixel to be restored, the sinogram can be restored by reflecting that the pixels in the region have a lower correlation with the pixel to be restored, thereby reducing the restoration error. Referring to FIGS. 4 and 5, since the light passes through a region unrelated to P as it is farther from the restoration target pixel P, the correlation with the restoration target pixel is lowered.

10 is a flowchart illustrating a sinogram restoration method according to an embodiment of the disclosed technology.

Referring to FIG. 10, in step 1010, the sinogram reconstruction apparatus reconstructs a sinogram to generate an initial image, and identifies a metal object in the initial image. For example, the sinogram reconstruction apparatus may generate an initial image by filtering the sinogram to be filtered backprojection (FBP). In addition, the sinogram reconstruction apparatus may use a threshold to identify metal material in the original image.

In step 1020, the sinogram reconstruction apparatus forward-projects the metal material to create a region corresponding to the metal material in the sinogram. Hereinafter, the region corresponding to the metal material will be referred to as a metal projection region.

In operation 1030, the sinogram reconstruction apparatus detects a peripheral area related to the pixel value inside the metal projection area among the peripheral areas outside the metal projection area, and obtains a pixel value inside the metal projection area from the detected peripheral area. Create a restored sinogram. Here, the sinogram reconstruction apparatus may generate a reconstructed sinogram by detecting a peripheral area for all pixels in the metal projection area, respectively.

In operation 1040, the sinogram reconstruction apparatus generates a reconstructed image by performing post-filtering reverse projection (FBP) on the reconstructed sinogram, and combines the metallic material with the reconstructed image.

The disclosed technique may have the following effects. It is to be understood, however, that the scope of the disclosed technology is not to be construed as limited thereby, as it is not meant to imply that a particular embodiment should include all of the following effects or only the following effects.

According to an embodiment, the sinogram reconstruction apparatus detects a peripheral area related to a pixel value inside a metal projection area among peripheral areas outside the metal projection area, and detects a pixel value inside the metal projection area from the detected peripheral area. By acquiring and restoring the sinogram, the restoration error can be reduced and the performance of metal artifacts (MA) can be improved.

According to an embodiment, the sinogram reconstruction apparatus restores the sinogram by acquiring pixel values inside the metal projection region from the weighted related peripheral region, thereby further reducing the restoration error, thereby greatly improving MAR performance. have.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the present invention as defined by the following claims It can be understood that

1 is a block diagram illustrating a configuration of a sinogram restoration apparatus according to an embodiment of the disclosed technology.

FIG. 2 is a diagram for describing an example in which the sinogram restoration apparatus of FIG. 1 restores a sinogram.

3 is a block diagram illustrating an example of a restoration unit of FIG. 1.

FIG. 4 is a diagram for describing an example of detecting, by an associated peripheral region detector of FIG. 3, a related peripheral region related to a restoration target pixel located in the center of a detector.

FIG. 5 is a diagram for describing an example of detecting, by an associated peripheral region detector of FIG. 1, a related peripheral region related to a restoration target pixel located at an edge of a detector.

6 and 7 are diagrams for describing an algorithm in which the related peripheral region detector of FIG. 3 detects an associated peripheral region.

8 is a block diagram illustrating another example of the restoration unit of FIG. 1.

FIG. 9 is a diagram for describing an example in which the weight recovery unit of FIG. 8 obtains a weight of peripheral pixels of a related peripheral area and obtains a binary mask.

10 is a flowchart illustrating a sinogram restoration method according to an embodiment of the disclosed technology.

Claims (14)

A metal material identification unit reconstructing a sinogram to generate an initial image corresponding to the sinogram, and identifying a metal object in the initial image; A metal region generating unit configured to forward-project the metal material to generate a metal projection area that is a region corresponding to the metal material; Detecting a peripheral region related to pixel values inside the metal projection region among the peripheral regions outside the metal projection region, and obtaining a pixel value inside the metal projection region from the detected peripheral region to obtain a restored sinogram. A restoration unit to generate; And Reconstructing the reconstructed sinogram to generate a reconstructed image corresponding to the reconstructed sinogram, and including an image generator for combining the identified metal material with the reconstructed image, The restoration unit An associated peripheral region detector for detecting a related peripheral region, which is a peripheral region associated with a restoration target pixel inside the metal projection region, through a forward projection method; And An interpolation unit for interpolating pixel values inside the metal projection area from the associated peripheral area, And the associated peripheral regions are peripheral pixels including attenuation coefficient information for an object region through which light reaching the restoration target pixel through the front projection passes. delete delete The method of claim 1 wherein the associated peripheral region is A sinogram reconstruction device, characterized in that the pixels belong to an area surrounded by two sinusoids. The method of claim 1, wherein the restoration unit And a weight restoring unit applying a Gaussian weight to the relevant peripheral area and obtaining a value of pixels in the metal projection area from the weighted related peripheral area. The method of claim 5, wherein the associated peripheral region detection unit And detecting related peripheral areas of different sizes according to positions on the sinogram of the pixel to be restored. A metal region generating unit generating a metal projection region, which is a region corresponding to a metal object on a sinogram; Detecting a surrounding peripheral area, which is a peripheral area related to pixel values inside the metal projection area, from among the surrounding areas outside the metal projection area, and obtaining and reconstructing a pixel value inside the metal projection area from the associated peripheral area A restoring unit generating a gram; And Reconstructing the restored sinogram to generate a reconstructed image, and including an image generator for coupling the metal material to the reconstructed image, The restoration unit An associated peripheral region detector for detecting a related peripheral region, which is a peripheral region associated with a restoration target pixel inside the metal projection region, through a forward projection method; And An interpolation unit for interpolating pixel values inside the metal projection area from the associated peripheral area, And the associated peripheral regions are peripheral pixels including attenuation coefficient information for an object region through which light reaching the restoration target pixel through the front projection passes. delete Reconstructing a sinogram to generate an initial image corresponding to the sinogram and identifying a metal object in the initial image; Forward projecting the metal material to produce a metal projection area that is an area corresponding to the metal material; Detecting a peripheral region related to pixel values inside the metal projection region among the peripheral regions outside the metal projection region, and obtaining a pixel value inside the metal projection region from the detected peripheral region to obtain a restored sinogram. Generating; And Reconstructing the reconstructed sinogram to generate a reconstructed image corresponding to the reconstructed sinogram, and combining the identified metal material with the reconstructed image, Generating the restored sinogram Detecting, by a forward projection method, an associated peripheral region, which is a peripheral region associated with a restoration target pixel inside the metal projection region, of the peripheral region; And Interpolating pixel values inside the metal projection area from the associated peripheral area, And the associated peripheral regions are peripheral pixels including attenuation coefficient information for an object region through which light reaching the restoration target pixel through the front projection passes. delete delete 10. The method of claim 9, wherein the associated peripheral region is A method for recovering a sinogram characterized in that the pixels belong to an area surrounded by two sinusoidal curves. The method of claim 9, wherein generating the restored sinogram And applying a Gaussian weight to the associated peripheral region, and obtaining values of pixels within the metal projection region from the weighted associated peripheral region. 14. The method of claim 13, wherein detecting the associated peripheral area comprises: And detecting the associated peripheral regions of different sizes according to positions on the sinogram of the pixel to be restored.

Images