Information processing device, inference model generation method, digitization method, and digitization program

Abstract

Numerical data of a biological signal is generated from image data of that biological signal, which is not in a special format. [Solution] The information processing device includes a waveform detection unit that detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization unit that generates numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection results of the waveform detection unit. The generated numerical data can be used, for example, for decision-making in treatment or diagnosis.

Description

[Technical Field]

[0001] The present disclosure relates to an information processing apparatus, an inference model generation method, a digitization method, and a digitization program. [Background Art]

[0002] Various biological signal data are used for various applications. For example, the following Patent Document 1 discloses an electrocardiographic signal analyzer capable of identifying the risk of aortic stenosis from electrocardiograms using deep learning. [Prior Art Documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Unexamined Patent Publication No. 2021-112479 [Summary of the Invention] [Problem to be Solved by the Invention]

[0004] The electrocardiographic signal analyzer of Patent Document 1 uses electrocardiogram data compliant with MFER (Medical waveform Format Encoding Rules). MFER is a standardized standard that can describe general medical waveforms including electrocardiograms, electroencephalograms, respiratory waveforms, and the like. One of the characteristics of medical waveforms in MFER format is that conversion between numerical values and images is possible.

[0005] However, many medical waveforms of various biological signals including electrocardiograms do not employ a special format like the MFER format, and for such image data, there is no other option but for humans to visually read the numerical values. This causes significant hindrance, for example, when numerical data of biological signals is used for deep learning.

[0006] This disclosure has been made in view of the above-mentioned problems, and one exemplary purpose thereof is to provide a technology that enables the automatic generation of numerical data representing the time-series changes of a biological signal, even if the image data showing the waveform of the biological signal is not in a special format. [Means for solving the problem]

[0007] An information processing device relating to an exemplary aspect of this disclosure includes a waveform detection means for detecting a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization means for generating numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection result of the waveform detection means.

[0008] In an illustrative aspect of the present disclosure, at least one processor performs a waveform detection process to detect a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization process to generate numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection results in the waveform detection process.

[0009] A digitization program relating to an exemplary aspect of this disclosure causes a computer to function as a waveform detection means for detecting a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization means for generating numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection results of the waveform detection means. [Effects of the Invention]

[0010] According to an illustrative aspect of this disclosure, one exemplary effect is that it is possible to provide a technology that enables the automatic generation of numerical data representing the time-series changes of a biological signal, even if the image data showing the waveform of the biological signal is not in a special format. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram showing the configuration of the information processing device related to this disclosure. [Figure 2] This flowchart shows the flow of the quantification method related to this disclosure. [Figure 3] This is a block diagram showing the configuration of other information processing devices related to this disclosure. [Figure 4] This figure shows an example of image data acquired by the data acquisition unit. [Figure 5] This figure shows the image data generated by applying a binarization process to the second region of the image data shown in Figure 4. [Figure 6] This figure shows examples of electrocardiogram waveforms that are subject to quantification. [Figure 7] Figure 3 is a flowchart illustrating an example of the processing performed by the information processing device shown. [Figure 8] This is a block diagram showing the configuration of a computer that functions as an information processing device related to this disclosure. [Modes for carrying out the invention]

[0012] The following are examples of embodiments of the present invention. However, the present invention is not limited to the exemplary embodiments shown below, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining some or all of the technologies (things or methods) employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, embodiments obtained by appropriately omitting some of the technologies employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. In addition, the effects mentioned in each of the exemplary embodiments shown below are examples of effects that can be expected in that exemplary embodiment and do not define the scope of the present invention. That is, embodiments that do not produce the effects mentioned in each of the exemplary embodiments shown below may also be included in the scope of the present invention.

[0013] [First Exemplary Embodiment] A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is a basic form of each exemplary embodiment described later. The scope of application of each technology employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, as long as no particular technical impediment arises. Furthermore, each technology shown in the drawings referenced for describing this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, as long as no particular technical impediment arises.

[0014] (Configuration of Information Processing Apparatus 1) The configuration of the information processing apparatus 1 according to this exemplary embodiment will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the information processing apparatus 1. As shown in FIG. 1, the information processing apparatus 1 includes a waveform detection unit 101 and a digitization unit 102.

[0015] The waveform detection unit 101 detects a pixel group corresponding to a waveform portion from image data representing a waveform of a biological signal. Here, the "biological signal" is a signal that originates from a living body and reflects the state or function of the living body or its organs. For example, signals such as electrocardiogram, electroencephalogram, magnetoencephalogram, heart rate, body temperature, and blood pressure can also be used as the biological signal. Time-series biological signals can be represented as waveforms, and the aforementioned "image data" is data representing such a waveform. In image data representing a waveform, the waveform portion and the background portion can be distinguished from each other, so the waveform detection unit 101 can detect the pixel group corresponding to the waveform portion from the image data. The image data only needs to be data that enables detection of the pixel group corresponding to the waveform portion from the image data, and does not need to be image data of a special format.

[0016] The digitization unit 102 generates numerical data representing time-series changes of the biological signal in a predetermined unit corresponding to the biological signal, based on the detection result of the waveform detection unit 101. Here, the "predetermined unit corresponding to the biological signal" only needs to be an appropriate unit for representing the biological signal. For example, if the biological signal is an electrocardiogram, a unit indicating electric potential (typically mV: millivolt) may be used as the predetermined unit. Further, for example, if the biological signal is a heart rate, a unit indicating the number of heartbeats per unit time (typically beats / minute) may be used as the predetermined unit. Also, the term "numerical data" as used herein means data containing numerical values as elements.

[0017] Note that generating numerical data based on the detection result of the waveform detection unit 101 means generating numerical data by directly or indirectly using the detection result of the waveform detection unit 101. For example, the digitization unit 102 may generate the aforementioned numerical data using data obtained by analyzing the detection result of the waveform detection unit 101.

[0018] As described above, in the information processing apparatus 1 according to the present exemplary embodiment, a configuration is adopted that includes: the waveform detection unit 101 that detects a pixel group corresponding to a waveform portion from image data representing a waveform of a biological signal; and the digitization unit 102 that generates numerical data representing time-series changes of the biological signal in a predetermined unit corresponding to the biological signal, based on the detection result of the waveform detection unit 101.

