Efficient identification method and system for airway secretions

Abstract

The invention provides an efficient identification method and system for airway secretion, and the method comprises the steps: obtaining the breath sound data of a user and the historical secretion cleaning information of the user, and recognizing the secretion quality of the user and the correlation information of the audio feature information of the user based on the historical secretion cleaning information; in a medical care database, querying body type structure data of the user, and constructing lung secretion distribution information of the user based on the body type structure data and the breath sound data of the user; and extracting current audio features of the breath sound data through an audio feature extraction network, and identifying airway secretion information of the user based on the associated information, the current audio features and the lung secretion distribution information. By adopting the scheme, the positioning efficiency and the secretion identification accuracy of the tracheal secretions of the user can be improved.

Description

Technical Field

[0001] The present invention relates to the technical fields of artificial intelligence and audio analysis, and particularly relates to an efficient recognition method and system for airway secretions. Background Art

[0002] During the medical care process, especially when dealing with the secretions of users, for users who are unable to handle them by themselves, medical staff need to manually clean the airway secretions of the users. Therefore, the importance of accurately positioning the location of the secretions is gradually increasing. Due to the high severity of ICU users, sedation, analgesia, and assisted ventilation are often used, and the clearance of their airway secretions is more frequent. Therefore, how to efficiently locate the secretions on the basis of accurately positioning the secretions is the current research focus.

[0003] Traditional techniques are used for positioning through imaging diagnosis, fiberoptic bronchoscopy, and auscultation of sonorous rales. However, imaging diagnosis is difficult to apply to the clearance of airway secretions every time in clinical work, with low efficiency and long diagnosis time. The auscultation of the user's sonorous rales mainly relies on the subjective judgment and work experience of nursing staff. The nursing staff take measures such as postural drainage and sputum suction based on the sonorous rales heard, and cannot accurately locate and judge the sputum suction position, resulting in poor positioning efficiency for the user's tracheal secretions. Summary of the Invention

[0004] The main purpose of the present invention is to provide an efficient recognition method and system for airway secretions, aiming to solve the problems in the prior art that imaging diagnosis is difficult to apply to the clearance of airway secretions every time in clinical work, with low efficiency and long diagnosis time. The auscultation of the user's sonorous rales mainly relies on the subjective judgment and work experience of nursing staff. The nursing staff take measures such as postural drainage and sputum suction based on the sonorous rales heard, and cannot accurately locate and judge the sputum suction position, resulting in poor positioning efficiency for the user's tracheal secretions.

[0005] To achieve the above purpose, the present invention provides an efficient recognition method for airway secretions, and the method includes:

[0006] Obtain the respiratory sound data of the user and the historical secretion clearance information of the user, and based on the historical secretion clearance information, identify the association information between the secretion quality of the user and the audio feature information of the user;

[0007] In the medical care database, query the body structure data of the user, and based on the body structure data and the respiratory sound data of the user, construct the pulmonary secretion distribution information of the user;

[0008] Extract the current audio features of the breath sound data through an audio feature extraction network, and identify the airway secretion information of the user based on the association information, the current audio features, and the pulmonary secretion distribution information.

[0009] Optionally, the identifying the association information between the secretion quality of the user and the audio feature information of the user based on the historical secretion cleaning information includes:

[0010] Based on the historical secretion cleaning information, identify the secretion information cleaned by the user each time and the audio information of the user each time;

[0011] Extract the audio feature information in each of the audio information through an audio feature extraction network, and identify the secretion feature values of each secretion feature type in each of the secretion information;

[0012] For each secretion feature type, construct the secretion-audio association distribution information of the secretion feature type based on the secretion feature values of the secretion feature type and each of the audio feature information;

[0013] Based on the secretion-audio association distribution information, identify the sub-association information between the secretion feature type and the audio feature information, and use all the sub-association information as the association information between the secretion quality of the user and the audio feature information of the user.

[0014] Optionally, the constructing the pulmonary secretion distribution information of the user based on the body shape structure data and the breath sound data of the user includes:

[0015] Construct the torso structure model of the user based on the body shape structure data, and divide the breath sound data into sub-breath sound data at each three-dimensional position point;

[0016] Based on the sub-breath sound data at each three-dimensional position point, generate the breath sound distribution information corresponding to each three-dimensional position point through the torso structure model, and perform torso audio fitting processing on the breath sound distribution information corresponding to each three-dimensional position point in the torso structure model to obtain the torso audio distribution information of the user;

[0017] Construct the pulmonary secretion distribution information of the user based on the torso audio distribution information of the user.

[0018] Optionally, the constructing the pulmonary secretion distribution information of the user based on the torso audio distribution information of the user includes:

[0019] Based on the torso audio distribution information, identify the sub-audio distribution information of each voice type of the user, and construct a three-dimensional sub-audio distribution map corresponding to each sub-audio distribution information;

[0020] For each three-dimensional sub-audio distribution map, through the voice anomaly analysis strategy of the voice type corresponding to the three-dimensional sub-audio distribution map, identify the abnormal voice range in the three-dimensional sub-audio distribution map and the abnormal degree distribution information of the abnormal voice range;

[0021] Based on the abnormal voice ranges of the three-dimensional sub-audio distribution maps, screen the target pulmonary secretion range of the user, and based on the abnormal degree distribution information of the abnormal voice ranges of the three-dimensional sub-audio distribution maps, generate the secretion concentration distribution information of the target pulmonary secretion range through a distribution fitting strategy;

[0022] Take the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range as the pulmonary secretion distribution information of the user.

