JPH08336524A - X-ray image processing device - Google Patents

Abstract

(57) [Summary] [Purpose] A lesion is automatically identified from an X-ray CT image to facilitate a doctor's diagnosis. An image obtained by the X-ray CT apparatus 1 is stored in an original image storage memory 2, a blood vessel extracting unit 12 extracts blood vessels from the image, and a pixel number specifying unit 14 extracts blood vessels along the blood vessels. The number of pixels constituting a blood vessel in each micro region is obtained by dividing the micro region, and the lesion part specifying unit 15 examines the change in the number of pixels in each micro region with respect to the arrangement of the micro regions along the blood vessel. Identify the department. Since the normal blood vessel becomes thinner from the base end to the end, the change in the number of pixels in each minute region arranged in that direction gradually decreases, but the tumor part that forms a blood vessel swells in a spherical shape, The change in the number of pixels over the tumor area does not gradually decrease but increases. The lesion is identified by paying attention to the change in the number of pixels. The identified lesion area is displayed on the monitor 6.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an X-ray image processing apparatus for processing an image obtained by an X-ray CT apparatus, and more particularly to an X-ray image processing apparatus.
The present invention relates to an X-ray image processing apparatus that identifies a lesion area from a CT image of a ray and provides the diagnosis assistance.

[0002]

2. Description of the Related Art Recently, in medical institutions, for early detection of lung cancer and the like, CT images of a large number of patients are taken by an X-ray CT apparatus, and the obtained CT images are sequentially diagnosed by a doctor. Introducing mass screenings to find patients with lung cancer. However, this group examination requires a large amount of labor because a doctor diagnoses a large number of images (CT images), and a system that supports the doctor's diagnosis is desired.

A conventional diagnostic support system of this kind is currently in the research stage and no established method has been established yet. As a system in some trial stages, for example, an X obtained by helical scanning of the chest is obtained. There is a support device that extracts blood vessels from a line CT image and three-dimensionally displays the extracted blood vessels to make it easier to find lung cancer.

[0004]

However, the prior art having such a structure has the following problems. In the support device according to the above-mentioned conventional example, a blood vessel is three-dimensionally displayed to facilitate the discovery of a lesion, but in the end, the doctor needs to read images from multiple directions, which does not reduce the labor of the doctor. .

The present invention has been made in view of the above circumstances, and an X-ray image processing apparatus capable of automatically identifying a lesion area from an X-ray CT image and facilitating a doctor's diagnosis. The purpose is to provide.

[0006]

The present invention has the following configuration to achieve the above object. That is, the present invention is an X-ray image processing apparatus that performs image processing for identifying a lesion portion from an image obtained by an X-ray CT apparatus, and comprises (a) an image obtained by the X-ray CT apparatus. Blood vessel extracting means for extracting blood vessels, and (b) the number of pixels for dividing the blood vessels into minute areas along the blood vessels extracted by the blood vessel extracting means and determining the number of pixels forming the blood vessels in each minute area. Specifying means, and (c) checking the change in the number of pixels in each of the micro-regions with respect to the arrangement of the micro-regions along the blood vessel, and determining the micro-region in a portion where the change is different from a predetermined condition. And a lesion part specifying means for specifying
(D) A lesion area output unit that outputs the lesion area identified by the lesion area identification unit.

[0007]

The operation of the present invention is as follows. The blood vessel extraction means extracts blood vessels from the image obtained by the X-ray CT apparatus. Next, the pixel number specifying means divides the blood vessel into minute areas along the blood vessel, and obtains the number of pixels forming the blood vessel in each minute area. Then, the lesion area identification means examines the change in the number of pixels in each micro area for the arrangement of each micro area along the blood vessel, and identifies the micro area of the area where the change is different from the predetermined condition as the lesion area. To do. A normal blood vessel becomes thinner from its proximal end to its distal end. Therefore, for example, when the change in the number of pixels in each minute region is examined from the base end to the end of the blood vessel, the number of pixels gradually decreases in a normal blood vessel. On the other hand, the tumor part formed in the blood vessel is swelling in a spherical shape, and when the change in the pixel number was examined by the above procedure, the change in the pixel number from the normal blood vessel to the tumor part did not gradually decrease but increased. To do. Focusing on this change in the number of pixels, the tumorous part, that is, the lesioned part is specified. The lesion area output means outputs the identified lesion area to notify the doctor.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. 1 is a block diagram showing a schematic configuration of an X-ray image processing apparatus according to an embodiment of the present invention. In this example, lung cancer is specified based on the X-ray CT image of the lung field.

