Recovery period activity identification method and device based on near-infrared image data

Abstract

The invention discloses a recovery period activity identification method and device based on near-infrared image data. The method comprises the following steps: acquiring the near-infrared image data; inputting the acquired near-infrared image data into the trained recovery period activity identification model, and outputting a recovery period activity identification result; wherein the training of the recovery period activity identification model comprises the following steps: acquiring near-infrared image sample data and forming a plurality of oscillograms according to the near-infrared image sample data; determining peaks on each oscillogram according to a hill climbing method; judging whether each peak on each oscillogram is a true peak one by one; when the time of the forward T gravity center of any true peak is greater than or equal to a first threshold value and the peak value appearance time of the current true peak is greater than a second threshold value, judging that the current waveform has a recovery period active characteristic; and recording the waveform with the recovery period active characteristic and outputting a recovery period active identification result.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and device for recognizing activity during a recovery period based on near-infrared imaging data. Background Art

[0002] Mental illness is one of the brain function diseases. With the rapid development of brain science and brain science industry, researchers from universities, hospitals, enterprises and other fields have carried out a lot of research on solving the objective, quantitative auxiliary diagnosis and identification of mental illness. These studies are largely focused on exploring and finding the neural biomarkers and imaging features of each disease based on its neural activity characteristics.

[0003] For schizophrenia, in the past 20 years, a lot of research has been done on its neural activity characteristics and neural biomarkers at home and abroad, especially Japan has conducted a series of studies on schizophrenia using near-infrared brain imaging technology. In the course of the study, researchers found that when using the fNIRS-VFT paradigm to measure the prefrontal cerebral blood flow activity pattern of schizophrenia patients, there is a typical imaging feature for more patients with mental retardation, that is, after the task is completed, the cerebral blood flow shows another rising activity feature, which rarely appears in the same task examination of normal people, depression, bipolar and other patients. However, after this feature was discovered, due to various difficulties, such as: data sample size issues, exclusion of confounding features caused by other diseases, etc., it has been difficult to overcome its automatic feature recognition algorithm. Summary of the invention

[0004] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a method and device for identifying recovery period activity based on near-infrared imaging data. By analyzing the processed near-infrared waveform, it is judged whether the waveform is recovery period activity according to the characteristics of the waveform and combined with near-infrared imaging theory, thereby directly telling whether the near-infrared blood oxygen image has the characteristics of recovery period activity, thereby providing a potential feasible method for identifying the brain functional activity characteristics of schizophrenia symptoms in patients in clinical related departments.

[0005] According to one aspect of the present invention, a method for recognizing activity during recovery period based on near infrared image data is provided, comprising: Acquire near infrared imaging data; The acquired near-infrared image data is input into the trained recovery period activity recognition model, and the recovery period activity recognition result is output; wherein the training of the recovery period activity recognition model includes: Acquire near infrared image sample data and form a plurality of waveform graphs according to the near infrared image sample data; Determine the peak on each waveform graph using the hill climbing method; Determine whether each peak on each waveform is a true peak one by one; When the time of the positive T gravity center of any true peak is greater than or equal to the first threshold and the peak occurrence time of the current true peak is greater than the second threshold, it is determined that the current waveform has the active feature of the recovery period; Record the waveform with recovery period activity characteristics and output the recovery period activity recognition result.

[0006] As a further technical solution, the peak on each waveform is determined according to the hill climbing method, including: S1, the peak is determined by adjacent valleys, where the two valleys are the start valley and the end valley of the peak; S2, if the current peak meets the peak condition, update the start valley and the end valley, and return to step S1; S3, if the current peak does not meet the peak condition, update the end valley forward and return to step S1.

[0007] As a further technical solution, whether each peak on each waveform graph is a true peak is determined one by one, including: when the height of the end valley of the current peak is lower than the first preset proportion of the waveform maximum value, and the ratio of the peak height to the amplitude is greater than the second preset proportion, and the peak value is greater than the third threshold, the current peak is determined to be a true peak.