[0023] Optionally, the extracting the current audio features of the breath sound data through the audio feature extraction network includes:

[0024] In the breath sound data, screen the target sub-breath sound data collected at the user's mouth, and based on the target sub-breath sound data, identify the voice data of each voice type;

[0025] Through the sub-audio feature extraction networks of each voice type, respectively extract the voice features in the voice data of each voice type, and take the voice features of all voice types as the current audio features of the breath sound data.

[0026] Optionally, the identifying the airway secretion information of the user based on the association information, the current audio features, and the pulmonary secretion distribution information includes:

[0027] Based on the current audio features and the audio feature information corresponding to each association information, through the cosine similarity recognition algorithm, calculate the similarity values between the current audio features and each audio feature information, and screen the association information corresponding to the audio feature information with the maximum similarity value as the target association information;

[0028] Based on the secretion information in the target association information, identify the secretion characteristics of the user and the secretion content range of the user, and generate the secretion image guidance information of the user based on the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range;

[0029] Take the secretion characteristics of the user, the range of the user's secretion content, and the secretion image guidance information of the user as the airway secretion information of the user.

[0030] In addition, to achieve the above object, the present invention also provides an efficient recognition system for airway secretions. The efficient recognition system for airway secretions includes:

[0031] An acquisition module, configured to acquire the breath sound data of the user and the historical secretion cleaning information of the user, and based on the historical secretion cleaning information, identify the association information between the secretion quality of the user and the audio feature information of the user;

[0032] A construction module, configured to query the body structure data of the user in the medical staff database, and based on the body structure data and the breath sound data of the user, construct the pulmonary secretion distribution information of the user;

[0033] An identification module, configured to extract the current audio features of the breath sound data through an audio feature extraction network, and based on the association information, the current audio features, and the pulmonary secretion distribution information, identify the airway secretion information of the user.

[0034] Optionally, the acquisition module is specifically configured to:

[0035] Based on the historical secretion cleaning information, identify the secretion information cleaned by the user each time and the audio information of the user each time;

[0036] Extract the audio feature information in each of the audio information through an audio feature extraction network, and identify the secretion feature values of each secretion feature type in each of the secretion information;

[0037] For each secretion feature type, based on the secretion feature values of the secretion feature type and each of the audio information, construct the secretion audio association distribution information of the secretion feature type;

[0038] Based on the secretion audio association distribution information, identify the sub-association information between the secretion feature type and the audio feature information, and use all the sub-association information as the association information between the secretion quality of the user and the audio feature information of the user.

[0039] Optionally, the construction module is specifically configured to:

[0040] Based on the body structure data, construct the torso structure model of the user, and divide the breath sound data into sub-breath sound data of each three-dimensional position point;

[0041] Based on the sub-breathing sound data of each of the three-dimensional position points, through the torso structure model, generate the breathing sound distribution information corresponding to each three-dimensional position point, and perform torso audio fitting processing on the breathing sound distribution information corresponding to each three-dimensional position point in the torso structure model to obtain the torso audio distribution information of the user;

[0042] Based on the torso audio distribution information of the user, construct the pulmonary secretion distribution information of the user.

[0043] Optionally, the construction module is specifically configured to:

[0044] Based on the torso audio distribution information, identify the sub-sound distribution information of each sound type of the user, and construct a three-dimensional sub-sound distribution map corresponding to each sub-sound distribution information;

[0045] For each three-dimensional sub-sound distribution map, through the sound anomaly analysis strategy of the sound type corresponding to the three-dimensional sub-sound distribution map, identify the abnormal sound range in the three-dimensional sub-sound distribution map and the abnormal degree distribution information of the abnormal sound range;

[0046] Based on the abnormal sound ranges of the three-dimensional sub-sound distribution maps, screen the target pulmonary secretion range of the user, and based on the abnormal degree distribution information of the abnormal sound ranges of the three-dimensional sub-sound distribution maps, generate the secretion concentration distribution information of the target pulmonary secretion range through a distribution fitting strategy;

[0047] Take the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range as the pulmonary secretion distribution information of the user.

[0048] Optionally, the identification module is specifically configured to:

[0049] In the breathing sound data, screen the target sub-breathing sound data collected at the user's mouth, and based on the target sub-breathing sound data, identify the sound data of each sound type;

[0050] Through the sub-audio feature extraction networks of each sound type, respectively extract the sound features in the sound data of each sound type, and take the sound features of all sound types as the current audio features of the breathing sound data.

[0051] Optionally, the identification module is specifically configured to:

[0052] Based on the current audio feature and the audio feature information corresponding to each piece of the associated information, through the cosine similarity recognition algorithm, calculate the similarity values between the current audio feature and each piece of the audio feature information, and screen out the associated information corresponding to the audio feature information with the maximum similarity value as the target associated information;

[0053] Based on the secretion information in the target associated information, identify the secretion characteristics of the user and the range of the user's secretion content, and generate the secretion image guidance information of the user based on the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range;

[0054] Use the secretion characteristics of the user, the range of the user's secretion content, and the secretion image guidance information of the user as the airway secretion information of the user.