[0019] According to the above configuration, an effect is obtained that even if the image data representing the waveform of the biological signal does not use a special format, it is possible to automatically generate numerical data representing time-series changes of the biological signal.

[0020] (Digitization Program) The functions of the information processing device 1 described above can also be implemented by a program. The digitization program according to this exemplary embodiment causes the computer to function as a waveform detection means for detecting a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization means for generating numerical data representing the time-series changes of the biological signal in predetermined units according to the biological signal, based on the detection results of the waveform detection means. With this digitization program, even if the image data showing the waveform of the biological signal does not have a special format applied, it is possible to automatically generate numerical data representing the time-series changes of the biological signal.

[0021] (Process of quantification) The flow of the digitization method according to this exemplary embodiment will be explained with reference to Figure 2. Figure 2 is a flowchart showing the flow of the digitization method. Note that the entity executing each step in this digitization method may be a processor provided in the information processing device 1, a processor provided in another device, or the entity executing each step may be a processor provided in a different device.

[0022] In S1 (waveform detection processing), at least one processor detects a group of pixels corresponding to the portion of the waveform from image data showing the waveform of a biological signal.

[0023] In S2 (digitalization processing), at least one processor generates numerical data representing the time-series changes of the biological signal in predetermined units corresponding to the biological signal, based on the detection results in S1.

[0024] As described above, the digitization method according to this exemplary embodiment employs a configuration in which at least one processor performs a waveform detection process to detect a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization process to generate numerical data representing the time-series changes of the biological signal in predetermined units according to the biological signal, based on the detection results in the waveform detection process. This digitization method has the effect of automatically generating numerical data representing the time-series changes of the biological signal even if the image data showing the waveform of the biological signal does not have a special format applied to it.

[0025] [Second exemplary embodiment] A second exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same function as those described in the above-described exemplary embodiment are denoted by the same reference numerals, and their descriptions are omitted as appropriate. The scope of application of each technology adopted in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology adopted in this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise. Furthermore, each technology shown in the drawings referenced to describe this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise.

[0026] (Configuration of Information Processing Device 1A) The configuration of the information processing device 1A according to this exemplary embodiment will be described with reference to Figure 3. Figure 3 is a block diagram showing the configuration of the information processing device 1A. The information processing device 1A is a device equipped with the function of generating numerical data from biological signal image data.

[0027] As shown in the figure, the information processing device 1A includes a control unit 10A that controls all parts of the information processing device 1A, and a storage unit 11A that stores various data used by the information processing device 1A. The information processing device 1A also includes a communication unit 12A for the information processing device 1A to communicate with other devices, an input unit 13A that receives input to the information processing device 1A, and an output unit 14A for the information processing device 1A to output data. The control unit 10A includes a waveform detection unit 101A, a digitization unit 102A, a data acquisition unit 103A, a preprocessing unit 104A, a baseline derivation unit 105A, and a scale determination unit 106A.

[0028] The waveform detection unit 101A, similar to the waveform detection unit 101 in Exemplary Embodiment 1, detects pixel groups corresponding to the waveform portion from image data showing the waveform of a biological signal. The type of biological signal image data to be used for waveform detection is arbitrary. The following explanation will focus on an example where electrocardiogram image data is used for waveform detection, but "electrocardiogram" in the following explanation can basically be read as any "biological signal".

[0029] The digitization unit 102A, similar to the digitization unit 102 in Exemplary Embodiment 1, generates numerical data representing the time-series change of the biological signal in predetermined units corresponding to the biological signal, based on the detection results of the waveform detection unit 101A. For example, if the biological signal is an electrocardiogram, the digitization unit 102A generates numerical data representing the time-series change of the electrocardiogram in units of mV (millivolts).

[0030] The data acquisition unit 103A acquires image data that shows the waveform to be detected by the waveform detection unit 101A. As will be explained in detail below, the waveform detection unit 101A detects the pixel group corresponding to the waveform from the image data acquired by the data acquisition unit 103A (hereinafter referred to as the original image data) after the preprocessing unit 104A has performed preprocessing to remove elements other than the waveform.

[0031] The preprocessing unit 104A performs preprocessing on the raw image data acquired by the data acquisition unit 103A to remove elements other than the waveform, thereby generating image data that the waveform detection unit 101A will use to detect pixel groups. Details of the preprocessing performed by the preprocessing unit 104A will be described later.

[0032] The baseline derivation unit 105A derives a baseline that serves as a reference for generating numerical data representing the time-series changes of a biological signal in a predetermined unit corresponding to the biological signal. For example, if the biological signal is an electrocardiogram, the baseline derivation unit 105A may detect a portion of the electrocardiogram waveform where the electrocardiogram potential is linear and derive a baseline based on the potential of that linear portion. Alternatively, the baseline derivation unit 105A may, for example, detect multiple points in the electrocardiogram waveform where the electrocardiogram potential is zero and derive a straight line passing through each of the detected points as the baseline.

[0033] The scale determination unit 106A determines the scale in the image data showing the waveform of a biological signal. The digitization unit 102A generates numerical data using the determined scale. The method for determining the scale and the method for generating numerical data using the determined scale will be described later.

[0034] As described above, the information processing device 1A according to this exemplary embodiment also employs a configuration similar to that of the information processing device 1, comprising: a waveform detection unit 101A that detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal; and a digitization unit 102 that generates numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection result of the waveform detection unit 101A. Therefore, the information processing device 1A also has the effect of being able to automatically generate numerical data representing the time-series change of the biological signal even if the image data showing the waveform of the biological signal does not have a special format applied to it.

[0035] (Example of preprocessing 1: Extraction of a portion of the original image data) An example of preprocessing by the preprocessing unit 104A will be explained with reference to Figure 4. Figure 4 is a diagram showing an example of image data acquired by the data acquisition unit 103A (i.e., the original image data described above). The original image data Img1 shown in Figure 4 is electrocardiogram image data. More specifically, the original image data Img1 is 12-lead electrocardiogram image data and includes 12 waveforms corresponding to each of the 12 types of leads.

[0036] As shown in the diagram, the image region of the original image data Img1 is divided into a first region Ar1, which displays electrocardiogram-related information such as the name of the person being measured, the date and time of measurement, the analysis results, and the comments of the medical technologist, and a second region Ar2, which shows the 12-lead electrocardiogram waveform.

