CN119720057B - Abnormal discharge detection method, device, program product and electronic equipment - Google Patents

Abstract

The embodiments of the present disclosure relate to abnormal discharge detection methods, devices, program products and electronic devices, and to the field of artificial intelligence technology. The above method includes: obtaining a biomedical signal to be detected from a subject, and determining a target category of the biomedical signal to be detected; determining a target abnormal discharge detection model corresponding to the target category from a plurality of pre-trained abnormal discharge detection models according to the target category; processing the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining an abnormal discharge detection result of the subject according to the processing result. The present disclosure improves the efficiency of abnormal discharge detection and recognition and the accuracy of automatic recognition of abnormal discharge detection results.

Description

Abnormal discharge detection method, device, program product and electronic equipment

Technical Field

The present disclosure relates to the field of artificial intelligence, and in particular, to an abnormal discharge detection method, an abnormal discharge detection apparatus, a computer program product, and an electronic device.

Background

This section is intended to provide a background or context for the embodiments of the disclosure recited in the claims, which description herein is not admitted to be prior art by inclusion in this section.

Electroencephalogram is an important basis for judging whether the brain function of a patient is abnormal or not, and has important value for clinical diagnosis and treatment and disease severity assessment. The rapid development of artificial intelligence technology provides possibility for the realization of electroencephalogram automatic analysis, in particular to a neural network technology, which shows better application potential in the aspect of detecting abnormal electroencephalogram discharge.

Disclosure of Invention

However, the automatic detection method of brain electrical anomalies in the related art has insufficient detection accuracy and reliability, and it is difficult to realize truly independent brain electrical anomalies detection.

For this reason, an abnormal discharge detection method is highly required to improve the accuracy and reliability of automatic detection of brain electrical abnormalities.

In this context, it is desirable for embodiments of the present invention to provide an abnormal discharge detection method, an abnormal discharge detection apparatus, a computer-readable storage medium, a computer program product, and an electronic device.

According to a first aspect of the embodiment of the disclosure, an abnormal discharge detection method is provided, which comprises the steps of obtaining a biomedical signal to be detected of a detected object, determining a target class of the biomedical signal to be detected, determining a target abnormal discharge detection model corresponding to the target class from a plurality of pre-trained abnormal discharge detection models according to the target class, processing the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining an abnormal discharge detection result of the detected object according to a processing result.

Optionally, acquiring the biomedical signals to be detected of the tested object comprises acquiring the acquired biomedical signals of the tested object, acquiring the biomedical signals of the tested object in a segmented mode according to preset duration, and determining at least one biomedical signal to be detected of the tested object.

Optionally, the processing the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining the abnormal discharge detection result of the detected object according to the processing result includes processing the biomedical signal to be detected based on the target abnormal discharge detection model, obtaining the abnormal discharge detection result of the biomedical signal to be detected according to the processing result, marking the biomedical signal to be detected when the abnormal discharge detection result of the biomedical signal to be detected is abnormal discharge, and obtaining the abnormal discharge detection result of the detected object according to the marked biomedical signal to be detected.

Optionally, when the target class is a first class, the target abnormal discharge detection model is a first abnormal discharge detection model, the processing of the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining the abnormal discharge detection result of the tested object according to the processing result comprise inputting the biomedical signal to be detected into a first multi-head attention sub-model in the first abnormal discharge detection model, obtaining a first feature of the biomedical signal to be detected according to the output of the first multi-head attention sub-model, inputting the first feature into a first deep convolutional neural network sub-model in the first abnormal discharge detection model, and obtaining the abnormal discharge detection result of the tested object according to the output of the first deep convolutional neural network sub-model.

Optionally, when the target class is a second class, the target abnormal discharge detection model is a second abnormal discharge detection model, the second abnormal discharge detection model comprises a second deep convolutional neural network sub-model, the processing of the biomedical signals to be detected based on the target abnormal discharge detection model is performed, and the obtaining of the abnormal discharge detection result of the detected object according to the processing result comprises the steps of performing form transformation on the biomedical signals to be detected to obtain image-like biomedical signal data corresponding to the biomedical signals to be detected, inputting the image-like biomedical signal data into the second deep convolutional neural network sub-model, and obtaining the abnormal discharge detection result of the detected object according to the output of the second deep convolutional neural network sub-model.

Optionally, the biomedical signals to be detected comprise a first matrix, the first matrix corresponds to a first matrix dimension, the biomedical signals to be detected of the detected object are subjected to form transformation to obtain image-like biomedical signal data corresponding to the biomedical signals to be detected, the first matrix is subjected to matrix transformation to obtain a second matrix, the second matrix corresponds to a second matrix dimension, the difference value between the number of rows and the number of columns of the second matrix dimension is smaller than a preset value, and the image-like biomedical signal data is determined according to the second matrix.

Optionally, any one of the deep convolutional neural network submodels includes any one of a deep-expansion convolutional network submodel and a multi-level convolutional neural network submodel.

According to a second aspect of the embodiment of the present disclosure, there is provided an abnormal discharge detection apparatus, including a target class determination module configured to acquire a biomedical signal to be detected of a subject, determine a target class of the biomedical signal to be detected, a model selection module configured to determine a target abnormal discharge detection model corresponding to the target class from a plurality of pre-trained abnormal discharge detection models according to the target class, and an abnormal detection module configured to process the biomedical signal to be detected based on the target abnormal discharge detection model, and obtain an abnormal discharge detection result of the subject according to a processing result.

According to a third aspect of the present disclosure, there is provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform the steps of the abnormal discharge detection method as in the first aspect.

According to a fourth aspect of embodiments of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the abnormal discharge detection method as described in the first aspect in the above-described embodiments.

According to a fifth aspect of embodiments of the present disclosure, there is provided an electronic device including a processor, and a storage device for storing one or more programs, which when executed by the one or more processors, cause the one or more processors to implement the abnormal discharge detection method as described in the first aspect in the above embodiment.

According to the abnormal discharge detection method, the abnormal discharge detection device, the computer-readable storage medium, the computer program product and the electronic equipment of the embodiment of the disclosure, by classifying biomedical signals, an appropriate target abnormal discharge detection model is determined according to the classification of the biomedical signals, and abnormal discharge identification is performed on the biomedical signals based on the target abnormal discharge detection model. On one hand, the method can realize automatic detection and identification of abnormal discharge of biomedical signals and improve the efficiency of abnormal discharge detection, on the other hand, the method determines the suspected abnormal biomedical signals and the suspected abnormal categories thereof according to the categories of the biomedical signals, selects a target abnormal discharge detection model corresponding to the suspected abnormal biomedical signals, carries out targeted abnormal identification processing on the suspected abnormal biomedical signals of different categories again, improves the accuracy of abnormal discharge detection of the biomedical signals, on the other hand, the abnormal discharge detection result obtained according to the method of the method can provide accurate reference information for doctors, so that the doctors do not need to carefully observe all data, only browse some key information, and then combine the abnormal discharge detection result to rapidly and accurately judge the diagnosis result of a doctor-seeing object, thereby saving the workload of the doctors and being convenient for wide application in clinical practice.