[0055] In a third aspect, the present application provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method described in any one of the first aspects are implemented.

[0056] In a fourth aspect, the present application provides a computer-readable storage medium. A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method described in any one of the first aspects are implemented.

[0057] In a fifth aspect, the present application provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the method described in any one of the first aspects are implemented.

[0058] The present invention provides an efficient identification method and system for airway secretions. The method includes: obtaining the breath sound data of a user and the user's historical secretion cleaning information, and based on the historical secretion cleaning information, identifying the association information between the secretion quality of the user and the user's audio feature information; querying the body structure data of the user in a medical database, and based on the body structure data and the user's breath sound data, constructing the pulmonary secretion distribution information of the user; extracting the current audio features of the breath sound data through an audio feature extraction network, and based on the association information, the current audio features, and the pulmonary secretion distribution information, identifying the airway secretion information of the user. This solution improves the accuracy of identifying the characteristics and properties of the user's airway secretions by combining the user's historical secretion cleaning information to identify the association information between the secretion quality and the audio feature information of the user. Then, this solution combines the user's body structure data and the collected breath sound data of the user to construct the pulmonary secretion distribution information of the user, so that it is not necessary for medical staff to judge and identify the location of pulmonary secretions through experience and intuition, and thus can directly guide the medical staff to the location information of the user's current pulmonary secretions, improving the accuracy of positioning the user's pulmonary secretions. Then, this solution extracts the secretion quality associated with the current audio features of the user by a factor of ten through feature extraction, and based on this secretion quality and the obtained pulmonary secretion distribution information, identifies the airway secretion information of the user. Without the need for scanning and identification by complex equipment or the calculation and processing of complex neural networks, the airway secretion information of the user can be accurately obtained, thus greatly improving the accuracy of identifying the secretion and the positioning efficiency of the user's tracheal secretions. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the solutions in the present application, the following will briefly introduce the drawings required for the description of the embodiments of the present application. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0060] Figure 1 is a flowchart of the efficient identification method for airway secretions provided by an embodiment of the present invention;

[0061] Figure 2 is a schematic structural diagram of the efficient identification system for airway secretions provided by an embodiment of the present invention;

[0062] Figure 3 is an internal structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] The efficient recognition method for airway secretions provided by the embodiments of the present invention is applied to an efficient recognition system for airway secretions. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above drawings are intended to cover non-exclusive inclusion. The terms "first", "second", etc. in the specification and claims of this application or the above drawings are used to distinguish different objects and not to describe a specific order.

[0064] Referring to "embodiments" herein means that specific features, structures, or characteristics described in connection with the embodiments may be included in at least one embodiment of this application. The phrase appearing in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0065] In order to enable those skilled in the technical field to better understand the solutions of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0066] The efficient recognition method for airway secretions provided by the embodiments of the present application can be applied to the application environment of efficient recognition of airway secretions. Among them, this method can be applied to a terminal, a server, or a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. Among them, the terminal can be, but is not limited to, various personal computers, laptop computers, etc. Among them, the terminal combines the historical secretion cleaning information of the user to identify the association information between the secretion quality of the user and the audio feature information, improving the recognition accuracy of the characteristics and properties of the airway secretions of the user. Then, this solution combines the body structure data of the user and the collected breath sound data of the user to construct the distribution information of the pulmonary secretions of the user, so that there is no need for medical staff to judge and identify the location of the pulmonary secretions through experience and intuition, and thus can directly guide the medical staff to the location information of the current pulmonary secretions of the user, improving the positioning accuracy of the pulmonary secretions of the user. Then, this solution extracts features to tenfold the secretion quality associated with the current audio features of the user, and based on this secretion quality and the obtained distribution information of the pulmonary secretions, identifies the airway secretion information of the user. Without the scanning and recognition of complex equipment and the calculation and processing of complex neural networks, the airway secretion information of the user can be accurately obtained, thus greatly improving the positioning efficiency and secretion recognition accuracy of the tracheal secretions of the user.

[0067] In one embodiment, as Figure 1 shown, a method for efficiently recognizing airway secretions is provided. Taking the application of this method to a terminal as an example, the method includes the following steps:

[0068] Step S101, obtain the breath sound data of the user and the historical secretion cleaning information of the user, and based on the historical secretion cleaning information, identify the association information between the secretion quality of the user and the audio feature information of the user.

