KR20220112930A - Method for uroflowmetry - Google Patents

Abstract

The present invention relates to a uroflowmetry method, and more specifically, to a uroflowmetry method, which can obtain a maximum urinary flow rate measurement value by minimizing errors through classification and learning of a maximum urinary flow rate. According to an embodiment of the present invention, a uroflowmetry method includes the steps of: capturing an image of a uroflowmetry result on an electronic medical record; transmitting urinary wave information extracted from the uroflowmetry result image to a urinary wave database; inputting the urinary wave information into a maximum urinary flow rate analysis model capable of learning by using the urinary flow wave information stored in the urinary wave database; and calculating a maximum urinary flow rate result value of the urinary flow waveform from the maximum urinary flow rate analysis model.

Description

UROFLOWMETRY {METHOD FOR UROFLOWMETRY}

The present invention relates to a urine flow test method, and to a urine flow test method capable of obtaining a maximum urine velocity measurement value by minimizing errors through maximum urine velocity classification and learning.

Urinary urination test is a non-invasive test method that can be performed for both children and adults, as a method of examining the urinary function of a patient such as an abnormal urine output in a patient complaining of urination disorder. The main guidelines for urology in Korea and abroad are recommending it as a diagnostic test for patients with voiding disorders such as benign prostatic hyperplasia, overactive bladder, and neurogenic bladder.

Urinalysis measures urine flow time, urine volume, flow rate, and maximum urine velocity. Among them, the urine velocity is measured as a volume per unit time (ml/s), and the maximum urine velocity is the maximum velocity of the urine velocity, and is generally treated as the most important variable in the urine flow test.

The urine flow curve of a person who urinates normally appears as a continuous and smooth arc, whereas the urine flow curve of a patient with a problem with urination is observed to have an irregular shape, such as flattened, asymmetrical, or fluctuating up and down. do.

The current clinically used urine flow test measurement equipment cannot measure the maximum urine velocity that reflects the measurement variables of the urine flow test other than the maximum urine velocity and the characteristics of the patient during the test, for example, the change in posture during urination.

That is, when the patient moves the body or buttocks or coughs during the urine flow test, the shape of the urine stream of the patient changes irregularly, and an abnormal increase in urine flow that cannot be regarded as the actual maximum urine velocity may occur. The current clinically used urine flow test and measurement equipment cannot measure the maximum urine velocity considering such a situation.

Therefore, when a specific situation occurs during the urine flow test, it is difficult to see the maximum urine velocity of the urine flow test as the patient's actual maximum urine velocity, so correction and interpretation by a urologist is required clinically. At this time, since interpersonal and intrapersonal variations may occur in the interpretation process, errors are still highly likely to occur.

The present invention was created in view of the above circumstances, and an object of the present invention is to provide an analysis model and algorithm capable of minimizing the measurement error of the maximum urine velocity, which is an important measurement variable of the urine flow test method.

A urine flow test method according to an embodiment of the present invention includes the steps of: capturing a urine flow test result image on an electronic medical record; transmitting urine flow waveform information extracted from the urine flow test result image to a urine flow waveform database; and inputting the yaw waveform information into a maximum yaw velocity analysis model that can be learned by using the yaw waveform information stored in the .

Here, the yaw flow waveform database may classify and store the yaw flow waveform information as normal or abnormal according to a preset criterion.

Here, the preset criteria for classifying the urine flow waveform information as normal or abnormal may be set by combining one or more information among the maximum urine velocity, average urine velocity, the shape of the urine flow curve, the amount of urine, the amount of residual urine after urination, and the time it takes to urinate. .

Here, the maximum urine velocity analysis model excludes the peak value in the singular state from the urine flow waveform including the urine velocity value in the singular state during the urine flow test based on the information on the normal or abnormal waveform stored in the urine flow waveform database, and thus the maximum urine velocity result can be calculated.

