Anomaly detection device

Abstract

To provide a technique for accurately detecting an abnormality of a blood flow state.SOLUTION: An abnormality detection device includes: an acquisition section for time-sequentially acquiring blood flow information indicating the flow rate of blood in a biological tissue; a prediction section for obtaining a prediction range of the blood flow information after a predetermined period based on the blood flow information within the predetermined period in the time-sequentially acquired blood flow information; an abnormality detection section for detecting occurrence of an abnormality when the blood flow information after the predetermined period is deviated from the prediction range; and a warning section for outputting warning information when the occurrence of the abnormality is detected.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an abnormality detection device for detecting abnormalities in blood flow conditions in living tissue. [Background technology]

[0002] When some tissues become cancerous, such as in cases of tongue or pharyngeal cancer, these tissues are removed. To replace the removed tissue, skin or subcutaneous tissue with a blood flow (hereinafter also referred to as a skin flap) from another area, such as the abdomen or back, is sometimes removed and transplanted into the defect. The skin flap is excised with its blood vessels (arteries and veins) attached. When transplanted into the defect, blood flow is restored by suturing the blood vessels of the flap to those of healthy tissue near the defect. If a blood clot forms in the transplanted tissue, causing a lack of blood flow, the transplanted tissue may die. For this reason, medical staff must frequently check day and night to ensure that blood flow is not impaired after transplantation, which places a significant burden on them.

[0003] Patent Document 1 describes a device that detects blood flow disorders in biological tissue by attaching a biological tissue sensor sheet and using blood flow information obtained by this sensor sheet. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Re-tabled publication No. 2017-026393 Summary of the Invention [Problem to be solved by the invention]

[0005] Because the blood flow condition of a skin flap is constantly fluctuating, even if it temporarily drops, if it continues to increase thereafter, the flap may survive without any problems. On the other hand, if the blood flow condition gradually drops, by the time it is determined that action is necessary, it may be too late and the flap may not survive. For this reason, even if blood flow is checked frequently, there is a problem in that abnormalities in blood flow cannot be accurately detected by simply judging the blood flow condition at the time of checking.

[0006] Therefore, an object of the present invention is to provide a technique for accurately detecting abnormalities in blood flow conditions. [Means for solving the problem]

[0007] In order to solve the above problems, the abnormality detection device of the present invention comprises: an acquisition unit that acquires blood flow information indicating a blood flow rate in a biological tissue in a time series; a prediction unit that calculates a predicted value of the blood flow information after a predetermined period based on the blood flow information within the predetermined period among the blood flow information acquired in time series; an abnormality detection unit that detects the occurrence of an abnormality when the blood flow information after the predetermined period deviates from the predicted value; an alarm unit that outputs warning information when the occurrence of the abnormality is detected; Equipped with.

[0008] In the abnormality detection device, the prediction unit may perform a time series analysis of the blood flow information within the predetermined period and obtain a confidence interval of a predetermined probability as a range of predicted values.

[0009] The abnormality detection device further includes: the acquiring unit acquiring the blood flow information including at least one of a systolic blood pressure, a mean blood pressure, a diastolic blood pressure, a pulse rate, and a respiratory rate in addition to the blood flow rate; The prediction unit may perform machine learning on the blood flow information within a predetermined period to obtain a mathematical model, and may obtain a predicted value of the blood flow information after the predetermined period based on the mathematical model.

[0010] The abnormality detection device includes: performing a stationary test on the blood flow information; If it is determined that there is stationary state, the blood flow information within the predetermined period is subjected to a time series analysis, and a confidence interval of a predetermined probability is calculated as a range of predicted values; If it is determined that there is no constancy, a predicted value of the blood flow information after the predetermined period may be calculated based on the mathematical model.