[0069] In this embodiment, the terminal collects the breathing sound data of the user at each three-dimensional position point on the user's chest, mouth, and back through an audio collection device. Among them, the three-dimensional position points are the position points preset in the terminal. The terminal obtains the sub-breathing sound data of each three-dimensional position point of the user by responding to the information upload operation of the staff, and obtains the breathing sound data of the user. Then, the terminal queries the sub-breathing sound data collected at the user's mouth and the secretion information of the user cleared each time the airway secretion is cleared in the user's medical database of the user, and obtains the historical secretion clearance information of the user. Among them, the secretion information is the characteristic content of the secretion recorded by the medical staff, for example, information such as the volume of the secretion, the color of the secretion, the weight of the secretion, the state of the secretion, the viscosity level of the secretion, and the smell of the secretion. Then, the terminal identifies the association information between the secretion quality of the user and the audio feature information of the user based on the historical secretion clearance information. Among them, the secretion quality is used to cover the characteristic data representing each secretion characteristic type of the secretion. The secretion characteristic types include, but are not limited to, the secretion volume characteristic type, the secretion color characteristic type, the secretion weight characteristic type, the secretion state characteristic type, the secretion viscosity level characteristic type, and the secretion smell characteristic type. And the audio feature information includes the sound characteristics of each sound type, and the sound types include, but are not limited to, the pitch type, the timbre type, and the loudness type. The specific identification process will be described in detail later.

[0070] Step S102, query the body structure data of the user in the medical database, and construct the pulmonary secretion distribution information of the user based on the body structure data and the breathing sound data of the user.

[0071] In this embodiment, the terminal queries the body structure data of the user in the medical database. Among them, the body structure data is the three-dimensional structure data corresponding to the user's torso and neck. Then, the terminal constructs the pulmonary secretion distribution information of the user based on the body structure data and the breathing sound data of the user. Among them, the pulmonary secretion distribution information includes the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range. The specific construction process will be described in detail later.

[0072] Step S103, extract the current audio features of the breathing sound data through an audio feature extraction network, and identify the airway secretion information of the user based on the association information, the current audio features, and the pulmonary secretion distribution information.

[0073] In this embodiment, the terminal extracts the current audio features of the breath sound data through an audio feature extraction network, and identifies the airway secretion information of the user based on the association information, the current audio features, and the pulmonary secretion distribution information. Among them, the audio feature extraction network includes sub-audio feature extraction networks for each sound type. Among them, the sub-audio feature extraction network is a feature extraction neural network based on a Recurrent Neural Network (RNN) trained with sample sound data of different sound types. The specific extraction process will be described in detail later. The airway secretion information includes the secretion characteristics of the user, the range of the user's secretion content, and the secretion image guidance information of the user. The specific identification process will be described in detail later.

[0074] Based on the above solution, by combining the user's historical secretion cleaning information, the association information between the secretion quality of the user and the audio feature information is identified, improving the recognition accuracy of the characteristics and properties of the user's airway secretions. Then, this solution combines the user's body structure data and the collected breath sound data of the user to construct the pulmonary secretion distribution information of the user, so that there is no need for medical staff to judge and identify the location of pulmonary secretions through experience and intuition, and thus the current pulmonary secretion location information of the user can be directly guided to the medical staff, improving the positioning accuracy of the user's pulmonary secretions. Then, through the feature extraction method, the secretion quality associated with the user's current audio features is multiplied by ten, and based on the secretion quality and the obtained pulmonary secretion distribution information, the airway secretion information of the user is identified. Without the scanning and identification of complex equipment and the calculation and processing of complex neural networks, the airway secretion information of the user can be accurately obtained, thus greatly improving the positioning efficiency and secretion recognition accuracy of the user's tracheal secretions.

[0075] Optionally, identifying the association information between the secretion quality of the user and the audio feature information of the user based on the historical secretion cleaning information includes: identifying the secretion information of each user cleaning and the audio information of each user cleaning based on the historical secretion cleaning information; extracting the audio feature information in each audio information through the audio feature extraction network, and identifying the secretion feature values of each secretion feature type in each secretion information; for each secretion feature type, constructing the secretion-audio association distribution information of the secretion feature type based on the secretion feature values of the secretion feature type and each audio feature information; based on the secretion-audio association distribution information, identifying the sub-association information between the secretion feature type and the audio feature information, and taking all the sub-association information as the association information between the secretion quality of the user and the audio feature information of the user.

[0076] In this embodiment, the terminal splits the historical secretion cleaning information into secretion information for each cleaning by medical staff for the user, and extracts the respiratory audio information of the user collected by the medical staff during each cleaning of the user. Since the mouth respiratory audio information in the user's respiratory audio information is clearer and more distinct, the respiratory audio information of the user collected by the medical staff screened here is the respiratory audio information collected from the user's mouth.

[0077] The terminal respectively performs feature extraction processing on the sound data of each sound type in the respiratory audio information through the sub-audio feature extraction network of each sound type to obtain the audio feature information in each audio information. Then, the terminal splits each secretion information into sub-secretion information of each secretion feature type, and extracts the corresponding secretion data in each sub-secretion information to obtain the secretion feature values of each secretion feature type. Among them, the secretion feature value, for example, the volume value of the secretion volume feature type, the color value of the secretion color feature type, the weight value of the secretion weight feature type, the state type of the secretion state feature type, the viscosity level of the secretion viscosity level feature type, and the odor type of the secretion odor feature type, etc.