Here, an AI-based maximum yaw velocity evaluation algorithm for classifying and learning a peak value of yaw velocity according to yaw time from waveform information stored in the yaw waveform database may be applied to the maximum yaw velocity analysis model.

Here, the method may further include defining a boundary of the captured urine flow test result image and extracting a urine flow waveform.

Here, the turbulence waveform may be extracted by applying an image processing technique through setting of RGB average values and deviations.

Here, the maximum yaw velocity analysis model may be corrected by comparing the input value and the output value with a loss function by using the yaw flow time as an input value and the peak value of the yaw velocity as an output value.

Here, the method may further include, when the resultant maximum yaw velocity of the yaw flow waveform is calculated, mapping and labeling the pixel value of the yaw waveform with the pixel value of the peak and storing it as training data in the yaw waveform database.

Here, the yaw waveform information stored in the yaw waveform database may be periodically or irregularly input as learning data.

The urine flow test method according to the present invention has the effect of providing a precise maximum urine velocity measurement value by reflecting the characteristics of the patient during urination.

The urine flow test method according to the present invention has the effect of excluding the peak value of the singular state through data collection, statistical analysis, and learning on the singular state of the urine flow waveform that may occur abnormally during the urine flow test.

1 is a flowchart of a urine flow test method according to an embodiment of the present invention.
2A is a urine flow test result image extracted using a capture program according to the present invention.
FIG. 2B is a diagram illustrating a method of defining a desired boundary region in the captured image of FIG. 2A .
FIG. 2C is a diagram illustrating a graph region extracted through boundary definition of FIG. 2B .
2D is a diagram illustrating a detected yaw wave image.
2E is a view for explaining a process of determining the position of the peak of the urine velocity.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to exemplary drawings. In adding reference numerals to components in each drawing, it should be noted that the same components are given the same reference numerals as much as possible even though they are indicated in different drawings. In addition, in describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

1 is a flowchart of a urine flow test method according to an embodiment of the present invention.

Referring to FIG. 1 , first, a urine flow test result image is captured from an electronic medical record (EMR) (S110). The electronic medical record records and preserves all medical records as electronic documents, and the medical staff can inquire and prescribe patient information and medical records in real time through the electronic medical record, thereby eliminating duplicate work and improving the efficiency of treatment. can be raised

In the present invention, a separate capture program may be used to capture the urine flow test result image. When the capture program is running, the area where the urine flow test result image is checked, the area where the captured image is output when the capture button is pressed, the area where the boundary of a limited range including graphs is detected when the boundary detection button is pressed, and the usage log Includes output area.

Then, the yaw wave information is extracted from the captured image, and the yaw wave information is transmitted and stored in the yaw wave database ( S120 ). The yaw waveform database may classify and store the yaw waveform information as normal or abnormal according to a preset criterion.

That is, according to the urination state of the patient, the urine flow waveform information may be classified as normal or abnormal information, and the preset criteria for classifying the urine flow waveform information as normal or abnormal are the maximum urine velocity, the average urine velocity, the shape of the urine flow curve, the amount of urine, and the urination. It may be set by combining one or more pieces of information among the residual urine amount and the time it takes to urinate.

Subsequently, an AI (Artificial Intelligence)-based maximum yaw velocity evaluation algorithm for classifying and learning the peak value of yaw velocity according to yaw time is generated and executed from the waveform information stored in the yaw waveform database (S130).

Then, when the yaw waveform information is input to the maximum yaw velocity analysis model that can be learned using the yaw waveform information stored in the yaw waveform database (S140), the maximum yaw velocity result value of the yaw waveform is calculated from the maximum yaw velocity analysis model (S150).

Here, the AI-based maximum yaw rate evaluation algorithm can be applied to the maximum yaw velocity analysis model, which is based on the information about normal or abnormal waveforms stored in the urgency waveform database. The maximum yaw velocity result can be calculated by excluding the peak value in the singular state from the yaw flow waveform.