[0011] In order to solve the above problem, the abnormality detection program of the present invention comprises: Acquiring blood flow information indicating the flow rate of blood in biological tissue in a time series manner; calculating a predicted value of the blood flow information after a predetermined period based on the blood flow information within the predetermined period among the blood flow information acquired in time series; detecting an occurrence of an abnormality when the blood flow information after the predetermined period deviates from the predicted value; outputting warning information when the occurrence of the abnormality is detected; The control unit executes the above. [Effects of the Invention]

[0012] The present invention can provide a technique for detecting abnormalities in blood flow conditions with high accuracy. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is a diagram illustrating a schematic configuration of an anomaly detection system. [Figure 2] FIG. 2 is a diagram showing the installation state and configuration of a sensor. [Figure 3] FIG. 4 is a diagram showing a procedure of processing executed by a control unit of the abnormality detection device according to a program. [Figure 4] FIG. 10 is a diagram showing blood flow information acquired in time series and a confidence interval calculated from this blood flow information. [Figure 5] FIG. 1 is an explanatory diagram of time series analysis. [Figure 6] FIG. 10 is a diagram illustrating a configuration of an abnormality detection device according to a second embodiment. [Figure 7] FIG. 10 is a diagram showing blood flow information according to the second embodiment. [Figure 8]FIG. 4 is a diagram showing a procedure of processing executed by a control unit of the abnormality detection device according to a program. [Figure 9] FIG. 11 is a diagram showing a procedure of processing executed by a control unit of an abnormality detection device according to a third embodiment in accordance with a program. DETAILED DESCRIPTION OF THE INVENTION

[0014] First Embodiment An abnormality detection system 100 according to a mode for carrying out the present invention (hereinafter referred to as an embodiment) will be described below with reference to the drawings. The configuration of the following embodiment is an example, and the present invention is not limited to the configuration of the embodiment. FIG. 1 is a diagram showing a schematic configuration of the abnormality detection system 100, and FIG. 2 is a diagram showing the installation state and configuration of a sensor 20. The example of FIG. 1 shows an example in which a portion of tissue on the face of a patient 3 is excised, and a skin flap 31 separated from another part is transplanted into this missing part. Note that the part to be transplanted is not limited to the face, and may be another part. Furthermore, the living body is not limited to a human.

[0015] System Configuration The abnormality detection system 100 monitors the blood flow state of a biological tissue (skin flap) transplanted into a living body (e.g., a patient) after transplantation, and outputs an alarm if an abnormality is detected. The abnormality detection system 100 has a sensor 20 that detects blood flow information, and an abnormality detection device 10 that detects an abnormality based on the blood flow information.

[0016] Sensor The sensor 20 comprises a pad 21 that is attached to the skin flap 31 to be measured, and a sensor main body 22 that is electrically connected to the pad 21 and the abnormality detection device 10. As shown in Fig. 2, the pad 21 is attached to the surface of the skin flap 31 and has a light-emitting unit 211 that emits an optical signal into the flap and a light-receiving unit 212 that receives the optical signal reflected within the flap. The light-receiving unit 212 converts the received optical signal into an electrical signal (detection signal) and transmits it to the sensor main body 22.

[0017] The skin flap 31 is excised with the artery 32 and vein 33 included, and when transplanted into the defect, the blood flow is resumed by suturing the artery and vein of healthy tissue near the defect to the artery 32 and vein 33 on the side of the skin flap 31. If a thrombus forms in the artery 32, the skin flap 31 becomes ischemic, and if a thrombus forms in the vein 33, the skin flap 31 becomes congested, resulting in abnormal blood flow. If this condition continues, oxygen will not be distributed to the cells that make up the skin flap 31, and the skin flap 31 may become necrotic. For this reason, if an abnormality occurs, the state of blood flow is detected by the sensor 20 so that measures can be taken promptly, such as re-anastomosis of the artery 32 and vein 33 to ensure blood flow.