[0078] For each secretion feature type, the terminal constructs the secretion-audio association distribution information of the secretion feature type based on the secretion feature values of the secretion feature type and the audio feature information. The construction method includes that the terminal, for each secretion feature type, based on the secretion feature value of the secretion feature type and its corresponding audio feature information, performs distribution arrangement processing on the secretion feature value collected each time and the audio feature information to obtain the secretion-audio association table, and uses this secretion-audio association table as the secretion-audio association distribution information.

[0079] The terminal identifies the association information between the change information of the feature value of the secretion feature type and the change information of the audio feature information based on the secretion-audio association distribution information to obtain the sub-association information between the secretion feature type and the audio feature information. Among them, the association information, for example, the higher the proportion of the solid state in the secretion state, the greater the loudness, the lower the pitch, and the more turbid the timbre. The higher the proportion of the liquid state in the secretion state, the smaller the loudness, the higher the pitch, and the clearer the timbre.

[0080] Finally, the terminal uses all the sub-association information as the association information between the secretion quality of the user and the audio feature information of the user.

[0081] Based on the above solution, by splitting and analyzing the historical secretion cleaning information, feature extraction, and association information recognition, the association information between the secretion quality of the user and the audio feature information of the user is obtained, which improves the comprehensiveness and accuracy of the recognition of the association information between the secretion and the breathing sound of the user.

[0082] Optionally, based on the body structure data and the user's breath sound data, the pulmonary secretion distribution information of the user is constructed, including: based on the body structure data, a torso structure model of the user is constructed, and the breath sound data is divided into sub-breath sound data at each three-dimensional position point; based on the sub-breath sound data at each three-dimensional position point, through the torso structure model, the breath sound distribution information corresponding to each three-dimensional position point is generated, and the breath sound distribution information corresponding to each three-dimensional position point is subjected to torso audio fitting processing in the torso structure model to obtain the torso audio distribution information of the user; based on the torso audio distribution information of the user, the pulmonary secretion distribution information of the user is constructed.

[0083] In this embodiment, the terminal constructs a torso structure model of the user through a three-dimensional modeling program based on the body structure data, and divides the breath sound data into sub-breath sound data at each three-dimensional position point. Among them, each three-dimensional position point is the epidermal position point of this position point on the user's body part. For example, the position point 15 cm to the left after 10 cm below the cervical vertebra point; the position point 20 cm to the right after 15 cm below the trachea. The terminal maps each epidermal position point of the user to the torso structure model based on the torso structure model of the user to obtain the three-dimensional position information corresponding to each epidermal position point.

[0084] Then, the terminal generates the breath sound distribution information corresponding to each three-dimensional position point through the torso structure model based on the sub-breath sound data at each three-dimensional position point. Specifically, the terminal takes each three-dimensional position point as the center and generates an audio attenuation distribution map (three-dimensional map) in the torso structure model according to a preset audio attenuation distribution strategy, and uses this audio attenuation distribution map as the breath sound distribution information corresponding to each three-dimensional position point. Among them, the audio attenuation distribution strategy is an audio visualization technology, that is, an audio attenuation distribution map strategy generated based on the audio signal collected at a single point according to the propagation attenuation ratio value of the sound during the propagation process in the human body (preset in the terminal based on the information collected by the staff). Based on this strategy, the audio distribution map within the target range can be identified through the audio signal at a single point. Then, the terminal superimposes the breath sound distribution information corresponding to each three-dimensional position point onto the torso structure model, and screens the maximum audio information at each position as the target audio information at each position to obtain the torso audio distribution information of the user.

[0085] Finally, the terminal constructs the pulmonary secretion distribution information of the user based on the torso audio distribution information of the user. The specific construction process will be described in detail later.

[0086] Based on the above solution, by using the user's torso structure model and the breath sound data collected at multiple points of the user, the torso audio distribution information of the user is generated, so as to construct the pulmonary secretion distribution information of the user, improving the recognition accuracy and comprehensiveness of the pulmonary secretion distribution information of the user.

[0087] Optionally, constructing the pulmonary secretion distribution information of the user based on the torso audio distribution information of the user includes: based on the torso audio distribution information, identifying the sub-sound distribution information of each sound type of the user, and constructing a three-dimensional sub-sound distribution map corresponding to each sub-sound distribution information; for each three-dimensional sub-sound distribution map, through the sound anomaly analysis strategy of the sound type corresponding to the three-dimensional sub-sound distribution map, identifying the abnormal sound range in the three-dimensional sub-sound distribution map and the abnormal degree distribution information of the abnormal sound range; based on the abnormal sound ranges of each three-dimensional sub-sound distribution map, screening the target pulmonary secretion range of the user, and based on the abnormal degree distribution information of the abnormal sound ranges of each three-dimensional sub-sound distribution map, generating the secretion concentration distribution information of the target pulmonary secretion range through the distribution fitting strategy; using the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range as the pulmonary secretion distribution information of the user.

[0088] In this embodiment, the terminal identifies the sub-sound distribution information of each sound type of the user based on the torso audio distribution information, and constructs a three-dimensional sub-sound distribution map corresponding to each sub-sound distribution information. Since the sound data of different sound types are different, for example, the pitch distribution information and the loudness distribution information are different, and the timbre distribution information and the pitch distribution information are different, the method of splitting different sound types for secretion distribution can improve the recognition accuracy of secretion distribution.

