KR102237009B1 - Method for determine progress of polycystic kidney or polycystic liver and medical electronic device thereof - Google Patents

Abstract

In the medical electronic device according to an embodiment disclosed herein, an input module for acquiring at least one test image, a plurality of training images, and information on an area corresponding to a height or a region corresponding to a liver among the plurality of training images A memory for storing a deep learning model based on, and the input module and at least one processor electrically connected to the memory, wherein the at least one processor acquires the at least one test image through the input module, , Based on the stored deep learning model, detects an area corresponding to a height or an area corresponding to a liver among the acquired at least one test image, obtains an area of the detected area, and at least based on the acquired area It may be characterized in that it is configured to estimate the volume of the kidney or liver included in the obtained at least one test image. In addition, various effects that are directly or indirectly recognized through the present invention may be provided.

Description

TECHNICAL FIELD TECHNICAL FIELD The method of determining the progression of the Da Nang kidney or Da Nang, and a medical electronic device for performing the same.

The embodiments disclosed in the present invention relate to a method of determining a progression state between a Danang kidney or a Danang, and a technology for a medical electronic device for performing the same.

Polycystic kidney disease is a genetic disorder in which the kidney is transformed into a honeycomb due to multiple cysts filled with fluid. As the disease progresses, the cyst gradually replaces the normal tissue, the size of the kidney increases, and the function gradually decreases, leading to end-stage renal failure. Some polycystic kidney disease can also be accompanied by polycystic liver disease.

There are approximately 12.5 million people with polycystic kidney disease worldwide, and it is a disease that occurs in approximately one in 500 people. Danang kidney disease has been an incurable disease in the past, but has received a lot of attention due to the recent development of new drugs.
(Patent Document 1) Korean Patent Application Publication No. KR 10-2018-0092797A

Meanwhile, in the diagnosis of polycystic kidney disease or polycystic liver disease, accurately measuring the volume of the kidney or liver may be very important pathophysiologically or medically. For example, the pathophysiological hypothesis that a pressure effect occurs due to an increase in the volume of the kidney and liver has not yet been evaluated properly due to the practical reason that'the volume of the liver cannot be accurately measured'. For another example, depending on the volume of the kidney or liver, the doctor may more accurately determine whether the patient has polycystic kidney disease or polycystic liver disease, or the degree of progression of the disease.

If the volume of the kidney or liver can be accurately measured, the hypothesis related to the decline in renal function can be more clearly evaluated, and the value can be recognized as an important research data for the functional evaluation of drugs to be released in the future. In addition, it can greatly contribute to enabling the doctor to determine the appropriate time of treatment so that the patient does not develop chronic renal failure.

There is a method of using a semiautomatic extraction program (2D stereology) to measure the volume of the kidney or liver, but the purchase cost of the semi-automatic extraction program is expensive with hundreds of millions of dollars, and the analysis time is also 1 to 2 hours per patient. there is a problem. In addition, when using a semi-automatic extraction program, there may be a problem that additional labor costs are incurred.

In addition, although region extraction technology such as 2D stereology is a technology that has been attempted in various areas, the degree of technical completion may be relatively low because it is difficult to grasp the anatomy in the case of the kidney or liver. For example, the accuracy of kidney area extraction using 2D stereology technology is reported to be about 86% at the maximum.

Various embodiments disclosed herein may provide a technique for solving the above problem.

In the medical electronic device according to an embodiment disclosed herein, an input module for acquiring at least one test image, a plurality of training images, and information on an area corresponding to a height or a region corresponding to a liver among the plurality of training images A memory for storing a deep learning model based on, and the input module and at least one processor electrically connected to the memory, wherein the at least one processor acquires the at least one test image through the input module, , Based on the stored deep learning model, detects an area corresponding to a height or an area corresponding to a liver among the acquired at least one test image, obtains an area of the detected area, and at least based on the acquired area It may be characterized in that it is configured to estimate the volume of the kidney or liver included in the obtained at least one test image.