[0008] As a further technical solution, when no true peak is found after judging each peak on the waveform diagram one by one, the third threshold is lowered and the true peak judgment process is repeated.

[0009] As a further technical solution, whether each peak on each waveform is a true peak is determined one by one, including: when the current peak is flat and reverse and the area of ​​the peak is greater than a fourth threshold, determining that the current peak is a true peak.

[0010] As a further technical solution, after judging whether each peak on each waveform diagram is a true peak one by one, the method further includes: correcting each true peak judged one by one.

[0011] According to one aspect of the present invention, a recovery period activity recognition device based on near infrared image data is provided, comprising: The first main module is used to obtain near-infrared image data; The second main module is used to input the acquired near-infrared image data into the trained recovery period activity recognition model and output the recovery period activity recognition result; wherein the training of the recovery period activity recognition model includes: Acquire near infrared image sample data and form a plurality of waveform graphs according to the near infrared image sample data; Determine the peak on each waveform graph using the hill climbing method; Determine whether each peak on each waveform is a true peak one by one; When the time of the positive T gravity center of any true peak is greater than or equal to the first threshold and the peak occurrence time of the current true peak is greater than the second threshold, it is determined that the current waveform has the active feature of the recovery period; Record the waveform with recovery period activity characteristics and output the recovery period activity recognition result.

[0012] As a further technical solution, the peak on each waveform is determined according to the hill climbing method, including: S1, the peak is determined by adjacent valleys, where the two valleys are the start valley and the end valley of the peak; S2, if the current peak meets the peak condition, update the start valley and the end valley, and return to step S1; S3, if the current peak does not meet the peak condition, update the end valley forward and return to step S1.

[0013] As a further technical solution, whether each peak on each waveform graph is a true peak is determined one by one, including: when the height of the end valley of the current peak is lower than the first preset proportion of the waveform maximum value, and the ratio of the peak height to the amplitude is greater than the second preset proportion, and the peak value is greater than the third threshold, the current peak is determined to be a true peak; or when the current peak is flat and reverse and the area of ​​the peak is greater than the fourth threshold, the current peak is determined to be a true peak.

[0014] According to one aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the recovery period activity recognition method based on near-infrared image data are implemented.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention analyzes the patient's near-infrared blood oxygen imaging data, determines the characteristics of the blood oxygen imaging data, and detects whether it has active characteristics in the recovery period. This can assist in identifying whether the patient's brain function activities show active recovery period activities, and further assist doctors in objectively and scientifically determining whether the patient has symptoms of schizophrenia. This provides a possibility for resolving the traditional error-prone approach of clinical departments that have relied on behavioral observations or doctors' intuition to determine whether patients have symptoms of schizophrenia and then conduct related treatments, or the inability to conduct effective treatments early due to reliance on behavioral observations and doctors' experience. This has great potential value and significance for the clinical work of clinical departments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic flow chart of a method for recognizing activity during recovery period based on near-infrared imaging data provided by an embodiment of the present invention.

[0017] Figure 2 A schematic diagram of a hill climbing method flow chart provided in an embodiment of the present invention.

[0018] Figure 3 This is a near-infrared hemoglobin concentration change curve provided by an embodiment of the present invention.

[0019] Figure 4 It is a waveform diagram for correction provided by an embodiment of the present invention.

[0020] Figure 5 It is a schematic diagram comparing the original near-infrared waveform and the recognized waveform provided by an embodiment of the present invention.

[0021] Figure 6 A schematic diagram of the structure of a recovery period activity recognition device based on near-infrared image data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] It should be noted that: Peak height: The height of a peak minus the height of the nearest adjacent valley equals the height of the peak.

[0023] Amplitude: The maximum value minus the minimum value of a waveform.

[0024] Peak value: The Y-axis value corresponding to the peak.

