Artifact correction method, device and storage medium for near infrared signal data

Abstract

The present application relates to an artifact correction method, device and storage medium for near-infrared signal data. The method may include dividing the received near-infrared signal data into small segments of signal data of a preset length, and calculating the fluctuation deviation parameters of the signal data of each small segment, extracting the fluctuation deviation parameters of the signal data of the first predetermined proportion before the sorting, and determining the artifact recognition threshold based on the representative fluctuation deviation parameters and threshold multiplier of the fluctuation deviation parameters of the signal data of the first predetermined proportion before the extracted, and comparing the deviation of the signal data before and after each small segment with the artifact recognition threshold. When the deviation exceeds the artifact recognition threshold, the signal data of the small segment is identified as containing motion artifacts. The artifact recognition threshold determined by the present application has a high correlation with the physiological oscillation part in the signal, and can accurately identify motion artifacts even when the amplitude of the real physiological signal is contaminated or even submerged by motion artifacts.

Description

Technical Field

[0001] The present application relates to the field of medical data processing, and more specifically, to a method, device and storage medium for artifact correction of near-infrared signal data. Background Art

[0002] At present, although fNIRS (near-infrared functional brain imaging technology) is widely used in the field of neonatal research. However, as a method with great application prospects in the neonatal population, it still faces a challenge when applied to daily clinical diagnosis - the interference of motion artifacts. In the process of clinical application of fNIRS, since the neonatal group has no behavioral control ability and cannot follow instructions, the newborns as subjects will randomly make uncontrollable movements in the process of collecting near-infrared signal data, and there will still be a lot of body movements even under sleeping conditions. The frequent relative movement between the probe of the near-infrared functional brain imaging device and the scalp of the newborn may cause the light signal collected by the device to be interfered and destroyed, thereby generating dense motion artifacts. In addition, the frequency of motion artifacts in the signal data of the newborn is higher and the degree is more serious. Secondly, the skull of the newborn is soft, the fontanelle is not closed, and other tissues around the brain are relatively thin. When collecting the signal data of the newborn, the probe can only be gently attached to the head of the newborn, and external force is not allowed to be applied to the skull. The inadvertent slight movement of the newborn will cause the slip between the probe and the scalp. These unavoidable problems will lead to motion artifacts contaminating the fNIRS signal data of the whole infant.

[0003] Because the neonatal brain and neurovascular coupling are still in the development stage, and the cortical excitability is low, the amplitude of physiological oscillations in the neonatal fNIRS signal data is small. However, the peak or offset amplitude of a large number of motion artifacts in the signal data is much higher than the change in the neonatal physiological oscillation, resulting in a large difference between the information reflected when the fNIRS signal data is contaminated and the actual activity of the neonatal cerebral cortex. Inappropriate preprocessing or uncorrected motion artifacts usually result in false positive or false negative results. Therefore, how to deal with neonatal signal data containing dense motion artifacts is a challenge faced by researchers. Summary of the invention

[0004] The present application is intended to solve the above technical problems. The present application is intended to provide a method, device and storage medium for artifact correction of near-infrared signal data, which is particularly suitable for continuously collected fNIRS signal data such as newborns containing dense and irregular motion artifacts, and can accurately identify motion artifacts in continuous signal data whose real physiological signal amplitude is contaminated or even submerged by motion artifacts, so as to perform artifact correction in a targeted manner.

[0005] In one embodiment, the present application relates to an artifact correction method for near-infrared signal data. The artifact correction method includes: receiving near-infrared signal data; dividing the received near-infrared signal data into small segments of signal data of a preset length, and calculating the fluctuation deviation parameters of the signal data of each small segment; sorting the calculated fluctuation deviation parameters of the signal data of each small segment from small to large to extract the fluctuation deviation parameters of the signal data of the first predetermined proportion in the sorting; determining the artifact recognition threshold based on the representative fluctuation deviation parameters and threshold multiplier of the fluctuation deviation parameters of the signal data of the first predetermined proportion extracted; comparing the deviation of the signal data before and after each small segment with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, identifying the signal data of the small segment as containing motion artifacts.

[0006] In another embodiment, the present application relates to an artifact correction method for photoelectric signal data reflecting physiological conditions, wherein the photoelectric signal data is continuously collected. The artifact correction method includes: receiving photoelectric signal data; dividing the received photoelectric signal data into small segments of signal data of a preset length, and calculating the fluctuation deviation parameters of the signal data of each small segment; sorting the calculated fluctuation deviation parameters of the signal data of each small segment from small to large to extract the fluctuation deviation parameters of the signal data of the first predetermined proportion in the sorting; determining the artifact recognition threshold based on the representative fluctuation deviation parameters and threshold multiplier of the fluctuation deviation parameters of the signal data of the first predetermined proportion extracted; comparing the deviation of the signal data before and after each small segment with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, identifying the signal data of the small segment as containing motion artifacts.

[0007] In another embodiment, the present application relates to an artifact correction device for near-infrared signal data of a newborn at rest. The artifact correction device includes: an interface and at least one first processor. The interface is configured to obtain near-infrared signal data of the brain of a newborn subject at rest. The first processor can be configured to receive near-infrared signal data. The first processor can also be configured to divide the received near-infrared signal data into small segments of signal data of a preset length, and calculate the fluctuation deviation parameters of the signal data of each small segment. The first processor can also be configured to sort the calculated fluctuation deviation parameters of the signal data of each small segment from small to large, so as to extract the fluctuation deviation parameters of the signal data of the first predetermined proportion in the sorting. The first processor can also be configured to determine the artifact recognition threshold based on the representative fluctuation deviation parameters and threshold multiplier of the fluctuation deviation parameters of the signal data of the first predetermined proportion extracted. The first processor can also be configured to compare the deviation of the signal data before and after each small segment with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, the signal data of the small segment is identified as containing motion artifacts.

[0008] In another embodiment, the present application relates to a non-transitory computer-readable medium storing instructions. The instructions implement an artifact correction method for near-infrared signal data when executed by a processor. The method may include dividing the received near-infrared signal data into small segments of signal data of a preset length, and calculating the fluctuation deviation parameters of the signal data of each small segment. The method may also include: sorting the calculated fluctuation deviation parameters of the signal data of each small segment from small to large to extract the fluctuation deviation parameters of the signal data of the first front predetermined proportion in the sorting. The method may also include determining an artifact recognition threshold based on a representative fluctuation deviation parameter and a threshold multiplier of the fluctuation deviation parameters of the signal data of the first front predetermined proportion extracted. The method may also include comparing the deviation of the signal data before and after each small segment with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, identifying the signal data of the small segment as containing motion artifacts.