Drawings

The above, as well as additional purposes, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description when read in conjunction with the accompanying drawings. Several embodiments of the present invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which:

FIG. 1 illustrates a schematic diagram of an exemplary system architecture to which embodiments of the present disclosure may be applied;

fig. 2 shows a flow diagram of an abnormal discharge detection method in an exemplary embodiment of the present disclosure;

FIG. 3 illustrates a flow diagram of a method of acquiring biomedical signals to be detected in an exemplary embodiment of the present disclosure;

FIG. 4 is a flow chart of a method for obtaining an abnormal discharge detection result of a test object in an exemplary embodiment of the present disclosure;

FIG. 5 illustrates a flow chart of a method of determining an abnormal discharge detection result of a first type of biomedical signal to be detected in an exemplary embodiment of the present disclosure;

fig. 6 illustrates a schematic structure of a first abnormal discharge detection model in an exemplary embodiment of the present disclosure;

FIG. 7 illustrates a flow chart of a method of determining an abnormal discharge detection result of a second class of biomedical signals to be detected in an exemplary embodiment of the present disclosure;

Fig. 8 shows a composition diagram of an abnormal discharge detection apparatus in an exemplary embodiment of the present disclosure;

fig. 9 shows a schematic structural diagram of an electronic device in an exemplary embodiment of the present disclosure.

In the drawings, the same or corresponding reference numerals indicate the same or corresponding parts.

Detailed Description

The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are presented merely to enable those skilled in the art to better understand and practice the invention and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

It will be appreciated by those skilled in the art that embodiments of the present invention may be implemented as a system, apparatus, device, medium, method or computer program product. Accordingly, the present invention may be embodied in the form of hardware entirely, software (including firmware, resident software, micro-code, etc.), or in a combination of hardware and software.

According to an embodiment of the present invention, there are provided an abnormal discharge detection method, an abnormal discharge detection device, a computer program product quality, and an electronic apparatus.

Any number of elements in the figures are for illustration and not limitation, and any naming is used for distinction only, and not for any limiting sense.

The principles and spirit of the present invention are described in detail below with reference to several representative embodiments thereof.

Summary of The Invention

The inventors of the present disclosure found that the abnormal discharge detection method in the related art has a problem that the accuracy and reliability of detection of abnormal discharge of biomedical signals such as brain electricity are low.

In view of the foregoing, a basic idea of the present disclosure is to provide an abnormal discharge detection method, apparatus, program product, and electronic device, by classifying biomedical signals, determining an appropriate target abnormal discharge detection model according to the classification of the biomedical signals, and performing abnormal discharge identification on the biomedical signals based on the target abnormal discharge detection model. According to the abnormal discharge detection result obtained by the method, more accurate reference information can be provided for doctors, so that the doctors do not need to carefully observe all data, only some key information is roughly browsed, and then the abnormal discharge detection result is combined, diagnosis results of a doctor can be rapidly and accurately judged, the workload of the doctor is saved, and the method is convenient to widely apply in clinical practice.

Having described the basic principles of the present invention, various non-limiting embodiments of the invention are described in detail below.

Application scene overview

It should be noted that the following application scenarios are only shown for facilitating understanding of the spirit and principles of the present invention, and embodiments of the present invention are not limited in this respect. Rather, embodiments of the invention may be applied to any scenario where applicable.

The embodiment of the disclosure can be applied to a human brain abnormal discharge detection scene, for example, when a subject feels uncomfortable in the head, such as headache, the brain electric signals of the subject can be collected, then the collected brain electric signals of the subject are classified and identified, a target abnormal discharge detection model corresponding to the target category is determined from a plurality of pre-trained abnormal discharge detections according to the determined target category of the brain electric signals of the subject, then the collected brain electric signals are input into the target abnormal discharge detection model, and an abnormal discharge detection result of the subject is obtained according to the output of the target abnormal discharge detection model. And feeding back the abnormal discharge detection result of the object to be diagnosed to a doctor so as to assist the doctor to quickly and accurately make clinical diagnosis on the object to be diagnosed according to the abnormal discharge detection result and other key index data, thereby reducing the workload of the doctor and improving the working efficiency of the doctor.

Exemplary System architecture

First, a system architecture of an exemplary application environment of the present disclosure will be described with reference to fig. 1.

As shown in fig. 1, the system architecture 100 may include a terminal device 110 and a server 120. The terminal device 110 may be a terminal device such as a smart phone, a tablet computer, a desktop computer, a notebook computer, or an intelligent wearable device, and the server 120 generally refers to a background system that provides services related to the abnormal discharge detection method in the present exemplary embodiment, and may be a server or a cluster formed by multiple servers. The terminal device 110 and the server 120 may form a connection through a wired or wireless communication link for data interaction.

In an exemplary embodiment, the abnormal discharge detection method described above may be performed by the server 120. Accordingly, the abnormal discharge detecting apparatus may be provided in the server 120 to implement the corresponding module function. For example, biomedical signals of a detected object are collected through an electroencephalograph, the collected biomedical signals are sent to the server 120, the server 120 performs segmentation processing on the received biomedical signals to obtain biomedical signals to be detected, then a target class of the biomedical signals to be detected is determined through a pre-trained classification model, then a target abnormal discharge detection model is determined according to the target class, the biomedical signals to be detected are input into the target abnormal discharge detection model, and an abnormal discharge detection result of the detected object is determined according to the output of the target abnormal discharge detection model. After determining the abnormal discharge detection result, the server 120 may also send the abnormal discharge detection result to a client where the relevant person is located, such as a doctor's client.

In another exemplary embodiment, the above-described abnormal discharge detection method may also be performed by the terminal device 110. Accordingly, the abnormal discharge detecting apparatus may be provided in the terminal device 110 to implement the corresponding module function. For example, the user uploads the collected biomedical signals of the tested object to the terminal device 110, the terminal device 110 determines a target class of the biomedical signals uploaded by the user according to a classification model pre-configured in the terminal device 110, determines a target abnormal discharge detection model from a plurality of pre-trained abnormal discharge detection models configured in the terminal device 110 according to the target class, inputs the biomedical signals to the target abnormal discharge detection model, determines an abnormal discharge detection result according to an output of the target abnormal discharge detection model, and displays the determined abnormal discharge detection result in a graphical user interface of the terminal device 110.

It should be understood that the number of terminal devices and servers in fig. 1 is merely illustrative. There may be any number of terminal devices and servers, as desired for implementation. For example, the server 120 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, basic cloud computing services such as big data and artificial intelligence platforms, and the like.

It is to be understood by those skilled in the art that the above application scenario is merely for example, and the present exemplary embodiment is not limited thereto.

Exemplary method

Fig. 2 shows a flow chart of an abnormal discharge detection method in an exemplary embodiment of the present disclosure, and referring to fig. 2, the method includes:

step S210, obtaining biomedical signals to be detected of a detected object, and determining target categories of the biomedical signals to be detected;

step S220, determining a target abnormal discharge detection model corresponding to the target category from a plurality of pre-trained abnormal discharge detection models according to the target category;

and step S230, processing the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining an abnormal discharge detection result of the detected object according to a processing result.

Next, first, a detailed description will be given of a specific embodiment of "step S210, acquiring a biomedical signal to be detected of a subject, and determining a target class of the biomedical signal to be detected".

In an exemplary embodiment, the biomedical signal is an active signal spontaneously generated by a physiological process of a living body, and can more accurately reflect a physiological state or a physical sign state of a person, so that whether the living body is abnormal or not can be determined through the biomedical signal. The living body in the present disclosure may include any living body such as a human body, an animal body, and the like, and the present exemplary embodiment is not particularly limited thereto.