[0025] Low and flat: The maximum value of the waveform is less than a certain value, where "a certain value" can be set according to actual needs.

[0026] Reverse: Most of the points in the waveform are negative values, where the range or proportion of "most of the points" can be set according to actual needs.

[0027] Peak area: Unless otherwise specified, the areas referred to in the present invention are all positive areas.

[0028] Mental illnesses have the following characteristics: (1) Complex causes, usually caused by genes, heredity, fetal pregnancy conditions, family growth environment, life stress or special events; (2) Diverse symptoms, multiple symptoms such as physiological, behavioral, and emotional symptoms coexist, and there are serious comorbidity and mutual transformation problems between different mental illnesses; (3) Difficult to cure. Various mental illnesses are particularly difficult to treat after reaching clinical moderate or severe levels, and the course of the disease is extremely long. A large number of diseases, such as major depression and schizophrenia, often have lifelong illness problems. These characteristics of mental illnesses make the research on their specific neuropathological mechanisms, clinical auxiliary diagnosis and treatment very difficult and lengthy. At the same time, it is precisely because of this that the basic and clinical application research on mental illnesses has attracted many researchers at home and abroad to carry out various studies tirelessly for nearly a hundred years, and has achieved clinical diagnostic standards for diseases based on behavior and symptoms, such as: DSM mental disorder diagnostic standards and methods, various authoritative behavioral detection scales with wide clinical application, and other results; and has achieved a variety of treatment methods such as drugs and physical therapy, which are currently also the common theories and methods for the diagnosis, treatment and evaluation of clinical mental illnesses. However, behavioral-based diagnostic criteria are subjective and difficult to quantify, which brings a lot of difficulties and ambiguities to clinical diagnosis and treatment, and ultimately affects patients' disease treatment and rehabilitation.

[0029] With the rapid development of brain science and brain science industry, a lot of research has been done on the neural biomarkers and imaging features of various diseases based on their neural activity characteristics. Among them, some researchers found that when the fNIRS-VFT paradigm was used to measure the prefrontal cerebral blood flow activity pattern of schizophrenia patients, there was a typical imaging feature for more schizophrenia patients, but this feature was difficult to automatically identify. To this end, the present invention studies and develops an algorithm for the automatic identification of this feature based on large sample normative data.

[0030] The present invention analyzes the processed near-infrared waveform, and judges whether the waveform is active in the recovery period according to the characteristics of the waveform and the near-infrared imaging theory. The method of the present invention can directly tell whether the near-infrared blood oxygen image has the characteristics of active recovery period, and thus provide a potential feasible method for identifying the brain function activity characteristics of patients with schizophrenia symptoms in clinical related departments.

[0031] The following will clearly and completely describe the technical solutions of various embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0032] In the description of the present invention, it should be understood that the features defined as "first threshold", "second threshold", "first preset ratio" and "second preset ratio" are only for illustrative purposes and are not intended to limit the thresholds themselves or the order of the thresholds. In the description of the present invention, "multiple" and "several" mean two or more, unless otherwise clearly and specifically defined.

[0033] The embodiment of the present invention provides a method for recognizing activity during recovery period based on near infrared image data. Figure 1 As shown, including: Step 1: Acquire near-infrared image data.

[0034] In the embodiment of the present invention, near infrared equipment is used to collect VFT paradigm data.

[0035] After data collection, the collected data is processed, including raw light intensity to OD (optical density), OD to concentration, moving average and rotation and translation. This processing flow converts the change of light intensity into the concentration change of oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (HbR) during brain neural activity. The moving average is used to remove motion artifacts, heart rate, breathing and other noise. Finally, the problem of linear drift of the signal is solved by rotation and translation.

[0036] After data processing, a waveform is drawn according to the brain region division. In the embodiment of the present invention, the brain region is divided into the frontal pole, the left temporal lobe and the right temporal lobe. These three brain regions can well reflect the brain activity characteristics of functional brain diseases.