[0009] In another embodiment, the present application relates to an auxiliary analysis device for brain injury conditions. The auxiliary analysis device may include: an acquisition module and a second processor, wherein the acquisition module of the auxiliary analysis device may be configured to acquire near-infrared signal data of the brain of a newborn subject in a resting state; and the second processor may be configured to receive the near-infrared signal data. The second processor may also be configured to divide the received near-infrared signal data into small segments of signal data of a preset length, and calculate the fluctuation deviation parameters of the signal data of each small segment. The second processor may also be configured to sort the calculated fluctuation deviation parameters of the signal data of each small segment from small to large, so as to extract the fluctuation deviation parameters of the signal data of the first predetermined proportion in the sorting. The second processor may also be configured to determine the artifact recognition threshold based on the representative fluctuation deviation parameters and the threshold multiplier of the fluctuation deviation parameters of the signal data of the first predetermined proportion extracted. The second processor may also be configured to compare the deviation of the signal data before and after each small segment with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, the signal data of the small segment is identified as containing motion artifacts. The second processor may also be configured to process and analyze the artifact-corrected near-infrared signal data to assist in analyzing the brain injury status of the neonatal subject.

[0010] Thus, the present application first divides the near-infrared signal data into small segments of signal data of a preset length, and then calculates the fluctuation deviation parameters of all the small segments and sorts them. The smaller the fluctuation deviation parameter, the less or no motion artifacts there are in the segment of signal data. Therefore, the representative fluctuation deviation parameter is calculated using the small segment of signal data with a smaller fluctuation deviation parameter, and the artifact recognition threshold is determined based on the representative fluctuation deviation parameter. The artifact recognition threshold determined in this way has a high correlation with the physiological oscillation part in the signal data, so that the motion artifact can be accurately identified in the near-infrared signal data whose real physiological signal amplitude is contaminated or even submerged by the motion artifact, so as to perform artifact correction in a targeted manner.

[0011] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In the drawings, which are not necessarily drawn to scale, the same reference numerals may describe similar components in different views. The same numbers with letter suffixes or different letter suffixes may represent different instances of similar components. The drawings generally illustrate various embodiments by way of example and not by way of limitation, and together with the specification and claims, are used to illustrate the disclosed embodiments. Such embodiments are illustrative and are not intended to be exhaustive or exclusive embodiments of the method, apparatus, system, or non-transitory computer-readable medium having stored thereon instructions for implementing the method.

[0013] Figure 1 A method for correcting artifacts of near infrared signal data according to an embodiment of the present application is shown.

[0014] Figure 2 The process of performing fitting correction on signal data of each small segment identified as containing motion artifacts according to an embodiment of the present application is shown.

[0015] Figure 3 The process of correcting each peak in each small segment of signal data identified as containing motion artifacts according to an embodiment of the present application is shown.

[0016] Figure 4 A schematic diagram showing an artifact correction process for near infrared signal data according to an embodiment of the present application.

[0017] Figure 5 A schematic diagram of an artifact waveform of near-infrared signal data according to an embodiment of the present application is shown.

[0018] Figure 6 A method for correcting artifacts of photoelectric signal data reflecting physiological conditions according to an embodiment of the present application is shown.

[0019] Figure 7 A schematic block diagram of a device for artifact correction of near-infrared signal data according to an embodiment of the present application is illustrated. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the technical solution of the present application, the present application is described in detail below in conjunction with the accompanying drawings and specific implementation examples. The embodiments of the present application are further described in detail below in conjunction with the accompanying drawings and specific implementation examples, but are not intended to limit the present application.

[0021] The words "first", "second" and similar terms used in this application do not indicate any order, quantity or importance, but are only used to distinguish parts. The words "include" or "comprise" and similar terms mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of including other elements. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0022] All terms (including technical terms or scientific terms) used in the present invention have the same meaning as those understood by ordinary technicians in the field to which the present invention belongs, unless otherwise specifically defined. It should also be understood that terms defined in general dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined herein. Technologies, methods and equipment known to ordinary technicians in the relevant field may not be discussed in detail, but where appropriate, the technologies, methods and equipment should be considered as part of the specification.

[0023] The expressions "first correction signal data", "second correction signal data", "third correction signal data" and "fourth correction signal data" herein are only used to distinguish the signal data for the purpose of description, and are not intended to limit the quantity. The "first correction signal data", "second correction signal data", "third correction signal data" and "fourth correction signal data" may also be the same signal data or different signal data. In the various processes described herein, the order of steps shown in the accompanying drawings is only used as an example. Without affecting the logical relationship of the various steps, the relevant steps can be flexibly executed in an order different from that shown in the drawings.

[0024] Figure 1 FIG. 4 shows a method for correcting artifacts of near infrared signal data according to an embodiment of the present application. Figure 1As shown, a method for correcting artifacts of near-infrared signal data may include, in step S11, receiving near-infrared signal data. For example, the near-infrared signal data may include, but is not limited to, OD signal (optical density (OD) signal) data converted from the collected near-infrared optical signal of the subject. Among them, the collected near-infrared optical signal of the subject may be single-channel signal data or multi-channel signal data, which is not specifically limited here. The subjects may include ordinary subjects, and may also include subjects with poor behavioral control ability and command obedience, such as newborns, elderly people with stroke, etc. The near-infrared signal data of such subjects may contain dense and irregular motion artifacts in the signal data due to a large number of uncontrollable random movements, and the motion artifacts may contaminate or even submerge the waveform of the real physiological signal.

[0025] In step S12, the received near-infrared signal data can be divided into small segments of signal data of a preset length, and the fluctuation deviation parameters of the signal data of each small segment are calculated. Compared with the robust fluctuation deviation parameters of near-infrared signal data of a longer duration, the fluctuation deviation parameters of small segments of signal data are more sensitive to sudden artifacts, and are less likely to be masked by operations on a longer whole segment of signal data, such as averaging within a time window. In a duration of fNIRS signal data, motion artifacts may generally occupy only a part of the time period. Among the several small segments of signal data into which the overall signal data is divided, there are artifacts in some small segments, resulting in significantly larger fluctuation deviation parameters, while there are no artifacts in the remaining small segments, resulting in significantly smaller fluctuation deviation parameters. Specifically, significantly larger fluctuation deviation parameters more often characterize the interference oscillation deviation caused by artifacts, while significantly smaller fluctuation deviation parameters more often characterize the true physiological oscillation deviation. In the calculated deviation set, a relatively small physiological oscillation deviation indicates that this segment of signal data does not contain artifacts or only contains fewer artifacts. A relatively large physiological oscillation deviation indicates that this segment of signal data contains more artifacts.