The biomedical signals may include any one or more of electrocardiosignals, electroencephalogram signals, electromyographic signals, electrooculographic signals, gastric signals and other physiological signals, or may also include body temperature, blood pressure, pulse, respiration and other non-electrogenic signals, which may be set by themselves according to practical situations, and the disclosure is not limited in particular.

The electrocardiograph may be an electrocardiogram acquired by a multichannel electrocardiograph, and the Electrocardiogram (ECG) refers to a graph in which the heart is excited by a pacing point, an atrium and a ventricle sequentially in each cardiac cycle, and along with the bioelectric change of the electrocardiogram, various potential changes are led out of the body surface by the electrocardiograph. Electrocardiography is an objective indicator of the occurrence, spread, and recovery process of cardiac excitation.

The electroencephalogram signal can be an electroencephalogram acquired by using a multichannel electroencephalogram machine, wherein the electroencephalogram (electroencephalogram, EEG) is a graph obtained by amplifying and recording spontaneous bioelectric potentials of brain cortex of brain complement from scalp through a precise instrument, and is spontaneous and rhythmic electric activities of brain cell groups recorded through electrodes. The electrical activity is a plan view of the relationship between the recorded potential and time, with the potential on the vertical axis and the time on the horizontal axis. The frequency (period), amplitude and phase of the brain waves constitute the fundamental features of the electroencephalogram.

Myoelectric signals (Electromyography, EMG) are a temporal and spatial superposition of the action potentials of the motor units in a multitude of muscle fibers, which can be obtained by applying myoelectric sensors to the skin.

The eye electrical signal (Electrooculogram, EOG) is a bioelectric signal caused by the potential difference between the cornea and retina of the eye, and is very convenient to collect and can be completed through a small number of electrodes.

Gastric electrical signals ((Electrogastrogram, EGG) are electrical signals generated by gastric muscle contractions, which can be collected on the surface of the human abdominal skin using electrodes.

In an alternative embodiment, the subject may include a patient with epilepsy or epileptic-like symptoms, and the subject may include other people with uncomfortable brain conditions for diagnosis, such as people with pain in the brain, and the exemplary embodiment is not particularly limited.

By way of example, fig. 3 shows a flow diagram of a method of acquiring biomedical signals to be detected in an exemplary embodiment of the present disclosure. Referring to fig. 3, the method may include steps S310 to S320. Wherein:

in step S310, an acquired biomedical signal of the subject is acquired.

For example, for each subject, a biomedical signal may be acquired for a long period of time, such as 10 minutes. Since the abnormal discharge is not continuous, the biomedical signals of the detected object can be processed in a segmented mode to obtain biomedical signals to be detected, whether the abnormal discharge exists in the detected object or not is determined according to the biomedical signals to be detected, and the efficiency and the accuracy of abnormal discharge detection are improved.

In step S320, biomedical signals of the tested object are obtained in segments according to a preset duration, and at least one biomedical signal to be detected of the tested object is determined.

Taking the preset duration of 4 seconds as an example, the biomedical signals collected in the step S210 may be segmented in units of 4 seconds, and each segmented signal with the length of 4 seconds is taken as a biomedical signal to be detected, so that 15 biomedical signals to be detected can be obtained in 1 minute, and 150 biomedical signals to be detected can be obtained in 10 minutes.

In another exemplary embodiment, the biomedical signal collected by the biomedical signal device may be directly used as the biomedical signal to be detected without segmentation, for example, the biomedical signal may be directly collected for 1 minute as required, and the collected biomedical signal for 1 minute is used as the biomedical signal to be detected, which is not particularly limited in this exemplary embodiment.

For example, the biomedical signal acquisition device directly acquires a biomedical signal digital sequence. For example, the multichannel electrocardiograph, the multichannel electroencephalograph and the multichannel electromyograph can be used for acquiring corresponding biomedical signals at a preset acquisition frequency, such as 500HZ, and the biomedical signal numerical sequence is obtained.

For example, the electroencephalogram signals collected by the electroencephalogram machine can be an electroencephalogram signal numerical sequence shown in the following formula (1).

x_t=(x_t1,x_t2,x_t3,x_t4,...,x_tn) (1)

In formula (1), x_t is an electroencephalogram signal numerical sequence. x_t1, x_t2, x_t3, x_t4, x_tn represents signal values of the electroencephalogram at different sampling moments, for example x_t1 represents signal values of the 1 st sampling moment in the sequence of electroencephalogram values x_t, and x_tn represents signal values of the n-th sampling moment in the sequence of electroencephalogram values x_t.

As can be seen from the formula (1), the form of the original biomedical signal acquired by the biomedical signal acquisition device can be regarded as a one-dimensional or multi-dimensional long matrix, for example, the biomedical signal to be detected is a 29-by-2000-dimensional long matrix, which is obtained by taking the multi-channel electroencephalogram signal as an example, and taking the number of channels of the multi-channel electroencephalogram signal as 29 and taking total sampling for 4 seconds as 2000 times. The abnormal discharge detection model is usually a neural network model, particularly a convolutional neural network, which has natural processing advantages on the image signal, and can extract deeper feature expression of the image signal when the image signal is processed, wherein the image signal is mostly a matrix of a plurality of rows and columns with basically equal or no great difference between the rows and columns, such as 256 times 256,512 times 512, and the like. In other words, the original biomedical signal resembles a rectangular input, and the image signal resembles a square input, which have a large difference in form, so that the neural network, especially the convolutional neural network, cannot fully exert the advantage of feature extraction in abnormal discharge detection.

That is, under the condition that the form of the input biomedical signal of the abnormal discharge detection model is similar to that of the image signal, the advantages of the neural network, particularly the convolutional neural network, in the aspect of feature extraction can be fully applied to the field of abnormal discharge detection, the biomedical signal features of a deeper level are extracted, and the accuracy of biomedical signal detection is improved.

The biomedical signal is time-dependent, as described above, and is a time sequence, the time information is lost after the biomedical signal is converted into a signal in an image form, and for some abnormal discharge signals with a large dependence on the time information, such as a phenomenon that some objects have continuous abnormal discharge during discharge, the abnormal discharge signals are converted into a signal in an image form, and although other deep features are extracted, the time information is lost, so that the accuracy of a detection result is reduced when the abnormal detection is performed by using a convolutional neural network or the like.

However, for some abnormal discharge signals with lower dependence on time information, such as sudden transient abnormal discharge of some objects, the degree of dependence on time is lower, and after the abnormal discharge signals are converted into signals in an image form, the influence on the abnormal discharge signals is not very great although the time information is lost, but the processing advantage of the convolutional neural network on the signals in the image form is extracted, so that the detection accuracy is improved.

Based on the above, in the disclosure, biomedical signals may be classified, a biomedical signal suspected to be abnormal and a type of suspected to be abnormal are determined according to the classification result, then, according to the characteristics of the suspected type of suspected to be abnormal biomedical signals, a proper abnormal discharge detection mode is selected, and abnormal confirmation is performed on the suspected to be abnormal biomedical signals, so that the accuracy and reliability of abnormal discharge detection are improved.

In an exemplary embodiment, a classification model of the biomedical signal may be trained in advance, the biomedical signal to be detected is input into the classification model, and a target class of the biomedical signal to be detected is determined according to an output of the classification model, so that whether the biomedical signal to be detected is a biomedical signal suspected to be abnormal and a class of the suspected abnormality are determined according to the target class of the biomedical signal to be detected.