[0037] When such raw image data Img1 is acquired, the preprocessor 104A can extract the second region Ar2 from the raw image data Img1. For example, if the ranges occupied by the first and second regions are determined in the raw image data Img1, the preprocessor 104A can extract the range of the second region Ar2 from the raw image data Img1. Alternatively, the preprocessor 104A may analyze the raw image data Img1 to identify the range of the second region Ar2 and extract the identified range from the raw image data Img1. The user may also be allowed to specify the extraction range, in which case the preprocessor 104A can extract the range specified by the user from the raw image data Img1 via the communication unit 12A or the input unit 13A.

[0038] Thus, the preprocessing unit 104A may perform a process to extract regions in the original image data where waveforms are drawn, as a preprocessing step to remove elements other than waveforms from the original image data. This eliminates the influence of regions in the original image data where waveforms are not drawn, making it easier or more accurate for waveform detection by the waveform detection unit 101A.

[0039] (Example of preprocessing 2: Removal of grid lines from the waveform background) In the original image data Img1 shown in Figure 4, a grid of lines indicating the scale of the electrocardiogram waveform is drawn in the background. Such grid lines can also interfere with waveform detection by the waveform detection unit 101A. For this reason, the preprocessing unit 104A may perform preprocessing to remove the grid lines in the background of the electrocardiogram waveform.

[0040] For example, the preprocessing unit 104A may remove the grid lines in the background of the electrocardiogram waveform by binarizing the original image data. Figure 5 shows image data Img2 generated by applying binarization to the second region Ar2 in the original image data Img1 shown in Figure 4. By appropriately setting the binarization threshold, image data Img2 is obtained from which the grid lines drawn in the background of the electrocardiogram waveform have been removed.

[0041] Furthermore, in addition to the 12-lead electrocardiogram waveforms, the image data Img2 also contains symbols such as "I" and "II" indicating the type of each lead, as well as waveforms called square waves or calibration waves. The preprocessor 104A may remove such areas where information other than waveforms is displayed.

[0042] For example, the preprocessor 104A may apply a mask to areas where information other than the waveform is displayed. The image data Img3 shown in Figure 5 is an example of image data generated by applying a mask to areas in the image data Img2 where information other than the waveform is displayed. Since the areas in which information other than the waveform is displayed are usually predetermined, the preprocessor 104A only needs to target the areas in which such information is displayed for masking.

[0043] Furthermore, the preprocessor 104A may detect characters or other elements other than waveforms using processing such as OCR (Optical Character Recognition) and remove the detected characters or other elements. The preprocessor 104A may also detect and remove shapes other than waveforms using pattern matching or similar methods. Additionally, if the display colors differ between the waveform and non-waveform areas, the preprocessor 104A may detect and remove the non-waveform areas based on their color.

[0044] In Figure 5, the image data Img3 shows the masked area in black, but the masking process may also be performed by making each pixel in the target area white (pixel value zero). Alternatively, instead of masking, the area targeted for detection by the waveform detection unit 101A may be narrowed down to an area where no information other than waveforms, such as symbols, is displayed. Furthermore, if the original image data contains multiple waveforms, the preprocessing unit 104A may perform preprocessing to cut out the areas in which each waveform is captured.

[0045] As described above, the information processing device 1A includes a preprocessing unit 104A that performs preprocessing on the source image data to remove elements other than waveforms, thereby generating image data that is the target of waveform detection by the waveform detection unit 101A. As a result, in addition to the effects of the information processing device 1, the influence of elements other than waveforms in the source image data is eliminated, making waveform detection by the waveform detection unit 101A easier or more accurate.

[0046] (Detection of pixel groups corresponding to the waveform portion) The waveform detection unit 101A detects pixel groups corresponding to the waveform portion of the electrocardiogram from the image data generated by the preprocessing unit 104A as described above. In detecting the pixel groups, the waveform detection unit 101A assigns coordinate values ​​to each pixel included in the image data. For example, the waveform detection unit 101A may assign coordinate values ​​to each pixel of the image data by setting the coordinates of the pixel at a reference position, such as the upper left corner of the image data, to (0,0), the rightward direction of the image data to the positive x-axis, the downward direction to the positive y-axis, and the interval between adjacent pixels to 1. This makes it possible to represent the pixels included in the image data with coordinate values ​​corresponding to their position and to detect pixels corresponding to the waveform portion using those coordinate values.

[0047] The method for detecting pixels corresponding to the waveform portion is not particularly limited. For example, the waveform detection unit 101A may detect a group of adjacent pixels in the image data as a group of pixels corresponding to the waveform portion of the electrocardiogram. If the image data is generated by preprocessing that removes all elements other than the waveform, as in the image data Img3 in Figure 5, it is possible to detect the group of pixels corresponding to the waveform portion of the electrocardiogram with such a simple process. Twelve pixel groups corresponding to each of the 12 leads are detected from the image data Img3.

[0048] Furthermore, the waveform detection unit 101A can also detect pixel groups corresponding to the waveform portion from image data that includes elements other than the waveform, as shown in the image data Img2 in Figure 5. In this case, the waveform detection unit 101A may detect pixel groups that satisfy conditions specific to pixel groups corresponding to the waveform portion from among the pixel groups consisting of adjacent pixels in the image data, and identify them as pixel groups corresponding to the waveform portion of the electrocardiogram. For example, the waveform detection unit 101A may identify pixel groups with a number of elements equal to or greater than a threshold as pixel groups corresponding to the waveform portion of the electrocardiogram. Alternatively, for example, the waveform detection unit 101A may identify pixel groups where the distance from the leftmost pixel to the rightmost pixel is equal to or greater than a threshold as pixel groups corresponding to the waveform portion of the electrocardiogram. These processes can exclude pixel groups corresponding to elements such as characters and symbols.

[0049] In the image data Img2 in Figure 5, a rectangular waveform is drawn at the left end of the 12 electrocardiogram waveforms. As mentioned above, this waveform is called a rectangular wave or calibration wave, and it is an object for calibration. In the image data Img2, the electrocardiogram waveforms intersect with the calibration wave, so if a group of pixels is detected from the image data Img2 using the condition "a group of pixels consisting of adjacent pixels," a group of pixels containing both the pixels that make up the electrocardiogram waveform and the pixels that make up the calibration wave will be detected.