The X-ray CT apparatus 1 photographs the chest of a patient (not shown) by a helical scan method, reconstructs a CT image of the lung field, and stores it in the original image storage memory 2. An X-ray CT image of the reconstructed lung field is shown in FIG. Each pixel forming this image has a CT value (-1000) corresponding to the internal tissue in the body.
~ +1000). The position of each pixel is specified by the XY coordinate values.

In the mass examination, a large number of patients are sequentially photographed, and the obtained images are stored in the original image storage memory 2
Are all stored in the order of shooting. The image stored in the original image storage memory 2 is controlled by the original image display controller 3 and displayed on the monitor 4. The original image display control unit 3 reads out the instructed image from the images stored in the original image storage memory 2 according to an instruction given by the doctor from the operation panel 5, and reads out the monitor 4
To be displayed. The original image storage memory 2 stores information about the patient (for example, the patient's ID number) together with the image, and the monitor 4 displays the information together with the image. It is configured to inform the doctor if it belongs to the patient.

On the basis of the image stored in the original image storage memory 2, the X-ray image processing apparatus 10 which is the main part of the present invention.
The lung cancer is identified by, and when the lung cancer is found, the image is displayed on the monitor 6.

The X-ray image processing apparatus 10 includes a blood vessel base end specifying section 11, a blood vessel extracting section 12, a blood vessel memory 13, a pixel number specifying section 14, a lesion section specifying section 15, a lesion section display data editing section 16, and a lesion. Display memory 17, lesion display control unit 1
It is composed of 8. The detailed configuration of each unit will be described below with reference to the flowcharts shown in FIGS. 3 and 4 according to the operation of the apparatus of this embodiment.

First, the outline of the operation of this apparatus will be described with reference to the flowchart shown in FIG. Prior to starting the following operations, the original image storage memory 2 has already stored the CT images of the lung fields of all patients who have undergone mass examination, and in the following operations, the operation panel by the doctor is used. According to the instruction from 5, the images stored in the original image storage memory 2 are sequentially displayed on the monitor 4, and it is checked whether or not the image displayed on the monitor 4 has a lesion, and the lesion is found. Is displayed on the monitor 6.

That is, first, the doctor waits for an instruction from the operation panel 5, and when the instruction is given, the process of step S2 is executed (step S1).

In step S2, the lesion display control unit 18
Erases the screen of monitor 6. Then, the original image display control unit 3 sets the image next to the image currently displayed on the monitor 4 (however, in the first process, the image stored at the beginning of the original image storage memory 2) to the original image storage memory. The data is read from No. 2 and displayed on the monitor 4 (step S3).

Next, a process for identifying a lesion (lung cancer) in the image displayed on the monitor 4 is executed (step S4). This lesion-specific process will be described later. If there is no lesion in the image, the process branches to step S6, and if there is a lesion, the process branches to step S7 (step S5).

In step S6, for example, the buzzer 7 is sounded once in order to notify the doctor that there is no lesion area when the lesion area identification process is completed. And
The process proceeds to step S10.

Further, in step S7, for example, the buzzer 7 is sounded twice in order to notify the doctor of the presence of the lesion when the lesion identification process is completed. Then, the lesion display data editing unit 16 edits the display data for displaying the discovered lesion on the monitor 6 and stores the edited display data in the lesion display memory 17 (step S8). Based on the display data stored in the part display memory 17, the lesion part is displayed on the monitor 6 (step S9), and the process proceeds to step S10.