In an exemplary embodiment, the categories of biomedical signals in the present disclosure may include a first category, a second category, and a third category. The first class characterizes biomedical signals with suspected abnormalities that have a high or a high degree of dependence on time information, the second class characterizes biomedical signals with suspected abnormalities that do not have a low or a high degree of dependence on time information, and the third class characterizes normal biomedical signals.

For example, the first data set, the second data set and the third data set may be determined by means of manual labeling, where the first data set corresponds to the first class, i.e. the biomedical signals in the first data set are all in the first class, the second data set corresponds to the second class, and the third data set corresponds to the third class, e.g. the class of biomedical signals is marked manually according to the discharge characteristics of the existing biomedical signals. If the biomedical signal with continuous abnormal discharge exists, the classification thereof is marked as a first classification, the biomedical signal with short-term abnormal discharge is marked as a second classification, and the biomedical signal without abnormal discharge is marked as a third classification, thereby obtaining a training data set of the classification model.

It should be noted that, the specific training manner of the classification model and the loss function in the training process may refer to the related content of the existing machine learning model capable of being used for classification, which is not described herein.

The classification model may include any machine learning model capable of implementing classification, such as a support vector machine, a decision tree model, and the like, which is not particularly limited in the present exemplary embodiment.

After the classification model is trained, the classification model is mainly used for classifying biomedical signals to be detected, and as mentioned above, the classified categories can include a first category, a second category and a normal category, wherein the first category and the second category belong to abnormal categories, but belong to different abnormal conditions. The biomedical signals similar to the normal biomedical signals or the abnormal biomedical signals can be rapidly determined through the classification model, so that the biomedical signals suspected to be abnormal and the suspected abnormal categories corresponding to the biomedical signals suspected to be abnormal are rapidly determined. If the biomedical signal to be detected is the first type, the biomedical signal to be detected is the biomedical signal suspected to be abnormal, the type of the biomedical signal suspected to be abnormal is the first type, the biomedical signal to be detected is the second type, the type of the biomedical signal suspected to be abnormal is the second type, and the biomedical signal to be detected is the third type, the biomedical signal to be detected is the normal biomedical signal.

In an exemplary embodiment, the classification model may be a simple machine learning model, and the main purpose of the classification model is to perform a rapid preliminary classification on biomedical signals, determine biomedical signals suspected of abnormalities according to the classification result, and determine the type of suspected abnormalities. In the subsequent steps, an abnormal discharge detection model corresponding to different suspected abnormal categories is used for extracting deeper features so as to further confirm whether the abnormality exists or not, and therefore the accuracy of abnormality detection of the biomedical signals is improved.

Next, a detailed description will be given of a specific embodiment of "step S220 in which the target abnormal discharge detection model corresponding to the target category is determined from a plurality of pre-trained abnormal discharge detection models according to the target category".

For example, a plurality of pre-trained abnormal discharge detection models may be derived from different training data. For example, a first abnormal discharge detection model may be obtained by performing supervised learning training on one machine learning model via the first data set and the third data set, and a second abnormal discharge detection model may be obtained by performing supervised learning training on the other machine learning model via the second data set and the third data set.

When the abnormal discharge detection model is trained, the labels of the biomedical signals in the first data set, the second data set and the third data set are all abnormal discharge labels, namely, the abnormal discharge labels are used for indicating whether abnormal discharge exists in the biomedical signals, if the biomedical signals in the first data set, the second data set and the third data set are all biomedical signals in the existing data set, the labels can be provided with the labels whether the abnormal discharge exists, and if the labels are not the biomedical signals in the existing data set, the abnormal discharge labels corresponding to the biomedical signals in the first data set, the second data set and the third data set can be marked in a manual marking mode, such as a doctor reading mode.

It should be noted that, the specific embodiment of training to obtain the first abnormal discharge detection model according to the biomedical signals in the first data set and the third data set and the corresponding abnormal discharge labels thereof, and training to obtain the second abnormal discharge detection model according to the biomedical signals in the second data set and the third data set and the corresponding abnormal discharge labels thereof may refer to the training mode of the existing machine learning model, and will not be described herein.

For example, the first abnormal discharge detection model is obtained by training according to the first data set and the third data set, the class label of the biomedical signal in the first data set is a first class, the second abnormal discharge detection model is obtained by training according to the second data set and the third data set, and the class label of the biomedical signal in the second data set is a second class, so the first class and the first abnormal discharge detection model can be associated in advance, and the second class and the second abnormal discharge detection model can be associated in advance, thereby obtaining a mapping relation table between the biomedical signal class and the pre-trained abnormal discharge detection model.

According to the mapping relation table, the target abnormal discharge detection model of the biomedical signals of the first category is determined to be the first abnormal discharge detection model, and the target abnormal discharge detection model of the biomedical signals of the second category is determined to be the second abnormal discharge detection model. In other words, when the target class is the first class, the first abnormal discharge detection model may be determined as the target abnormal discharge detection model based on the map table, and when the target class is the second class, the second abnormal discharge detection model may be found as the target abnormal discharge detection model based on the map table.

In one exemplary embodiment, the abnormal discharge detection model in the present disclosure is determined from a convolutional neural network. For example, the first abnormal discharge detection model corresponding to the first category is a first convolutional neural network, and the second abnormal discharge detection model corresponding to the second category is a second convolutional neural network.

In another exemplary embodiment, since the biomedical signals of the first category have a strong dependency on the time information, the time information features of the biomedical signals of the first category may be extracted through the multi-headed attention model. Based on this, the first abnormal discharge detection model may be composed of a first multi-headed attention sub-model and a first deep convolutional neural network sub-model, and the second abnormal discharge detection model may include a second deep convolutional neural network. Since the biomedical signals of the second category have a low dependency on time information, the second abnormal discharge detection model may be composed of only one second deep convolutional neural network submodel.

In an exemplary embodiment, any one of the deep convolutional neural network submodels includes any one of a deep-inflated convolutional network submodel and a multi-level convolutional neural network submodel. That is, both the first deep neural network sub-model and the second deep convolutional neural network sub-model may include any one of a deep-expansion convolutional network sub-model and a multi-level convolutional neural network sub-model. Of course, any deep convolutional neural network submodel may also include other convolutional network types, which the exemplary embodiment does not specifically limit.

Next, a detailed description will be given of a specific embodiment of "step S230, in which the biomedical signal to be detected is processed based on the target abnormal discharge detection model, and the abnormal discharge detection result of the test object is obtained according to the processing result".

Illustratively, fig. 4 shows a flowchart of a method for obtaining an abnormal discharge detection result of a test object in an exemplary embodiment of the present disclosure. Referring to fig. 4, the method may include steps S410 to S430. Wherein:

in step S410, the biomedical signal to be detected is processed based on the target abnormal discharge detection model, and the abnormal discharge detection result of the biomedical signal to be detected is obtained according to the processing result.

For example, the biomedical signal to be detected may be input into the target abnormal discharge detection model, the biomedical signal to be detected is processed, and an abnormal discharge detection result of whether the biomedical signal to be detected has abnormal discharge or not is obtained according to the output of the target abnormal discharge detection model.

In step S420, in the case that the abnormal discharge detection result of the biomedical signal to be detected is abnormal discharge, the biomedical signal to be detected is marked.