[0089] For each three-dimensional sub-sound distribution map, the terminal identifies the abnormal sound range in the three-dimensional sub-sound distribution map and the abnormal degree distribution information of the abnormal sound range through the sound anomaly analysis strategy of the sound type corresponding to the three-dimensional sub-sound distribution map. Specifically, the terminal presets the normal sound data range of each sound type, and based on the three-dimensional sub-sound distribution map of each sound type, screens the distribution range that exceeds the normal sound data range as the abnormal sound range of the user, where the abnormal sound range can be one or more. Then, for each abnormal sound range, the terminal identifies the sound data distribution information of the sound type in the abnormal sound range and the characteristic distribution information of each secretion characteristic type corresponding thereto based on the association information between the secretion quality of the user and the audio feature information of the user, and uses the characteristic distribution information of all secretion characteristic types as the abnormal degree distribution information of the abnormal sound range.

[0090] Then, based on the abnormal sound ranges of the sub-audio three-dimensional distribution maps, the terminal filters the target pulmonary secretion range of the user. Among them, the target pulmonary secretion range is the range included in the abnormal sound ranges of each sound type. That is, the terminal performs range deduplication on the abnormal sound ranges of all sound types to obtain the target pulmonary secretion range of the user.

[0091] Based on the distribution information of the abnormal degree of the abnormal sound ranges of the sub-audio three-dimensional distribution maps, the terminal generates the secretion concentration distribution information of the target pulmonary secretion range through a distribution fitting strategy. Specifically, the terminal sorts the feature distribution information of each secretion feature type corresponding to each abnormal sound range in descending order to obtain the secretion concentration distribution information of each target pulmonary secretion range.

[0092] Finally, the terminal takes the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range as the pulmonary secretion distribution information of the user.

[0093] Based on the above solution, after classifying and analyzing different sound types, the secretion feature analysis and distribution fitting are carried out by different secretion feature types, so as to obtain the secretion concentration distribution information of the target pulmonary secretion range, improving the recognition accuracy of the secretion concentration distribution information.

[0094] Optionally, through an audio feature extraction network, the current audio features of the breath sound data are extracted, including: screening the target sub-breath sound data collected at the user's mouth in the breath sound data, and identifying the sound data of each sound type based on the target sub-breath sound data; respectively extracting the sound features in the sound data of each sound type through the sub-audio feature extraction networks of each sound type, and taking the sound features of all sound types as the current audio features of the breath sound data.

[0095] In this embodiment, the terminal screens the target sub-breath sound data collected at the user's mouth in the breath sound data, and identifies the sound data of each sound type based on the target sub-breath sound data. Then, the terminal respectively extracts the sound features in the sound data of each sound type through the sub-audio feature extraction networks of each sound type, and takes the sound features of all sound types as the current audio features of the breath sound data.

[0096] Based on the above solution, by collecting the target sub-breath sound data at the mouth, the recognition accuracy of the current audio features of the user is improved.

[0097] Optionally, based on the associated information, the current audio features, and the pulmonary secretion distribution information, the airway secretion information of the user is identified, including: based on the current audio features and the audio feature information corresponding to each associated information, through the cosine similarity recognition algorithm, calculating the similarity values between the current audio features and each audio feature information, and screening the associated information corresponding to the audio feature information with the maximum similarity value as the target associated information; based on the secretion information in the target associated information, identifying the secretion characteristics of the user and the secretion content range of the user, and based on the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range, generating the secretion image guidance information of the user; taking the secretion characteristics of the user, the secretion content range of the user, and the secretion image guidance information of the user as the airway secretion information of the user.

[0098] In this embodiment, the terminal calculates the similarity values between the current audio features and each audio feature information through the cosine similarity recognition algorithm based on the current audio features and the audio feature information corresponding to each associated information, and screens the associated information corresponding to the audio feature information with the maximum similarity value as the target associated information.

[0099] Then, the terminal identifies the characteristic data of each secretion characteristic type based on the secretion information in the target associated information, and based on the characteristic data of each secretion characteristic type, identifies each secretion characteristic of the user and the secretion content range of the user. Then, the terminal generates the secretion image guidance information of the user based on the target pulmonary secretion range and the secretion concentration distribution information of the target pulmonary secretion range. The secretion image guidance information is to mark the secretion concentration distribution information of each target pulmonary secretion range in each target pulmonary secretion range of the user's torso structure model according to the marking method corresponding to each secretion characteristic type to obtain the three-dimensional secretion marking information of the user.

[0100] Finally, the terminal takes the secretion characteristics of the user, the secretion content range of the user, and the secretion image guidance information of the user as the airway secretion information of the user.

[0101] Based on the above solution, by means of feature similarity recognition, the secretion characteristic information of the user is identified, and the secretion image guidance information of the user is generated, which not only reduces the time for medical staff to analyze the secretion characteristics of the user, but also effectively locates the position for medical staff to extract secretions and identifies the secretion distribution data, thus improving the positioning efficiency and secretion recognition accuracy of the user's tracheal secretions.