An image showing a lesion on the blood vessel image is displayed on the monitor 6. The lesion display data editing unit 16 includes a blood vessel image stored in a blood vessel storage memory 13, which will be described later, position information of the lesion given from the lesion specifying unit 15, and patient information attached to the image currently being processed. The display data is edited based on (reading from the original image storage memory 2). The blood vessel image stored in the blood vessel storage memory 13 is a binary image (created by the blood vessel extraction unit 12 to be described later) in which “1” is added to the pixels forming the blood vessel and “0” is added to the pixels forming the background. is there. In the lesion area display data editing unit 16, the pixel of the lesion area is specified in the binary image based on the position information of the lesion area, and the pixel of the lesion area in the binary image is assigned, for example, “2”, The patient information is added and stored in the lesion display memory 17.

The lesion display control unit 18 receives the instruction from the lesion display data editing unit 16 (the instruction to notify that the edited display data is stored in the lesion display memory 17). Based on the ternary image (background “0”, blood vessel “1”, lesion “2”) stored in, for example, the background is white, the blood vessel is black, the lesion is blinking, the background, blood vessel and lesion are If the display density is different, or if the monitor 6 is a color monitor, different colors are displayed for the background, the blood vessel, and the lesion, and the lesion is displayed.
Further, the patient information is displayed on the monitor 6 to inform the doctor of the patient information.

The lesion display data editing unit 16, lesion display memory 17, lesion display control unit 18, monitor 6
Corresponds to the lesion output means in this invention.

In step S10, the original image storage memory 2
For all the images stored in
It is determined whether or not the process of S9 is completed, and if it is completed, the operation is completed. On the other hand, if there is an unprocessed image, the process returns to step S1 and waits for an instruction from the doctor. For example, if there is no lesion in the previous image, the doctor knows that the processing of the previous image has ended due to the buzzer 7 ringing once, and instructs the operation panel 5 to perform the processing of the next image. If there is a lesion in the previous image, the doctor knows that the buzzer 7 has sounded twice and the processing of the previous image is completed, and the monitor 6
The display data of the lesion area displayed on the screen and the original image displayed on the monitor 4 are observed, and when the display is completed, an instruction is given from the operation panel 5. When instructed from the operation panel 5, the same processing as above is performed on the next image.

Next, details of the lesion area specifying process in step S4 will be described with reference to the flowchart shown in FIG. First, when there is an instruction from the original image display control unit 3 (an instruction to notify that the original image has been displayed on the monitor 4), the blood vessel proximal end identifying unit 11 determines whether the image stored in the original image storage memory 2 is displayed. Of these, the same image as the image displayed on the monitor 4 (this image is referred to as a processed image) is read out, and the base end of the blood vessel is specified based on the processed image (step S1).
1). Specifically, muscles and bones are removed from the processed image to extract the lung region, the center of gravity in the lung region is obtained, and the center of the body is specified. The center of this body approximately coincides with the proximal end of the blood vessel.
In order to extract the lung region, for example, the CT value of each pixel of the processed image is divided into two by a predetermined threshold value (threshold value for removing muscles and bones) ((“0” or more? A binary image in which pixels less than “0” are marked with “1” and pixels less than “0” are marked with “0” is created as shown in FIG. Then, the boundary between “0” and “1” is examined, the contour RL of the lung region HA is extracted, and the center of gravity within this contour RL (corresponding to the lung region) is specified. The position information (XY coordinate values) of the base end of is given to the pixel number specifying unit 14.

On the other hand, when the original image display control unit 3 gives an instruction, the blood vessel extraction unit 12 reads the processed image from the original image storage memory 2, extracts the blood vessel in the processed image, and outputs "1" for the blood vessel. A binary image with "0" added to the background is stored in the blood vessel storage memory 13 (step S12). Specifically,
"0" when the CT value of each pixel of the processed image is "-900" or more
Pixels less than 1 are blood vessels, and "1" is added to those pixels,
"0" is added to the other pixels having CT values to create a binary image in which blood vessels are extracted. The created binary image of the blood vessel is
The data is stored in the blood vessel storage memory 13, and the completion of processing is notified to the pixel number specifying unit 14. The blood vessel extraction unit 12
Corresponds to the blood vessel extraction means in the present invention.

