KR102495365B1 - System for analysising body joint score of bone scan image using artificial intelligence and method thereof - Google Patents

Abstract

The present invention relates to a joint analysis system and method using artificial intelligence.
According to the present invention, in the joint analysis method using an artificial intelligence-based whole body joint analysis system, pre-stored whole body bone scan images and lesions according to joint values in the whole body bone scan images are determined through artificial intelligence. Creating a learning model by learning, scanning the whole body bones of a measurement subject to whom a radioactive material was administered, and obtaining a whole body bone scan image for the whole body bones, a region of interest (ROI, Region Of) in the whole body bone scan image Interest), and extracting each joint value corresponding to the region of interest, applying each extracted joint value to the learning model to determine whether the subject's joint has a lesion and the prognosis of the subject. The step of determining whether the joint has a lesion and the joint value of the joint is displayed and provided on a whole-body bone scan image of the subject to be measured.
As described above, according to the present invention, standardized score data for joints of the whole body can be accumulated, and when a lesion exists, data comparison and analysis can be performed through numerical values. In addition, through this, it can be used not only for clinical use, but also as data for artificial intelligence reading through machine learning.

Description

Joint analysis system and method using artificial intelligence {SYSTEM FOR ANALYSISING BODY JOINT SCORE OF BONE SCAN IMAGE USING ARTIFICIAL INTELLIGENCE AND METHOD THEREOF}

It relates to a joint analysis system and method using artificial intelligence, and relates to a joint analysis system and method using artificial intelligence for tracking a joint lesion or prognosis using a score of a user's joint.

The joint is an important part of the human body where bones are connected. It acts as an axis for various movements of the spine and limbs, enabling various movements. Movement becomes difficult, and in severe cases, it can lead to death.

In particular, according to data researched by the "Health Insurance Review and Assessment Service", the mortality rate within one year after hip fracture in Korea over the age of 50 is very high at 14.8% for women and 20.9% for men. Moreover, due to the aging society, the number of hip fracture patients is steadily increasing, and the corresponding social cost is also significantly increasing by more than 1 trillion won per year.

These technologies for early diagnosis of osteoporosis and risk assessment for osteoporotic fractures are being actively researched according to the recent development of medical imaging devices, image processing techniques, and biomedical engineering, and 3D quantitative computed tomography images and finite element analysis The bone strength evaluation method of a patient linked with finite element analysis (finite element analysis) has recently received a lot of attention in the academic and medical world because it can quantitatively estimate the structural strength of the target skeletal system based on the bone density distribution of an individual patient.

However, in order to evaluate bone strength based on finite element analysis (FEM), it is essential to secure an image segmentation technology capable of constructing a finite element model by segmenting only joint parts from complex medical image information composed of muscles and bones.

Automatic image segmentation of joints is still difficult due to the morphological characteristics of joints and technical limitations of medical imaging devices, and since joints are usually located deep in the body, low signal-to-noise ratio (SNR) In addition, thin cortical bone and articular cartilage of the ball head were vulnerable to the partial volume effect due to the low spatial resolution of in vivo medical imaging devices.

In order to solve these problems, radiopharmaceuticals containing radioactive isotopes such as PET (Positron Emission Tomograhpy, Positron Emission Tomography) and SPECT (Single Phothon Emission Computed Tomography) are administered to the human body, and then By counting the radiation emitted from a specific organ and imaging the counted radiation data, a diagnostic imaging method has been derived that allows medical professionals to check the distribution of radiation in the organ and evaluate biochemical changes or functional problems through imaging. .

However, in order to interpret such a nuclear image, there is a problem in that the interpretation direction may be changed according to medical experts.

Therefore, in order to solve this problem, a method of extracting values from the corresponding image, learning and providing the extracted values and lesions through artificial intelligence has become necessary.

The background technology of the invention is disclosed in Republic of Korea Patent Registration No. 10-1331504 (2013.11.26. notice).

A technical problem to be achieved by the present invention is to provide a joint analysis system and method using artificial intelligence for tracking a joint lesion or prognosis using a score of a user's joint.

