A lung respiratory cloud image and index positioning method and system based on multi-source fusion

Abstract

The application provides a lung respiratory cloud map and index positioning method and system based on multi-source fusion, which comprises the following steps: collecting respiratory sound signals; obtaining respiratory mechanics parameters; determining the corresponding collection points of different lung regions based on the chest wall standardized grid template, the double lung plane template and the positions of the collection points; extracting the respiratory sound characteristics of the respiratory sound signals; calculating the local respiratory sound indexes of each collection point, and inversely obtaining the local small airway pressure index and the local alveolar ventilation index in combination with the respiratory mechanics parameters; calculating the respiratory cloud index based on each local index in combination with the regional heterogeneity index and the respiratory phase delay index; mapping the respiratory mechanics parameters, each local index and each respiratory cloud index into a double lung region dynamic cloud map and a chest wall grid positioning map, respectively; and outputting the abnormal prompt information of the corresponding lung region and / or the positioning result of the abnormal region on the chest wall surface. Thus, the continuous, dynamic and regional visualization and abnormal positioning of the ventilation state of the lung region are realized.

Description

Technical Field

[0001] This application relates to the field of medical respiratory monitoring and processing technology, and in particular to a method and system for lung respiratory cloud map and index localization based on multi-source fusion. Background Technology

[0002] Lung regional ventilation status, airway pressure changes, and abnormal breath sounds are crucial sources of information for clinical assessment of ventilation homogeneity, airway obstruction, lung recruitment response, and the risk of overinflation. In intensive care, intraoperative anesthesia, and respiratory support scenarios, healthcare professionals need to monitor the ventilation status of different lung regions in real time, continuously, and regionally to promptly identify abnormalities and adjust ventilation strategies. Traditional auscultation relies on healthcare professionals using handheld stethoscopes to auscultate point by point, which suffers from drawbacks such as the influence of healthcare professional experience on auscultation results, strong subjectivity, difficulty in quantification and comparison, inability to achieve continuous monitoring, and inability to accurately locate abnormalities on the chest wall surface. To address these shortcomings, breath sound acquisition and analysis systems have become relatively mature alternatives to traditional auscultation methods. For example, these systems can acquire breath sounds and identify respiratory cycles, respiratory phases, and respiratory quality characteristics; some systems can visualize lung regional ventilation delays, tidal images, and regional ventilation distribution; and some systems disclose the setting of pressure monitoring pathways to measure airway pressure.

[0003] However, existing breath sound acquisition and analysis systems have the following technical problems: First, breath sound display, regional ventilation assessment, and airway pressure monitoring are isolated from each other. The breath sound protocol lacks mechanical parameter constraints, making it impossible to infer the internal airway pressure status; ventilator pressure monitoring provides overall airway parameters, making it difficult to intuitively reflect differences between different regions of the left and right lungs. Since each modality's data is presented independently, healthcare professionals must manually interpret the data from multiple devices, resulting in low clinical response efficiency.

[0004] Second, there is a lack of a unified spatiotemporal fusion framework. Existing solutions do not spatiotemporally fuse multiple breath sounds, respiratory mechanics parameters, and regional ventilation status on the same standardized chest wall grid and a dual-lung plane template, resulting in the inability to synchronously locate and output the chest wall surface position and the internal lung lobe position of abnormal areas.

[0005] Third, there is a lack of a quantifiable regional-level comprehensive evaluation system. Existing solutions mostly present regional conditions using a single parameter or a single image, lacking a quantitative index system that integrates multi-source information into a comparable and continuously trackable system, making it difficult to support bedside dynamic monitoring and efficacy comparison.

[0006] Fourth, there is a lack of auxiliary diagnostic capabilities for unilateral pleural cavity abnormalities. Although existing regional ventilation monitoring protocols can show differences in ventilation between the left and right lungs, they cannot further indicate the risk of pleural cavity lesions such as pneumothorax and hemothorax based on the spatial distribution characteristics of unilateral abnormalities.