Next, the pixel number specifying unit 14 divides the blood vessel image into minute areas along the blood vessel image stored in the blood vessel storage memory 13 (from the proximal end to the end of the blood vessel), and each minute area is divided. Then, the number of pixels forming the blood vessel is calculated (step S
13). In the process of segmenting the blood vessel image into minute regions, as shown in FIG. 6, the blood vessel image KZ is thinned, and along the thin line SL, minute regions ... S ₁ , S ₂ ,.
S _K , ... Is classified. Then, each of these minute regions ... S
_The number of pixels (corresponding to the area of each minute region) of the pixels (pixels having “1”) forming the blood vessel in ₁ , S ₂ , ..., _SK , ... Is calculated. Each minute area ..., S ₁ , S ₂ , ..., S
_K, the number of pixels in _{... ..., G 1, G 2} , ..., G K, ... to. This result is given to the lesion area specifying unit 15. The pixel number specifying unit 14 corresponds to the pixel number specifying means in this invention.

Then, the lesion area specifying unit 15 determines the position of the lesion area (X by changing the number of pixels in each given micro area).
The Y coordinate value) is specified (step S14). In a normal blood vessel, the blood vessel becomes thinner from the proximal end to the distal end. Therefore, for example, from the base end to the end of the blood vessel image, the number of pixels of each of the minute regions ..., S ₁ , S ₂ , ..., S _K , ..., G ₁ , G ₂ , ..., G _K ,. If you examine the changes,
If the blood vessel is normal, the number of pixels gradually decreases. That is, the condition of ...> G ₁ > G ₂ >...> G _K >. On the other hand, the tumor part formed in the blood vessel is, as shown in FIG.
It is swelled in a spherical shape, and when the change in the number of pixels is examined by the above procedure, the change in the number of pixels from the normal blood vessel KZ to the tumor site CS does not gradually decrease but increases. That is,…> G
_S1 > G _S2 <G _C1 <G _C2 <... In addition, ..., G _S1 , G
_{S2, G C1, G C2,} ... is small region of FIG. 7 ..., S _S1,
It is the number of pixels in S _S2 , S _C1 , S _C2 , .... Focusing on this change in the number of pixels, the tumor site, that is, the lesion site, is specified,
The XY coordinate values of the minute area of the lesion area are given to the lesion area display data editing section 16 as the location of the lesion area. The lesion part specifying unit 15 corresponds to the lesion part specifying means in the present invention.

As described above, in this embodiment, the X-ray C
Since the lesion (lung cancer) is automatically detected and notified to the doctor based on the image obtained by the T device 1, the burden on the doctor can be significantly reduced.

By the way, the blood vessels are dendritic,
The above device erroneously detects this bifurcation as a lesion.
Therefore, it is preferable to detect this bifurcation and reflect the detected bifurcation in the lesion area identification processing. FIG. 8 shows an apparatus that detects such a blood vessel bifurcation and reflects the detected bifurcation in the lesion area identification processing, and the configuration thereof will be described below. Note that the parts denoted by the same reference numerals as those in FIG. 1 have the same configuration as the above-described embodiment apparatus, and therefore detailed description thereof is omitted here.

This modification is characterized in that a branching portion detecting section 21 is provided and the branching portion detected by the branching portion detecting portion 21 is reflected in the identification of the lesion portion by the lesion portion specifying portion 22.

As shown in FIG. 9, the branching section identifying section 21 thins the blood vessel image stored in the blood vessel storage memory 13 to obtain the position of the intersection point CP of the branching section. Then, the branch area BA is set according to the thickness of the blood vessels in the vicinity thereof.
Further, the branch area BA is a branch point (BP in FIG. 9).
1 and 2 points of BP2) are included. In addition,
The branching portion detection processing by the branching portion detection unit 21 is performed between steps S13 and S14 in the flowchart of FIG.

In the lesion part specifying part 22, the branch part detecting part 21
The minute area included in the branch area BA detected in step 1 or the minute area including one part in the branch area BA is excluded from the above-described processing for checking the change in the number of pixels due to the comparison of the number of pixels. Prevents false detection of a part as a lesion. For example, in the case shown in FIG. 10, a minute area S _B1 that contains 1 part bifurcation area BA, microscopic regions S _B2 included in the branching region BA, ..., minute regions other than S _BK ..., S _T1 , S _T2 , ..., S _T3 , ..., S _T4 , S _T5 ,.
The change in the number of pixels of S _T6 , ... _Is examined. Further, with respect to the branched blood vessel BK, the same branching portion specifying process as that described above with the branching portion area BA as a base end is separately performed. The process performed by the lesion area specifying unit 22 is performed in step S14 in the flowchart of FIG.