[0037] Step 2: Input the acquired near-infrared image data into the trained recovery period activity recognition model, and output the recovery period activity recognition result.

[0038] After obtaining the waveform diagram of each brain region, each waveform is sequentially input into the trained recovery period activity recognition model, and each waveform is judged by the trained recovery period activity recognition model to determine whether the waveform has the recovery period activity feature.

[0039] In the embodiment of the present invention, the training of the recovery period activity recognition model includes: Step 2.1, obtaining near-infrared image sample data and forming a plurality of waveform graphs according to the near-infrared image sample data.

[0040] Specifically, the collected data is divided into a training set and a test set, for example, there are 1203 cases in the training set and 3291 cases in the test set. The main purpose of the training set is to study the data characteristics, so as to make accurate predictions on new data, and to help select appropriate parameters to improve the accuracy of the model algorithm. The main purpose of the test set is to test the accuracy of the algorithm, which helps to determine whether the performance of the algorithm meets the needs of practical applications. Since the test set is not touched during the model training and parameter adjustment process, its evaluation results can provide accurate information about the actual performance of the algorithm.

[0041] The embodiment of the present invention first studies and organizes the training set, integrates the characteristics of the active recovery period in the training set data, and sets corresponding conditions according to the characteristics of the active recovery period, so as to determine whether the waveform has the characteristics of the active recovery period.

[0042] Step 2.2, determine the peak on each waveform graph according to the hill climbing method.

[0043] according to Figure 2 The process of hill climbing shown in Figure 3 The peak is found by using the waveform of the near-infrared hemoglobin concentration change shown in the figure. The specific process is as follows: S1, the peak is determined by adjacent valleys, where the two valleys are the start valley and the end valley of the peak; S2, if the current peak meets the peak condition, update the start valley and the end valley, and return to step S1; S3, if the current peak does not meet the peak condition, update the end valley forward and return to step S1.

[0044] Combination Figure 3 From the waveform diagram, we can see that the troughs of the waveform are 1, 2, 3, 4, and 5. Start step S1, determine valley 1 and valley 2 as the starting valley and the ending valley respectively, and form a peak. Judge the peak as a false peak, update the ending valley, and re-judge the ending valley as valley 3. If the valley formed by valley 1 and valley 3 is a true peak, update valley 3 and valley 4 as the starting valley and the ending valley respectively, and continue to judge in the same way.

[0045] Step 2.3, determine whether each peak on each waveform graph is a true peak one by one.

[0046] As a preferred embodiment, the process of determining whether each peak on each waveform graph is a true peak one by one may be as follows: when the height of the end valley of the current peak is lower than the first preset proportion of the waveform maximum value, and the ratio of the peak height to the amplitude is greater than the second preset proportion, and the peak value is greater than the third threshold, the current peak is determined to be a true peak.

[0047] In this embodiment, if a peak is a true peak, it needs to meet two conditions: the end valley must drop completely, and the height of the peak must be high enough.

[0048] It should be noted that the end valley must drop completely, which can be regarded as: in order to ensure the integrity of the peak and not be affected by fluctuations, the height of the end valley must be lower than a certain ratio (decline_ratio, ranging from 0.2-0.5) of the maximum value of the waveform. This ratio is the first preset ratio mentioned above and can be set according to actual needs.

[0049] The peak height must be high enough to be considered: when the peak height / amplitude>Ratio (a second preset ratio, ranging from 0.1 to 0.5) and the peak value>Th (a third threshold, ranging from 0.1 to 0.2 mmol / mm), the peak is considered to be high enough.

[0050] As a preferred embodiment, the process of determining whether each peak on each waveform is a true peak may be as follows: if the peak is flat, reversed, and the area of ​​the peak is greater than Th_LF_SIn (the fourth threshold, ranging from 0.8-1), the peak is determined to be a true peak.