[0026] In some embodiments, the fluctuation deviation parameter includes a standard deviation value (STD).

[0027] Step S12 can be specifically implemented as follows. First, the received near-infrared signal data is divided into small segments of signal data of a preset length using a sliding window method. The duration of the sliding window can be set to 4s, and the interval can be set to 2s, that is, a small segment of signal data with a signal data length of 4s is taken every 2s. Then, the standard deviation (STD) of the signal data of each small segment can be calculated. STD can use various formulas. For example, the average value of n OD signal data in a time period of length n can be calculated, and then the square of the difference between each OD signal data and the average value is calculated, and then all the squares of the differences are summed to obtain the sum of squared differences, and the sum of squared differences is averaged for the length n, and the root of the mean is used as STD. As another example, the standard deviation (STD) of each small segment of signal data can also be calculated according to formula (1).

[0028]

[0029] Among them, N1 is the sequence length of the data in the sliding window, x (t) is the tth OD signal data. When the sampling rate of the fNIRS signal data is 10 Hz, the duration of the 4S window as described above can allow the signal data to meet the requirements for calculating the STD of a small segment of signal data.

[0030] In step S13, the calculated fluctuation deviation parameters of the signal data of each small segment are sorted from small to large to extract the fluctuation deviation parameters of the signal data of a first predetermined proportion in the sorting.

[0031] By sorting the set of fluctuation deviation parameters (such as but not limited to STD) of the calculated signal data of each small segment, the fluctuation deviation parameters of the small segment containing artifacts will be placed at the back, while the fluctuation deviation parameters of the small segment containing few or even no artifacts will be placed at the front. Extracting the signal data of the first predetermined ratio in the sorting can extract the signal data of those small segments containing few or even no artifacts, and the representative fluctuation deviation parameters calculated based on this are more similar to the fluctuation deviation parameters caused by the real physiological oscillation, and can eliminate the interference of the amplitude and frequency of the artifacts, and are relatively more stable. It is more accurate and has stronger anti-interference ability when used as a reference for artifact identification. In some embodiments, the first predetermined ratio in the sorting can be set according to the specific situation of the proportion of the time period in which the artifact occurs in the near-infrared signal data. For example, for the near-infrared signal data of the resting state of the newborn, the first predetermined ratio in the sorting can be set to the first 20%-first 35% (for example, 30%) in the sorting. For another example, for the near-infrared signal data collected during the rehabilitation exercise of the elderly with stroke, the first predetermined ratio in the sorting can be changed to adapt to the specific situation of the proportion of the time period in which the artifact occurs in the application scenario. In this way, the small segment at the front with “no artifact interference” can be efficiently screened out, and the deficiency of the small segment at the front with “no artifact interference” that causes the inability to correctly and comprehensively characterize the real physiological oscillation can be avoided.

[0032] Step S14, based on the representative fluctuation deviation parameter and threshold multiple of the fluctuation deviation parameter of the extracted first front predetermined proportion of signal data, determine the artifact recognition threshold. The representative fluctuation deviation parameter may include any one of the mean value, median value, root of the mean of the sum of squares, etc. For example, the fluctuation deviation parameter of the extracted first front predetermined proportion of signal data may be averaged to obtain the representative fluctuation deviation parameter. For the first front predetermined proportion of 30%, the fluctuation deviation parameter of the smaller first 30% of signal data is averaged to obtain the representative fluctuation deviation parameter to mark and identify motion artifacts. Further accounting for the threshold multiple on the basis of the representative fluctuation deviation parameter can improve the detection robustness of the artifact. The threshold multiple can be set according to the specific situation of the multiple difference between the fluctuation deviation parameter caused by the artifact in the near-infrared signal data and the fluctuation deviation parameter caused by the real physiological oscillation, so that it can ensure the detection rate of the artifact segment and reduce the false positive rate of the artifact segment. In some embodiments, for the near-infrared signal data of the newborn in the resting state, the threshold multiple can be set to 4-8, for example but not limited to 6.8.

[0033] In some embodiments, the representative fluctuation deviation parameter obtained is multiplied by the threshold multiple to obtain the artifact recognition threshold. Since the present application first divides the entire segment of near-infrared signal data into small segments of signal data, the fluctuation deviation parameters of all small segments of signal data are calculated. In this deviation set, the fluctuation deviation parameters of the smaller first 30% of the signal data are averaged to obtain the representative fluctuation deviation parameter, and further obtain the artifact recognition threshold. The fluctuation deviation parameters of the smaller first 30% of the signal data represent the deviation (STD) of the (basically) noise-free part of the physiological oscillation in the signal. Therefore, the smaller part is averaged to replace the deviation of the overall signal data and multiplied by the threshold multiple to calculate the final artifact recognition threshold. In this way, not only is the representative fluctuation deviation parameter closer to the fluctuation deviation parameter caused by the real physiological oscillation, it can also eliminate the interference of the amplitude and frequency of the artifact, and it is also relatively more stable, so the threshold used for artifact recognition is more accurate and has stronger anti-interference. On this basis, combined with the threshold multiple that meets the specific situation, it can ensure the detection rate of the small segment of the artifact and reduce the false positive rate of the small segment of the artifact.

[0034] In some embodiments, before being divided into small segments of signal data of a preset length, or before step S12, the received near-infrared signal data may be subjected to high-pass filtering within a preset frequency range. Specifically, the preset frequency may be a cutoff frequency of high-pass filtering, and the cutoff frequency range may be 0.1 Hz-1 Hz, or may be other frequency ranges determined according to the received near-infrared signal data, which are not specifically limited herein. The purpose of high-pass filtering is to retain the physiological oscillation information in the near-infrared signal data and the motion artifact information to be removed, and to remove other irrelevant noise. The filtered signal data is used to estimate the required artifact recognition threshold. Before being divided into small segments of signal data of a preset length, the received near-infrared signal data is first subjected to high-pass filtering, and the artifact recognition threshold is determined based on the signal data after high-pass filtering, which can reduce the influence of other irrelevant noises on the artifact recognition threshold, and retain the physiological oscillation information in the near-infrared signal data and the motion artifact information to be removed, so that a more accurate artifact recognition threshold can be obtained.

[0035] In step S15, the deviation of the signal data before and after each small segment may be compared with the artifact recognition threshold. When the deviation exceeds the artifact recognition threshold, the signal data of the small segment is identified as containing motion artifacts.