For example, in the case that the biomedical signal to be detected has abnormal discharge, an abnormal mark, such as a first mark, may be added to the biomedical signal to be detected, and since the biomedical signal to be detected is obtained in a segmented manner, the abnormal discharge of the detected object in which time period can be located through the abnormal mark, so that the doctor can conveniently perform subsequent diagnosis and confirmation.

For example, in the case where the abnormal discharge detection result of the biomedical signal to be detected is a normal discharge, it may not be marked at all, or a normal mark such as a second mark may be added thereto.

In step S430, an abnormal discharge detection result of the test object is obtained according to the labeled biomedical signal to be detected.

For example, the time periods corresponding to the marked biomedical signals to be detected can be summarized to obtain each time period when the detected object has abnormal discharge, so that each time period when the detected object has abnormal discharge is used as an abnormal discharge detection result and fed back to the client where the doctor is located. The doctor can quickly and accurately position the time period in which the abnormal discharge is possible according to the abnormal discharge detection result, and research and read the biomedical signal image in the time period in which the abnormal discharge is possible to confirm whether the abnormal discharge is true or not. Therefore, a doctor does not need to read the whole biomedical signal image, and only needs to confirm the abnormality of the biomedical signal image with the abnormal time period summarized in the detection result, thereby improving the working efficiency of the doctor.

By way of example, fig. 5 shows a flow chart of a method of determining an abnormal discharge detection result of a first class of biomedical signals to be detected in an exemplary embodiment of the present disclosure. Referring to fig. 5, the method may include steps S510 to S520.

In step S510, a biomedical signal to be detected is input into a first multi-headed attention sub-model in a first abnormal discharge detection model, and a first characteristic of the biomedical signal to be detected is obtained from an output of the first multi-headed attention sub-model.

For example, in the case that the biomedical signal to be detected is of the first type, the target abnormal discharge detection model is the first abnormal discharge detection model described above, and as described above, the first abnormal discharge detection model includes a first multi-head attention sub-model and a first deep convolutional neural network sub-model. The original biomedical signal to be detected may be input into a first multi-headed attention sub-model in the first abnormal discharge detection model, and the first feature of the biomedical signal to be detected may be extracted through the first multi-headed attention sub-model.

In an exemplary embodiment, QKV of the first multi-headed attention sub-model are identical, i.e. are both the original signals of the biomedical signal to be detected. Wherein Q, K, V correspond to a Query, key and Value matrix in the multi-headed attention model, respectively.

The specific structure of the first multi-headed attention sub-model in the present disclosure may refer to the structure of a multi-headed attention model in the existing natural language processing field, and will not be described herein.

After the original signal of the biomedical signal to be detected passes through the first multi-head attention sub-model, the first characteristic output by the first multi-head attention sub-model is a shallow characteristic, and the shallow biomedical signal characteristic comprises a time information characteristic but cannot represent the deeper signal meaning of the biomedical signal.

In step S520, a first feature of the biomedical signal to be detected is input into a first deep convolutional neural network sub-model in a first abnormal discharge detection model, and an abnormal discharge detection result of the tested object is obtained according to an output of the first deep convolutional neural network sub-model.

In one exemplary embodiment, as previously described, the first deep convolutional neural network submodel may include any one of a first deep-inflated convolutional network submodel and a first multi-level convolutional neural network submodel.

In the case that the first deep convolutional neural network model is a first deep convolutional neural network sub-model, since the original biomedical signal, such as the above-mentioned electroencephalogram signal 29 times 2000, is less sensitive to convolution, the convolution expansion rates of the layers in the first deep convolutional neural network sub-model may be the same, such as 1, i.e., the first deep convolutional neural network sub-model may be a common multi-layer convolutional neural network sub-model, so that the processing efficiency of the model may be improved while the processing effect is ensured. Of course, the expansion rates of the layers of the first deep-expansion convolutional network submodel may also be different, which is not particularly limited in the present exemplary embodiment.

The number of layers and the expansion rate of each layer of the first deep expansion convolutional network submodel can be determined according to requirements or experimental results, for example, experimental results show that when the number of layers is 3 and the expansion rate of each layer is 1, the training result is the best, and then the number of layers of the first deep expansion convolutional network submodel can be 3 and the expansion rate of each layer is 1.

In the case where the first deep convolutional neural network sub-model is a first multi-level convolutional neural network sub-model, the convolutional kernels of each layer of the first multi-level convolutional neural network sub-model may be the same, but the number of convolutional kernels of each layer is different. The convolution kernels of each layer of the first multi-level convolutional neural network submodel may also be different, which is not particularly limited in the present exemplary embodiment.

An embodiment of step S520 may include inputting a first feature of the biomedical signal to be detected into a first deep convolutional neural network submodel in a first abnormal discharge detection model, obtaining a second feature of the biomedical signal to be detected according to an output of the first deep convolutional neural network submodel, inputting the second feature of the biomedical signal to be detected into a first full-connection layer, obtaining the first full-connection feature according to an output of the first full-connection layer, processing the first full-connection feature through a preset activation function, and obtaining an abnormal discharge detection result of the tested object according to a processing result.

For example, the structure of the first abnormal discharge detection model may be as shown in fig. 6, that is, the first abnormal discharge detection model is composed of a first multi-head attention sub-model 61, a first deep convolutional neural network sub-model 62, and a first full connection layer 63.

In actual abnormal discharge detection, a first type of biomedical signal to be detected may be input into the first multi-head attention sub-model 61, the biomedical signal to be detected is processed through the first multi-head attention sub-model, a shallow feature of the biomedical signal to be detected is obtained according to the output of the first multi-head attention sub-model, the shallow feature is input into the first deep convolutional neural network sub-model 62, a deeper biomedical signal feature is extracted based on the first deep convolutional neural network sub-model, a deep feature is obtained according to the output of the first deep convolutional neural network sub-model, then the first full-connection layer 63 is used for dimension reduction of the deep feature, the feature after dimension reduction of the full-connection layer is mapped through a preset activation function, such as a binary activation function sigmoid, so as to obtain an abnormal probability, when the abnormal probability is greater than a preset value, abnormal discharge exists in the biomedical signal to be detected, otherwise, abnormal discharge is determined not to exist, so as to obtain an abnormal discharge detection result of the biomedical signal to be detected. As described above, the abnormal discharge detection result of the test object is obtained according to the abnormal discharge detection result of the biomedical signal to be detected according to the method shown in fig. 4.

In an exemplary embodiment, when the first abnormal discharge detection model determines that abnormal discharge exists, it is determined that abnormal discharge exists in the biomedical signal to be detected, that is, an abnormal discharge detection result of the biomedical signal to be detected is abnormal, and when the first abnormal discharge detection model determines that abnormal discharge does not exist, it is determined that the biomedical signal to be detected is suspected abnormal discharge, a suspected abnormal discharge label is not added, and a doctor can perform manual abnormal confirmation on the biomedical signal to be detected of the suspected abnormal discharge label.

For example, when the biomedical signal to be detected is in the second category, the abnormal discharge detection result of the test object may be obtained by the method shown in fig. 7. Referring to fig. 7, a method of determining an abnormal discharge detection result of a second class of biomedical signals to be detected may include steps S710 to S730. Wherein:

In step S710, the biomedical signal to be detected is subjected to form transformation to obtain image-like biomedical signal data corresponding to the biomedical signal to be detected.