The unusual condition during the urine flow test refers to a situation in which the urine stream of the patient changes irregularly, such as when the patient moves or coughs during the urine flow test. According to the present invention, when such a singular state occurs, by excluding the peak value of the urgency waveform that cannot be viewed as the actual maximum urgency, it is possible to minimize the error of the inspection and increase the precision.

FIG. 2A is a urine flow test result image extracted using a capture program according to the present invention, and FIG. 2B is a diagram illustrating a method of defining a desired boundary area in the captured image of FIG. 2A.

Referring to FIGS. 2A and 2B , the X-axis is time (s) and the Y-axis is urine rate (ml/s), indicating the amount of urine per unit time. In addition, the waveform represents the change in the amount of urine over time.

In order to extract a waveform from the captured image, preprocessing is performed to extract the region containing the graph where the turbulence is displayed. If the boundary is not set as a limited area, an unintended graph or image pattern may be extracted, so the area including the waveform is extracted using the boundary detection technique. In order to define the boundary, as shown in FIG. 2B , the rectangular region may be extracted in the order of left, top, right, and bottom. Here, the order of limiting the boundary detection in the left, top, right, and bottom order may be changed or may be performed simultaneously.

FIG. 2C is a diagram illustrating a graph region extracted through boundary definition of FIG. 2B, and FIG. 2D is a diagram illustrating a detected turbulence waveform image.

In FIGS. 2C and 2D, in order to extract the turbulence waveform, if pixels whose RGB values are less than or equal to a preset value are selected and output as an image, the graph area is highlighted on the capture program as shown in FIG. 2D ( highlighted) and displayed.

RGB of a computer is R (Red), G (Green), and B (Blue), each having a value between 0 and 255, and the corresponding value means the amount or concentration of a color. As for the RGB value, the closer to 0, the darker the color, the closer to 255, the brighter the color. Various colors can be expressed by combining the three values.

In the image output from the capture program, the graph and various horizontal and vertical axes are displayed in black, and the graph is closer to black than other colors. Therefore, in order to extract the curve graph at the center of the image, for example, it may be set to extract a value having an average RGB value of 40 or less and a deviation of R-G-B color values of 5 or less. It goes without saying that the RGB average values and deviations may be variously modified depending on the type of capture program.

2E is a view for explaining a process of determining the position of the peak of the urine velocity.

Referring to FIG. 2E , in the graph area extracted in FIG. 2D , a horizontal line for determining the peak position is generated, and the position of the graph contacting the horizontal line and the uppermost end is determined as the position of the peak while moving the horizontal position up and down. When the peak is determined, the corresponding waveform can be stored as training data in a database by mapping and labeling the pixel value of the waveform and the pixel value of the peak. The yaw wave information stored in the yaw wave database may be periodically or irregularly input as learning data.

Also in FIG. 2E , the X-axis is time (s), and the Y-axis is urine rate (ml/s), indicating the amount of urine per unit time. In addition, the waveform represents the change in the amount of urine over time.

2E shows a urine flow waveform including a urine velocity value in a specific state due to a change in a patient's posture during a urine flow test. In such a situation, the existing program determines the highest point as the peak of the urine velocity, but it cannot be clinically determined as the maximum urine velocity.

The urine flow test method of the present invention excludes the peak value of the singular state based on the information about the waveform in the singular state stored in the urine flow waveform database so that the maximum urine velocity analysis model does not simply judge the highest point of the graph as the maximum urine velocity. Extract the maximum yaw velocity result. These corrected results are stored in the yaw wave database again as data for machine learning and artificial intelligence learning to form big data. In this learning process, the secondary correction of the peak value and data storage process by the user (physician) may be performed.

In this case, the waveform information stored in the yaw waveform database is classified and stored as normal or abnormal according to a preset reference based on yaw time, average yaw velocity, maximum yaw velocity, and the like.

Next, by using the classified waveform information in the yaw waveform database as training data, an analysis evaluation model is constructed using a Recurrent Neural Network (RNN) algorithm among supervised classification techniques on the AI analysis framework. Here, as the RNN algorithm, for example, Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), etc. may be used, but is not limited thereto.