[0051] Furthermore, when this step determines that the waveform has no true peak, the threshold Th is lowered and the judgment in step 2.3 is performed again to prevent the signal from missing a low-level situation.

[0052] For example, when judging whether it is a true peak, if Th=1 at this time, if there is a true peak, this step ends. If there is no true peak, then Th is lowered to re-judge whether there is a true peak. At this time, Th=0.5. When Th=0.5, the true peak judgment process is repeated. If there is a true peak, the next step is performed. If not, the current waveform is output without the active feature of the recovery period.

[0053] As a preferred embodiment, after determining whether each peak on each waveform diagram is a true peak, the method further includes: correcting each true peak determined one by one to obtain true peak data of better quality.

[0054] like Figure 4 As shown, there are two peaks on the graph, located at 38 seconds and 110 seconds respectively. However, the peak at 110 seconds is too small compared to the overall area, so the peak at 110 seconds can be determined as a false peak during the correction process.

[0055] It should be noted that, compared with the overall area, being too small can be regarded as: the result of the peak area / the waveform area is too low, wherein the specific threshold or range of “too low” can be set according to actual needs.

[0056] Step 2.4: when the time of the positive T gravity center of any true peak is greater than or equal to the first threshold and the peak occurrence time of the current true peak is greater than the second threshold, it is determined that the current waveform has the active feature of the recovery period.

[0057] In the embodiment of the present invention, the positive T centroid can be regarded as: if the positive areas on the left and right sides of the peak can be made the same at a certain moment, then this moment is the positive T centroid.

[0058] In this step, the second threshold may be 70 seconds, that is, the peak value of the true peak must appear in the waveform after 70 seconds. The range of the first threshold is 80-85.

[0059] Step 2.5, record the waveform with the recovery period activity characteristics and output the recovery period activity recognition result.

[0060] like Figure 5 As shown, the left figure is the original near-infrared waveform, and the right figure is the result of recovery period activity algorithm identification. The asterisk represents the peak and the cross represents the trough. The dotted line at x=109 in the right figure is the T center time of the corresponding true peak. This peak meets the recovery period activity conditions and is determined to have recovery period activity characteristics.

[0061] It should be noted that as long as there is a true peak in the waveform that meets the recovery period activity condition, the waveform can be considered to have the recovery period activity feature.

[0062] The implementation basis of each embodiment of the present invention is to implement programmed processing through a device with a processor function. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention are encapsulated into various modules. Based on this reality, on the basis of the above embodiments, an embodiment of the present invention provides a recovery period activity recognition device based on near-infrared image data, and the device is used to execute a recovery period activity recognition device method based on near-infrared image data in the above method embodiment.

[0063] See also Figure 6 The device includes: a first main module for acquiring near-infrared image data; a second main module for inputting the acquired near-infrared image data into a trained recovery period activity recognition model and outputting a recovery period activity recognition result; wherein the training of the recovery period activity recognition model includes: acquiring near-infrared image sample data and forming a plurality of waveform graphs according to the near-infrared image sample data; determining the peak on each waveform graph according to the hill climbing method; judging whether each peak on each waveform graph is a true peak one by one; judging that the current waveform has the recovery period activity feature when the time of the positive T center of gravity of any true peak is greater than or equal to a first threshold and the peak occurrence time of the current true peak is greater than a second threshold; recording the waveform with the recovery period activity feature and outputting the recovery period activity recognition result.

[0064] The embodiment of the present invention provides a recovery period activity recognition device based on near-infrared imaging data. The traditional method of clinical departments relying on behavioral observation or doctor's intuition to judge whether a patient has symptoms of schizophrenia and then conducting relevant treatment is prone to errors or cannot conduct effective treatment early due to relying on behavioral observation and doctor's experience. Figure 6 Several modules in the system can analyze the patient's near-infrared blood oxygen imaging data, judge the characteristics of the blood oxygen imaging data, and detect whether it has the characteristics of recovery period activity. This can help identify whether the patient's brain function activities show recovery period activity, and then assist doctors in objectively and scientifically judging whether the patient has symptoms of schizophrenia.