In an exemplary embodiment, the biomedical signal to be detected comprises a first matrix, which corresponds to the first matrix dimension. For example, the biomedical signal to be detected may be represented by a first matrix, which may be a matrix of N rows and M columns, such as the above-mentioned matrix of 29 rows by 2000 columns. N represents the number of channels of the biomedical signal, 2000 represents the number of samples per channel of the biomedical signal to be detected.

Based on this, an exemplary embodiment of step S710 may include performing matrix deformation on the first matrix to obtain a second matrix, where the second matrix corresponds to a second matrix dimension, and a difference between a number of rows and a number of columns of the second matrix dimension is smaller than a preset value, and determining image-like biomedical signal data according to the second matrix.

For example, the first matrix may be converted into the second matrix of K rows and L columns by a matrix conversion function, such as reshape function, where the absolute value of the difference between K and L is smaller than a preset value, the preset value may be determined by user definition according to the requirement, such as 0,50,100, etc., where the preset value cannot be set too large, otherwise, the difference between the converted matrix and the image data matrix may be too large, and the preset value is set to make the converted matrix more suitable for performing convolution operation, such as converting the first matrix of 29 times 2000 into the second matrix of 290 times 200, so that it may be ensured that the form of the second matrix is similar to the form of the image signal, and the second matrix is used as the image-like biomedical signal of the biomedical signal to be detected.

It should be noted that the process of converting from the first matrix to the second matrix is only that the row-column dimension of the matrix is changed, but the elements in the matrix are not changed.

In step S720, the image-like biomedical signal data is input into the second deep convolutional neural network submodel, and the abnormal discharge detection result of the tested object is obtained according to the output of the second deep convolutional neural network submodel.

In an exemplary embodiment, in the case that the biomedical signal to be detected is of the second type, the target abnormal discharge detection model is the second abnormal discharge detection model described above, and as described above, the second abnormal discharge detection model includes a second deep convolutional neural network submodel. The biomedical signal data of the similar image can be input into the second depth convolution neural network submodel through the second depth convolution neural network submodel, deep feature extraction is carried out on the biomedical signal data of the similar image, and whether the biomedical signal to be detected really has abnormality or not is determined according to the deeper features.

In one exemplary embodiment, as previously described, the second deep convolutional neural network submodel may include any one of a second deep-inflated convolutional network submodel and a second deep multi-level convolutional neural network submodel.

In the case that the second deep convolutional neural network sub-model is a second deep convolutional network sub-model, the second deep convolutional network sub-model may be a multi-layer convolutional neural network with different expansion rates of each layer, such as a 3-layer convolutional neural network with expansion rates of 1, 2, and 3, respectively. Thus, through the second deep expansion convolutional network submodel, deep feature extraction can be performed on the image-like biomedical signal data, and richer biomedical signal features are obtained.

In the case where the second deep convolutional neural network submodel is a second multi-level convolutional neural network submodel, the convolutional kernels of each layer of the second multi-level convolutional neural network submodel may be different, for example, the first layer includes 3 convolutional kernels 1, the second layer includes 2 convolutional kernels 2, and the third layer includes 1 convolutional kernel 3, wherein the convolutional kernels 1,2, and 3 are different from each other, and the present exemplary embodiment is not particularly limited thereto.

An exemplary embodiment of step S730 may include inputting the image-like biomedical signal data into a second deep convolutional neural network sub-model, obtaining a third characteristic of the image-like biomedical signal data according to an output of the second deep convolutional neural network sub-model, inputting the third characteristic of the image-like biomedical signal data into a second full-connection layer, obtaining a second full-connection characteristic according to an output of the second full-connection layer, processing the second full-connection characteristic through a preset activation function, and obtaining an abnormal discharge detection result of the object under test according to a processing result.

For example, other embodiments of step S720 may refer to the related content of step S520, which is not described herein.

For example, when the second abnormal discharge detection model determines that the biomedical signal to be detected has abnormal discharge, it determines that the biomedical signal to be detected has abnormal discharge, when the second abnormal discharge detection model determines that the biomedical signal to be detected does not have abnormal discharge, it determines that the biomedical signal to be detected is suspected to have abnormal discharge, a suspected abnormal discharge label is added to the biomedical signal to be detected, and a doctor can perform artificial abnormal confirmation on the biomedical signal to be detected of the suspected abnormal discharge label, that is, the doctor determines whether the biomedical signal to be detected of the suspected abnormal label has abnormal actually, the doctor considers that the biomedical signal has abnormal actually, and the doctor considers that the biomedical signal does not have abnormal, so that the biomedical signal does not have abnormal.

In another exemplary embodiment, for the biomedical signal to be detected, which is determined to be the third category in step S210, i.e., the normal biomedical signal to be detected, it may be directly determined that there is no abnormal discharge. Because the dependence information of the normal biomedical signal to be detected on time is low, the normal biomedical signal to be detected can be converted into similar image data and then input into a second abnormal discharge detection model, whether the normal biomedical signal to be detected is true or not is determined through the second abnormal discharge detection model, if the normal biomedical signal to be detected is determined to be true through the second abnormal discharge detection model, otherwise, the normal biomedical signal to be detected is determined to be suspected to be normal, a suspected normal label is added to the normal biomedical signal to be detected, and a doctor can manually and normally confirm the biomedical signal to be detected, to which the suspected normal label is added.

In the method, biomedical signals are classified according to the characteristics of the biomedical signals, biomedical signals suspected to be abnormal and the types of the suspected abnormalities can be primarily and rapidly determined according to the classification results, the types of the biomedical signals which are more suitable for feature extraction by using the deep neural network can be converted into image-like biomedical signal data and then processed by using a corresponding abnormal discharge detection model, so that the advantages of the deep neural network model in biomedical signal processing are fully excavated, abnormal conditions are confirmed more accurately, the detection results of abnormal discharge of the biomedical signals are more accurate and reliable, and more accurate auxiliary diagnosis information is provided for doctors.

Furthermore, it should be noted that the above-described figures are merely illustrative of the processes involved in the method according to the exemplary embodiment of the present invention, and are not intended to be limiting. It will be readily appreciated that the processes shown in the above figures do not indicate or limit the temporal order of these processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, for example, among a plurality of modules.

Exemplary apparatus

The exemplary embodiments of the present disclosure also provide an abnormal discharge detection apparatus. Referring to fig. 8, the abnormal discharge detection apparatus 800 may include a target class determination module 810 configured to acquire a biomedical signal to be detected of a subject, determine a target class of the biomedical signal to be detected, a model selection module 820 configured to determine a target abnormal discharge detection model corresponding to the target class from a plurality of pre-trained abnormal discharge detection models according to the target class, and an abnormal detection module 830 configured to process the biomedical signal to be detected based on the target abnormal discharge detection model, and obtain an abnormal discharge detection result of the subject according to a processing result.

In an exemplary embodiment, acquiring biomedical signals to be detected of a tested object comprises acquiring the acquired biomedical signals of the tested object, acquiring the biomedical signals of the tested object in sections according to preset time length, and determining at least one biomedical signal to be detected of the tested object.