[0050] Therefore, the waveform detection unit 101A may distinguish and detect the electrocardiogram waveform and the calibration wave. Since the electrocardiogram waveform and the calibration wave have clearly different shapes, it is possible to distinguish and detect the electrocardiogram waveform and the calibration wave based on the difference in shape. For example, the waveform detection unit 101A may detect the pixel group constituting the calibration wave from the image data by pattern matching or the like. In this case, the waveform detection unit 101A can remove each pixel from the pixel group constituting the calibration wave from the pixel group that includes the pixels constituting the electrocardiogram waveform and the pixels constituting the calibration wave, and detect the remainder as a pixel group corresponding to the part of the electrocardiogram waveform.

[0051] It should be noted that the detection of the pixel group constituting the calibration wave does not necessarily have to be performed by the waveform detection unit 101A. A separate detection unit may be provided to detect the pixel group constituting the calibration wave, in addition to the waveform detection unit 101A.

[0052] (Method for generating numerical data) The method for generating numerical data by the digitization unit 102A will be explained with reference to Figure 6. Figure 6 shows an example of an electrocardiogram waveform to be digitized. In the waveform W1 shown in Figure 6, waveforms called P waves, R waves, and T waves are repeated periodically. In addition, a calibration wave W2 is drawn corresponding to waveform W1. As shown in the figure, the calibration wave W2 is drawn with a size such that its height H is 1mV.

[0053] In generating numerical data, the baseline is derived by the baseline derivation unit 105A and the scale is determined by the scale determination unit 106A. The methods for deriving the baseline and determining the scale will be explained below.

[0054] (Method for deriving the baseline) In an electrocardiogram, the baseline is a straight line that represents a state where there is no electrical activity in the myocardium. The baseline derivation unit 105A may, for example, detect each pixel located at the beginning of the P wave among the pixels detected by the waveform detection unit 101A, calculate the average value μ of the y-coordinates of those pixels, and derive a straight line represented by the formula y=μ as the baseline. The pixels located at the beginning of the P wave can be detected by analyzing the coordinate values ​​of each pixel included in the pixel group detected by the waveform detection unit 101A.

[0055] Alternatively, for example, the baseline derivation unit 105A may detect a pixel located at the beginning of one P wave and a pixel located at the beginning of the next P wave, and derive a straight line passing through these points as the baseline. In this case, the baseline derivation unit 105A may detect the coordinate values ​​of the pixel located at the beginning of one P wave and the coordinate values ​​of the pixel located at the beginning of the next P wave from the group of pixels detected by the waveform detection unit 101A, and derive the equation of the straight line connecting these detected coordinate values ​​as the baseline. For example, in the example in Figure 6, the baseline derivation unit 105A can derive a straight line connecting the two points as the baseline L1 by detecting the coordinate values ​​of the beginning point P1 of the first P wave included in the waveform W1 and the coordinate values ​​of the beginning point P2 of the second P wave.

[0056] Alternatively, for example, the baseline derivation unit 105A may detect a section in which the numerical data is constant or nearly constant, calculate the average value μ of the y-coordinate in the detected section, and derive a straight line represented by the formula y=μ as the baseline. Since the value of the y-coordinate on the baseline of the waveform is constant or nearly constant, it is also possible to derive the baseline in this way. For example, the baseline derivation unit 105A may use a small bandwidth h and calculate the value of the y-coordinate of the numerical data as y i Alternatively, for each of i=0,1,...,N, the value of the following formula can be calculated.

[0057]

number

[0058] The value of the above formula is y iSince this indicates the magnitude of fluctuations in the numerical data in the vicinity, the baseline derivation unit 105A only needs to detect the section where the calculated value is small as the section where the numerical data is constant or nearly constant. Then, the baseline derivation unit 105A calculates the average value μ of the y coordinate in the detected section and derives the straight line represented by the equation y=μ as the baseline.

[0059] (Method for determining scale) In an electrocardiogram waveform, the vertical direction (y-axis direction) indicates the strength of the potential (which can also be called amplitude). Normally, a potential of 1mV is represented by 10mm, but depending on the amplitude of the waveform, a potential of 1mV may be represented by 5mm, 2.5mm, or even 20mm. If the scale for representing a potential of 1mV with 10mm is 1x, then the scale for representing a potential of 1mV with 5mm is 1 / 2x, the scale for representing a potential of 1mV with 2.5mm is 1 / 4x, and the scale for representing a potential of 1mV with 20mm is 2x. By considering such scales, the coordinates of each pixel corresponding to the part of the waveform detected by the waveform detection unit 101A can be converted into appropriate numerical data.

[0060] To determine the scale, a reference object is used, which is drawn so that a predetermined numerical value expressed in a predetermined unit corresponds to a predetermined size. As mentioned above, an electrocardiogram displays a waveform called a calibration wave. By using this calibration wave as the reference object, the scale can be determined.

[0061] For example, in the example in Figure 6, calibration wave W2 is plotted in correspondence with waveform W1, while calibration wave W4 is plotted in correspondence with waveform W3. Calibration wave W2 has a height of H, while calibration wave W4 has a height of H / 2. Both of these calibration waves represent 1mV. In other words, waveform W1, to which calibration wave W2 is plotted, is displayed at a scale where 1mV is H. On the other hand, waveform W3, to which calibration wave W4 is plotted, is displayed at a scale where 1mV is H / 2. In other words, waveform W3 is displayed at half the scale of waveform W1 in the y-axis direction.

[0062] When a calibration wave is used as the reference object, the scale determination unit 106A determines the height of the calibration wave in the image data. For example, the scale determination unit 106A may determine the maximum and minimum values ​​from the y-coordinate values ​​of each pixel constituting the calibration wave, and determine the height of the calibration wave by subtracting the minimum value from the maximum value. Then, the scale determination unit 106A determines how many mV the waveform is drawn at per unit height (1 pixel) by dividing 1 mV by the determined height. For example, the calibration wave W2 drawn corresponding to waveform W1 in Figure 6 has a height of H, so it is determined that it is drawn at a scale of "1 / H" mV per unit height in the y-axis direction. On the other hand, the calibration wave W4 is determined to be drawn at a scale of "2 / H" mV per unit height in the y-axis direction.

[0063] Furthermore, the scale in the y-axis direction can also be determined from the distance between two predetermined points on the waveform. Here, the two predetermined points are those for which the potential difference is known. For example, in the example in Figure 6, if the peak value of the P wave is known, the scale in the y-axis direction can be determined using the height of the P wave and the known peak value.