[0036] Various methods can be used to calculate the deviation of the signal data before and after each small segment. Note that the so-called deviation of the signal data before and after each small segment, "before and after" can be within each small segment, or outside each small segment (including the front and back endpoints of the time window of each small segment). In some embodiments, the STD of the signal data in each small segment can be calculated as the deviation of the signal data before and after each small segment. For example, the deviation between adjacent signal data can be calculated in sequence as the deviation of the signal data before and after each small segment. In this way, the STD of the signal data of each small segment that has been calculated in step S12 can be reused, thereby simplifying the calculation. For another example, the deviation between the signal data of the first segment and the last segment of a small segment data can be calculated as the deviation of the signal data before and after each small segment. In some other embodiments, it can be calculated whether the difference between the signal data of the front position and the signal data of the rear position before and after each small segment exceeds the artifact recognition threshold. If it exceeds, it will be marked as a motion artifact. For example, the front position and the rear position can just span the time window of the corresponding small segment, or it can be obtained by extending a certain margin forward and backward from the time window. For another example, the so-called front and rear positions can be a single position or multiple positions, and the signal data of multiple positions can be averaged and then the difference between the front and rear positions can be calculated. In other words, the signal data of the front position and the signal data of the rear position can be single-point signal data or representative signal data obtained from multiple points.

[0037] After detecting the motion contaminated portion of the near infrared signal data (that is, the small segment containing the artifact) based on the artifact recognition threshold, the obtained near infrared signal data can be expressed by the following formula (2):

[0038] x(t)={x FM,1 (t),x MA,1 (t),x FM,2 (t),x MA,2 (t),…,x FM,L (t),x MA,L (t)} Formula (2) Where x(t) is the time series of all channels, x FM represents a small segment of signal data that is noise-free (i.e., identified as not containing motion artifacts), and x MA represents a small segment of signal data containing motion artifacts, and L represents the channel.

[0039] After identifying the artifact, the motion artifact needs to be corrected. This artifact correction can be performed on the original near-infrared signal data of a single channel of the brain, and then subsequent integrated analysis can be performed; or it can be performed on the fused signal data after the near-infrared signal data of multiple channels are fused, such as by averaging and other methods, and then performing artifact correction. In the following, for the small segments of signal data identified as containing motion artifacts, spikes (such as Figure 5 51, 53), slow drift, baseline shift (such as Figure 5 As an example, the present application uses a fitting method such as a spline interpolation method to correct slow drift, baseline offset, etc. in motion artifacts, and then corrects the peaks in motion artifacts as an example for explanation. However, this is not a limitation, and the order of the correction steps can also be changed.

[0040] In some embodiments, when near-infrared signal data is acquired using a near-infrared brain function imaging device, near-infrared signal data of multiple channels can be acquired simultaneously. Preferably, step S11 receives near-infrared signal data of a single channel for motion artifact recognition, that is, the artifact correction method described in each embodiment of the present application is to perform motion artifact recognition on the near-infrared signal data of each channel respectively, divide the near-infrared signal data of each channel into small segments of signal data of a preset length, and sort the calculated fluctuation deviation parameters of the signal data of each small segment from small to large, and determine the artifact recognition threshold for each channel based on the representative fluctuation deviation parameter and threshold multiplier of the fluctuation deviation parameter of the first front predetermined ratio of the extracted signal data, so that the recognition accuracy of motion artifacts is higher.

[0041] In other embodiments, what is received in step S11 may also be near-infrared signal data of multiple channels simultaneously acquired by a near-infrared brain function imaging device, and this application does not make any specific limitation on this.

[0042] Figure 2 FIG. 2 shows a process of fitting and correcting the signal data of each small segment identified as containing motion artifacts according to an embodiment of the present application. Figure 2As shown, the artifact correction method may include: step S21, the signal data of each small segment identified as containing motion artifacts can be fitted to obtain fitted signal data. For example, the fitting can adopt various fitting algorithms, such as but not limited to spline interpolation method, polynomial curve fitting, etc., so as to fit and simulate the slow drift, baseline offset and other slowly changing artifact components in the motion artifact. Taking spline interpolation as an example, modeling can be performed by cubic spline interpolation. According to the parameter p in the specified range [0,1], a spline function of an appropriate degree is selected to specify the accuracy of the interpolation. For example, when p=0, the least squares straight line fitting is applied, and when p=1, the natural cubic spline interpolation is performed. In the present application, p=0.99 is used to correct the identified motion artifacts.

[0043] In step S22, the corresponding fitting signal data can be subtracted from the signal data of each small segment identified as containing motion artifacts to obtain the first corrected signal data of each small segment. The signal data of each small segment identified as containing motion artifacts can be subtracted from the signal data fitted by cubic spline interpolation to obtain the artifact-free signal data.

[0044] In step S23, the first correction signal data of each small segment can be balanced and adjusted based on the previous signal data adjacent to each small segment to obtain the second correction signal data of the small segment. Balanced adjustment: The baseline data of the previous signal data adjacent to each small segment can be referred to, and the baseline data of the first correction signal data can be pulled back to the same or similar position as the baseline data of the previous signal data adjacent to each small segment to ensure the continuity and smoothness of the front and rear signal data. The amplitude level (for example but not limited to the average value) before and after the motion artifact segment in the signal data of each small segment can be balanced and adjusted to reconstruct the entire time series. The reconstructed signal data can be expressed by the following formula (3):

[0045] y(t)={x FM,1 (t),x SP,1 (t)+a1,x FM,2 (t),x SP,2 (t)+a2,…,x FM,L (t),x SP,L (t)+a L} Formula (3)

[0046] Among them, y(t) represents the signal data after equalization adjustment, x SP represents the signal data of the corresponding segment minus the fitted signal data, a represents the parameter used for equalization adjustment, and L represents the channel.

[0047] Figure 3The process of determining the artifact recognition threshold based on the representative fluctuation deviation parameter of the fluctuation deviation parameter of the second predetermined ratio of the signal data extracted and the threshold multiplier, identifying each peak in each small segment of signal data containing motion artifacts, and correcting it is shown. Figure 3 As shown, the artifact correction method may include step S31, for each peak in each small segment of signal data identified as containing motion artifacts, fitting corresponding Gaussian white noise signal data based on the extracted signal data of the second previous predetermined ratio. Specifically, the first previous predetermined ratio and the second previous predetermined ratio may be the same previous predetermined ratio, or different previous predetermined ratios, which are not specifically limited here. In a preferred embodiment, the first previous predetermined ratio and the second previous predetermined ratio may be the same previous predetermined ratio, so that the corresponding Gaussian white noise signal data fitted based on the extracted signal data of the second previous predetermined ratio is closer to the signal data of the first previous predetermined ratio that does not contain motion artifacts or contains fewer motion artifacts, so that when the identified peaks are replaced with the corresponding Gaussian white noise signal data, there will be no significant deviation from the signal data of the first previous predetermined ratio, and the reliability is high.