In an exemplary embodiment, the abnormality detection module 830 may be specifically configured to process a biomedical signal to be detected based on the target abnormal discharge detection model, obtain an abnormal discharge detection result of the biomedical signal to be detected according to a processing result, mark the biomedical signal to be detected if the abnormal discharge detection result of the biomedical signal to be detected is abnormal discharge, and obtain the abnormal discharge detection result of the tested object according to the marked biomedical signal to be detected.

In an exemplary embodiment, when the target class is a first class, the target abnormal discharge detection model is a first abnormal discharge detection model, the processing the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining the abnormal discharge detection result of the tested object according to the processing result includes inputting the biomedical signal to be detected into a first multi-head attention sub-model in the first abnormal discharge detection model, obtaining a first feature of the biomedical signal to be detected according to the output of the first multi-head attention sub-model, inputting the first feature into a first deep convolutional neural network sub-model in the first abnormal discharge detection model, and obtaining the abnormal discharge detection result of the tested object according to the output of the first deep convolutional neural network sub-model.

In an exemplary embodiment, when the target class is the second class, the target abnormal discharge detection model is a second abnormal discharge detection model, the second abnormal discharge detection model comprises a second deep convolutional neural network sub-model, the processing of the biomedical signal to be detected based on the target abnormal discharge detection model is performed, the obtaining of the abnormal discharge detection result of the detected object according to the processing result comprises the steps of performing form transformation on the biomedical signal to be detected to obtain image-like biomedical signal data corresponding to the biomedical signal to be detected, inputting the image-like biomedical signal data into the second deep convolutional neural network sub-model, and obtaining the abnormal discharge detection result of the detected object according to the output of the second deep convolutional neural network sub-model.

In an exemplary embodiment, the biomedical signals to be detected include a first matrix, the first matrix corresponds to a first matrix dimension, the biomedical signals to be detected of the detected object are subjected to form transformation to obtain image-like biomedical signal data corresponding to the biomedical signals to be detected, the first matrix is subjected to matrix transformation to obtain a second matrix, the second matrix corresponds to a second matrix dimension, the difference value between the number of rows and the number of columns of the second matrix dimension is smaller than a preset value, and the image-like biomedical signal data is determined according to the second matrix.

In an exemplary embodiment, any of the deep convolutional neural network submodels includes any of a deep-inflated convolutional network submodel and a multi-level convolutional neural network submodel.

The specific details of each part of the apparatus are already described in the above embodiments of the corresponding method part, and the details not disclosed can be referred to the above embodiments of the method part, so that they will not be repeated.

Exemplary storage Medium

A storage medium according to an exemplary embodiment of the present invention will be described below.

In the present exemplary embodiment, the above-described method may be implemented by a program product, such as a portable compact disc read only memory (CD-ROM) and including program code, and may be run on a device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of a readable storage medium include an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The computer readable signal medium may include a data signal propagated in baseband or as part of a carrier wave with readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RE, etc., or any suitable combination of the foregoing.

Program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., connected via the Internet using an Internet service provider).

Exemplary computer program product

Exemplary embodiments of the present disclosure also provide a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the abnormal discharge detection method described above.

In one embodiment, the computer program product may be a tangible product comprising a computer program, such as a computer readable storage medium having the computer program stored thereon. The readable storage medium may be a storage medium based on signals of electric, magnetic, optical, electromagnetic, infrared, etc., including, but not limited to, random Access Memory (RAM), read Only Memory (ROM), magnetic tape, floppy disk, flash memory (Flash), mechanical hard disk (HDD), solid State Disk (SSD), etc. By way of example, the computer program product may be embodied as a non-volatile storage medium, such as read-only memory, nand flash memory (NAND FLASH), or the like, in which the computer program is stored.

In one embodiment, the computer program product may be an intangible product containing a computer program. By way of example, the computer program product may be embodied as a virtual digital product, such as a digital file, an executable file storing a computer program, an installation package, or the like.

The code of the computer program may be written in one or more programming languages. Programming languages such as C language, java, c++, python, etc. The program code may execute entirely on the user's computing device, or partly on the user's computing device, or as a stand-alone software package, or partly on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, such as a Local Area Network (LAN), wide Area Network (WAN), etc., or may be connected to an external computing device (e.g., an Internet connection provided by an operator).

The computer program may be carried or transmitted by signals of electronic, magnetic, optical, electromagnetic, infrared, etc. The electronic device may convert a signal carrying the computer program into a digital signal, thereby running the computer program. When the computer program runs on the electronic device, the code of the computer program is used for enabling the electronic device to execute (more specifically, enabling a processor of the electronic device to execute) the method steps of various exemplary embodiments of the present disclosure, such as the method for detecting abnormal discharge, can be executed, and the method comprises the steps of obtaining a biomedical signal to be detected of a detected object, determining a target class of the biomedical signal to be detected, determining a target abnormal discharge detection model corresponding to the target class from a plurality of pre-trained abnormal discharge detection models according to the target class, processing the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining an abnormal discharge detection result of the detected object according to a processing result.

The method comprises the steps of executing the steps of the method through a computer program, on one hand, automatically detecting and identifying abnormal discharge of biomedical signals according to an abnormal discharge detection model, improving the efficiency of abnormal discharge detection, on the other hand, determining biomedical signals with suspected abnormalities according to the categories of the biomedical signals, selecting a target abnormal discharge detection model corresponding to the categories of the biomedical signals with suspected abnormalities, carrying out targeted processing on biomedical signals of different categories, improving the accuracy of abnormal discharge detection of the biomedical signals, on the other hand, providing accurate reference information for doctors, enabling the doctors to not need to carefully observe all data, only approximately browsing some key information, combining with the abnormal discharge detection results, rapidly and accurately judging the diagnosis results of the doctor, saving the workload of the doctor, and being convenient for wide application in clinical practice.

Exemplary electronic device

An electronic device of an exemplary embodiment of the present disclosure is described with reference to fig. 9. The electronic device is the terminal device 110 or the server 120 described above. The electronic device may include a processor and a memory. The memory stores executable instructions of the processor, such as may be a computer program. The processor performs the method steps of the various exemplary embodiments of the present disclosure by executing the executable instructions. The electronic device may further comprise a display for displaying the graphical user interface.

An electronic device is illustrated in the form of a general purpose computing device with reference to fig. 9. It should be understood that the electronic device 900 illustrated in fig. 9 is merely an example and should not be construed as limiting the functionality and scope of use of embodiments of the present disclosure.

As shown in FIG. 9, electronic device 900 may include a processor 910, memory 920, bus 930, I/O (input/output) interface 940, network adapter 950, and display 960.

The memory 920 may include volatile memory, such as RAM 921, a cache unit 922, and may also include nonvolatile memory, such as ROM 923. Memory 920 may also include one or more program modules 924, such program modules 924 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment. For example, program modules 924 may include the modules in the apparatus described above.

The Processor 910 may include one or more processing units, for example, the Processor 910 may include an AP (Application Processor ), modem Processor, GPU (Graphics Processing Unit, graphics Processor), ISP (IMAGE SIGNAL Processor ), controller, encoder, decoder, DSP (DIGITAL SIGNAL Processor ), baseband Processor and/or NPU (Neural-Network Processing Unit, neural network Processor), etc.

The processor 910 may be configured to execute executable instructions stored in the memory 920, for example, may execute the above-mentioned abnormal discharge detection method, where the method includes the steps of obtaining a biomedical signal to be detected of a detected object, determining a target class of the biomedical signal to be detected, determining a target abnormal discharge detection model corresponding to the target class from a plurality of pre-trained abnormal discharge detection models according to the target class, processing the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining an abnormal discharge detection result of the detected object according to a processing result.