[0064] (Generating numerical data using baselines and scales) The digitization unit 102A calculates numerical data for a pixel by multiplying the distance in the y-direction between the pixel constituting the waveform and the baseline derived by the baseline derivation unit 105A by the scale determined by the scale determination unit 106A. Since the baseline is usually parallel to the x-axis, the distance in the y-direction from the baseline can be calculated by the difference between the y-coordinate value initially assigned to each pixel and the y-coordinate value of the baseline. Therefore, if the coordinates of the pixel constituting the waveform are (x1, y1), the value of the y-intercept of the baseline is y2, and the scale is s, the numerical data corresponding to the y-coordinate of this pixel is expressed as (y1-y2)×s. Thus, the process of generating numerical data can also be described as a process of transforming the coordinate values ​​initially assigned to each pixel based on the baseline and the scale.

[0065] For example, the distance from the pixel located at the peak P3 of the first peak in waveform W1 in Figure 6 to the baseline L1 is HA This is the difference between the y-coordinate of vertex P3 and the y-intercept of baseline L1. Furthermore, as mentioned above, the scale of this waveform W1 is specified as 1 / H. Therefore, the numerical data corresponding to the y-coordinate of the pixel located at vertex P3 is (1 / H) × H. A =( H A The value is calculated as ( / H)mV. This value represents the peak value of the first P wave in waveform W1. On the other hand, the distance from the pixel located at the peak P4 of the first peak in waveform W3 to the baseline L3 is H B Therefore, the numerical data corresponding to the y-coordinate of the pixel located at vertex P4 is (2 / H) × H B =(2H B It is calculated as ( / H)mV. This value represents the peak value of the first P wave in waveform W3.

[0066] (Regarding the x-coordinate value) The digitization unit 102A can similarly convert the x-coordinate values ​​of the pixels that make up the waveform into numerical data. In an electrocardiogram waveform, the x-axis direction represents time, and its unit is usually seconds. Therefore, the digitization unit 102A converts the x-coordinate values ​​of the pixels that make up the waveform into numerical data in units of "seconds". A baseline perpendicular to the x-axis and a scale in the x-axis direction are used to convert the x-coordinate values.

[0067] For example, the baseline derivation unit 105A may derive a straight line perpendicular to the x-axis that passes through the pixel with the smallest x-coordinate value among the pixels detected by the waveform detection unit 101A as the baseline. The baseline L2 in waveform W1 and the baseline L4 in waveform W3 in Figure 6 can be derived in this way. Alternatively, the user may be allowed to specify the position of the baseline. In this case, the baseline derivation unit 105A should derive a straight line perpendicular to the x-axis that passes through the position specified by the user as the baseline.

[0068] The x-axis scale can be determined by using a reference object drawn such that a predetermined numerical value expressed in a predetermined unit becomes a predetermined size in the x-axis direction. For example, if a grid drawn in the background of the waveform is detected as a reference object, the scale determination unit 106A can determine the x-axis scale from the width of the detected grid. For example, if the width of one grid corresponding to 0.04 seconds is w, the x-axis scale is determined to be (0.04 / w). Note that the waveform detection unit 101A may perform the detection of the grid, or a separate detection unit for detecting the grid may be provided in addition to the waveform detection unit 101A.

[0069] Furthermore, the x-axis scale can also be determined from the distance between two predetermined points on the waveform. Here, the two predetermined points are those for which the time interval is known. For example, in the example in Figure 6, the time from point P1 to point P2 is called the PP time. If this PP time is known, the x-axis scale can be determined using the x-axis distance between points P1 and P2 and the PP time.

[0070] As described above, the information processing device 1A may include a detection unit that detects a reference object drawn from image data such that a predetermined numerical value expressed in a predetermined unit is of a predetermined size. The digitization unit 102A may then generate numerical data representing the time-series changes of the biological signal in a predetermined unit, based on the size (more precisely, the width or height on the image data) of the detected reference object. This provides the effect of automatically generating numerical data with appropriate values ​​for waveforms of various scales, in addition to the effects of the information processing device 1. As mentioned above, the waveform detection unit 101A may be used to detect the reference object, or a detection unit that detects the reference object separately from the waveform detection unit 101A may be provided.

[0071] Furthermore, as described above, the information processing device 1A may include a baseline derivation unit 105A that derives a baseline passing through multiple points in the waveform where the biological signal reaches a predetermined reference value. The digitization unit 102A may then generate numerical data representing the time-series change of the biological signal in predetermined units, based on the distance from the pixel detected by the waveform detection unit 101A to the baseline. This provides the added benefit of automatically generating numerical data with reasonable values ​​based on the baseline, in addition to the effects of the information processing device 1.

[0072] Furthermore, baseline detection and scale determination do not necessarily need to be performed automatically. In other words, it is not essential to provide a baseline derivation unit 105A, nor is it essential to provide a scale determination unit 106A. For example, the user may specify either the baseline or the scale, or both. In that case, the digitization unit 102A should generate numerical data by applying the specified baseline and / or scale.

[0073] (Variations in the method of generating numerical data) As described above, after assigning coordinate values ​​to each pixel of the image data, it is possible to generate numerical data by transforming the coordinate values ​​of each pixel in the pixel group corresponding to the waveform portion. Furthermore, it is also possible to generate numerical data representing the time-series changes of a biological signal in predetermined units corresponding to the biological signal without performing such transformations.

[0074] For example, the digitization unit 102A can also generate numerical data by using a pixel in the pixel group corresponding to the waveform portion of the biological signal as a reference, for which a value in a predetermined unit corresponding to the biological signal is known, and assigning a numerical value to each pixel according to its positional relationship with the reference pixel.

[0075] For example, suppose the coordinates of the leftmost point in waveform W1 in Figure 6 are (0 seconds, 0 mV), and it is known that one pixel in the x-axis direction is t seconds and one pixel in the y-direction is v mV. This information may be entered by the user or determined by analyzing the image data.

[0076] In this case, the digitization unit 102A uses the coordinates of the leftmost point in the waveform W1 as a reference point and assigns numerical values ​​to the other pixels in the pixel group corresponding to the portion of the biological signal waveform, according to their positional relationship with the reference point. Specifically, the digitization unit 102A assigns numerical values ​​for the x-coordinate such that it increases by t for every pixel moved one pixel away from the reference point in the x-axis direction, and assigns numerical values ​​for the y-coordinate such that it increases by v for every pixel moved one pixel away from the reference point in the y-axis direction. For example, the digitization unit 102A can assign the value (10t, 0) to a point that is 10 pixels away from the reference point in the x-axis direction. By assigning numerical values ​​to each point on the waveform W1 in this way, the digitization unit 102A can generate numerical data without performing the above-mentioned transformations.