According to an embodiment, the at least one processor is configured to determine a progression between a kidney or a kidney of a kidney or a liver of a kidney included in the acquired at least one test image based on the estimated volume. I can.

According to an embodiment, the at least one processor stores the acquired at least one test image and information on the detected area in the memory, and the acquired at least one test image and the detected area are It may be characterized in that it is configured to update the at least one deep learning model based at least on information.

According to an embodiment, the input module may include a camera module, and the at least one processor may be configured to acquire the at least one test image using the camera module.

According to an embodiment, the input module includes a communication module for wirelessly or wired communication with an external electronic device, and the at least one processor uses the communication module to transmit the at least one test image from the external electronic device. It can be configured to obtain.

According to an embodiment, the at least one test image includes a first tomography image and a second tomography image, and the at least one processor uses the input module to correspond to the first tomography image. A distance between a first tomography and a second tomography corresponding to the second tomography image is acquired, and a kidney or liver included in the acquired at least one test image based at least on the acquired spacing and the acquired area. It may be characterized in that it is configured to estimate the volume.

According to an embodiment, an output module electrically connected to the at least one processor may be further included, and the at least one processor may be configured to display the estimated volume through the output module.

According to an embodiment, the at least one processor may be configured to perform preprocessing of deleting at least one identification information included in the at least one test image.

According to an embodiment, an area corresponding to the height and an area between the plurality of training images may be specified with the same color. According to an embodiment, an area corresponding to the height and an area between the plurality of training images may be specified in different colors.

In addition, the method of determining the progression between the Danang kidney or the Danang according to an embodiment disclosed herein includes obtaining at least one test image, the acquired at least one test based on a deep learning model stored in a memory. Detecting an area corresponding to a height or a region corresponding to a liver among the images, acquiring an area of the detected area, and a height included in the acquired at least one test image based at least on the acquired area It may be characterized in that it comprises the step of estimating the volume of the liver.

According to an embodiment, the method may further include determining a progression of a kidney or a kidney or a liver between the kidney or the liver included in the acquired at least one test image based on the estimated volume.

According to an embodiment, the method includes storing the obtained at least one test image and information on the detected region in the memory, and the obtained at least one test image and information on the detected region. It may further include updating the deep learning model based on at least.

According to an embodiment, the at least one test image includes a first tomography image and a second tomography image, and the step of estimating the volume of the kidney or liver included in the at least one test image 1 obtaining a distance between a first tomography corresponding to a tomography image and a second tomography corresponding to the second tomography image, and the obtained at least one based at least on the obtained distance and the obtained area It may include estimating the volume of the kidney or liver included in the test image.

According to an embodiment, the method may further include displaying the estimated volume through an output module.

According to an embodiment, the method may further include performing preprocessing of deleting at least one identification information included in the at least one test image.

According to an embodiment, the method may further include acquiring the deep learning model using a plurality of training images and storing it in the memory.

In an embodiment, the obtaining of the deep learning model and storing it in the memory may include specifying an area corresponding to a height or an area corresponding to a liver among the plurality of training images.

In an embodiment, an area corresponding to the height and an area between the plurality of training images may be specified with the same color. In an embodiment, an area corresponding to the height and an area between the plurality of training images may be specified in different colors.

According to various embodiments disclosed herein, it is possible to more accurately and quickly extract the volume of the kidney or liver from the image including the kidney or liver. Through this, the user can quickly grasp the progression between the Da Nang kidney or the Da Nang with respect to the patient, and provide a more accurate diagnosis for the patient.

In addition, various effects that are directly or indirectly recognized through the present application may be provided.

1 is a block diagram of a medical electronic device according to an exemplary embodiment.
2 is a flowchart illustrating a method of estimating a volume of a kidney or liver by a medical electronic device according to an exemplary embodiment.
3 is a flowchart illustrating a method of estimating a volume of a kidney or liver and updating a deep learning model by a medical electronic device according to an exemplary embodiment.
4 is a diagram illustrating a result of extracting and displaying a kidney or liver region from among at least one test image by a medical electronic device according to an exemplary embodiment.