[0102] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are sequentially shown according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0103] Based on the same inventive concept, an embodiment of the present application further provides a high-efficiency recognition system for airway secretions for implementing the above-mentioned high-efficiency recognition method for airway secretions. The solution provided by this system to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the high-efficiency recognition system for airway secretions provided below can refer to the limitations on the high-efficiency recognition method for airway secretions in the above text, and will not be repeated here.

[0104] Further reference Figure 2 , as an implementation of the above Figure 1 shown method, an embodiment of the present application provides a high-efficiency recognition system 200 for airway secretions. The high-efficiency recognition system for airway secretions includes an acquisition module 210, a construction module 220, and an identification module 230, where:

[0105] The acquisition module 210 is used to acquire the breath sound data of the user and the historical secretion cleaning information of the user, and based on the historical secretion cleaning information, identify the association information between the secretion quality of the user and the audio feature information of the user;

[0106] The construction module 220 is used to query the body structure data of the user in the medical database, and based on the body structure data and the breath sound data of the user, construct the pulmonary secretion distribution information of the user;

[0107] The identification module 230 is used to extract the current audio features of the breath sound data through an audio feature extraction network, and based on the association information, the current audio features, and the pulmonary secretion distribution information, identify the airway secretion information of the user.

[0108] Optionally, the acquisition module 210 is specifically used for:

[0109] Based on the historical secretion cleaning information, identify the secretion information cleaned by the user each time and the audio information of each cleaning by the user;

[0110] Through the audio feature extraction network, extract the audio feature information in each audio information, and identify the secretion feature values of each secretion feature type in each secretion information;

[0111] For each secretion feature type, based on the secretion feature values of the secretion feature type and each audio information, construct the secretion-audio association distribution information of the secretion feature type;

[0112] Based on the secretion-audio association distribution information, identify the sub-association information between the secretion feature type and the audio feature information, and use all the sub-association information as the association information between the secretion quality of the user and the audio feature information of the user.

[0113] Optionally, the construction module 220 is specifically configured to:

[0114] Based on the body structure data, construct the torso structure model of the user, and divide the breath sound data into sub-breath sound data of each three-dimensional position point;

[0115] Based on the sub-breath sound data of each three-dimensional position point, through the torso structure model, generate the breath sound distribution information corresponding to each three-dimensional position point, and perform torso audio fitting processing on the breath sound distribution information corresponding to each three-dimensional position point in the torso structure model to obtain the torso audio distribution information of the user;

[0116] Based on the torso audio distribution information of the user, construct the pulmonary secretion distribution information of the user.

[0117] Optionally, the construction module 220 is specifically configured to:

[0118] Based on the torso audio distribution information, identify the sub-sound distribution information of each sound type of the user, and construct a sub-sound three-dimensional distribution map corresponding to each sub-sound distribution information;

[0119] For each sub-sound three-dimensional distribution map, through the sound anomaly analysis strategy of the sound type corresponding to the sub-sound three-dimensional distribution map, identify the abnormal sound range in the sub-sound three-dimensional distribution map and the abnormal degree distribution information of the abnormal sound range;

[0120] Based on the abnormal sound ranges of the respective sub - voice three - dimensional distribution maps, screen the target lung secretion range of the user, and based on the abnormal degree distribution information of the abnormal sound ranges of the respective sub - voice three - dimensional distribution maps, generate the secretion concentration distribution information of the target lung secretion range through a distribution fitting strategy;

[0121] Take the target lung secretion range and the secretion concentration distribution information of the target lung secretion range as the lung secretion distribution information of the user.

[0122] Optionally, the recognition module 230 is specifically configured to:

[0123] In the breath sound data, screen the target sub - breath sound data collected at the user's mouth, and based on the target sub - breath sound data, identify the sound data of each of the sound types;

[0124] Through the sub - audio feature extraction networks of each of the sound types, extract the sound features in the sound data of each of the sound types respectively, and take the sound features of all sound types as the current audio features of the breath sound data.

[0125] Optionally, the recognition module 230 is specifically configured to:

[0126] Based on the current audio features and the audio feature information corresponding to each of the association information, calculate the similarity values between the current audio features and the audio feature information through a cosine similarity recognition algorithm, and screen the association information corresponding to the audio feature information with the maximum similarity value as the target association information;

[0127] Based on the secretion information in the target association information, identify the secretion characteristics of the user and the range of the user's secretion content, and generate the secretion image guidance information of the user based on the target lung secretion range and the secretion concentration distribution information of the target lung secretion range;

[0128] Take the secretion characteristics of the user, the range of the user's secretion content, and the secretion image guidance information of the user as the airway secretion information of the user.

[0129] Each module in the above - mentioned efficient recognition system for airway secretions can be implemented in whole or in part by software, hardware, and their combinations. The above - mentioned modules can be embedded in the processor of a computer device in hardware form or be independent of it, or can be stored in the memory of a computer device in software form so that the processor can call and execute the operations corresponding to the above - mentioned modules.

[0130] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structural diagram may be as shown in Figure 3 . The computer device includes a processor, a memory, a communication interface, a display screen, and an input system connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be implemented through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for efficient recognition of airway secretions. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input system of the computer device can be a touch layer covering the display screen, or buttons, a trackball, or a touchpad provided on the computer device housing, or an external keyboard, touchpad, or mouse, etc.

