KR102582658B1 - Operating method for scoliosis diagnostic device - Google Patents

Abstract

The operating method of the scoliosis diagnosis device of the present invention includes receiving a radiographic image of a subject from a radiography device, analyzing the radiographic image through an artificial intelligence algorithm to recognize vertebral body images. A step of displaying rectangular vertebral blocks generated corresponding to the sizes of each of the vertebral body images on the vertebral body images, and displaying two vertebral blocks located at both ends of the consecutive vertebral blocks. Displaying a reference line connecting center points on the radiographic image, rotating the radiographic image so that a preset vertical line in the radiographic image is parallel to the reference line, and determining a distance between the center point of each of the spinal blocks and the reference line. Measuring the separation distance, extracting blocks to be inspected among the spine blocks whose separation distance is greater than or equal to a reference distance, and selecting first and second reference blocks for measuring the Corpse angle among the blocks to be inspected. It includes a step of determining and calculating the Corpse angle by calculating an angle formed by a vertical line segment with an extension line of one side of the first and second reference blocks.

Description

{OPERATING METHOD FOR SCOLIOSIS DIAGNOSTIC DEVICE}

The present invention relates to a method of operating a scoliosis diagnosis device.

Scoliosis refers to an abnormal spine condition in which the spine is curved in an S or C shape, unlike the normal spine that is straight. Scoliosis is caused by various causes such as posture and lack of exercise, and if left untreated, it can cause disc disease, spinal stenosis, joint pain, arthritis, etc.

Scoliosis is diagnosed by measuring Cobb's Angle using an X-ray. Corpse angle refers to the angle at the point of intersection by finding the vertebra where tilt begins and the vertebra that is most tilted among the vertebrae, drawing a horizontal line for each vertebra, and drawing a line perpendicular to the horizontal line.

Corpse angle measurement is based on X-ray images, and an error of approximately 10° may occur depending on the person measuring the Corpse angle and the measurement method. Such errors may cause problems in making a diagnosis of scoliosis to patients with scoliosis.

The technical task to be achieved by the present invention is to provide a method of operating a scoliosis diagnosis device that analyzes radiological images through an artificial intelligence algorithm and automatically measures the Corpse angle required for scoliosis examination based on the analysis results.

A method of operating a scoliosis diagnosis device according to an embodiment of the present invention includes the steps of receiving a radiographic image of a subject from a radiography device, and analyzing the radiographic image through an artificial intelligence algorithm to determine the vertebral body. Recognizing images, displaying rectangular spine blocks generated corresponding to the size of each of the vertebral body images on the vertebral body images, and two steps located at both ends of the consecutive spine blocks. Displaying a reference line connecting the center points of the spinal blocks on the radiographic image, rotating the radiographic image so that the reference line is parallel to a vertical line preset in the radiographic image, and the central point of each of the spinal blocks Measuring the separation distance between the reference lines, extracting blocks to be inspected among the spine blocks whose separation distance is greater than or equal to the reference distance, and first and second blocks for measuring the Corpse angle among the blocks to be inspected. It includes determining two reference blocks and calculating the Corpse angle by calculating an angle formed by line segments perpendicular to an extension line of one side of the first and second reference blocks.

According to an embodiment, the step of determining the first and second reference blocks includes dividing the inspection target blocks located at the top and bottom into the first group and A step of dividing into a second group, and dividing each of the most distorted inspection target blocks among the inspection target blocks of the first group and the most distorted inspection target blocks among the inspection target blocks of the second group into the first reference block. and selecting the second reference block.

According to an embodiment, the step of selecting the first and second reference blocks includes measuring the angle formed by one side of the blocks to be inspected included in the first and second groups with the horizontal line, and setting the inspection target block with the largest angle among the inspection target blocks included as the first reference block; and setting the inspection target block with the largest angle among the inspection target blocks included in the second group as the first reference block. 2 It may include the step of setting it as a reference block.

Depending on the embodiment, at least one inspection target block may be arranged between the first and second reference blocks.

Depending on the embodiment, the step of determining the first and second reference blocks may determine the first and second blocks only when at least three consecutive blocks to be inspected exist.

Depending on the embodiment, the method of operating the scoliosis diagnosis device may further include diagnosing a subject in the radiographic image as scoliosis when the calculated Corpse angle exceeds 10°.

Depending on the embodiment, each of the vertebral body images may be placed inside each corresponding vertebral block.

According to the operating method of the scoliosis diagnosis device according to an embodiment of the present invention, radiological images can be analyzed through an artificial intelligence algorithm, and the Cobb's angle required for scoliosis examination can be precisely measured based on the analysis results.