Hereinafter, various embodiments of the present invention will be described with reference to the accompanying drawings. However, this is not intended to limit the present invention to a specific embodiment, it should be understood to include various modifications, equivalents, and/or alternatives of the embodiments of the present invention. In connection with the description of the drawings, similar reference numerals may be used for similar elements.

In this application, expressions such as "have", "may have", "include", or "may include" indicate the presence of a corresponding feature (eg, a number, a function, an action, or a component such as a part). Points, and does not exclude the presence of additional features.

As used herein, expressions such as “A or B”, “at least one of A or/and B”, or “one or more of A or/and B” may include all possible combinations of the items listed together. For example, "A or B", "at least one of A and B", or "at least one of A or B" includes (1) at least one A, (2) at least one B, Or (3) it may refer to all cases including both at least one A and at least one B.

As used herein, expressions such as "first", "second", "first", or "second" may modify various elements regardless of their order and/or importance, and one element may be another element. It is used to distinguish it from an element, but does not limit the corresponding elements. For example, a first user device and a second user device may represent different user devices regardless of order or importance. For example, without departing from the scope of the rights described in the present invention, a first component may be referred to as a second component, and similarly, a second component may be renamed to a first component.

Some component (eg, the first component) is “(functionally or communicatively) coupled with/to)” to another component (eg, the second component) or “ When it is referred to as "connected", it should be understood that a certain component may be directly connected to another component or may be connected through another component (eg, a third component). On the other hand, when a component (eg, a first component) is referred to as being “directly connected” or “directly connected” to another component (eg, a second component), the component and the It may be understood that no other component (eg, a third component) exists between the different components.

The expression "configured to (or set)" as used herein is, depending on the situation, for example, "suitable for", "having the ability to", "designed to", "modified to", " It can be used interchangeably with "made to" or "can do". The term “configured to (or set) to” may not necessarily mean only “specially designed” in hardware. Instead, in some situations, the expression "a device configured to" may mean that the device "can" along with other devices or parts. For example, the phrase "a processor configured (or configured) to perform A, B, and C" means a dedicated processor (eg, an embedded processor) for performing the corresponding operation, or executing one or more software programs stored in a memory device. By doing so, it may mean a generic-purpose processor (eg, a CPU or an application processor) capable of performing corresponding operations.

Terms used herein are merely used to describe a specific embodiment, and may not be intended to limit the scope of other embodiments. Singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as commonly understood by a person of ordinary skill in the technical field described in the present invention. Among the terms used in the present invention, terms defined in a general dictionary may be interpreted as having the same or similar meaning as the meaning in the context of the related technology, and unless explicitly defined in the present invention, the terms in an ideal or excessively formal meaning It is not interpreted. In some cases, even terms defined in the present invention cannot be interpreted to exclude embodiments of the present invention.

1 is a block diagram of a medical electronic device according to an exemplary embodiment.

Referring to FIG. 1, the medical electronic device 100 may include an input module 110, a memory 120, and a processor 130. According to various embodiments, the configuration of the medical electronic device 100 may not be limited to that illustrated in FIG. 1. For example, the medical electronic device 100 may omit some of the configurations illustrated in FIG. 1, or may additionally include a configuration not illustrated in FIG. 1. For example, the medical electronic device 100 may further include an output module electrically connected to at least the processor 130.

The input module 110 may acquire at least one test image. The test image may include, for example, a tomographic screen including at least a part of a human body, for example, a kidney or a liver. The test image may be an image in which the kidney or liver region is not specified. For example, the test image may be an image obtained by photographing a tomography including a kidney or liver of a patient requiring measurement of the volume of the kidney or liver.