[0018] The optical signal emitted from the light-emitting unit 211 has a wavelength that is easily absorbed by hemoglobin in the blood, so when the amount of blood in the flap changes, the amount of absorbed optical signal also changes, and the movement of blood is reflected in the detection signal. As a result, the sensor main body 22 detects the blood flow rate (hereinafter also referred to as blood flow rate) from the phase difference between the detection signal detected by the light-receiving unit 212 and the signal emitted by the light-emitting unit 211. The sensor main body 22 inputs the detected blood flow rate value to the abnormality detection device 10 as blood flow information.

[0019] <Anomaly detection device> The abnormality detection device 10 is an information processing device (computer) that performs processing to output an alarm when it is determined that an abnormality has occurred in the blood flow state of the skin flap 31 based on blood flow information detected by the sensor 20. The abnormality detection device 10 includes a control unit 11, a storage unit 12, and an input / output unit 13. The control unit 11 comprehensively executes various types of calculation processing in the abnormality detection device 10. The control unit 11 includes a CPU (Central Processing Unit), a DSP (Digital Signal Processor), Alternatively, it is a processing means such as an FPGA (Field-Programmable Gate Array).

[0020] The storage unit 12 includes, for example, a main storage unit 121 and an auxiliary storage unit 122. The main storage unit 121 includes a main storage unit such as a RAM (Random Access Memory) or a ROM (Read Only Memory), and stores information to be processed by the control unit 11, for example. The main storage unit 121 may be formed integrally with the control unit 11.

[0021] The auxiliary storage unit 122 is configured from storage media such as volatile memory such as RAM, non-volatile memory such as ROM, erasable programmable ROM (EPROM), hard disk drive (HDD), removable media, etc. The removable media is an externally attachable and computer-readable recording medium such as a USB (Universal Serial Bus) memory or a memory card.

[0022] The auxiliary storage unit 122 can store an operating system (OS), various programs, various tables, various databases, user data, and the like, for executing the operations of the protection device 1.

[0023] The input / output unit 13 is, for example, an interface that inputs blood flow information from the sensor 20 and outputs information (control signals, etc.) to the sensor 20 or other devices. The input / output unit 13 also functions as a communication interface that inputs (receives) information from other devices and outputs (transmits) information to other devices. Furthermore, the input / output unit 13 may be a user interface that inputs operation information by the user using an operation button, a touch panel, or the like, and outputs (display, sound output, etc.) to the user using a display, a speaker, or the like.

[0024] The control unit 11 reads out a program stored in the auxiliary storage unit 122 into the working area of ​​the main storage unit 121, executes the program, and controls means such as the sensor 20 and the display via the input / output unit 13, thereby functioning as predetermined functional units, for example, a blood flow information acquisition unit 111, a prediction unit 112, an abnormality detection unit 113, and an output control unit 114. Note that these functional units are not limited to those realized based on a program (software), and some or all of them may be configured by hardware circuits such as a processor, an integrated circuit, and a logic circuit.

[0025] The blood flow information acquiring unit 111 acquires blood flow information indicating the blood flow rate in biological tissue in chronological order from the sensor 20 via the input / output unit 13. For example, the blood flow information acquiring unit 111 sequentially acquires blood flow rate values ​​measured at a predetermined interval from the sensor 20, and stores the blood flow information in the storage unit 12 together with information indicating the measurement time in an array according to the measurement time.

[0026] The prediction unit 112 calculates a predicted range of blood flow information after a predetermined period based on blood flow information within a predetermined period of time among blood flow information acquired in time series. For example, the prediction unit 112 calculates a predicted range of blood flow information to be acquired next based on blood flow information from the start of measurement to the present time. The prediction unit 112 of this embodiment calculates a confidence interval of a predetermined probability, for example, a 50% confidence interval or a 95% confidence interval, as a predicted range within which a predicted value can fall, by time series analysis.

[0027] The abnormality detection unit 113 detects the occurrence of an abnormality when the blood flow information after a predetermined period of time falls outside the predicted range. For example, the abnormality detection unit 113 determines that no abnormality has occurred if the blood flow information acquired next after the predetermined period of time is within the predicted range, and determines that an abnormality has occurred if the blood flow information acquired next falls outside the predicted range.