With this configuration, it is possible to prevent the branch portion from being erroneously detected as a lesion portion.

In the above-described embodiment and its modification, the change in the number of pixels in the micro area from the base end to the end of the blood vessel was examined. Even if the change in the number of pixels is examined, the change in the number of pixels in the lesion is different from the condition of the change in the number of pixels in the normal blood vessel (in this case, increase in removal). Part can be specified.

Further, in the above-mentioned embodiment and its modification, the case where the lesion part is specified from the two-dimensional image has been described. For example, the three-dimensional image is reconstructed based on the data obtained by the helical scan, Based on the three-dimensional image, the blood vessels are three-dimensionally divided into minute areas (minute spaces),
The change in the number of pixels in the micro space (corresponding to the volume in the micro space) is examined, and the property that the number of pixels in the micro space gradually decreases as the blood vessel goes from the proximal end to the distal end (from the distal end to the proximal end The device may be configured to identify the lesion part by utilizing the property that the number of pixels in the minute space gradually increases toward the position.

[0035]

As is apparent from the above description, according to the present invention, a blood vessel is extracted from an X-ray CT image, the blood vessel is divided into minute regions, and a change in the number of pixels in each minute region is detected. Since each unit for investigating along the blood vessel to identify the lesion portion is provided, the lesion portion can be automatically identified from the X-ray CT image. Therefore, in mass examination and the like, the burden on the doctor can be reduced when sequentially identifying lesions in a large number of images.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an X-ray image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing an X-ray CT image of a reconstructed lung field.

FIG. 3 is a flowchart showing the operation of the apparatus of the embodiment.

FIG. 4 is a flowchart showing a lesion part specifying process of the embodiment apparatus.

FIG. 5 is a diagram for explaining a process of extracting a lung region from an X-ray CT image.

FIG. 6 is a diagram for explaining a process of segmenting a blood vessel image into minute regions.

FIG. 7 is a diagram for explaining the principle of identifying a lesion.

FIG. 8 is a block diagram showing a schematic configuration of a modified example of the embodiment apparatus.

FIG. 9 is a diagram for explaining a process of detecting a branch portion.

FIG. 10 is a diagram for explaining processing of a lesion part specifying unit according to a modified example.

[Explanation of symbols]

 1 ... X-ray CT device 4, 6 ... Monitor 10 ... X-ray image processing device 11 ... Blood vessel base end identification part 12 ... Blood vessel extraction part 13 ... Blood vessel memory 14 ... Pixel number identification part 15 ... Lesion part identification part 16 ... Lesion Part display data editing unit 17 ... Lesion display memory 18 ... Lesion display control unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yukio Sato 1 Nishinokyo Kuwabara-cho, Nakagyo-ku, Kyoto City Stock company Shimadzu Sanjo factory (72) Inventor Harufumi Kato 1-30-15-305 Kita-Shinjuku, Shinjuku-ku, Tokyo (72) Inventor Makoto Saito 1-33-18 Kichijoji Honcho, Musashi City, Tokyo 504 Mansion 504

Claims (1)

[Claims] 1. An X-ray image processing apparatus for performing image processing for identifying a lesion from an image obtained by an X-ray CT apparatus, comprising: (a) a blood vessel from the image obtained by the X-ray CT apparatus. And (b) segmenting the blood vessel into minute regions along the blood vessel extracted by the blood vessel extracting means,
A pixel number specifying means for obtaining the number of pixels forming the blood vessel in each micro region; and (c) a change in the number of pixels in each micro region with respect to the arrangement of the micro regions along the blood vessel, Lesion part specifying means for specifying, as a lesion part, a minute region in which the change is different from a predetermined condition;
An X-ray image processing apparatus, comprising: a lesion area output unit that outputs the lesion area identified by the lesion area identifying unit.