1 is a diagram of a scoliosis diagnosis system according to an embodiment of the present invention.
Figure 2 is a flowchart for explaining the operation method of the scoliosis diagnosis device according to an embodiment of the present invention.
3 to 6 are diagrams for explaining a method of measuring the Corpse angle by a scoliosis diagnosis device according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the embodiments described below are only intended to provide a detailed explanation so that a person skilled in the art can easily implement the invention, and this does not limit the scope of protection of the present invention. doesn't mean In describing various embodiments of the present invention, the same reference numerals will be used for components having the same technical features.

As used in the specification, the term “unit” or “module” refers to a hardware component such as software, FPGA, or ASIC, and the “unit” or “module” performs certain roles. However, “part” or “module” is not limited to software or hardware. A “unit” or “module” may be configured to reside on an addressable storage medium and may be configured to run on one or more processors. Thus, as an example, a “part” or “module” refers to components such as software components, subject-oriented software components, class components and task components, processes, functions, properties, Includes procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functionality provided within components and “parts” or “modules” can be combined into smaller components and “parts” or “modules” or into additional components and “parts” or “modules”. Could be further separated.

1 is a diagram of a scoliosis diagnosis system according to an embodiment of the present invention.

Referring to FIG. 1, the scoliosis diagnosis system 10 according to an embodiment of the present invention includes a radiography device 100 and a scoliosis diagnosis device 200.

The radiography device 100 is a device that generates a radiation image by detecting radiation that has passed through the subject (HU). It radiates radiation to the radiation detection area and creates a radiation image (IM) based on the radiation that has passed through the subject (HU). can be created. The radiography apparatus 100 may transmit a radiographic image (IM) generated through wired or wireless communication to the scoliosis diagnosis apparatus 200. Radiographic image (IM) refers to an image of the front of the spine for diagnosing scoliosis.

The scoliosis diagnosis device 200 can receive a radiological image (IM), measure Cobb's Angle, which determines the degree of curvature of the spine, and diagnose scoliosis based on the measured Cobb's angle.

The scoliosis diagnosis device 200 can extract only the spinal region included in the radiological image (IM) using an artificial intelligence model. The scoliosis diagnosis device 200 may use a multi-layered perceptron (MLP) as an artificial intelligence algorithm to acquire an image of the spinal region, that is, the vertebral body, but is a type of convolutional neural network (CNN). Artificial intelligence models of various architectures, such as ResNet, U-net, and RCNN, may be used.

The scoliosis diagnosis device 200 can perform the learning process of the artificial neural network in advance and extract only the vertebral body image from the radiological image (IM) using the pre-trained artificial neural network.

Specifically, the scoliosis diagnosis device 200 adjusts the image size and removes unnecessary areas through a preprocessing process for the radiological image (IM) to find the desired region of interest and calculates the number of bits set in the artificial intelligence model. Conversion work can be done with . At this time, if there is noise in the radiation image (IM), a process of removing it may also be accompanied.

In addition, the scoliosis diagnosis device 200 can detect the spine area using a pre-trained artificial intelligence model. The image converted through the pre-processing process is input to the pre-trained artificial intelligence model, and the artificial intelligence model produces the learning result. An output image according to can be obtained. If the output image is normally output during this process, the output image may be displayed with low brightness where the prediction result is low and with high brightness where the prediction result is high, depending on the intensity of the spinal region.

The scoliosis diagnosis device 200 may perform threshold value processing to clearly separate images of each spinal region, that is, the vertebral body. Next, blob detection processing can be performed to obtain the coordinates of a clear spine region. The scoliosis diagnosis device 200 can acquire the size and position of the vertebral body image in the coordinate plane after block detachment processing.

After the scoliosis diagnosis device 200 recognizes the vertebral body image, it can display rectangular spine blocks on each vertebral body image, and each vertebral block can be set so that the vertebral body image is placed inside. there is. In other words, the spine block can be set in a form that does not protrude the vertebral body image to the outside.

The scoliosis diagnosis device 200 can determine first and second reference blocks for measuring the Corpse angle by analyzing the arrangement of the spinal blocks, and the line segments perpendicular to the extension of one side of the first and second reference blocks form You can calculate the Corpse angle by calculating the angle.

The scoliosis diagnosis device 200 can diagnose scoliosis for the subject (HU) in the radiological image when the calculated Corpse angle exceeds 10°.

Depending on the embodiment, the scoliosis diagnosis device 200 may be a personal computer (PC), a smart phone, a tablet PC, a mobile internet device (MID), an internet tablet, or an internet tablet (IoT). of things) device, IoE (internet of everything) device, desktop computer, laptop computer, workstation computer, Wibro (Wireless Broadband Internet) terminal, and PDA (Personal Digital Assistant). You can.

Figure 2 is a flowchart for explaining the operation method of the scoliosis diagnosis device according to an embodiment of the present invention.