[0007] Therefore, there is currently a lack of a dynamic multi-parameter display scheme that can spatiotemporally integrate multiple breath sounds, respiratory mechanics parameters, and regional ventilation status of the chest wall in a unified dual-lung plane template and a standardized chest wall grid template, and form a scheme that can be located, quantified, and alarmed. Summary of the Invention

[0008] To address the aforementioned problems, one objective of one or more embodiments of this application is to provide a lung respiratory cloud map and index localization method based on multi-source fusion. This method constructs a dynamic cloud map and respiratory cloud index system for the lung region through multi-source fusion inversion of chest wall breath sound signals and respiratory mechanics parameters, achieving continuous, dynamic, and regionalized visualization and anomaly localization of the ventilation status of the lung region. Another objective of one or more embodiments of this application is to provide a lung respiratory cloud map and index localization system based on multi-source fusion. Through the coordinated operation of various functional modules, the above method is executed to achieve collaborative functions such as breath sound acquisition, parameter acquisition, feature extraction, fusion inversion, index calculation, dynamic cloud map generation, anomaly localization, and assisted diagnosis.

[0009] To achieve one of the objectives, in a first aspect, the present invention provides a lung respiratory cloud map and index localization method based on multi-source fusion, the method comprising: Multiple collection points were set up in different areas of the chest wall of the subject to collect respiratory sound signals from multiple chest wall areas. Acquire respiratory mechanics parameters, which include at least one of airway pressure parameters, ventilation parameters, chest wall movement parameters, and regional impedance change parameters; Based on the standardized chest wall grid template, the double lung plane template, and the spatial coordinates of multiple acquisition points set in each chest wall region, at least one acquisition point corresponding to each of the different lung regions is determined. Extract the respiratory sound features of the respiratory sound signal collected at each of the acquisition points; The local breath sound index for each lung region corresponding to each of the acquisition points is calculated based on the breath sound characteristics, and the local small airway pressure index and local alveolar ventilation index are obtained by inversion with the respiratory mechanics parameters. Based on the local breath sound index, the local small airway pressure index, and the local alveolar ventilation index, combined with regional heterogeneity indicators and respiratory phase delay indicators, the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index are calculated. The respiratory mechanics parameters, local indices, and respiratory cloud indices are mapped to dynamic cloud maps of the two lung regions and chest wall grid positioning maps according to the correspondence between each collection point and the grid unit and the lung region, respectively. Based on the changing trends of continuous time frames, abnormal prompt information corresponding to the lung regions is output according to the dynamic cloud map of the bilateral lung regions, and / or the location result of the abnormal region on the chest wall surface is output according to the chest wall grid positioning map.

[0010] Optionally, multiple collection points are set up in different areas of the chest wall of the subject to collect breath sound signals from multiple chest wall areas, including: Each grid cell of the standardized chest wall grid template is assigned a unique location number; Each of the aforementioned grid units is projected onto different chest wall regions, and multiple acquisition points corresponding to each of the aforementioned grid units are deployed. An acoustic sensor is set at each of the acquisition points, and the acquisition point where each acoustic sensor is located corresponds to the location number of at least one of the grid cells; The respiratory sound signals of the corresponding chest wall region where the grid unit is located are collected by each of the acoustic sensors.

[0011] Optionally, assigning a unique location number to each grid cell of the standardized chest wall grid template includes: The standardized chest wall grid template is divided into anterior chest grid area, lateral chest grid area and back grid area, and the multiple grid units contained in each grid area are respectively labeled with the position number in sequence; Based on the divided grid areas, at least one of the following chest wall regions is selected: left anterior chest, left lateral chest, left posterior chest, right anterior chest, right lateral chest, and right posterior chest. The step of projecting each of the grid cells onto different chest wall regions and setting up multiple acquisition points corresponding to each of the grid cells includes: Each of the acquisition points located on the anterior chest of the left and right lungs corresponds to at least one of the grid cells in the anterior chest grid area; Each of the acquisition points located on the left and right lateral chest lungs corresponds to at least one grid cell in the lateral chest grid area; Each of the acquisition points located on the back of the left lung and the back of the right lung corresponds to at least one grid cell in the back grid area.