According to an embodiment of the present invention for achieving this technical problem, in the joint analysis method using a whole body joint analysis system based on artificial intelligence, a pre-stored whole body bone scan image and joint values in the whole body bone scan image Creating a learning model by learning whether or not there is a lesion according to artificial intelligence, scanning the whole body bones of the measurement subject to whom radioactive material is administered, and acquiring a whole body bone scan image of the whole body bones, the whole body bone scan Setting a region of interest (ROI) in the image and extracting each joint value corresponding to the region of interest; Determining whether or not there is a lesion and a prognosis of the subject to be measured, and displaying and providing the whether or not the joint has a lesion and the joint value of the corresponding joint on a whole-body bone scan image of the subject to be measured.

The joint value has different values depending on the severity of the lesion of the measurement subject, and a joint value of a bone with the lesion may have a greater value than a joint value of a bone without the lesion.

The region of interest may be set to a part corresponding to a joint of the measurement subject in the scanned whole-body bone image.

The displaying and providing of the joint values may include overlaying a maximum value, a minimum value, and an average value of the joint values corresponding to the ROI on the whole body bone scan image.

The displaying and providing of the joint values may include calculating and providing a ratio between an average joint value in the region of interest and an average joint value of normal bones around the region of interest.

According to another embodiment of the present invention, in the joint analysis system based on artificial intelligence for whole body joint analysis, it is artificially determined whether or not there is a lesion according to a pre-stored whole body bone scan image and joint values in the whole body bone scan image. A learning unit that learns through intelligence to create a learning model, a scan image acquisition unit that scans the whole body bones of a measurement subject to whom radioactive material has been administered, and acquires a whole body bone scan image of the whole body bone, in the whole body bone scan image A joint value extractor for setting a region of interest (ROI) and extracting each joint value corresponding to the region of interest, and applying each extracted joint value to the learning model to apply the joint value of the subject to be measured. A determination unit for determining whether or not there is a lesion and a prognosis of the measurement subject, and a display unit for displaying and providing the presence or absence of a lesion of the joint and the joint value of the corresponding joint on a whole-body bone scan image of the measurement subject.

As described above, according to the present invention, standardized score data for joints of the whole body can be accumulated, and when a lesion exists, data comparison and analysis can be performed through numerical values.

In addition, through this, it can be used not only for clinical use, but also as data for artificial intelligence reading through machine learning.

1 is a configuration diagram for explaining the configuration of a whole body joint analysis system according to an embodiment of the present invention.
Figure 2 is a flow chart for explaining a whole body joint analysis method using the whole body joint analysis system according to an embodiment of the present invention.
FIG. 3 is a diagram for explaining step S220 of FIG. 2 .
FIG. 4 is a diagram for explaining step S230 of FIG. 2 .
5A and 5B are diagrams for explaining step S250 of FIG. 2 .

Hereinafter, embodiments of the present invention will be described in detail so that those skilled in the art can easily practice with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. In addition, in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a certain component is said to "include", it means that it may further include other components without excluding other components unless otherwise stated.

Then, with reference to the accompanying drawings, an embodiment of the present invention will be described in detail so that those skilled in the art can easily practice it.

1 is a configuration diagram for explaining the configuration of a whole body joint analysis system according to an embodiment of the present invention.

As shown in FIG. 1, the whole body joint analysis system 100 includes a learning unit 110, a scan image acquisition unit 120, a joint value extraction unit 130, a determination unit 140, and a display unit 150. .

First, the learning unit 110 creates a learning model by learning pre-stored whole body bone scan images and whether there are lesions according to joint values in the whole body bone scan images through artificial intelligence.

At this time, artificial intelligence enables a computer or machine to imitate human learning, reasoning, perception ability, and natural language understanding ability. create a model

Next, the scan image acquisition unit 120 scans the whole body bones of the measurement subject to which the radioactive material is administered, and acquires a whole body bone scan image of the whole body bones of the measurement subject.

Next, the joint value extraction unit 130 sets a region of interest (ROI) in the whole body bone scan image, and extracts each joint value corresponding to the region of interest.