[0065] It should be noted that the device embodiments provided by the present invention are used to implement the methods in the above method embodiments as well as the methods in other method embodiments provided by the present invention. The only difference is that the corresponding functional modules are set, and the principle is basically the same as the principle of the above device embodiments provided by the present invention. As long as the technical personnel in the field refer to the specific technical solutions in other method embodiments on the basis of the above device embodiments, obtain the corresponding technical means and the technical solutions composed of these technical means by combining the technical features, and on the premise of ensuring the practicality of the technical solutions, improve the modules in the above device embodiments to obtain the corresponding device class embodiments, which are used to implement the methods in other method class embodiments. For example: Based on the content of the above device embodiment, as a preferred embodiment, a recovery period activity identification device based on near-infrared image data provided in the embodiment of the present invention determines the peak on each waveform graph according to the hill climbing method, and also includes: S1, the peak is determined by adjacent valleys, where the two valleys are the start valley and the end valley of the peak; S2, if the current peak meets the peak condition, update the start valley and the end valley, and return to step S1; S3, if the current peak does not meet the peak condition, update the end valley forward and return to step S1.

[0066] Based on the content of the above device embodiment, as a preferred embodiment, a recovery period activity identification device based on near-infrared image data provided in the embodiment of the present invention determines whether each peak on each waveform graph is a true peak one by one, and also includes: When the height of the end valley of the current peak is lower than the first preset ratio of the waveform maximum value, the ratio of the peak height to the amplitude is greater than the second preset ratio, and the peak value is greater than the third threshold, the current peak is determined to be a true peak.

[0067] Based on the contents of the above-mentioned device embodiments, as a preferred embodiment, a recovery period activity identification device based on near-infrared imaging data is provided in the embodiments of the present invention, which determines whether each peak on each waveform graph is a true peak one by one, and also includes: when the current peak is low and reverse and the area of ​​the peak is greater than the fourth threshold, it is determined that the current peak is a true peak.

[0068] Based on the same technical concept as the above embodiment, the embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the recovery period activity recognition method based on near-infrared image data are implemented, including: Acquire near infrared imaging data; The acquired near-infrared image data is input into the trained recovery period activity recognition model, and the recovery period activity recognition result is output; wherein the training of the recovery period activity recognition model includes: Acquire near infrared image sample data and form a plurality of waveform graphs according to the near infrared image sample data; Determine the peak on each waveform graph using the hill climbing method; Determine whether each peak on each waveform is a true peak one by one; When the time of the positive T gravity center of any true peak is greater than or equal to the first threshold and the peak occurrence time of the current true peak is greater than the second threshold, it is determined that the current waveform has the active feature of the recovery period; Record the waveform with recovery period activity characteristics and output the recovery period activity recognition result.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for recuperation activity recognition based on near-infrared imaging data, characterized in that: include: Acquire near infrared image data, wherein the near infrared image data includes original light intensity; The acquired near-infrared image data is input into the trained recovery period activity recognition model, and the recovery period activity recognition result is output; wherein the training of the recovery period activity recognition model includes: Acquire near-infrared image data, wherein the near-infrared image data includes original light intensity, convert the original light intensity into light density, convert the light density into concentration, perform moving average and rotation and translation processing on the acquired near-infrared image data, and draw the processed data according to brain area division to form a plurality of waveform graphs; Determine the peak on each waveform graph using the hill climbing method; The method of determining whether each peak on each waveform graph is a true peak one by one includes: when the height of the end valley of the current peak is lower than a first preset ratio of the maximum value of the waveform, and the ratio of the peak height to the amplitude is greater than a second preset ratio, and the peak value is greater than a third threshold, determining that the current peak is a true peak, wherein the first preset ratio is in the range of 0.2-0.5, the second preset ratio is in the range of 0.1-0.5, and the third threshold is in the range of 0.1-0.2 millimoles per millimeter; When the time of the positive T gravity center of any true peak is greater than or equal to the first threshold and the peak occurrence time of the current true peak is greater than the second threshold, it is determined that the current waveform has the active feature of the recovery period, wherein the first threshold ranges from 80 to 85 seconds and the second threshold is 70 seconds; Record the waveform with recovery period activity characteristics and output the recovery period activity recognition result.