[0012] Optionally, the bilateral lung planar template includes at least one of the lung regions selected from the left upper lobe, left lower lobe, right upper lobe, right middle lobe, and right lower lobe; The determination of at least one collection point corresponding to each of the different lung regions includes: Each of the grid units in the chest wall standardized grid template is mapped onto the bilateral lung plane template according to the anatomical projection relationship, so that the multiple grid units correspond to at least one of the lung regions of the left upper lobe, the left lower lobe, the right upper lobe, the right middle lobe, and the right lower lobe; Based on the lung region corresponding to each of the grid cells and the grid cell corresponding to each collection point, the mapping relationship between each collection point and each of the lung regions is determined.

[0013] Optionally, the ventilation parameters include airflow parameters and / or tidal volume parameters; and / or, the breath sound characteristics include breath sound intensity characteristics, frequency characteristics, time-frequency characteristics, and respiratory phase characteristics; and / or, the local breath sound index includes a local breath sound intensity index and a local breath sound frequency index; wherein, The inversion yields the local small airway pressure index and the local alveolar ventilation index, including: The local small airway pressure index is calculated based on the joint inversion of at least two types of information: the time-frequency characteristics, the respiratory phase characteristics, the airway pressure parameters, the airflow parameters, the tidal volume parameters, the thoracic motion parameters, and the regional impedance change parameters. The local alveolar ventilation index is calculated based on the joint inversion of at least two types of information: the temporal variation of the breath sound intensity characteristics, the respiratory phase characteristics, the temporal variation of the tidal volume parameters, the temporal variation of the chest wall motion parameters, and the temporal variation of the regional impedance parameters.

[0014] Optionally, the calculation of the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud temporal delay sub-index includes: The total respiratory cloud index, the heterogeneity sub-index of respiratory cloud, and the phase delay sub-index of respiratory cloud are each preset to a normal reference range; The actual values ​​of the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index are compared with their corresponding normal reference intervals to obtain the deviation of each respiratory cloud index from its corresponding normal reference interval. Calculate the individual score for each of the respiratory cloud indices based on the deviation amount; The individual scores of the total respiratory cloud index, the individual scores of the respiratory cloud heterogeneity index, and the individual scores of the respiratory cloud phase delay index are weighted and summed according to preset weighting coefficients to obtain a comprehensive evaluation value.

[0015] Optionally, mapping the respiratory mechanics parameters, local indices, and respiratory cloud indices to dynamic cloud maps of the bilateral lung regions and chest wall grid positioning maps according to the correspondence between each collection point and the grid cells and the lung regions, respectively, includes: The local small airway pressure index, the local alveolar ventilation index, the local small airway pressure index, and the local alveolar ventilation index are located at the positions corresponding to the grid units on the standardized chest wall grid template according to the correspondence between each collection point and each grid unit; and according to the correspondence between each collection point and each lung region, they are located at the positions corresponding to each lung region on the bilateral lung plane template. Each of the grid cells and each of the local indices on each of the lung regions are rendered using preset visual encoding rules to generate the chest wall grid positioning map and the dynamic cloud map of the bilateral lung regions. The respiratory cloud total index, the respiratory cloud heterogeneity sub-index, the respiratory cloud phase delay sub-index, and at least one of the respiratory mechanics parameters are displayed in association with a preset information area in the dynamic cloud map of the bilateral lung regions and / or the chest wall grid positioning map.

[0016] Optionally, the step of outputting abnormality alert information corresponding to the lung region based on the change trend of continuous time frames and the dynamic cloud map of the bilateral lung regions, and / or outputting the location result of the abnormal region on the chest wall surface based on the chest wall grid positioning map, includes: Based on the changes of the respiratory mechanics parameters, the local indices, and the respiratory cloud indices between consecutive time frames, the grid cells and lung regions corresponding to the collection points that meet the preset abnormal conditions are identified as abnormal regions. The abnormal areas are marked on the grid cells corresponding to each of the collection points on the chest wall grid positioning map, and the positioning results of the abnormal areas on the chest wall surface are output. And / or, mark the abnormal areas on the lung regions corresponding to each of the collection points in the dynamic cloud map of the bilateral lung regions, and output the abnormal prompt information corresponding to the lung regions; When the grid cells corresponding to multiple adjacent acquisition points in a unilateral chest wall region continuously meet the preset abnormal conditions, auxiliary diagnostic prompts for unilateral pneumothorax, unilateral hemothorax, and / or unilateral hemopneumothorax are output. Specifically, when each of the local indices and / or the respiratory cloud indices in the abnormal region reaches a preset alarm threshold, an audible alarm and / or a visual alarm are triggered.