Here, the region of interest (ROI) is one of the image processing methods and is used to set the range of an organ or joint of a subject to be measured, and is used to extract joint values in the present invention.

In this case, the joint value has different values depending on the severity of the lesion of the measurement subject, and the joint value of the bone with the lesion has a larger value than the joint value of the bone without the lesion.

Next, the determination unit 140 applies each of the extracted joint values to the learning model to determine whether or not the subject's joint has a lesion and the prognosis of the subject to be measured.

Next, the display unit 150 displays and provides the presence or absence of joint lesions and the joint value of the corresponding joint on the whole body bone scan image of the subject to be measured.

At this time, the display unit 150 overlays the maximum value, the minimum value, and the average value of the joint values corresponding to the region of interest on the whole body bone scan image and provides them.

Also, the display unit 150 calculates and provides a ratio between the average joint value in the region of interest and the average joint value of normal bones around the region of interest.

Hereinafter, a whole body joint analysis method according to an embodiment of the present invention will be described using FIGS. 2 to 5B.

Figure 2 is a flow chart for explaining a whole body joint analysis method using the whole body joint analysis system according to an embodiment of the present invention.

First, the learning unit 110 learns whether or not there is a lesion according to joint values in a pre-stored whole body bone scan image and whole body bone scan image through artificial intelligence to create a learning model (S210).

At this time, the learning unit 110 may learn the whole body bone scan image taken in real time and whether or not there is a lesion according to the joint value of the corresponding whole body bone scan image by applying artificial intelligence.

Next, the scan image acquisition unit 120 scans the whole body bones of the measurement subject to which the radioactive material is administered, and acquires a whole body bone scan image for the whole body bones (S220).

FIG. 3 is a diagram for explaining step S220 of FIG. 2 .

That is, as shown in FIG. 3 , the scan image acquisition unit 120 acquires a whole body bone scan image of the whole body bones of the measurement subject.

At this time, the whole body bone scan image is divided into front and back images of the subject to be measured and scanned.

In addition, a radioactive material is a material that generates energy by itself, and is measured in the form of a normal distribution curve with energy having a spectrum of a certain size.

In addition, the amount of energy generated by the radioactive material varies depending on the presence or absence of lesions in the organ or bone.

Next, the joint value extraction unit 130 sets a region of interest (ROI) in the acquired whole-body bone scan image, and extracts each joint value corresponding to the region of interest (S230).

FIG. 4 is a diagram for explaining step S230 of FIG. 2 .

That is, as shown in FIG. 4 , the joint value extraction unit 130 sets a region of interest (ROI).

At this time, image A in FIG. 4 is a whole body bone scan image of the subject to be measured, image B is a pelvic bone scan image of the subject to be measured, image C is a scan image of the hand bones of the subject to be measured, and image D is a scan of the bones of the foot of the subject to be measured. means video.

In addition, in FIG. 4 , the joint value extractor 130 sets a region of interest (ROI) using different colors according to each joint.

Then, the joint value extraction unit 130 extracts each joint value for each corresponding joint.

Next, the determination unit 140 applies each extracted joint value to the learning model to determine whether or not the subject's joint has a lesion and the prognosis of the subject (S240).

Here, the lesion of the joint means each disease name that may appear in the joint, such as arthritis, osteoporosis, or carpal tunnel syndrome.

In addition, when the joint value of the measurement subject is measured at the previous time point, the determination unit 140 compares the joint value at the previous time point and the current time point to determine a prognosis.

That is, if the current knee joint value is higher than the knee joint value at the previous time of the measurement subject, the determination unit 140 determines that the measurement subject's current knee has an abnormality.

Next, the display unit 150 displays and provides the presence or absence of joint lesions and the joint value of the corresponding joint on the whole body bone scan image of the measurement subject (S250).

5A and 5B are diagrams for explaining step S250 of FIG. 2 .

As shown in FIGS. 5A and 5B , the display unit 150 displays and provides the presence or absence of joint lesions and the joint value of the corresponding joint on a whole-body bone scan image of the subject to be measured.