Specifically, by setting the RNN algorithm to input the yaw time on the X-axis and output the peak point of the yaw on the Y-axis, a maximum yaw velocity analysis model is generated through data learning. An input value and an output value of the maximum yaw velocity analysis model may be compared with a loss function to correct the accuracy of the analysis model.

The urine flow test method of the present invention is for checking whether a peak value of urine flow that can reflect the actual patient's maximum urine velocity is present, and it is possible to reduce the error of the test by finding an accurate maximum urine velocity result value.

In order to finally determine whether a patient's urination status is normal or abnormal through a urine flow test, it is necessary to consider various factors such as maximum urine velocity, average urine velocity, shape of a urine flow curve, urine output, residual urine volume after urination, and time taken to urinate. Among these various factors, the maximum urine velocity is the most important and basic element for determining the urination state of a patient, and the present invention can provide an accurate maximum urine velocity result value by excluding errors that may occur in the existing urine flow test method. .

The maximum yaw velocity analysis model may be improved through regular or irregular learning by using yaw waveform information classified and stored in the database.

The calibrated maximum urine velocity analysis model can be integrated with a program that executes real-time urine flow test result image capture and analysis on a computer (PC) running an electronic medical record (EMR).

The integrated program presents the maximum yaw velocity result according to the analysis result through real-time image capture.

Although the present invention has been described in detail with reference to preferred embodiments so far, the present invention is not limited to the above-described embodiments, and without departing from the gist of the present invention as claimed in the following claims, the technical field to which the present invention pertains It will be said that the technical spirit of the present invention extends to the extent that any person with ordinary skill in the art can make various modifications or corrections.

Claims (10)

capturing an image of the urine flow test result on the electronic medical record;
transmitting the urine flow waveform information extracted from the urine flow test result image to a urine flow waveform database;
inputting the yaw waveform information into a maximum yaw velocity analysis model that can be learned by using the yaw waveform information stored in the yaw waveform database; and
and calculating a maximum urine velocity result value of the urine flow waveform from the maximum urine velocity analysis model. The urine flow test method according to claim 1, wherein the urine flow waveform database classifies and stores the urine flow waveform information as normal or abnormal according to a preset criterion. The method according to claim 1 or 2, wherein the preset criterion for classifying the urine flow waveform information as normal or abnormal is at least one of a maximum urine velocity, an average urine velocity, a shape of a urine flow curve, an amount of urine, an amount of residual urine after urination, and time taken to urinate. A urine flow test method, characterized in that it is set by combining information. According to claim 1 or 2, wherein the maximum urine velocity analysis model is based on the information on the normal or abnormal waveform stored in the urine flow waveform database, from the urine flow waveform including the urine velocity value of the singular state during the urine flow test in a singular state. A urine flow test method, characterized in that the maximum urine velocity result value is calculated by excluding the peak value of The urine flow test according to claim 1, wherein an AI-based maximum yaw velocity evaluation algorithm for classifying and learning a peak value of yaw velocity according to urgency time is applied to the yaw flow analysis model from the waveform information stored in the yaw waveform database. Way. The urine flow test method according to claim 1, further comprising defining a boundary of the captured urine flow test result image and extracting a urine flow waveform. The method of claim 1 or 6, wherein the urine flow waveform is extracted by applying an image processing technique through setting of RGB average values and deviations. The method according to claim 1, wherein the maximum urine velocity analysis model uses urine flow time as an input value and a peak value of urine velocity as an output value, and compares the input value and the output value with a loss function and corrects it. . The method of claim 1, further comprising: when the resultant maximum yaw velocity of the yaw waveform is calculated, mapping and labeling the pixel value of the yaw waveform with the pixel value of the peak and storing the yaw waveform database as training data A method of testing for urine flow with The urine flow test method according to claim 1 or 9, wherein the urine flow waveform information stored in the urine flow waveform database is periodically or irregularly input as learning data.

Images