[0017] To achieve the second objective, the present invention provides a lung respiratory cloud map and index localization system based on multi-source fusion, the system comprising: The breath sound acquisition module is used to set up multiple acquisition points in different areas of the chest wall of the subject and acquire breath sound signals from multiple chest wall areas. A respiratory mechanics parameter acquisition module is used to acquire respiratory mechanics parameters, which include at least one of airway pressure parameters, ventilation parameters, chest wall motion parameters, and regional impedance change parameters. The lung region mapping module is used to determine at least one of the acquisition points corresponding to different lung regions based on the standardized chest wall grid template, the dual lung plane template, and the spatial coordinates of multiple acquisition points set in each chest wall region. The feature extraction module is used to extract the respiratory sound features of the respiratory sound signal collected at each of the acquisition points; The region inversion module is used to calculate the local breath sound index of each lung region corresponding to each of the acquisition points based on the breath sound characteristics, and to invert the local small airway pressure index and local alveolar ventilation index in combination with the respiratory mechanics parameters. The index calculation module is used to calculate the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index based on the local breath sound index, the local small airway pressure index, and the local alveolar ventilation index, combined with regional heterogeneity indicators and respiratory phase delay indicators. The dynamic cloud map generation module is used to map the respiratory mechanics parameters, local indices, and respiratory cloud indices into dynamic cloud maps of the two lung regions and chest wall grid positioning maps respectively, according to the correspondence between each collection point and the grid unit and the lung region. The anomaly localization and alert module is used to output anomaly alert information corresponding to the lung region based on the change trend of the continuous time frame and the dynamic cloud map of the bilateral lung regions, and / or output the localization result of the abnormal region on the chest wall surface based on the chest wall grid localization map.

[0018] Optionally, the respiratory sound acquisition module includes multiple acoustic sensors disposed in various regions of the chest wall; The respiratory mechanics parameter acquisition module includes at least one of the following sensing units: The pressure sensing unit includes a fiber optic pressure sensor and / or a MEMS pressure sensor disposed on the endotracheal tube, for acquiring the airway pressure parameters. A ventilator sensing unit is installed on the ventilator tubing to acquire the ventilation parameters; A thoracic motion monitoring unit, mounted on the chest wall, is used to acquire the thoracic motion parameters; An impedance imaging unit, mounted on the chest wall, is used to acquire impedance change parameters of the region.

[0019] Compared with the prior art, this application has the following advantages: The method provided in this application first collects breath sound signals and at least one respiratory mechanics parameter from multiple regions of the chest wall. Second, it performs standardized grid division of the entire chest and establishes a mapping relationship between the chest wall grid and lung lobe regions. Third, it performs regional inversion based on breath sound characteristics and respiratory mechanics parameters to obtain local breath sound intensity index, local breath sound frequency index, local small airway pressure index, and local alveolar ventilation index. Then, it further calculates the total respiratory cloud index, respiratory cloud heterogeneity sub-index, and respiratory cloud temporal delay sub-index at the lung lobe and chest wall grid levels. Finally, it maps the above parameters and indices to a dynamic cloud map of both lungs and a chest wall grid location map, and outputs a location prompt when an anomaly reaches a threshold.

[0020] By fusing multiple breath sounds from the chest wall with respiratory mechanics parameters on the same left and right lung planes, and further outputting the total respiratory cloud index, respiratory cloud heterogeneity sub-index, and respiratory cloud phase delay sub-index, the originally dispersed regional parameters are transformed into easily comparable and continuously trackable quantitative results. This facilitates bedside monitoring, efficacy comparison, and parameter adjustment, enabling continuous, bedside, and dynamic visualization of the lung region's condition, rather than just a single waveform or single image display. By gridding the entire chest and establishing a correspondence with the chest wall breath sound acquisition locations, not only can lobar-level visualization be provided, but abnormal areas can also be further located to the grid position on the chest wall surface, outputting abnormal areas such as potential lesion areas, ventilation abnormal areas, and / or pressure / ventilation mismatch areas.