At this time, the display unit 150 displays whether or not the corresponding joint has a lesion and the joint value of the corresponding joint on the whole body bone scan image of the subject to be measured through a terminal of a medical staff or a Dicom Viewer.

Here, the display unit 150 overlays the maximum value, the minimum value, and the average value of the joint values corresponding to the region of interest on the whole body bone scan image and provides them.

Also, the display unit 150 provides a ratio of an average joint value in the region of interest and an average joint value of normal bones around the region of interest.

At this time, if the joint values of the measurement subject at the previous time point are stored, the display unit 150 displays the joint values at the previous time point together on the whole body bone scan projection and provides them.

As described above, according to an embodiment of the present invention, standardized score data for joints of the whole body can be accumulated, and when a lesion exists, comparative analysis of the data is possible through numerical values.

In addition, through this, it can be used not only for clinical use, but also as data for artificial intelligence reading through machine learning.

Although the present invention has been described with reference to the embodiments shown in the drawings, this is only exemplary, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true technical scope of protection of the present invention should be determined by the technical spirit of the appended claims.

100: whole body joint analysis system, 110: learning unit,
120: scan image acquisition unit, 130: joint value extraction unit,
140: determination unit, 150: display unit

Claims (10)

In the joint analysis method using the whole body joint analysis system based on artificial intelligence,
Creating a learning model by learning pre-stored whole body bone scan images and whether there are lesions according to joint values in the whole body bone scan images through artificial intelligence;
Acquiring a whole body bone scan image for the whole body bone by scanning the whole body bone of the measurement subject to whom the radioactive material was administered;
Setting a region of interest (ROI) in the whole body bone scan image and extracting joint values corresponding to the region of interest;
Applying each of the extracted joint values to the learning model to determine whether or not the subject's joint has a lesion and the prognosis of the subject; and
Including the step of displaying and providing the presence or absence of the joint lesion and the joint value of the corresponding joint on a whole-body bone scan image of the measurement subject,
In the step of displaying and providing the joint values, the joint analysis method provides a maximum value, a minimum value, and an average value of the joint values corresponding to the region of interest by overlaying them on the whole body bone scan image. According to claim 1,
The joint value is,
Joint analysis method having different values depending on the severity of the lesion of the measurement subject, and the joint value of the bone in which the lesion exists has a greater value than the joint value of the bone in which the lesion does not exist. According to claim 1,
The area of interest is
Joint analysis method that is set to a part corresponding to the joint of the measurement subject from the scanned whole body bone image. delete According to claim 1,
The step of displaying and providing the joint value,
A joint analysis method for calculating and providing a ratio between an average of joint values in the region of interest and an average of joint values of normal bones around the region of interest. In the joint analysis system based on artificial intelligence for whole body joint analysis,
A learning unit for generating a learning model by learning pre-stored whole body bone scan images and whether there are lesions according to joint values in the whole body bone scan images through artificial intelligence;
A scan image acquisition unit that scans the whole body bones of the measurement subject to whom the radioactive material has been administered and acquires a whole body bone scan image for the whole body bones;
a joint value extractor configured to set a region of interest (ROI) in the whole body bone scan image and extract joint values corresponding to the region of interest;
A determination unit for applying each of the extracted joint values to the learning model to determine whether or not a joint of the measurement subject has a lesion and a prognosis of the measurement subject; and
A display unit for displaying and providing the presence or absence of the joint lesion and the joint value of the corresponding joint on a whole-body bone scan image of the subject to be measured,
the display unit,
A joint analysis system for providing a maximum value, a minimum value, and an average value of joint values corresponding to the region of interest by overlaying them on the whole body bone scan image. According to claim 6,
The joint value is,
Joint analysis system having different values depending on the severity of the lesion of the measurement subject, and the joint value of the bone where the lesion exists has a greater value than the joint value of the bone where the lesion does not exist. According to claim 6,
The area of interest is
Joint analysis system that is set to a part corresponding to the joint of the measurement subject from the scanned whole body bone image. delete According to claim 6,
the display unit,
A joint analysis system for calculating and providing a ratio of an average of joint values in the region of interest and an average of joint values of normal bones around the region of interest.

Images