[0077] (Process flow) The processing flow performed by the information processing device 1A will be explained with reference to Figure 7. Figure 7 is a flowchart showing an example of the processing performed by the information processing device 1A. The flowchart in Figure 7 includes each process of the digitization method according to this exemplary embodiment.

[0078] In S11, the data acquisition unit 103A acquires image data showing the waveform to be detected by the waveform detection unit 101A. The image data acquired here is before preprocessing by the preprocessing unit 104A and corresponds to the "original image data" described above. For example, the data acquisition unit 103A may acquire original image data that includes display areas for elements other than the waveform, such as the original image data Img1 in Figure 4. The data acquisition unit 103A may also acquire image data that has been preprocessed by the information processing device 1A or other devices, in which case the processing in S12 and S13 is omitted.

[0079] In S12, the preprocessing unit 104A extracts the region in which the waveform to be digitized is drawn from the image data (original image data) acquired in S11. For example, if the image data acquired in S11 is the original image data Img1 shown in Figure 4, the preprocessing unit 104A extracts the second region Ar2 from the original image data Img1.

[0080] In S13, the preprocessing unit 104A binarizes the region extracted in S12 from the original image data. This generates image data that can be detected by the waveform detection unit 101A, from which elements such as grid lines have been removed. For example, the preprocessing unit 104A may generate image data like the image data Img2 in Figure 5. Alternatively, for example, the preprocessing unit 104A may apply masking to the region in the image data Img2 where characters or symbols are drawn to generate image data like the image data Img3 in Figure 5. In this exemplary embodiment, an example is shown in which one preprocessing unit 104A performs multiple types of preprocessing, but a separate preprocessing unit may be provided for each type of preprocessing.

[0081] In S14 (waveform detection processing), the waveform detection unit 101A assigns coordinate values ​​to each pixel in the image data generated in S13, and then detects a group of pixels from the image data that corresponds to the waveform portion of the biological signal. Since each pixel is assigned a coordinate value, it can also be said that S14 detects a group of coordinate values ​​that represent the waveform of the biological signal.

[0082] Furthermore, if the image data generated in S13 includes multiple waveforms, such as the image data Img2 and Img3 in Figure 5, in S14, the waveform detection unit 101A detects the pixel groups corresponding to each of those multiple waveforms. In addition, in S14, the waveform detection unit 101A may also detect reference objects (more precisely, pixel groups corresponding to reference objects) that are drawn such that predetermined numerical values ​​expressed in predetermined units become a predetermined size, such as the calibration waves W2 and W4 shown in Figure 6.

[0083] In S15, the baseline derivation unit 105A derives a baseline that serves as a reference for generating numerical data representing the time-series change of the biological signal in a predetermined unit corresponding to the biological signal. For example, the baseline derivation unit 105A may detect multiple pixels at the starting point of the P wave from the pixels detected in S14 and derive a straight line connecting them as the baseline (the baseline that serves as the reference for generating numerical data in the y-axis direction). Alternatively, for example, the baseline derivation unit 105A may detect the pixel with the smallest x-coordinate among the pixels detected in S14 and derive a straight line passing through the detected pixel and perpendicular to the x-axis as the baseline (the baseline that serves as the reference for generating numerical data in the x-axis direction). If multiple pixel groups were detected in S14, the baseline derivation unit 105A derives a baseline for each pixel group (which can also be rephrased as each waveform).

[0084] In S16, the scale determination unit 106A determines the scale (both x-axis and y-axis) of the image data generated in S13. For example, if a reference object is detected in S14, the scale determination unit 106A can determine the scale from the size of the detected reference object. As mentioned above, a grid drawn on the background of the waveform may also be used as the reference object. In that case, the grid should be detected from the original image data in which the grid has not been removed, and the scale should be determined from the size (width) of the detected grid.

[0085] If multiple pixel groups are detected in S14, the scale determination unit 106A determines the scale for each pixel group (which can also be called each waveform). However, for pixel groups with a common scale, it is sufficient to determine the scale for only one of them. For example, if the scale in the x-axis direction is common to multiple pixel groups, the scale in the x-axis direction is determined for only one pixel group, and the same scale is applied to the other pixel groups.

[0086] In S17 (digitization processing), the digitization unit 102A generates numerical data representing the time-series change of the biological signal in predetermined units corresponding to the biological signal, based on the detection result in S14, the baseline derived in S15, and the scale specified in S16. Specifically, the digitization unit 102A calculates the distance between the x and y coordinates of each pixel included in the pixel group corresponding to one waveform and the baseline, and multiplies the calculated distance by the scale specified in S16. By performing this process for each pixel included in the pixel group, numerical data representing the time-series change of the biological signal in predetermined units corresponding to the biological signal is generated. If multiple pixel groups were detected in S14, the digitization unit 102A generates numerical data for each pixel group (which can also be called each waveform). This completes the process shown in Figure 7.

[0087] [Regarding the use of generated numerical data] Numerical data generated by the information processing device 1A can be used for a variety of purposes. For example, the information processing device 1A may include a function derivation unit that derives a function representing the waveform of a biological signal using numerical data generated from image data representing the waveform of that biological signal. By using the function representing the waveform, the value of the biological signal at any given time can be calculated.

[0088] Furthermore, the generated numerical data can be used, for example, for decision-making in treatment or diagnosis. From the perspective of use in treatment and diagnosis, for example, the information processing device 1A may be equipped with an index value calculation unit that calculates various index values ​​used for the diagnosis of a subject from the numerical data generated from the subject's biological signals. For example, if the biological signals are an electrocardiogram, the index values ​​may include the width and height of the P wave, the PQ interval (the interval between the P wave and the Q wave), the RR interval (the interval between the peaks of two consecutive R waves on the electrocardiogram), and the width of the QRS complex (from the beginning of the Q wave to the end of the S wave).