[0021] In addition, for unilateral abnormal grid clusters, it can also provide auxiliary diagnosis and early warning prompts for unilateral pneumothorax, unilateral hemothorax and / or unilateral hemopneumothorax, thereby improving the accuracy of abnormal location, the intuitiveness of bedside operation and the efficiency of clinical response.

[0022] The advantages of the system and the method described above over the prior art are the same, and will not be repeated here. Attached Figure Description

[0023] To more clearly illustrate the technical solution of this application, the drawings used in the description of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart of the steps of the lung respiratory cloud map and index localization method based on multi-source fusion according to an embodiment of this application; Figure 2 This is a schematic diagram showing the distribution of collection points according to an embodiment of this application; Figure 3 This is a schematic diagram of the division and abnormal location of the standardized grid template for the chest wall according to an embodiment of this application; Figure 4 This is a display interface diagram of the dynamic cloud map of the bilateral lung regions described in an embodiment of this application; Figure 5 This is a functional module diagram of a lung respiratory cloud map and index positioning system based on multi-source fusion, as described in one embodiment. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] It should be noted that the "breath sound signal" mentioned in this application refers to the acoustic vibration signal collected by an acoustic sensor placed on the surface of the chest wall, which belongs to the acoustic mode; the "breathing mechanical parameters" refer to the mechanical quantity measurement values ​​directly output by existing equipment such as pressure sensors, flow meters, motion sensors, or impedance electrodes, which belong to the mechanical mode. The "breath sound characteristics" refer to the signal attributes extracted from the original signal through an algorithm. The "inversion" refers to the process of inferring internal, unmeasurable states from surface observation data.

[0027] To ensure logical clarity throughout the text, the following description follows the order of "system overall structure - data acquisition and grid establishment - lung region mapping - feature extraction and local index calculation - multi-source fusion inversion - respiratory cloud index calculation - dynamic cloud map construction - anomaly identification and localization - unilateral auxiliary diagnosis - audible and visual alarms", and provides specific explanations of the technical solutions provided by one or more embodiments.

[0028] Example 1: Reference Figure 5 This embodiment provides a lung respiratory cloud map and index localization system based on multi-source fusion, including: a respiratory sound acquisition module, used to deploy multiple acquisition points in different regions of the chest wall of the subject to acquire respiratory sound signals from multiple chest wall regions. Exemplarily, the respiratory sound acquisition module includes a chest wall multi-point respiratory sound acquisition array, consisting of multiple acoustic sensors deployed in the anterior chest, lateral chest, and back regions according to the grid positions of a standardized chest wall grid template.

[0029] A respiratory mechanics parameter acquisition module is used to acquire respiratory mechanics parameters. These parameters include at least one of airway pressure parameters, ventilation parameters, chest wall movement parameters, and regional impedance change parameters. These parameters are provided by existing medical equipment and read by the system through a communication interface.

[0030] The lung region mapping module is used to determine at least one acquisition point corresponding to each lung region based on a standardized chest wall grid template, a dual-lung plane template, and the spatial coordinates of multiple acquisition points set for each chest wall region. The core function of the lung region mapping module is to establish the mapping relationship between acquisition points and lung lobe regions, the specific mechanism of which is described in detail in Example 2.

[0031] The feature extraction module is used to extract the respiratory sound features of the respiratory sound signal acquired at each acquisition point. After the respiratory sound signal is preprocessed (bandpass filtering, noise suppression, heart sound separation, and artifact correction), the feature extraction module extracts respiratory sound intensity features, frequency features, time-frequency features, and respiratory phase features from the processed signal.

[0032] The regional inversion module is used to calculate the local breath sound index for each lung region corresponding to each acquisition point based on the breath sound characteristics, and to invert the local small airway pressure index and local alveolar ventilation index by combining respiratory mechanics parameters. The local breath sound index includes the local breath sound intensity index and the local breath sound frequency index, which respectively reflect the amplitude level and frequency distribution characteristics of the breath sounds in the corresponding acquisition area. During the inversion process, the breath sound characteristics and respiratory mechanics parameters observed on the chest wall surface are used as constraints to estimate the small airway pressure state and alveolar ventilation state that cannot be directly measured within each lung region. It should be noted that the concept and principle of "inversion" have been applied in fields such as electronic control, and this embodiment will not elaborate on them further.