[0028] The output control unit 114 outputs the blood flow information acquired by the blood flow information acquisition unit 111 and warning information indicating that an abnormality has been detected by the abnormality detection unit 113 to an output means such as a display or a speaker, and causes display output or sound output. That is, the output control unit 114 of this embodiment is one form of an alarm unit.

[0029] <Operation of the abnormality detection device> 3 is a diagram showing the procedure of processing executed by the control unit 11 of the abnormality detection device 10 according to a program. The control unit 11 starts and repeatedly executes the processing of FIG. 3 when the power is turned on or when an instruction to start processing is received. In addition, in parallel with the processing of FIG. 3, the control unit 11 sequentially acquires blood flow information from the sensor 20 and stores it in the memory unit 12 as time-series data. This allows the control unit 11 to acquire past blood flow information from the memory unit 12 in time series.

[0030] In step S10, the control unit 11 acquires from the storage unit 12 blood flow information for a predetermined period, such as from the start of measurement to the present time.

[0031] In step S20, the control unit 11 calculates a prediction range for blood flow information after a predetermined period of time based on the blood flow information acquired in step S10. For example, the prediction unit 112 performs a time series analysis based on the blood flow information from the start of measurement to the present time, and calculates a 50% confidence interval to set the prediction range.

[0032] In step S30, the control unit 11 determines whether or not an abnormality has occurred based on the blood flow information after a predetermined period of time and the prediction range obtained in step S20. If the blood flow information acquired next after the predetermined period falls outside the predicted range, it is determined that an abnormality has occurred (positive determination), and the process proceeds to step S40.

[0033] In step S40, the control unit 11 outputs warning information indicating that an abnormality has occurred to a display, a speaker, or other device via the input / output unit 13. For example, the control unit 11 outputs a warning sound from the speaker and displays a warning message on the display. The control unit 11 also transmits the warning information to information terminals such as mobile phones and smartphones carried by medical staff. Furthermore, the warning information may be transmitted to other systems for notifying abnormalities, such as a nurse call control device.

[0034] On the other hand, if the determination in step S30 is negative, the control unit 11 ends the processing of FIG. 3 without outputting any warning information.

[0035] As a result, the abnormality detection device 10 constantly monitors the state of blood flow in the skin flap 31 and issues a warning only when an abnormality occurs.

[0036] <Anomaly prediction processing> Figure 4 shows blood flow information acquired over time and the confidence interval calculated from this blood flow information. In Figure 4, the horizontal axis represents the time elapsed since the start of measurement, and the vertical axis represents blood flow rate [ml / min], showing changes in blood flow rate.

[0037] The control unit 11 performs time-series analysis of this blood flow information to determine a confidence interval for a predetermined probability. While various models such as the AR model, MA model, and SARIMA model can be used for time-series analysis, the ARMA model is used in this embodiment. Figure 5 is a diagram illustrating the ARMA model.

[0038] For example, as shown in Figure 5, a model of time series data Y(t) made up of an autoregressive (AR) part and a moving average (MA) part is called an ARMA model. Yt=aYt-1+b It is defined as follows.

[0039] The composite function of the generalized AR(p) process and MA(q) process is expressed as ARMA(p,q). Figure 4 In the example, ARMA(3,5) was obtained.