Referring to FIG. 2 , the scoliosis diagnosis apparatus 200 may receive a rectangular radiographic image (IM) of a subject (HU) from the radiography apparatus 100 (S100).

The scoliosis diagnosis device 200 can recognize vertebral body images by analyzing radiographic images (IM) using an artificial intelligence algorithm (S110).

The scoliosis diagnosis apparatus 200 may display rectangular vertebral blocks generated to correspond to the size of each vertebral body image on the vertebral body images (S120).

The scoliosis diagnosis device 200 may display a reference line connecting the center points of two spinal blocks located at both ends of the continuous spinal blocks on the radiological image (S130). Here, the reference line is used to convert the position of the radiological image (IM) so that the center of the vertebrae is located at the center of the radiological image, and the scoliosis diagnosis device 200 determines the spine located first from the top in the radiological image (IM). A reference line can be created by connecting the center point of the block and the last spinal block, and displayed on the radiological image (IM).

The scoliosis diagnosis device 200 may rotate the radiographic image (IM) so that the preset vertical line and reference line in the radiological image (IM) are parallel to each other (S140).

The scoliosis diagnosis device 200 can calculate the distance between the reference line and the center of each spinal block, and extract spinal blocks with a separation distance greater than the reference distance among the spinal blocks as blocks to be examined (S160). . In addition, the scoliosis diagnosis device 200 sets the test target block whose center point is the most distant from the baseline as the peak block, sets the test target block located at the top based on the peak block as the first group, and sets the test target block located at the bottom as the first group. The target block can be set to the second group.

The scoliosis diagnosis apparatus 200 may determine first and second reference blocks for measuring the Corpse angle among the blocks to be examined (S170), where the first reference block is the highest among the blocks to be examined in the first group. It refers to the most distorted spinal block, and the second reference block refers to the most distorted spinal block among the blocks to be inspected in the second group.

The scoliosis diagnosis apparatus 200 may determine that the first and second reference blocks are normally determined only when at least one test target block exists between the first and second reference blocks.

Additionally, the scoliosis diagnosis apparatus 200 can determine the first and second reference blocks only when there are at least three consecutive blocks to be tested.

Depending on the embodiment, the scoliosis diagnosis device 200 may measure the angle formed by one side of the blocks to be tested with the horizontal line in order to determine the degree of distortion of the blocks to be tested. If the vertebrae are straight and not distorted, one side of the block to be tested will be parallel to the horizontal line, but if the vertebrae are twisted, one side of the block to be tested will be at a certain angle with the horizontal line.

The scoliosis diagnosis device 200 sets the test object block with the largest angle among the test object blocks of the first group as the first reference block, and sets the test object block with the largest angle among the test object blocks of the second group as the first reference block. A block can be set as a second reference block.

The scoliosis diagnosis device 200 may set an extension line on one side of the first and second reference blocks and calculate the Cobb's angle by calculating the angle formed by line segments perpendicular to the extension line (S180). Here, one side of the first and second reference blocks means a side corresponding to the cross section of the vertebra in contact with the intervertebral disc.

If the calculated Corpse angle exceeds 10°, the scoliosis diagnosis device 200 can diagnose scoliosis for the subject (HU) in the radiological image.

3 to 6 are diagrams for explaining a method of measuring the Corpse angle by a scoliosis diagnosis device according to an embodiment of the present invention.

Referring to FIG. 3, the scoliosis diagnosis device 200 can extract only the spinal region included in the radiological image (IM) using an artificial intelligence model. That is, the scoliosis diagnosis device 200 can generate a radiology image (IM') by extracting only the vertebral body image (BI) from the radiology image (IM) using a pre-trained artificial neural network.

After recognizing the vertebral body image (BI) included in the radiological image (IM'), the scoliosis diagnosis device 200 can display rectangular vertebral blocks (BL) on each vertebral body image (BI). Each vertebral block (BL) may be set so that a vertebral body image (BI) is placed therein. That is, the spine block BL may be set in a form that does not protrude the vertebral body image BI to the outside.

Referring to FIG. 4, the scoliosis diagnosis device 200 uses a radiographic image (SL) connecting the center points (PO1, PO2) of two consecutive spinal blocks (BL1 to BL19) located at both ends. IM').

The reference line (SL) is used to convert the position of the radiographic image (IM') so that the center of the vertebrae is located at the center of the radiographic image (IM'), and the scoliosis diagnosis device 200 uses the radiographic image (IM') ') A reference line (SL) is created by connecting the center point (PO1) of the first spinal block (BL1) and the center point (PO2) of the last spinal block (BL19) from the top of the bone, and this is used on the radiographic image (IM'). It can be displayed in .

The scoliosis diagnosis device 200 may rotate the radiographic image IM' so that the preset vertical line VL and the reference line in the radiological image IM' are parallel or coincident.