[0048] For example, the signal data of Gaussian white noise fitting is shown in the following formula (4):

[0049] y GS (t)={STD1×Wgn(t),STD2×Wgn(t),…,STD L ×Wgn(t)} Formula (4)

[0050] Among them, y GS (t) represents the signal data of all channels after Gaussian white noise fitting, STD represents the representative fluctuation deviation parameter, Wgn(t) represents the Gaussian white noise series, L represents the channel, and STD1×Wgn(t) refers to the fitted Gaussian white noise signal data obtained by multiplying the representative fluctuation deviation parameter of the first channel with the Gaussian white noise series.

[0051] In step S32, each identified peak may be replaced with corresponding Gaussian white noise signal data to obtain third corrected signal data of each small segment.

[0052] Fitting methods for continuous curves such as spline interpolation can simulate continuous slowly changing artifact components, but cannot fit sudden spikes. The spike elimination operation can be further performed on the small segments of signal data that have eliminated slowly changing artifact components such as slow drift and baseline offset, thereby comprehensively removing artifact components of various distribution frequencies and distribution spans. It should be noted that only the spikes in motion artifacts will be corrected, and will not affect normal signal data and signal data after spline interpolation artifacts have been removed.

[0053] In some embodiments, after the peak is replaced with the corresponding Gaussian white noise signal data to obtain the signal data with the peak eliminated, the signal data with the peak eliminated may have a baseline drift relative to the signal data of its front and rear segments. In this case, the signal data with the peak eliminated needs to be balanced again to ensure the continuity and smoothness of the signal data of the front and rear segments.

[0054] In some embodiments, the artifact correction method further includes: step S33, performing equalization adjustment on the third correction signal data of each small segment based on the previous segment signal data adjacent to the small segment to obtain the fourth correction signal data of the small segment.

[0055] In other embodiments, the artifact correction method also includes: after subtracting the corresponding fitting signal data from each small segment of signal data identified as containing motion artifacts to obtain the first corrected signal data of each small segment, directly performing the step of fitting the corresponding Gaussian white noise signal data based on the extracted second predetermined proportion of signal data, and using the Gaussian white noise signal data to replace each peak in the signal data. That is to say, in this scheme, only one equalization adjustment is performed after the peak is replaced with the Gaussian white noise signal data.

[0056] Step S34, based on the fourth correction signal data of each small segment, the subsequent signal data adjacent to each small segment is balanced and adjusted to complete the artifact correction. The balanced adjustment can refer to the baseline data of the fourth correction signal data, for example, the balanced adjustment parameter is determined based on the amplitude average value of the fourth correction signal data, and the baseline data of the subsequent signal data adjacent to the fourth correction signal data is pulled to the same or similar value as the baseline data of the fourth correction signal data, so as to ensure the continuity and smoothness of the previous and next signal data.

[0057] After replacement, artifact correction can be accomplished by adjusting the signal segments before and after the spike noise.

[0058] The signal data reconstructed in this way can be expressed by formula (5):

[0059] z (t) ={y FM,1 (t),y GS,1 (t)+b1,y FM,2 (t),y GS,2 (t)+b2,…,y FM,L (t),y GS,L (t)+b L} Formula (5)

[0060] Among them, z (t) Represents the signal data of all channels after correction, y FMRepresents the signal data without spikes in the signal data, y GS It represents the signal data with the peak part eliminated, b represents the parameter used for equalization adjustment, and L represents the channel.

[0061] Figure 4 FIG. 1 is a schematic diagram showing an artifact correction process of near infrared signal data according to an embodiment of the present application. Figure 4 As shown, a method for correcting artifacts of near-infrared signal data includes: collecting near-infrared optical signals ( Figure 4 41) and display the waveform on the computer ( Figure 4 42). After the collected near-infrared optical signal of the subject is converted into OD signal (optical density (OD) signal data), it is high-pass filtered using a high-pass filter with a cutoff frequency of 0.75 Hz, and then divided into small segments of signal data of a preset length ( Figure 4 43), and calculate the fluctuation deviation parameter of the signal data of each small segment ( Figure 4 44). Sort the calculated fluctuation deviation parameters of the signal data of each small segment from small to large ( Figure 4 45) to extract the fluctuation deviation parameters of the first 30% of the signal data in the sorting. The artifact recognition threshold is determined based on the representative fluctuation deviation parameters of the fluctuation deviation parameters of the first 30% of the signal data extracted and the threshold multiplier 6.8. The deviation of the signal data before and after each small segment is compared with the artifact recognition threshold. When the deviation exceeds the artifact recognition threshold, the signal data of the small segment is identified as containing motion artifacts ( Figure 4 The cubic spline interpolation method is used to correct the slow drift, baseline drift and other slow-changing trajectory components in the signal data of each small segment (such as Figure 5 52). Then identify the peaks in the motion artifacts ( Figure 4 47), and the identified peaks were corrected by replacing them with Gaussian white noise ( Figure 4 48), and the near-infrared signal data with motion artifacts eliminated was obtained ( Figure 4 of 49).

[0062] Although the above description uses near-infrared signal data as an example, it can also be applied to other continuously collected photoelectric signal data reflecting physiological conditions, such as but not limited to at least one channel of EEG signal data, at least one channel of ECG signal data, etc. Figure 6 The present invention shows an artifact correction method for photoelectric signal data reflecting physiological conditions that is continuously collected. Figure 6 As shown, the artifact correction method may include the following steps:

[0063] Step 61 may be to receive photoelectric signal data based on a representative fluctuation deviation parameter of the fluctuation deviation parameter of the first front predetermined proportion of signal data extracted in order to extract the fluctuation deviation parameter of the first front predetermined proportion of signal data in the sorting and a threshold magnification.

[0064] Step 62 may be to divide the received photoelectric signal data into small segments of signal data of a preset length, and calculate the fluctuation deviation parameter of each small segment of signal data.

[0065] Step 63 may be to sort the calculated fluctuation deviation parameters of the signal data of each small segment from small to large, so as to extract the fluctuation deviation parameters of the signal data of a first predetermined proportion in the sorting.

[0066] In step 64, an artifact recognition threshold may be determined based on a representative fluctuation deviation parameter of the fluctuation deviation parameter of the signal data of the first predetermined proportion extracted and a threshold multiplier.

[0067] Step 65 may be to compare the deviation of the signal data before and after each small segment with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, identify the signal data of the small segment as containing motion artifacts.