[0089] For example, when the index value calculation unit calculates the RR interval as an index value, it may detect an extreme value that satisfies predetermined conditions from the numerical data of the electrocardiogram and estimate the position where the numerical data of the electrocardiogram becomes that extreme value as the position of the peak of the R wave. In this way, the index value calculation unit can calculate various index values ​​with simple calculations by using the numerical data generated by the digitization unit 102A.

[0090] Furthermore, if an index value calculation unit is provided, the information processing device 1A may further include a display control unit that displays the calculated index value together with image data showing the waveform of the biological signal. This allows medical professionals to simultaneously recognize both the visual characteristics of the biological signal waveform and the index value, which can aid in rapid and highly accurate treatment and diagnosis.

[0091] Furthermore, for example, the information processing device 1A may include a diagnostic unit that determines the health status and presence or absence of disease of the subject based on the calculated index values. In this case, the index value calculation unit should be made to calculate the index values ​​necessary for the diagnostic unit's determination. For example, if the diagnostic unit determines the presence or absence of arrhythmia, the index value calculation unit should calculate the index values ​​necessary for determining arrhythmia.

[0092] Furthermore, for example, the information processing device 1A may include a training data generation unit that generates training data for machine learning of an inference model that uses the generated numerical data to infer the subject's state, such as their health status and the presence or absence of disease, from the subject's biological signals. The training data generation unit may also generate training data using both the generated numerical data and the image data from which it was derived. By using the training data generated in this way, it becomes possible to generate an inference model that can perform inferences that take into account both the numerical and waveform characteristics of the biological data. Furthermore, for example, the information processing device 1A may include a learning unit that generates the above inference model by performing machine learning using the above training data.

[0093] Furthermore, for example, in the flowchart of S7, a training data generation process may be performed after S17, in which the training data generation unit generates training data. In addition, an inference model generation process may be performed after the training data generation process, in which the learning unit generates an inference model.

[0094] In other words, the training data generation method according to this exemplary embodiment includes: a waveform detection process that detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal; a digitization process that generates numerical data representing the time-series change of the biological signal in predetermined units corresponding to the biological signal, based on the detection results in the waveform detection process; and a training data generation process that uses the generated numerical data to generate training data used for machine learning of an inference model that infers the state of a subject from the numerical data of the subject's biological signal. This training data generation method has the effect of making it possible to generate training data that includes the numerical data of the biological signal as an explanatory variable, using image data showing the waveform of the biological signal.

[0095] Furthermore, the inference model generation method according to this exemplary embodiment includes: a waveform detection process that detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal; a digitization process that generates numerical data representing the time-series change of the biological signal in predetermined units corresponding to the biological signal, based on the detection results in the waveform detection process; and an inference model generation process that generates an inference model that infers the state of a subject from the numerical data of the subject's biological signal using machine learning with training data generated from the numerical data. This inference model generation method has the effect of making it possible to generate an inference model that includes the numerical data of the biological signal as an explanatory variable, using image data showing the waveform of the biological signal.

[0096] Furthermore, numerical data generally requires less data than image data. Therefore, by storing numerical data generated by the information processing device 1A instead of image data of the biological signal waveform, the storage capacity of the biological signal data can be reduced. In this case, the information processing device 1A may include an image generation unit that generates image data of the biological signal waveform from the numerical data of the biological signal. This makes it possible to provide image data of the biological signal waveform as needed while reducing the storage capacity of the biological signal data.

[0097] [Variation] The execution entities for each process described in the exemplary embodiments above are arbitrary and not limited to the examples given. For example, a system having the same functions as the information processing devices 1 and 1A can be constructed using multiple devices that can communicate with each other. Furthermore, the execution entities for each process shown in the flowchart of Figure 7, and the execution entities for each process of the training data generation method and inference model generation method described above, may be a single device (which can also be called a processor) or multiple devices (which can also be called processors).

[0098] [Examples of implementation using software] Some or all of the functions of the information processing devices 1,1A (hereinafter also referred to as "each of the above devices") may be implemented by hardware such as integrated circuits (IC chips) or by software.

[0099] In the latter case, each of the above devices is implemented, for example, by a computer that executes instructions for a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as Computer C) is shown in Figure 8. Figure 8 is a block diagram showing the hardware configuration of Computer C, which functions as each of the above devices.

[0100] Computer C comprises at least one processor C1 and at least one memory C2. Memory C2 stores a program (digitalization program) P that causes computer C to operate as each of the above-mentioned devices. In computer C, processor C1 reads program P from memory C2 and executes it, thereby realizing each of the above-mentioned functions of the devices.

[0101] For processor C1, for example, a CPU (Central Processing Unit), GPU (Graphic Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating Point Number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination thereof can be used. For memory C2, for example, flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof can be used.

[0102] Computer C may also be equipped with RAM (Random Access Memory) for loading program P at runtime and for temporarily storing various data. Furthermore, computer C may be equipped with communication interfaces for sending and receiving data with other devices. Additionally, computer C may be equipped with input / output interfaces for connecting input / output devices such as keyboards, mice, displays, and printers.

[0103] Furthermore, program P can be recorded on a non-temporary, tangible recording medium M that is readable by computer C. Such a recording medium M could be, for example, tape, disk, card, semiconductor memory, or programmable logic circuitry. Computer C can acquire program P via such a recording medium M. Program P can also be transmitted via a transmission medium. Such a transmission medium could be, for example, a communication network or broadcast waves. Computer C can also acquire program P via such a transmission medium.

[0104] Furthermore, each of the above functions of each of the above devices may be implemented by a single processor in a single computer, by multiple processors in a single computer working together, or by multiple processors in each of multiple computers working together. In addition, the programs for implementing each of the above functions in each of the above devices may be stored in a single memory in a single computer, distributed and stored in multiple memories in a single computer, or distributed and stored in multiple memories in each of multiple computers.

[0105] [Additional Notes] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0106] (Note A1) An information processing device comprising: a waveform detection means for detecting a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal; and a digitization means for generating numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection results of the waveform detection means.

[0107] (Appendix A2) The information processing apparatus according to Appendix A1, comprising: detection means for detecting a reference object drawn from the image data such that a predetermined numerical value expressed in a predetermined unit is of a predetermined size, and the digitization means for generating numerical data representing the time-series change of the biological signal in a predetermined unit based on the size of the detected reference object.