Referring to FIG. 5, the scoliosis diagnosis device 200 calculates the distance between the reference line (SL) and the center point of each spinal block and examines the spinal blocks BL10 to BL17 whose separation distance is greater than or equal to the reference distance. It can be extracted as target blocks (SBL).

In addition, the scoliosis diagnosis device 200 sets the test block (SBL) whose center point is the most distant from the baseline (SL) as the peak block (PB), and sets the test block located at the top based on the peak block (PB). (SBL) can be set to the first group (GR1), and the inspection target block (SBL) located at the bottom can be set to the second group (GR2).

Referring to FIG. 6 , the scoliosis diagnosis apparatus 200 may determine the first and second reference blocks SB1 and SB2 among the blocks to be examined (SBL). The first reference block (SB1) refers to the most distorted spinal block among the inspection target blocks (SBL) of the first group (GR1), and the second reference block (SB2) refers to the inspection target block of the second group (GR2). It refers to the most twisted spinal block among (SBL).

The scoliosis diagnosis device 200 can measure the angle θ formed by one side BL of the blocks SBL to be tested with the horizontal line HL. If the vertebrae are straight and not distorted, one side (BL) of the block to be tested will be parallel to the horizontal line (HL). However, if the vertebrae are twisted, one side (BL) of the block to be tested will be parallel to the horizontal line (HL). It will form an angle.

The scoliosis diagnosis device 200 sets the test block (SBL) with the largest angle in the first group (GR1) as the first reference block (SB1), and sets the largest angle in the second group (GR2) as the first reference block (SB1). The inspection target block (SBL) may be set as the second reference block (SB2).

The scoliosis diagnosis device 200 sets extension lines (SBL1, SBL2) on one side of the first and second reference blocks (SB1, SB2), and line segments (CL1) perpendicular to each of the extension lines (SBL1, SBL2). , CL2) can be calculated to calculate the Corpse angle (θ _C ).

Additionally, the scoliosis diagnosis device 200 can diagnose scoliosis for the subject (HU) in the radiological image when the Corpse angle exceeds 10°.

The steps of the method or algorithm described in connection with the embodiment of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or a combination thereof. The software module may be RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), Flash Memory, hard disk, removable disk, CD-ROM, or It may reside on any type of computer-readable recording medium well known in the art to which the present invention pertains.

Although the embodiments of the present invention have been described above, it is understood that those skilled in the art can make various modifications without departing from the scope of the claims of the present invention.

10: Scoliosis diagnosis system
100: Radiography device
200: Scoliosis diagnostic device

Claims (7)

Receiving a radiation image of a subject from a radiography device;
Recognizing vertebral body images by analyzing the radiographic images using an artificial intelligence algorithm;
displaying rectangular vertebral blocks generated to correspond to the sizes of each of the vertebral body images on the vertebral body images;
Displaying a reference line connecting the center points of two spinal blocks located at both ends of the consecutive spinal blocks on the radiographic image;
rotating the radiation image so that a preset vertical line in the radiation image is parallel to the reference line;
Measuring a separation distance between the center point of each of the spinal blocks and the reference line;
extracting blocks to be inspected from among the spinal blocks whose separation distance is greater than or equal to a reference distance;
determining first and second reference blocks for measuring Corpse angle among the blocks to be inspected; and
Comprising the step of calculating the Corpse angle by calculating an angle formed by line segments perpendicular to an extension line of one side of the first and second reference blocks,
The step of determining the first and second reference blocks is:
Classifying inspection target blocks located at the top and bottom into a first group and a second group based on the inspection target block having the largest separation distance among the extracted inspection target blocks; and
Selecting the most distorted inspection target block among the inspection target blocks of the first group and the most distorted inspection target block among the inspection target blocks of the second group as the first reference block and the second reference block, respectively. Includes steps,
The step of selecting the first and second reference blocks is,
measuring an angle formed by one side of blocks to be inspected included in the first and second groups with a horizontal line;
Setting an inspection target block with the largest angle among inspection target blocks included in the first group as the first reference block; and
Comprising the step of setting the inspection target block with the largest angle among the inspection target blocks included in the second group as the second reference block,
A method of operating a scoliosis diagnosis device in which at least one test target block is disposed between the first and second reference blocks. delete delete delete The method of claim 1, wherein determining the first and second reference blocks comprises:
A method of operating a scoliosis diagnosis device that determines the first and second blocks only when there are at least three consecutive blocks to be tested. According to paragraph 1,
A method of operating a scoliosis diagnosis device further comprising diagnosing a subject in the radiographic image as scoliosis when the calculated Corpse angle exceeds 10°. According to paragraph 1,
A method of operating a scoliosis diagnosis device in which each of the vertebral body images is disposed inside each corresponding vertebral block.

Images