2. The method for recuperation period activity recognition based on near infrared image data according to claim 1, characterized in that: Determine the peaks on each waveform using the hill climbing method, including: S1, the peak is determined by adjacent valleys, where the two valleys are the start valley and the end valley of the peak; S2, if the current peak meets the peak condition, update the start valley and the end valley, and return to step S1; S3, if the current peak does not meet the peak condition, update the end valley forward and return to step S1.

3. The method for recuperation activity recognition based on near infrared image data according to claim 1, characterized in that: When no true peak is found after judging each peak on the waveform diagram one by one, the third threshold is lowered and the true peak judgment process is repeated.

4. The method for recuperation activity recognition based on near infrared image data according to claim 2, characterized in that: The method of judging whether each peak on each waveform diagram is a true peak one by one may include: when the current peak is flat and reverse and the area of ​​the peak is greater than a fourth threshold, judging that the current peak is a true peak.

5. The method for recuperation period activity recognition based on near infrared image data according to claim 1, characterized in that: After determining whether each peak on each waveform diagram is a true peak one by one, the method further includes: correcting each true peak determined one by one.

6. A recovery period activity recognition device based on near infrared image data, characterized in that: include: A first main module is used to obtain near-infrared image data, wherein the near-infrared image data includes original light intensity; The second main module is used to input the acquired near-infrared image data into the trained recovery period activity recognition model and output the recovery period activity recognition result; wherein the training of the recovery period activity recognition model includes: Acquire near-infrared image data, wherein the near-infrared image data includes original light intensity, convert the original light intensity into light density, convert the light density into concentration, perform moving average and rotation and translation processing on the acquired near-infrared image data, and draw the processed data according to brain area division to form a plurality of waveform graphs; Determine the peak on each waveform graph using the hill climbing method; The method of determining whether each peak on each waveform graph is a true peak one by one includes: when the height of the end valley of the current peak is lower than a first preset ratio of the maximum value of the waveform, and the ratio of the peak height to the amplitude is greater than a second preset ratio, and the peak value is greater than a third threshold, determining that the current peak is a true peak, wherein the first preset ratio is in the range of 0.2-0.5, the second preset ratio is in the range of 0.1-0.5, and the third threshold is in the range of 0.1-0.2 millimoles per millimeter; When the time of the positive T gravity center of any true peak is greater than or equal to the first threshold and the peak occurrence time of the current true peak is greater than the second threshold, it is determined that the current waveform has the active feature of the recovery period, wherein the first threshold ranges from 80 to 85 seconds and the second threshold is 70 seconds; Record the waveform with recovery period activity characteristics and output the recovery period activity recognition result.

7. The recovery period activity recognition device based on near infrared image data according to claim 6, characterized in that: Determine the peaks on each waveform using the hill climbing method, including: S1, the peak is determined by adjacent valleys, where the two valleys are the start valley and the end valley of the peak; S2, if the current peak meets the peak condition, update the start valley and the end valley, and return to step S1; S3, if the current peak does not meet the peak condition, update the end valley forward and return to step S1.

8. The recovery period activity recognition device based on near infrared image data according to claim 7, characterized in that: The method of judging whether each peak on each waveform diagram is a true peak one by one may include: when the current peak is flat and reverse and the area of ​​the peak is greater than a fourth threshold, judging that the current peak is a true peak.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the recovery period activity recognition method based on near-infrared imaging data described in any one of claims 1 to 5 are implemented.

Images