[0040] As shown in FIG. 4, the control unit 11 defines a predetermined period A0 from the start of measurement to the time when blood flow information is acquired in step S10, and calculates confidence intervals A1 and A2 from the blood flow information during this predetermined period A0. In FIG. 4, symbol A1 indicates a 50% confidence interval, and symbol A2 indicates a 95% confidence interval. If normal blood flow continues after surgery, the newly acquired blood flow information will fall within the confidence intervals A1 and A2. However, if the blood flow rate drops abnormally due to a thrombus or other cause, the newly acquired blood flow information will fall outside the confidence intervals A1 and A2. Therefore, the control unit 11 can determine the occurrence of an abnormality based on whether the blood flow information falls outside the confidence interval A1 or A2 in step S30. Note that while FIG. 4 shows examples of a 50% confidence interval and a 95% confidence interval, the probability (confidence coefficient) of the confidence interval is not limited to these and may be other values. For example, a medical staff member may select an arbitrary confidence coefficient from a predetermined range (e.g., 50% to 95%) at the start of measurement. Furthermore, when calculating multiple confidence intervals with different confidence coefficients, the timing of the alert may be varied depending on the confidence coefficient of the confidence interval that has fallen outside the 95% confidence interval, such as outputting an alert immediately if the value falls outside the 95% confidence interval, or outputting an alert when the value falls outside the 50% confidence interval and the value remains outside the 95% confidence interval for a predetermined period of time (e.g., 3 minutes), or when the value falls outside the 50% confidence interval a predetermined number of times (e.g., 3 consecutive times, 5 times per unit time, etc.) after repeated determination.

[0041] Effects of the embodiment As described above, according to this embodiment, the postoperative blood flow information is analyzed in a time series manner to determine the predicted range (confidence interval) of the blood flow information, and the occurrence of an abnormality is detected based on whether the subsequently acquired blood flow information falls outside the predicted range. In this way, the abnormality detection device of this embodiment determines the predicted range that serves as the basis for judgment based on the change in blood flow information over a predetermined period, so the occurrence of an abnormality can be detected with high accuracy. This allows the blood flow status to be monitored automatically, reducing the burden on medical staff.

[0042] Second Embodiment FIG. 6 is a diagram showing the configuration of an abnormality detection device 10A according to a second embodiment, and FIG. 7 is a diagram showing blood flow information according to the second embodiment. In the first embodiment described above, the predicted range of blood flow information was calculated using the ARMA model, but this is not limiting. In this embodiment, an example is shown in which the predicted range of blood flow information is calculated using AI. Note that other configurations are substantially the same as those in the first embodiment described above, and therefore, the same elements are designated by the same reference numerals, and repeated explanations are omitted.

[0043] As shown in Fig. 6, the abnormality detection device 10A is connected to a blood pressure sensor 24, a pulse sensor 25, and a respiratory rate sensor 26 via the input / output unit 13. The abnormality detection device 10A acquires the systolic blood pressure, mean blood pressure, and diastolic blood pressure of the patient (living body) to which the skin flap 31 has been transplanted from the blood pressure sensor 24. The abnormality detection device 10A also acquires the pulse rate and respiratory rate of the patient (living body) to which the skin flap 31 has been transplanted from the pulse sensor 25 and the respiratory rate sensor 26. Note that these sensors 24 to 26 are well-known sensors, and therefore detailed description thereof will be omitted.

[0044] In this way, the control unit 11 of the abnormality detection device 10A, as the blood flow information acquisition unit 111, acquires blood flow volume, systolic blood pressure, mean blood pressure, diastolic blood pressure, pulse rate, and respiratory rate from the sensors 20, 24 to 26, and sets these as blood flow information. In Fig. 7, symbol D1 indicates the systolic blood pressure, symbol D2 indicates the mean blood pressure, symbol D3 indicates the diastolic blood pressure, symbol D4 indicates the pulse rate, symbol D5 indicates the respiratory rate, and symbol D6 indicates the blood flow volume of the skin flap.

[0045] Furthermore, the control unit 11 of this embodiment, as the prediction unit 112, creates a mathematical model by machine learning from the blood flow information and determines the predicted range of the blood flow information based on the mathematical model. That is, the prediction unit 112 determines the predicted range of the blood flow information by AI.

[0046] 8 is a diagram showing the procedure of processing executed by the control unit 11 of the abnormality detection device 10A according to a program. The control unit 11 starts the processing of FIG. 8 when the power is turned on or when an instruction to start processing is received, and executes it repeatedly. In addition, in parallel with the processing of FIG. 8, the control unit 11 sequentially acquires blood flow information from the sensors 20, 24 to 26 and stores it in the memory unit 12 as time-series data. This allows the control unit 11 to acquire past blood flow information from the memory unit 12 in time series.