[0068] The various steps of the artifact correction method for near-infrared signal data described in the various embodiments of the present application can also be applied to other continuously collected photoelectric signal data, and will not be described repeatedly here. For multi-channel continuous photoelectric signal data, by using the artifact correction method of each embodiment of the present application to perform artifact correction on the photoelectric signal data of each channel (especially channel-by-channel artifact correction), the artifacts of each channel can be efficiently removed, so that more accurate diagnostic and analysis results can be obtained when diagnosing and analyzing diseases based on the corrected data, especially when calculating functional connection parameters, the false correlation between the signal data of each channel can be reduced to avoid interference with the analysis conclusion.

[0069] verify

[0070] In order to verify the effectiveness of the present method, the applicant also used simulation data and real collected near-infrared signal data of newborns at rest to verify the present method. Specifically, the following two indicators were used to compare the actual HbO (oxyhemoglobin) concentration change and the corrected HbO concentration change: (1) Pearson correlation coefficient R; (2) root mean square error RMSE, where the calculation formula of RMSE is:

[0071]

[0072] For the simulation data, a(t i) is the oxygenated hemoglobin concentration change data after the i-th original artifact-free signal data is processed, and N2 represents the sequence length of the oxygenated hemoglobin concentration change data.

[0073] For the resting near-infrared signal data of newborns, a(t i ) is the oxygen and hemoglobin concentration change data of the newborn after manual selection of artifacts and processing. b(t i ) is the oxygenated hemoglobin concentration change data obtained after processing the signal data with different artifact processing methods. The Pearson correlation coefficient R is used to evaluate the similarity between different processing methods and the established standards. The closer R is to 1, the better the effect of artifact processing. The smaller the root mean square error RMSE, the smaller the difference between the results obtained after processing with the artifact processing method used and the standard results.

[0074] In order to compare the effectiveness of the present method with other traditional methods, the applicant also used wavelet algorithm and spline interpolation combined with wavelet method to perform artifact processing on the simulated signal data and the collected real neonatal resting near-infrared signal data, and compared the artifact processing analysis results with the artifact processing analysis results of the method of the present application. The results are as follows: for the simulated signal data, the artifact processing analysis results of the present method are: R = 0.860 ± 0.037; RMSE = 0.196 ± 0.02, while the artifact processing analysis results of the wavelet algorithm are: R = 0.746 ± 0.069; RMSE = 0.314 ± 0.048, and the artifact processing analysis results of the spline interpolation combined with wavelet are: R = 0.724 ± 0.096; RMSE = 0.282 ± 0.053. For the collected real neonatal resting near-infrared signal data, the artifact processing analysis results of this method are: R = 0.745 ± 0.142; RMSE = 0.356 ± 0.339, while the artifact processing analysis results of the wavelet algorithm are: R = 0.542 ± 0.222; RMSE = 0.811 ± 0.520, and the artifact processing analysis results of spline interpolation combined with wavelet are: R = 0.555 ± 0.046; RMSE = 0.661 ± 0.373. It can be seen that the method of the present application has a more ideal artifact processing effect than the traditional method.

[0075] Since a large number of artifacts in the signal data of newborns will affect the calculation of the recognition threshold, the strategy of the present application is to eliminate this influence by dividing the entire signal data into small segments of signal data and sorting the fluctuation deviation parameters of the signal data of each small segment, so as to obtain a more accurate artifact recognition threshold. When faced with the near-infrared signal data of newborns in resting state containing a large number of artifacts, the physiological oscillation of the noise-free part in the signal data is adaptively estimated by this strategy, and the artifact is objectively identified to the greatest extent based on the artifact recognition threshold determined by it. And in this way, the threshold multiple will also be fixed in a smaller range, so that the signal data containing motion artifacts can be more accurately identified. In addition, in order to solve the problem that the artifacts are still residual after correction, the present application adopts different processing methods for different types of artifacts. This method first uses the spline interpolation method to remove the baseline drift, and then uses Gaussian replacement to correct the spike. In this way, the Gaussian replacement process will not be affected by the baseline drift in the signal data, thereby producing a better artifact correction effect.

[0076] Figure 7 FIG. 1 is a block diagram illustrating an exemplary apparatus for correcting artifacts of near-infrared signal data of a newborn baby at rest according to an embodiment of the present application, as shown in FIG. Figure 7 As shown, an artifact correction device for near-infrared signal data of a newborn at rest includes an interface 707 and a first processor 701. The interface 707 can be configured to obtain near-infrared signal data of the brain of a newborn subject at rest. The first processor 701 can be configured to execute the artifact correction method for near-infrared signal data according to various embodiments of the present application.

[0077] Through the interface 707, the artifact correction device for the resting near-infrared signal data of the newborn can be connected to a network (not shown), such as but not limited to a local area network or the Internet in a hospital. However, the communication method implemented by the interface 707 is not limited to the network, and may include NFC, Bluetooth, WIFI, etc.; it can be a wired connection or a wireless connection. Taking the network as an example, the interface 707 can connect the artifact correction device for the resting near-infrared signal data of the newborn with external devices such as the resting near-infrared signal database 708 of the newborn and the near-infrared signal data acquisition device 709.

[0078] In some embodiments, the artifact correction device can be a dedicated intelligent device or a general intelligent device. For example, the artifact correction device can be a computer customized for near infrared signal data acquisition and near infrared signal data processing tasks, or a server placed in the cloud. For example, the artifact correction device can be integrated into the signal acquisition device.

[0079] The artifact correction apparatus may include a first processor 701 and a memory 704 , and may further include at least one of an input / output 702 and an image display 703 .