[0108] (Note A3) The information processing apparatus according to Appendix A1 or A2, comprising a baseline derivation means for deriving a baseline that passes through a plurality of points in the waveform where the biological signal is a predetermined reference value, and the digitization means generates numerical data representing the time-series change of the biological signal in predetermined units based on the distance from the pixel detected by the waveform detection means to the baseline.

[0109] (Note A4) An information processing apparatus according to any one of appendices A1 to A3, comprising preprocessing means for generating the aforementioned image data by applying preprocessing to the source image data that will be the basis of the aforementioned image data, to remove elements other than the waveform.

[0110] (Note B1) A digitization method comprising: a waveform detection process in which at least one processor detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal; and a digitization process that generates numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection results in the waveform detection process.

[0111] (Note B2) The digitization method according to Appendix B1, wherein the at least one processor includes a detection process for detecting a reference object drawn from the image data such that a predetermined numerical value expressed in a predetermined unit is of a predetermined size, and in the digitization process, the at least one processor generates numerical data representing the time-series change of the biological signal in a predetermined unit based on the size of the detected reference object.

[0112] (Note B3) The digitization method according to Appendix B1 or B2, wherein the at least one processor includes a baseline derivation process for deriving a baseline that passes through a plurality of points in the waveform where the biological signal is a predetermined reference value, and in the digitization process, the at least one processor generates numerical data representing the time-series change of the biological signal in predetermined units based on the distance from the pixel detected in the waveform detection process to the baseline.

[0113] (Note B4) The digitization method according to any one of appendices B1 to B3, wherein at least one processor includes a preprocessing step to generate the image data by removing elements other than the waveform from the source image data that will be the basis of the image data.

[0114] (Note C1) A digitization program that causes a computer to function as a waveform detection means for detecting a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization means for generating numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection results of the waveform detection means.

[0115] (Note C2) The digitization program described in Appendix C1, wherein the computer functions as a detection means for detecting a reference object drawn from the image data such that a predetermined numerical value expressed in a predetermined unit is of a predetermined size, and the digitization means generates numerical data representing the time-series change of the biological signal in a predetermined unit based on the size of the detected reference object.

[0116] (Note C3) The computer functions as a baseline derivation means for deriving a baseline that passes through a plurality of points in the waveform where the biological signal is a predetermined reference value, and the digitization means generates numerical data representing the time-series change of the biological signal in predetermined units based on the distance from the pixel detected by the waveform detection means to the baseline, as described in Appendix C1 or C2.

[0117] (Note C4) A digitization program according to any one of appendices C1 to C3, wherein the computer functions as a preprocessing means for generating image data by applying preprocessing to the source image data that will be the basis of the image data, to remove elements other than the waveform.

[0118] (Note D1) An information processing device comprising at least one processor, wherein the at least one processor performs a waveform detection process to detect a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization process to generate numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection result of the waveform detection process.

[0119] The information processing device may also include memory. Furthermore, the memory may store a numerical program for causing at least one processor to execute each of the aforementioned processes.

[0120] (Note D2) The information processing apparatus according to Appendix D1, wherein at least one processor performs a detection process to detect a reference object drawn from the image data such that a predetermined numerical value expressed in a predetermined unit is of a predetermined size, and in the digitization process, the at least one processor generates numerical data representing the time-series change of the biological signal in a predetermined unit based on the size of the detected reference object.

[0121] (Note D3) The information processing apparatus according to Appendix D1 or D2, wherein at least one processor performs a baseline derivation process to derive a baseline passing through a plurality of points in the waveform where the biological signal is a predetermined reference value, and in the digitization process, the at least one processor generates numerical data representing the time-series change of the biological signal in predetermined units based on the distance from the pixel detected in the waveform detection process to the baseline.

[0122] (Note D4) The information processing apparatus according to any one of appendices D1 to D3, wherein at least one processor performs preprocessing on the source image data that will be the basis of the image data, removing elements other than the waveform to generate the image data.

[0123] (Note E) A non-temporary recording medium that records a digitization program that causes a computer to perform a waveform detection process that detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and a digitization process that generates numerical data representing the time-series change of the biological signal in predetermined units according to the biological signal, based on the detection results of the waveform detection process. [Explanation of Symbols]

[0124] 1. Information Processing Device 101 Waveform detection unit (waveform detection means) 102 Numericalization Unit (Numericalization Means) 1A Information Processing Device 101A Waveform detection unit (waveform detection means / detection means) 102A Numericalization Unit (Numericalization Means) 104A Pre-processing unit (pre-processing means) 105A Baseline derivation section (baseline derivation means)

Claims

1. A waveform detection means for detecting a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, An information processing apparatus comprising: a digitization means that generates numerical data representing the time-series change of the biological signal in a predetermined unit corresponding to the biological signal, based on the detection result of the waveform detection means.

2. The system includes a detection means for detecting a reference object drawn from the image data such that a predetermined numerical value expressed in a predetermined unit becomes a predetermined size, The information processing apparatus according to claim 1, wherein the digitization means generates numerical data representing the time-series change of the biological signal in predetermined units based on the size of the detected reference object.

3. The waveform includes a baseline derivation means for deriving a baseline that passes through a plurality of points where the biological signal reaches a predetermined reference value, The information processing apparatus according to claim 1 or 2, wherein the digitization means generates numerical data representing the time-series change of the biological signal in predetermined units based on the distance from the pixel detected by the waveform detection means to the baseline.

4. The information processing apparatus according to claim 1 or 2, further comprising preprocessing means for generating the image data by applying preprocessing to the source image data that is the basis of the image data, to remove elements other than the waveform.

5. At least one processor, Waveform detection processing that detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, A digitization process that generates numerical data representing the time-series change of the biological signal in a predetermined unit corresponding to the biological signal, based on the detection results in the waveform detection process, An inference model generation method that performs an inference model generation process to generate an inference model that infers the state of a subject from numerical data of the subject's biological signals by machine learning using training data generated from the aforementioned numerical data.

6. At least one processor, Waveform detection processing that detects a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, A digitization method that performs a digitization process to generate numerical data representing the time-series changes of the biological signal in predetermined units according to the biological signal, based on the detection results in the waveform detection process.

7. Computers, Waveform detection means for detecting a group of pixels corresponding to a portion of a waveform from image data showing the waveform of a biological signal, and A digitization program that functions as a digitization means for generating numerical data representing the time-series changes of a biological signal in predetermined units corresponding to the biological signal, based on the detection results of the waveform detection means.

Images