According to various embodiments of the present disclosure, there may be a plurality of test images, and the plurality of test images may be a plurality of tomographic screens photographed at a specified interval with respect to at least one target. For example, the first test image may be a screen photographing a first tomography of at least one object, and the second test image is a second test image parallel to at least one object at a specified interval in the first tomography. It may be a tomographic image. According to various embodiments, the input module 110 may sequentially or simultaneously acquire at least one test image. For example, the input module 110 may acquire a first test image and a second test image sequentially or simultaneously.

According to an embodiment, the input module 110 may acquire information on a region corresponding to a height or a region among a plurality of training images and a plurality of training images. The plurality of training images may be, for example, a plurality of sample data for obtaining at least one deep learning model. In an embodiment, the plurality of training images may be images in which the specification of the kidney or liver region has been completed. For example, the training image may be an image in which a kidney or liver region or volume measurement has been completed among images obtained by taking a tomography including a kidney or liver of at least one patient.

According to an embodiment, the training image may be an image in which region specification for a plurality of objects has been completed. For example, the training image may be an image in which region specification for both a region for a liver and a region for a kidney has been completed. According to various embodiments of the present disclosure, accuracy of a deep learning model obtained by using a training image obtained by performing region specification on a plurality of objects may be further improved. For example, when a deep learning model is acquired using training images that have performed both region specification for a plurality of objects, for example, liver and kidney, the accuracy of region specification for liver and kidney among the test images may be further improved. have.

According to an embodiment, region specification for the plurality of objects may be performed through a single signal processing. For example, for the training image, the region specification for the liver and the region specification for the height may be displayed with a single and the same color. According to another embodiment, region specification for the plurality of objects may be performed through different signal processing. For example, for the training image, the region specification for the liver and the region specification for the height may be displayed in different colors. When region specification for a plurality of objects is made of different signals, a deep learning model with higher accuracy can be expected compared to a case made of a single signal. According to various embodiments, different signal processing for the plurality of objects may be performed using two or more different signals.

According to various embodiments of the present disclosure, the input module 110 may sequentially or simultaneously acquire information on a specific region among a plurality of training images or a plurality of training images. For example, the input module 110 may acquire information on a specific area among a plurality of training images of a first group or a plurality of training images of a first group, and then, a plurality of training images of a second group or Information on a specific region among a plurality of training images of the first group may be additionally obtained. For another example, the input module 110 simultaneously acquires information on a specific area among a plurality of training images of a first group, a plurality of training images of a second group, and a plurality of training images of a first group and a second group. You may.

According to various embodiments, the input module 110 may include at least one of a camera module, a scan module, a communication module, and a user input module. For example, the input module 110 includes a camera module, and the camera module may take a test image. In an embodiment, the captured test image may be transferred to the processor 130 or the memory 120. For another example, the input module 110 includes a scan module, and the scan module scans an already captured test image and converts it into an electrical signal. In an embodiment, the converted electrical signal may be transmitted to the processor 130 or the memory 120. For another example, the input module 110 includes a communication module, and the communication module may acquire a test image from an external electronic device through wireless communication or wired communication. In an embodiment, the acquired test image may be transferred to the processor 130 or the memory 120.

According to an embodiment, the input module 110 may acquire a user's input as well as a test image. For example, the input module 110 may include a user input module, and the user input module may obtain various inputs for processing a test image from a user. For example, the input module 110 may obtain an interval between tomography corresponding to each test image for a plurality of test images from a user.

The memory 120 may include a volatile memory or a nonvolatile memory. Volatile memory may be composed of, for example, random access memory (RAM) (eg, DRAM, SRAM, or SDRAM). Non-volatile memory is, for example, OTPROM (one time programmable read-only memory; ROM), PROM (programmable read-only memory), EPROM (erasable programmable read-only memory), EEPROM (electrically erasable programmable read-only memory). ), mask ROM, flash ROM, flash memory, hard drive, or solid state drive (SSD).