[0131] Those skilled in the art can understand that Figure 3 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0132] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, it implements the steps of the method described in any one of the first aspects.

[0133] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, it implements the steps of the method described in any one of the first aspects.

[0134] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, it implements the steps of the method described in any one of the first aspects.

[0135] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data that have been authorized by the user or fully authorized by all parties.

[0136] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0137] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0138] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. An efficient recognition method for airway secretions, characterized in that The method includes: Obtaining the breath sound data of the user and the historical secretion cleaning information of the user, and based on the historical secretion cleaning information, identifying the association information between the secretion quality of the user and the audio feature information of the user; Querying the body structure data of the user in the medical staff database, and based on the body structure data and the breath sound data of the user, constructing the pulmonary secretion distribution information of the user; Extracting the current audio features of the breath sound data through an audio feature extraction network, and based on the association information, the current audio features, and the pulmonary secretion distribution information, identifying the airway secretion information of the user.

2. The method according to claim 1, wherein The identifying the association information between the secretion quality of the user and the audio feature information of the user based on the historical secretion cleaning information includes: Based on the historical secretion cleaning information, identifying the secretion information cleaned by the user each time and the audio information of the user each time; Extracting the audio feature information in each piece of audio information through an audio feature extraction network, and identifying the secretion feature values of each secretion feature type in each piece of secretion information; For each secretion feature type, based on the secretion feature values of the secretion feature type and each piece of audio information, constructing the secretion-audio association distribution information of the secretion feature type; Based on the secretion-audio association distribution information, identifying the sub-association information between the secretion feature type and the audio feature information, and using all the sub-association information as the association information between the secretion quality of the user and the audio feature information of the user.

3. The method according to claim 1, characterized in that, The constructing the pulmonary secretion distribution information of the user based on the body structure data and the breath sound data of the user includes: Based on the body structure data, constructing the torso structure model of the user, and dividing the breath sound data into sub-breath sound data of each three-dimensional position point; Based on the sub-breath sound data of each three-dimensional position point, through the torso structure model, generating the breath sound distribution information corresponding to each three-dimensional position point, and performing torso audio fitting processing on the breath sound distribution information corresponding to each three-dimensional position point in the torso structure model to obtain the torso audio distribution information of the user; Based on the torso audio distribution information of the user, constructing the pulmonary secretion distribution information of the user.

4. The method according to claim 3, wherein The constructing the pulmonary secretion distribution information of the user based on the torso audio distribution information of the user includes: Based on the torso audio distribution information, identifying the sub-sound distribution information of each sound type of the user, and constructing a sub-sound three-dimensional distribution map corresponding to each sub-sound distribution information; For each sub-sound three-dimensional distribution map, through the abnormal sound analysis strategy of the sound type corresponding to the sub-sound three-dimensional distribution map, identifying the abnormal sound range in the sub-sound three-dimensional distribution map and the abnormal degree distribution information of the abnormal sound range; Based on the abnormal sound range of each of the sub - sound three - dimensional distribution maps, screen the target lung secretion range of the user, and based on the abnormal degree distribution information of the abnormal sound range of each of the sub - sound three - dimensional distribution maps, generate the secretion concentration distribution information of the target lung secretion range through a distribution fitting strategy; Use the target lung secretion range and the secretion concentration distribution information of the target lung secretion range as the lung secretion distribution information of the user.

5. The method according to claim 1, wherein The extraction of the current audio features of the breath sound data by the audio feature extraction network includes: In the breath sound data, screen the target sub - breath sound data collected at the user's mouth, and based on the target sub - breath sound data, identify the sound data of each sound type; Through the sub - audio feature extraction networks of each of the sound types, extract the sound features in the sound data of each of the sound types respectively, and use the sound features of all sound types as the current audio features of the breath sound data.

6. The method according to claim 4, wherein The identification of the airway secretion information of the user based on the association information, the current audio features, and the lung secretion distribution information includes: Based on the current audio features and the audio feature information corresponding to each of the association information, calculate the similarity values between the current audio features and each of the audio feature information through a cosine similarity recognition algorithm, and screen the association information corresponding to the audio feature information with the maximum similarity value as the target association information; Based on the secretion information in the target association information, identify the secretion characteristics of the user and the secretion content range of the user, and based on the target lung secretion range and the secretion concentration distribution information of the target lung secretion range, generate the secretion image guidance information of the user; Use the secretion characteristics of the user, the secretion content range of the user, and the secretion image guidance information of the user as the airway secretion information of the user.

7. An efficient recognition system for airway secretions, characterized in that, The system includes: An acquisition module, configured to acquire the breath sound data of the user and the historical secretion cleaning information of the user, and based on the historical secretion cleaning information, identify the association information between the secretion quality of the user and the audio feature information of the user; A construction module, configured to query the body type structure data of the user in the medical staff database, and based on the body type structure data and the breath sound data of the user, construct the lung secretion distribution information of the user; An identification module, configured to extract the current audio features of the breath sound data through an audio feature extraction network, and based on the association information, the current audio features, and the lung secretion distribution information, identify the airway secretion information of the user.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When this computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 6.

Images