[0080] The first processor 701 may be a processing device including one or more general-purpose processing devices, such as a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), etc. More specifically, the first processor 701 may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor running other instruction sets, or a processor running a combination of instruction sets. The first processor 701 may also be one or more special-purpose processing devices, such as an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a system on a chip (SoC), etc. As will be appreciated by those skilled in the art, in some embodiments, the first processor 701 may be a special-purpose processor rather than a general-purpose processor. The first processor 701 may include one or more known processing devices, such as a microprocessor from the Pentium TM, Core TM, Xeon TM, or Itanium series manufactured by Intel TM, a Turion TM, Athlon TM, Sempron TM, Opteron TM, FX TM, Phenom TM series manufactured by AMD TM, or various processors manufactured by Sun Microsystems. The first processor 701 may also include a graphics processing unit, such as a graphics processing unit from GPU manufactured by Nvidia TM series, GMA manufactured by Intel TM, Iris TM series, or Radeon TM series manufactured by AMD TM. The first processor 701 may also include an accelerated processing unit, such as Desktop A-4 (6, 6) series manufactured by AMD TM, Xeon Phi TM series manufactured by Intel TM. The disclosed embodiments are not limited to any type of processor or processor circuit, which is otherwise configured to receive near-infrared signal data; divide the received near-infrared signal data into small segments of signal data of a preset length, and calculate the fluctuation deviation parameters of the signal data of each small segment; sort the calculated fluctuation deviation parameters of the signal data of each small segment from small to large to extract the fluctuation deviation parameters of the signal data of the first predetermined proportion in the sorting; determine the artifact recognition threshold based on the representative fluctuation deviation parameters and threshold multiplier of the fluctuation deviation parameters of the signal data of the first predetermined proportion extracted; compare the deviation of the signal data before and after each small segment with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, identify the signal data of the small segment as containing motion artifacts, or manipulate any other type of signal data consistent with the disclosed embodiments. Additionally, the term "processor" or "image processor" may include more than one processor, for example, a multi-core design or multiple processors, each processor having a multi-core design. The first processor 701 may execute computer program instruction sequences stored in the memory 704 to perform various operations, processes and methods disclosed herein.

[0081] The first processor 701 can be communicatively coupled to the memory 704 and configured to execute computer executable instructions stored therein. The memory 704 may include a read-only memory (ROM), a flash memory, a random access memory (RAM), a dynamic random access memory (DRAM) such as a synchronous DRAM (SDRAM) or a Rambus DRAM, a static memory (e.g., a flash memory, a static random access memory), etc., on which the computer executable instructions are stored in any format. In some embodiments, the memory 704 may store computer executable instructions of one or more medical treatment programs 705. The computer program instructions may be accessed by the first processor 701, read from the ROM or any other suitable memory location, and loaded into the RAM for execution by the first processor 701. For example, the memory 704 may store one or more software applications. The software applications stored in the memory 704 may include, for example, an operating system (not shown) for a general computer system and an operating system for a soft control device.

[0082] Furthermore, the memory 704 may store the entire software application or only a portion of the software application (eg, the medical processing program 705) executable by the first processor 701. Furthermore, the memory 704 may store signal data after a plurality of correction steps.

[0083] In addition, the memory 704 may store signal data generated / buffered when executing a computer program, for example, near-infrared signal data 706, including near-infrared signal data sent from a neonatal resting near-infrared signal database 708, a near-infrared signal data acquisition device 709, etc. In some embodiments, the near-infrared signal data 706 may be neonatal resting near-infrared signal data, and the medical processing program 705 may implement the artifact correction method of the near-infrared signal data in various embodiments of the present application, such as dividing it into small segments of signal data of a preset length, calculating the fluctuation deviation parameters of the signal data of each small segment, sorting the calculated fluctuation deviation parameters, determining an artifact recognition threshold, etc.

[0084] In some embodiments, a near infrared signal data acquisition device 709 may be provided to exchange near infrared signal data with a neonatal resting state near infrared signal database 708, and the memory 704 may communicate with the neonatal resting state near infrared signal database 708 to obtain near infrared signal data to be corrected for artifacts. For example, the near infrared signal data of the neonatal resting state acquired by the near infrared signal data acquisition device 709 may be transmitted and stored in the neonatal resting state near infrared signal database 708, and the artifact correction device may obtain the near infrared signal data from the neonatal resting state near infrared signal database 708 and perform artifact correction on the near infrared signal data.

[0085] In some embodiments, the memory 704 may communicate with the neonatal resting near-infrared signal database 708 to transmit and save the determined artifact recognition threshold together with the artifact correction result of the obtained near-infrared signal data into the neonatal resting near-infrared signal database 708 .

[0086] In addition to displaying the near infrared signal data, the image display 703 can also display other information, such as the second corrected signal data obtained after spline interpolation correction, and the fourth corrected signal data obtained after the peak is replaced by Gaussian white noise. In some embodiments, the image display 703 can be an LCD, CRT or LED display.

[0087] Input / output 702 may be configured to allow the artifact correction device to receive and / or send signal data. Input / output 702 may include one or more digital and / or analog communication devices that allow the artifact correction device to communicate with a user or other machines and devices. For example, input / output 702 may include a keyboard and mouse that allow a user to provide input.

[0088] In some embodiments, the image display 703 may present a user interface so that the user can conveniently and intuitively modify (such as edit, move, modify, etc.) the generated signal data labels by communicating with the user interface using the input / output 702 .

[0089] The interface 707 may include a network adapter, a cable connector, a serial connector, a USB connector, a parallel connector, a high-speed data transmission adapter such as an optical fiber, USB 6.0, Lightning, a wireless network adapter such as a Wi-Fi adapter, a telecommunication (6G, 4G / LTE, etc.) adapter. The artifact correction device may be connected to a network through the interface 707. The network may provide a local area network (LAN), a wireless network, a cloud computing environment (e.g., software as a service, a platform as a service, an infrastructure as a service, etc.), a client-server, a wide area network (WAN), etc.

[0090] In other embodiments, the present application also provides an auxiliary analysis device for brain injury conditions, which is used to assist in the analysis of brain injury conditions in newborns, and the auxiliary analysis device includes: an acquisition module and a second processor, wherein the acquisition module can be configured to acquire near-infrared signal data of the brain of a newborn subject in a resting state. The acquisition module can be implemented in various ways, such as but not limited to using a near-infrared signal data acquisition device. Part of the functions of the second processor are similar to the first processor 701 of the artifact correction device for near-infrared signal data of a newborn in a resting state, and both execute the artifact correction method to perform artifact correction on the collected near-infrared signal data. In addition, the second processor can also process and analyze the near-infrared signal data that has completed artifact correction to assist in the analysis of the brain injury condition of the newborn subject.

[0091] Various operations or functions are described herein, which may be implemented as or defined as software code or instructions. Such content may be source code or differential code ("delta" or "patch" code) that can be directly executed ("object" or "executable" form). Software codes or instructions may be stored in a computer-readable storage medium, and when executed, may cause a machine to perform the functions or operations described, and include any mechanism for storing information in a form accessible to a machine (e.g., a computing device, an electronic system, etc.), such as a recordable or non-recordable medium (e.g., a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium, an optical storage medium, a flash memory device, etc.).

[0092] The foregoing description has been presented for purposes of illustration. It is not exhaustive and is not limited to the precise forms or embodiments disclosed. Modifications and adaptations of the embodiments will be apparent from consideration of the specification and practice of the disclosed embodiments.