According to various embodiments, the memory 120 may store, for example, at least one other software component of the medical electronic device 100, for example, a command, information, or data related to a program. In various embodiments, the memory 120 may store information obtained from the input module 110 or information processed from the processor 130. In various embodiments, the information obtained from the input module 110 may include, for example, a test image, a training image, information on a specific region among the training images, for example, information on a region corresponding to a height or liver, or a user input. . In various embodiments, the information processed by the processor 130 is, for example, information on a specific region of the test image, for example, information on a region corresponding to a height or liver, a training image, and a dip obtained from information on a specific region of the training image. It may include a running model and the like.

The processor 130 may include one or more of a central processing unit (CPU), an application processor (AP), or a communication processor (CP). In various embodiments, the processor 130 drives an operating system or an application program to drive at least one other component (eg, a hardware or software component) of the medical electronic device 100 connected to the processor 130. ) Can be controlled, and various data processing and operations can be performed. For example, the processor 130 may perform an operation on a command or data received from at least one of other components (eg, the input module 110) and store the result in the memory 120.

According to an embodiment, the processor 130 includes at least one deep learning model based on information on a specific region, for example, a height or a liver, among a plurality of training images and a plurality of training images stored in the memory 120. Can be obtained. In an embodiment, the processor 130 may update the deep learning model. For example, the processor 130 acquires information on a region corresponding to the kidney or liver among the new training image and the new training image, and based on the acquired image and information, the processor 130 uses the deep learning model stored in the memory 120. Can be updated.

According to an embodiment, the processor 130 acquires at least one test image through the input module 110, and based on the deep learning model stored in the memory 120, the area corresponding to the height of the at least one test image Alternatively, a region corresponding to the liver can be detected.

According to an embodiment, the processor 130 may estimate the volume of the kidney or liver included in at least one test image. For example, the processor 130 may acquire the area of the detected area among at least one test image, and estimate the height or the volume of the liver based on the acquired area. For example, the processor 130 is an interval at which at least one test image is photographed from a user through the input module 110, for example, the interval between the tomography in which the first test image is photographed and the tomography in which the second test image is photographed. And the volume of the kidney or liver may be estimated based on the obtained area. According to an embodiment, the processor 130 may determine the progression of the kidney or the liver with respect to the kidney or liver corresponding to the at least one test image based on the estimated volume of the kidney or liver.

According to various embodiments, the medical electronic device 100 may further include an output module. The output module may include, for example, at least one display. The output module is electrically connected to the processor 130 and may output data transmitted from the processor 130 in a form that can be recognized by a user. In various embodiments, the form that the user can recognize may include at least one of text, voice, image, or video. According to various embodiments of the present disclosure, the processor 130 may display a region corresponding to the kidney or liver, or the volume of the kidney or liver among the test images using an output module. According to another embodiment, the processor 130 may display a progression between the kidney or the liver and the kidney or the liver using the output module.

In the present application, the description in FIG. 1 may be the same for configurations having the same reference numerals as those included in the medical electronic device 100.

2 is a flowchart illustrating a method of estimating a volume of a kidney or liver by a medical electronic device according to an exemplary embodiment.

Referring to FIG. 2, a method 200 of estimating the volume of the kidney or liver by the medical electronic device 100 may include steps 201 to 209. According to various embodiments, the method 200 may omit at least one of steps 201 to 209 or may further include steps not shown in FIG. 2. According to an embodiment, steps 201 to 209 may be understood as methods executed by the medical electronic device 100 or the processor 130 illustrated in FIG. 1.

In step 201, the medical electronic device 100 may acquire at least one deep learning model. In an embodiment, the at least one deep learning model may be acquired based on information on a kidney or liver region among a plurality of training images and a plurality of training images acquired through the input module 110.

According to an embodiment, the at least one deep learning model may be a model for extracting a kidney or liver region from an image including a kidney or liver. For example, the at least one deep learning model may be a model for inducing a relationship between an image and a region, for example, weights for a plurality of elements included in the image. In various embodiments, a plurality of factors and weights may be different based on a training method or data of a deep learning model. According to an embodiment, the training method may include v-net training.