[0047] In step S110, the control unit 11 acquires blood flow information from the storage unit 12, such as information from the start of measurement to the present time.

[0048] In step S120, control unit 11 performs machine learning using some of the blood flow information from the start of measurement to the present as training data to create a mathematical model. For example, control unit 11 sets the period from the start of measurement to 80% of the period A3 (FIG. 7) from the start of measurement to the present as a first period, and the period from after first period A4 to the present as a second period A5, and sets the blood information of first period A3 as training data.

[0049] In step S130, the control unit 11 performs a calculation based on the mathematical model created in step S120. This allows predicting blood flow information, such as blood flow rate.

[0050] In step S140, the control unit 11 compares the blood flow information (actual measurement value) for the second period A4 with the blood flow rate (predicted value) predicted in step S130, and determines whether an abnormality has occurred. For example, if the blood flow rate (actual measurement value) for the second period A4 deviates from the predicted value, the control unit 11 determines that an abnormality has occurred (positive determination), and proceeds to step S150.

[0051] In step S150, the control unit 11 outputs warning information indicating that an abnormality has occurred to a display, a speaker, or other device via the input / output unit 13. On the other hand, if the determination in step S140 is negative, the control unit 11 ends the processing in FIG. 8 without outputting any warning information.

[0052] In this way, according to this embodiment, blood flow information is predicted by machine learning, and the occurrence of an abnormality is detected based on whether the subsequently acquired blood flow information deviates from the predicted value. In this way, the abnormality detection device of this embodiment obtains the prediction range that serves as the basis for judgment based on the change in blood flow information over a predetermined period (first period), so the occurrence of an abnormality can be detected with high accuracy.

[0053] Third Embodiment 9 is a diagram showing the procedure of processing executed by the control unit 11 of the abnormality detection device 10A according to the third embodiment in accordance with a program. In the first embodiment, blood flow information was predicted by time series analysis, and in the second embodiment, blood flow information was predicted by AI. However, in this embodiment, an example is shown in which the prediction method is switched depending on the constancy of the blood flow information. Note that other configurations are substantially the same as those of the first or second embodiment, and therefore, the same elements are designated by the same reference numerals, and repeated explanations are omitted.

[0054] As in the second embodiment, the abnormality detection device 10A can acquire blood flow volume, systolic blood pressure, mean blood pressure, diastolic blood pressure, pulse rate, and respiratory rate from the sensors 20, 24 to 26 as blood flow information.

[0055] When comparing predictions by time series analysis with predictions by AI, if the change in blood flow rate is stationary, the accuracy of the prediction by time series analysis is high, and if the change in blood flow rate is not stationary (low stationarity), the accuracy of the prediction by AI tends to be high. For this reason, the abnormality detection device 10A of this embodiment tests the acquired blood flow rate for stationarity, and if stationarity is found, makes a prediction by time series analysis, and if stationarity is not found, makes a prediction by AI.

[0056] The control unit 11 of the abnormality detection device 10A according to this embodiment starts and repeatedly executes the process of Fig. 9 when the power is turned on or when an instruction to start the process is received. In addition, in parallel with the process of Fig. 9, the control unit 11 sequentially acquires blood flow information from the sensors 20, 24 to 26 and stores it as time-series data in the storage unit 12. This allows the control unit 11 to acquire past blood flow information from the storage unit 12 in time series.

[0057] In step S3, the control unit 11 checks the test result, and if the test has not yet been performed, proceeds to step S5, if the test has already been performed to determine that the signal is stationary, proceeds to step S10, and if the test has already been performed to determine that the signal is not stationary, proceeds to step S110.

[0058] When the process proceeds to step S5, the control unit 11 determines whether or not a predetermined time (e.g., 30 minutes) has elapsed since the start of measurement, during which testing is possible. If the determination in step S5 is negative, the process in FIG. 9 is terminated, and if the determination is positive, the process proceeds to step S7.