According to the method, on one hand, automatic detection and identification of abnormal discharge of biomedical signals can be achieved according to the abnormal discharge detection model, the abnormal discharge detection efficiency is improved, on the other hand, according to the categories of biomedical signals, a target abnormal discharge detection model corresponding to the categories of biomedical signals is selected, the biomedical signals of different categories are processed in a targeted mode, the accuracy of abnormal discharge detection of the biomedical signals is improved, on the other hand, accurate reference information can be provided for doctors, the doctors do not need to carefully observe all data, only key information is generally browsed, and then diagnosis results of a doctor can be judged quickly and accurately according to the abnormal discharge detection results, so that the method is convenient to use widely in clinical practice.

The bus 930 is used to facilitate connections between the different components of the electronic device 900 and may include a data bus, an address bus, and a control bus.

The electronic device 900 may communicate with one or more external devices 1000 (e.g., keyboard, mouse, external controller, etc.) through an I/O interface 940.

The electronic device 900 may communicate with one or more networks through a network adapter 950, e.g., the network adapter 950 may provide a mobile communication solution such as 3G/4G/5G, or a wireless communication solution such as wireless local area network, bluetooth, near field communication, etc. The network adapter 950 may communicate with other modules of the electronic device 900 via the bus 930.

The electronic device 900 may display a graphical user interface, such as a graphical user interface displaying the abnormal discharge detection result, etc., through the display 960.

Although not shown in FIG. 9, other hardware and/or software modules may also be provided in electronic device 900, including, but not limited to, microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

It is to be understood that the disclosure is not limited to the particular process steps or structures described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. Other embodiments will readily occur to those skilled in the art based on the specific embodiments provided by this disclosure. Accordingly, the detailed description provided herein is exemplary only, and the scope and spirit of the present disclosure is indicated by the appended claims, to encompass any modifications, uses, or adaptations of the disclosure following the general principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains.

Claims (9)

1. A method for detecting abnormal discharge, comprising: Acquire a biomedical signal to be detected from a subject, and determine a target category of the biomedical signal to be detected; According to the target category, determining a target abnormal discharge detection model corresponding to the target category from a plurality of pre-trained abnormal discharge detection models; Based on the target abnormal discharge detection model, the biomedical signal to be detected is processed, and an abnormal discharge detection result of the detected object is obtained according to the processing result; The target categories include a first category and a second category, the first category characterizing suspected abnormal biomedical signals whose to-be-detected biomedical signals are dependent on time information, and the second category includes suspected abnormal biomedical signals whose to-be-detected biomedical signals are not dependent on time information; Among them, the processing of the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining the abnormal discharge detection result of the object under test according to the processing result includes: when the target category is the first category, directly inputting the biomedical signal to be detected into the first abnormal discharge detection model, and obtaining the abnormal discharge detection result of the object under test according to the output of the first abnormal discharge detection model, and the first abnormal discharge detection model includes a first multi-head attention sub-model and a first deep convolutional neural network sub-model; when the target category is the second category, transforming the form of the biomedical signal to be detected to obtain image-like biomedical signal data corresponding to the biomedical signal to be detected, inputting the image-like biomedical signal data into the second abnormal discharge detection model, and obtaining the abnormal discharge detection result of the object under test according to the output of the second abnormal discharge detection model, and the second abnormal discharge detection model includes a second deep convolutional neural network sub-model. 2. The method according to claim 1, wherein obtaining the biomedical signal to be detected of the subject comprises: Acquiring the collected biomedical signals of the subject; The biomedical signals of the subject are acquired in segments according to a preset time length, and at least one biomedical signal to be detected of the subject is determined. 3. The method according to claim 2, characterized in that the processing of the biomedical signal to be detected based on the target abnormal discharge detection model and obtaining the abnormal discharge detection result of the detected object according to the processing result comprises: Based on the target abnormal discharge detection model, the biomedical signal to be detected is processed, and an abnormal discharge detection result of the biomedical signal to be detected is obtained according to the processing result; When the abnormal discharge detection result of the biomedical signal to be detected is abnormal discharge, marking the biomedical signal to be detected; The abnormal discharge detection result of the detected object is obtained according to the marked biomedical signal to be detected. 4. The method according to claim 1, characterized in that: The step of directly inputting the biomedical signal to be detected into a first abnormal discharge detection model and obtaining an abnormal discharge detection result of the detected object according to an output of the first abnormal discharge detection model comprises: Inputting the biomedical signal to be detected into a first multi-head attention sub-model in the first abnormal discharge detection model, and obtaining a first feature of the biomedical signal to be detected according to an output of the first multi-head attention sub-model; The first feature is input into a first deep convolutional neural network sub-model in the first abnormal discharge detection model, and an abnormal discharge detection result of the object under test is obtained according to an output of the first deep convolutional neural network sub-model. 5. The method according to claim 4, characterized in that the biomedical signal to be detected includes a first matrix, the first matrix corresponds to a first matrix dimension, and the form transformation of the biomedical signal to be detected to obtain the image-like biomedical signal data corresponding to the biomedical signal to be detected includes: Performing matrix deformation on the first matrix to obtain a second matrix, where the second matrix corresponds to a second matrix dimension, and a difference between the number of rows and the number of columns of the second matrix dimension is less than a preset value; The image-like biomedical signal data is determined according to the second matrix. 6. The method according to claim 1 is characterized in that any of the deep convolutional neural network sub-models includes any one of a deep dilated convolutional network sub-model and a multi-layer convolutional neural network sub-model. 7. An abnormal discharge detection device, comprising: A target category determination module is configured to obtain a biomedical signal to be detected from a subject and determine a target category of the biomedical signal to be detected; A model selection module is configured to determine, according to the target category, a target abnormal discharge detection model corresponding to the target category from a plurality of pre-trained abnormal discharge detection models; an abnormality detection module, configured to process the biomedical signal to be detected based on the target abnormal discharge detection model, and obtain an abnormal discharge detection result of the detected object according to the processing result; The target categories include a first category and a second category, the first category characterizing suspected abnormal biomedical signals whose to-be-detected biomedical signals are dependent on time information, and the second category includes suspected abnormal biomedical signals whose to-be-detected biomedical signals are not dependent on time information; Among them, the processing of the biomedical signal to be detected based on the target abnormal discharge detection model, and obtaining the abnormal discharge detection result of the object under test according to the processing result includes: when the target category is the first category, directly inputting the biomedical signal to be detected into the first abnormal discharge detection model, and obtaining the abnormal discharge detection result of the object under test according to the output of the first abnormal discharge detection model, and the first abnormal discharge detection model includes a first multi-head attention sub-model and a first deep convolutional neural network sub-model; when the target category is the second category, transforming the form of the biomedical signal to be detected to obtain image-like biomedical signal data corresponding to the biomedical signal to be detected, inputting the image-like biomedical signal data into the second abnormal discharge detection model, and obtaining the abnormal discharge detection result of the object under test according to the output of the second abnormal discharge detection model, and the second abnormal discharge detection model includes a second deep convolutional neural network sub-model. 8. A computer program product, comprising a computer program, characterized in that when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented. 9. An electronic device, comprising: one or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, enables the one or more processors to implement the method according to any one of claims 1 to 6.