In an embodiment, information on a height or liver region among the plurality of training images and the plurality of training images may be a set of previously acquired data. For example, a plurality of learning images may be acquired from data previously accumulated in a medical institution or the like. In an embodiment, information on a region corresponding to the height or liver among the plurality of training images may be obtained through pre-processing of the plurality of training images. The preprocessing may include, for example, performing an operation of directly displaying an area on the plurality of training images in an existing medical institution or the like with respect to the plurality of training images. The accuracy of at least one deep learning model may be improved as the number and type of the plurality of training images increase.

In various embodiments, step 201 may be omitted. For example, when there is at least one deep learning model already stored in the memory 120, the medical electronic device 100 may skip step 201 and perform step 203. In this case, the medical electronic device 100 may use at least one deep learning model already stored in the memory 120 in steps 203 to 209 below.

In step 203, the medical electronic device 100 may acquire a test image. The test image may be acquired through at least one of the input module 110, for example, a camera module, a scan module, and a communication module. In an embodiment, the test image may be a 2D tomographic image of a patient's body for which a kidney or liver volume measurement is required.

In step 205, the medical electronic device 100 may detect a region corresponding to the kidney or liver among the test images acquired in step 203. For example, the medical electronic device 100 may apply the deep learning model acquired in step 201 to the acquired test image. In an embodiment, the deep learning model may include weight information for a plurality of elements, and the weights may be applied to a plurality of elements included in the test image. Through this, the medical electronic device 100 may detect a specific region corresponding to the deep learning model, for example, a region corresponding to the kidney or liver, among the test images.

In step 207, the medical electronic device 100 may obtain the area of the area detected in step 205. For example, the medical electronic device 100 may acquire the area based on the size of the test image, the magnification of the test image, and/or the ratio of the detected area among the test images.

In step 209, the medical electronic device 100 may estimate the volume of the kidney or liver included in the test image acquired in step 203. For example, the medical electronic device 100 may estimate the volume of the kidney or liver based on the specified interval when the test image is continuously photographed with respect to one object, for example, a human body including the kidney or liver. I can. For example, the test image may include a plurality of images continuously photographed at a specified interval for one object. The medical electronic device 100 may continuously perform step 207 for each of the plurality of images, and may estimate the volume of the kidney or liver based on the results of step 207 and a specified interval.

According to various embodiments of the present disclosure, the medical electronic device may further include an output module, and may display the volume of the kidney or liver estimated in step 209 on the output module. Through this, the user can check the progression between the Danang kidney or the Danang with respect to the patient corresponding to the test image.

Through steps 201 to 209, the medical electronic device 100 may accurately and quickly estimate the height or volume included in the test image.

3 is a flowchart illustrating a method of estimating a volume of a kidney or liver and updating a deep learning model by a medical electronic device according to an exemplary embodiment.

Referring to FIG. 3, a method 300 of estimating the volume of the kidney or liver and updating the deep learning model by the medical electronic device 100 may include steps 301 to 307. According to various embodiments, the method 300 may omit at least one of steps 301 to 307 or may further include steps not shown in FIG. 3. According to an embodiment, steps 301 to 307 may be understood as a method executed by the medical electronic device 100 or the processor 130 illustrated in FIG. 1. In the description of FIG. 3, content overlapping with the description of FIG. 2 may be omitted.

In step 301, the medical electronic device 100 may estimate the volume of the kidney or liver included in the test image by using the test image. For example, the medical electronic device 100 may estimate the volume of the kidney or liver included in the test image by performing steps 201 to 209 shown in FIG. 2.

In step 303, the medical electronic device 100 may determine a progression state of the kidney or the liver of the danang based on the volume of the kidney or liver estimated in the step 301. For example, the medical electronic device 100 may obtain information on a reference volume of a patient's height or liver corresponding to the test image through the input module 110. According to various embodiments, the reference volume may be a normal kidney or liver volume for a patient, or a kidney or liver volume measured several months before the patient.