[0093] The exemplary methods described herein may be at least partially machine or computer-implemented. Some examples may include a computer-readable medium or machine-readable medium encoded with instructions, which are operable to configure an electronic device to perform the method as described in the above example. The implementation of this method may include software code, such as microcode, assembly language code, higher-level language code, etc. Various programs or program modules may be created using various software programming techniques. For example, Java, Python, C, C++, assembly language, or any known programming language may be used to design program segments or program modules. One or more such software parts or modules may be integrated into a computer system and / or a computer-readable medium. Such software code may include computer-readable instructions for performing various methods. The software code may form a part of a computer program product or a computer program module. In addition, in one example, the software code may be tangibly stored on one or more volatile, non-temporary, or non-volatile tangible computer-readable media, such as during execution or at other times. Examples of these tangible computer-readable media may include, but are not limited to, hard disks, removable disks, removable optical disks (e.g., optical disks and digital video disks), cassettes, memory cards or sticks, random access memory (RAM), read-only memory (ROM), etc.

[0094] In addition, although illustrative embodiments are described herein, the scope includes any and all embodiments with equivalent elements, modifications, omissions, combinations (e.g., combinations of schemes across various embodiments), adjustments, or changes based on the present application. The elements in the claims are to be interpreted broadly based on the language used in the claims, and are not limited to the examples described in this specification or during the continuation of this application, which examples are to be interpreted as non-exclusive. In addition, the steps of the disclosed methods may be modified in any way, including by reordering steps or inserting or deleting steps. Therefore, it is intended that the description be considered as an example only, and the true scope is represented by the following claims and their full range of equivalents.

[0095] The above description is intended to be illustrative rather than restrictive. For example, the above examples (or one or more of their solutions) may be used in combination with each other. Other embodiments may be used by those of ordinary skill in the art when viewing the above description. Moreover, in the above detailed description, various features may be combined together to simplify the application. This should not be interpreted as intending to make the disclosed features that are not required to be protected indispensable for any claim. Therefore, the following claims are thus incorporated into the specific implementation as examples or embodiments, wherein each claim is independent as a separate embodiment, and it is conceivable that these embodiments may be combined with each other in various combinations or permutations. The scope of the present invention should be determined with reference to the appended claims and the full scope of equivalents to these claims.

Claims

1. A method for correcting artifacts of near infrared signal data, characterized in that: include: Receiving near infrared signal data; Divide the received near-infrared signal data into small segments of signal data of a preset length, and calculate the fluctuation deviation parameter of the signal data of each small segment; Sorting the calculated fluctuation deviation parameters of the signal data of each small segment from small to large, so as to extract the fluctuation deviation parameters of the signal data of a first predetermined proportion in the sorting; The artifact recognition threshold is obtained by multiplying the representative fluctuation deviation parameter of the fluctuation deviation parameter of the signal data of the first predetermined proportion extracted by the threshold multiplier; The deviation of the signal data before and after each small segment is compared with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, the signal data of the small segment is recognized as containing motion artifacts.

2. The artifact correction method according to claim 1, characterized in that: Also includes: Fitting the signal data of each small segment identified as containing motion artifacts to obtain fitting signal data; Subtracting corresponding fitted signal data from signal data of each small segment identified as containing motion artifacts to obtain first corrected signal data of each small segment; Based on the preceding segment signal data immediately adjacent to each segment, the first correction signal data of the segment is equalized and adjusted to obtain the second correction signal data of the segment.

3. The artifact correction method according to claim 1 or 2, characterized in that: Also includes: For each peak in the signal data of each small segment identified as containing motion artifacts, corresponding Gaussian white noise signal data is fitted based on the extracted second predetermined proportion of signal data; Each identified peak is replaced with corresponding Gaussian white noise signal data to obtain third corrected signal data of each small segment.

4. The artifact correction method according to claim 3, characterized in that: The artifact correction method further comprises: Performing equalization adjustment on the third correction signal data of each small segment based on the previous segment signal data immediately adjacent to each small segment, so as to obtain the fourth correction signal data of the small segment; Based on the fourth correction signal data of each small segment, equalization adjustment is performed on the subsequent signal data adjacent to each small segment to complete the artifact correction.

5. The artifact correction method according to claim 1, characterized in that: The representative fluctuation deviation parameters of the fluctuation deviation parameters of the signal data of the first predetermined proportion extracted include: The fluctuation deviation parameters of the extracted first predetermined proportion of the signal data are averaged to obtain the representative fluctuation deviation parameter.

6. The artifact correction method according to claim 1, characterized in that: The fluctuation deviation parameter includes a standard deviation value (STD), and the first predetermined proportion in the ranking is the first 20%-first 35% in the ranking.

7. The artifact correction method according to claim 1, characterized in that: Also included before dividing the signal data into small segments of preset length: The received near-infrared signal data is subjected to high-pass filtering within a preset frequency range.

8. A method for correcting artifacts of photoelectric signal data reflecting physiological conditions, wherein the photoelectric signal data is continuously collected, characterized in that: The artifact correction method comprises: Receive photoelectric signal data; Divide the received photoelectric signal data into small segments of signal data of a preset length, and calculate the fluctuation deviation parameter of the signal data of each small segment; Sorting the calculated fluctuation deviation parameters of the signal data of each small segment from small to large, so as to extract the fluctuation deviation parameters of the signal data of a first predetermined proportion in the sorting; The artifact recognition threshold is obtained by multiplying the representative fluctuation deviation parameter of the fluctuation deviation parameter of the signal data of the first predetermined proportion extracted by the threshold multiplier; The deviation of the signal data before and after each small segment is compared with the artifact recognition threshold, and when the deviation exceeds the artifact recognition threshold, the signal data of the small segment is recognized as containing motion artifacts.

9. An artifact correction device for resting near-infrared signal data of a newborn, characterized in that: include: An interface configured to acquire near infrared signal data of a brain of a neonatal subject in a resting state; as well as At least one first processor, configured to execute the artifact correction method according to any one of claims 1-7.

10. A non-transitory computer-readable medium storing instructions, which, when executed by a processor, executes the artifact correction method for near infrared signal data according to any one of claims 1 to 7.

11. A device for assisting analysis of brain injury conditions, used for assisting analysis of brain injury conditions of newborns, characterized in that: The auxiliary analysis device comprises: An acquisition module configured to acquire near-infrared signal data of the brain of a neonatal subject in a resting state; a second processor configured to perform artifact correction on the near infrared signal data by executing the artifact correction method according to any one of claims 1 to 7, and The artifact-corrected near-infrared signal data is processed and analyzed to assist in analyzing the brain injury status of the neonatal subjects.

Images