[0059] In step S7, the control unit 11 performs a constancy test on the blood flow rate of the blood flow information acquired from the start of measurement to the present time. The test method is not particularly limited, but in this embodiment In this state, the control unit 11 performs an extended Dickey-Fuller test.

[0060] In step S9, the control unit 11 determines whether or not there is stationarity based on the test result of step S7. For example, the control unit 11 determines that there is stationarity if the test result of step S7 is greater than a predetermined value (a rejection value), and determines that there is no stationarity if the test result is equal to or less than the predetermined value (a rejection value).

[0061] If the determination in step S9 is affirmative, the control unit 11 proceeds to step S10 and performs the same processing as in the first embodiment described above. On the other hand, if the determination in step S9 is negative, the control unit 11 proceeds to step S110 and performs the same processing as in the second embodiment described above.

[0062] As described above, according to this embodiment, by selecting a prediction method depending on the constancy of the blood flow and obtaining a predicted value using an appropriate prediction method, abnormalities can be determined with even greater accuracy. [Explanation of symbols]

[0063] 1 Protective device 3 patients 10,10A Abnormality Detection Device 11 Control section 12 Storage section 13 Input / output section 20,24~26 Sensor 21 Pad 22 Sensor body 31 Skin flap 32 arteries 33 Veins 100 Anomaly Detection System 111 Blood flow information acquisition unit 112 Prediction Department 113 Abnormality detection unit 114 Output control section 121 Main memory 122 Auxiliary storage 211 Light-emitting part 212 Light receiving part

Claims

1. an acquisition unit that acquires blood flow information in a time series, including at least one of a systolic blood pressure, a mean blood pressure, a diastolic blood pressure, a pulse rate, and a respiratory rate in addition to a blood flow rate in a biological tissue; a prediction unit that performs machine learning on the blood flow information acquired in time series as training data to obtain a mathematical model, and that obtains a prediction range of the blood flow information after a predetermined period from the start of measurement of the blood flow information based on the mathematical model; an abnormality detection unit that detects the occurrence of an abnormality when the blood flow information after the predetermined period of time falls outside the prediction range; an alarm unit that outputs warning information when the occurrence of the abnormality is detected; An abnormality detection device comprising:

2. The prediction unit performing a stationary test on the blood flow information; When it is determined that there is stationary state, the blood flow information within the predetermined period is subjected to a time series analysis, and a confidence interval of a predetermined probability is obtained as a prediction range of the blood flow information; if it is determined that there is no steadiness, a prediction range of the blood flow information after the predetermined period is calculated based on the mathematical model; The abnormality detection device according to claim 1 .

3. 3. The abnormality detection device according to claim 1, wherein the biological tissue is a skin flap transplanted into a living body.

4. Acquiring blood flow information in a time series, including at least one of systolic blood pressure, mean blood pressure, diastolic blood pressure, pulse rate, and respiratory rate in addition to blood flow rate in a biological tissue; obtaining a mathematical model by machine learning the blood flow information acquired in time series as training data, and obtaining a prediction range of the blood flow information after a predetermined period from the start of measurement of the blood flow information based on the mathematical model; detecting an occurrence of an abnormality when the blood flow information after the predetermined period falls outside the prediction range; outputting warning information when the occurrence of the abnormality is detected; An abnormality detection program for causing the control unit to execute the above.

5. The control unit: performing a stationary test on the blood flow information; When it is determined that there is stationary state, the blood flow information within the predetermined period is subjected to a time series analysis, and a confidence interval of a predetermined probability is obtained as a prediction range of the blood flow information; if it is determined that there is no steadiness, a prediction range of the blood flow information after the predetermined period is calculated based on the mathematical model; The abnormality detection program according to claim 4 .

6. An abnormality detection program as described in claim 4 or 5, wherein the biological tissue is a skin flap transplanted into a living organism.

Images