According to an embodiment, the medical electronic device 100 may calculate a level, for example, how many times the volume of the kidney or liver obtained in step 301 is compared to the reference volume. Based on the calculated result, the medical electronic device 100 may determine a progression state of the patient's Da Nang kidney or Da Nang.

In operation 305, the medical electronic device 100 may store a test image and information on a detected region of the test image in the memory 120. When an area corresponding to the kidney or liver is detected in the test image through step 301, the test image and the detected area may be used as information on a specific area among the new training image and the new training image. Since the accuracy of the deep learning model may increase as the number of training images increases, the medical electronic device 100 may store a new training image and information on a specific region among the new training images in the memory 120.

In step 307, the medical electronic device 100 may update (or re-acquire) the deep learning model. Since the medical electronic device 100 obtains information on a specific area among the new training image and the new training image through step 305, the plurality of training images and the specific area among the plurality of training images previously stored in the memory 120 are obtained. It is possible to reacquire a new deep learning model or update an existing deep learning model along with information about. Through this, the medical electronic device 100 may develop a deep learning model and more accurately detect a region corresponding to the kidney or liver among the test images.

Through steps 301 to 307, the medical electronic device 100 may accurately and quickly grasp the progression between the Danang kidney or the Danang, and may improve the accuracy of the deep learning model.

4 is a diagram illustrating a result of extracting and displaying a kidney or liver region from among at least one test image by a medical electronic device according to an exemplary embodiment.

Referring to FIG. 4, a first test image 410 and a second test image 420 are shown. The first test image 410 and the second test image 420 may be images obtained by photographing a tomography of a part of the body, for example, including the kidney. Referring to the first test image 410 and the second test image 420, it may be difficult to clearly determine the boundary between the kidney and other parts with the naked eye.

In one embodiment, when the deep learning model according to the present invention is applied to the first test image 410 and the second test image 420, respectively, the first processed image 411 and the second processed image 421 are obtained. Can be. The first processed image 411 and the second processed image 421 are a first area 412 and a second area 422 corresponding to the height of the first test image 410 and the second test image 420, respectively. May be the displayed image.

According to an embodiment, the test image, for example, the first test image 410 or the second test image 420 may include at least one piece of identification information. For example, the identification information may include at least one of personal information about a patient, a photographing date, a photographing time, a photographing location, a tomographic interval, or scale information. According to an embodiment, before applying the deep learning model according to the present invention, a pre-processing process, for example, a process of deleting the identification information from a test image (410 or 420) or the identification information from a test image (410 or 420) A process of extracting the remaining areas except for may be performed. The pre-processing process may be performed, for example, by the processor 130 shown in FIG. 1.

In an embodiment, when a pre-processing process is performed on the test image 410 or 420, the test image 410 or 420 may be de-identified, and extraction accuracy for a target region may be improved. . That is, the pre-processing process may improve the accuracy of region extraction of the deep learning model.

According to an embodiment, the first region 412 and the second region 422 may be separated and extracted from the first processed image 411 and the second processed image 421. The medical electronic device may acquire the extracted areas of the first area 412 and the second area 422, and through this, the volume of the kidney included in the first test image 410 and the second test image 420 Can be estimated.

According to various embodiments disclosed herein, it is possible to more accurately and quickly extract the volume of the kidney or the liver from the image including the kidney or liver. Through this, the user can quickly grasp the progression between the Da Nang kidney or the Da Nang with respect to the patient, and provide a more accurate diagnosis for the patient.

According to various embodiments, the invention disclosed herein has been written to be mainly applied to the kidney or liver, but may not be limited thereto. For example, the invention disclosed herein can be extended and applied to each area of the body.

Claims (20)

In the medical electronic device,
An input module for obtaining at least one test image;
A memory for storing a deep learning model based on information on an area corresponding to a height and an area corresponding to a liver among a plurality of training images and the plurality of training images; And
Including; at least one processor electrically connected to the input module and the memory,
The at least one processor,
Obtaining the at least one test image through the input module,
Detecting an area corresponding to a height and an area corresponding to a liver among the acquired at least one test image based on the stored deep learning model,
Obtaining the area of the detected area,
Estimating the volume of the kidney and liver included in the acquired at least one test image based at least on the acquired area,
Acquiring information on the reference volume of the kidney and liver corresponding to the test image,
The reference volume is a volume measured before the test image,
The at least one processor is configured to determine a progression between the kidney and the polyangular kidney with respect to the kidney and liver included in the acquired at least one test image based on the estimated volume and the reference volume,
An electronic device for medical use, wherein an area corresponding to the height and an area corresponding to the liver among the plurality of training images are specified in different colors. delete The method of claim 1,
The at least one processor stores the obtained at least one test image and information on the detected region in the memory,
The medical electronic device, configured to update the at least one deep learning model based at least on the acquired at least one test image and information on the detected region. The method of claim 1,
The input module includes a camera module,
The at least one processor is configured to acquire the at least one test image using the camera module. The method of claim 1,
The input module includes a communication module for wirelessly or wired communication with an external electronic device,
The at least one processor is configured to obtain the at least one test image from the external electronic device using the communication module. The method of claim 1,
The at least one test image includes a first tomography image and a second tomography image,
The at least one processor,
Acquiring a distance between a first tomography image corresponding to the first tomography image and a second tomography image corresponding to the second tomography image using the input module,
The medical electronic device, configured to estimate the volume of the kidney and liver included in the acquired at least one test image based at least on the acquired spacing and the acquired area. The method of claim 1,
Further comprising an output module electrically connected to the at least one processor,
The at least one processor is configured to display the estimated volume through the output module. The method of claim 1,
The at least one processor is configured to perform preprocessing of deleting at least one identification information included in the at least one test image. delete delete In the method of determining the progression between the Danang kidney and the Danang by a processor,
Acquiring at least one test image;
Detecting an area corresponding to a height and an area corresponding to a liver among the acquired at least one test image based on a deep learning model stored in a memory;
Obtaining an area of the detected area;
Estimating the volume of kidney and liver included in the acquired at least one test image based at least on the acquired area;
Acquiring information on a reference volume of a kidney and a liver corresponding to the test image, wherein the reference volume is a volume measured before the test image; And
Determining a progression state between the kidney and the polyangular kidney with respect to the kidney and liver included in the acquired at least one test image based on the estimated volume and the reference volume;
Including,
A method of determining a progression state between a kidney and a danang, wherein an area corresponding to the kidney and an area corresponding to the liver among the plurality of learning images are specified in different colors. delete The method of claim 11,
Storing the acquired at least one test image and information on the detected area in the memory; And
Updating the deep learning model based at least on the acquired at least one test image and information on the detected region; further comprising, a method of determining a progress state between Da Nang and Da Nang. The method of claim 11,
The at least one test image includes a first tomography image and a second tomography image,
Estimating the volume of the kidney and liver included in the obtained at least one test image;
Acquiring a gap between a first tomographic layer corresponding to the first tomographic image and a second tomographic layer corresponding to the second tomographic image; And
Estimating the volume of the kidney and liver included in the acquired at least one test image based at least on the acquired spacing and the acquired area; including, a method for determining a progression between the polyangular kidney and the danang. The method of claim 11,
Displaying the estimated volume through an output module; further comprising, a method of determining a progress state between the Danang new and the Danang. The method of claim 11,
Performing pre-processing of deleting at least one identification information included in the at least one test image; further comprising, a method of determining a progress state between the Danang new and the Danang. The method of claim 11,
Acquiring the deep learning model using a plurality of training images and storing the deep learning model in the memory; further comprising, a method of determining a progress state between the Danang new and the Danang. The method of claim 17,
Acquiring the deep learning model and storing it in the memory; determining a region corresponding to a kidney and a region corresponding to a liver among the plurality of training images; How to. delete delete

Images