[0033] The index calculation module is used to calculate the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index based on the local breath sound index, the local small airway pressure index, and the local alveolar ventilation index, combined with regional heterogeneity indicators and respiratory phase delay indicators. The three respiratory cloud indices quantify and assess lung ventilation status from three dimensions: overall status, regional differences, and temporal synchronization.

[0034] The dynamic cloud map generation module is used to render various local indices and respiratory cloud indices into dynamic cloud maps of the bilateral lung regions and chest wall grid positioning maps, respectively, according to the correspondence between each collection point and the chest wall grid unit and the lung region. The two maps are updated synchronously, using the bilateral lung planar template and the standardized chest wall grid template as spatial references, respectively.

[0035] The anomaly localization and alert module is used to identify abnormal regions that meet preset abnormality conditions based on the changes of various local indices and / or respiratory cloud indices across consecutive time frames, and outputs the location results of the abnormal regions on the chest wall surface and / or the corresponding abnormality alert information for the lung regions. When multiple adjacent acquisition points in a unilateral chest wall region continuously meet the preset abnormality conditions, the anomaly localization and alert module outputs auxiliary diagnostic alert information for unilateral pneumothorax, unilateral hemothorax, and / or unilateral hemopneumothorax.

[0036] Optionally, the system also includes a timing synchronization module for synchronizing the breath sound signal with respiratory mechanics parameters, ensuring that multi-source data are fused on the same time reference. The timing synchronization module is connected to both the breath sound acquisition module and the respiratory mechanics parameter acquisition module.

[0037] Optionally, the system also includes an audible and visual alarm module and a display terminal. The audible and visual alarm module is used to trigger an audible alarm and / or a visual alarm when the local index and / or respiratory cloud index in the abnormal area reaches a preset alarm threshold. The display terminal is used to display dynamic cloud maps of the bilateral lung regions, chest wall grid positioning maps, abnormal prompts, and / or auxiliary diagnostic prompts.

[0038] Among them, the breathing sound acquisition module and the breathing mechanics parameter acquisition module are connected to the time synchronization module, the time synchronization module is connected to the feature extraction module, the feature extraction module is connected to the region inversion module, the region inversion module is connected to the index calculation module and the dynamic cloud map generation module, and the dynamic cloud map generation module is connected to the anomaly location and prompting module, the audible and visual alarm module and the display terminal.

[0039] Preferably, the respiratory sound acquisition module includes multiple acoustic sensors disposed in various regions of the chest wall. The respiratory mechanics parameter acquisition module includes at least one of the following sensing units: Pressure sensing unit: This includes a fiber optic pressure sensor and / or a MEMS (Micro-Electro-Mechanical System) pressure sensor mounted on the endotracheal cannula. The fiber optic pressure sensor preferably employs a fiber Bragg grating (FBG) sensor, which reflects pressure changes by detecting changes in the wavelength reflected by the grating. One or more fiber optic arrays are placed on the wall, proximal end, and / or distal end of an endotracheal cannula. The proximal end refers to the end closest to the ventilator interface, and the distal end refers to the end inserted into the trachea. By placing pressure sensing units at both the proximal and distal ends, the proximal and distal airway pressures can be simultaneously acquired for assessing the airway pressure gradient.

[0040] Ventilator sensor unit: Located on the ventilator tubing, the raw flow and pressure signals are processed internally by the ventilator. Ventilation parameters output by the ventilator, including peak airway pressure, positive end-expiratory pressure, tidal volume, and airflow rate, can be read through the ventilator's data communication interface.

[0041] Thoracic motion monitoring unit: Located on the chest wall, it includes motion sensors (such as MEMS accelerometers, gyroscopes, or strain gauges) attached to the chest wall surface or a sensing band surrounding the chest. Thoracic motion parameters include chest wall displacement, acceleration, or chest circumference changes, reflecting the degree of chest wall expansion.

[0042] Impedance imaging unit: Located on the chest wall, it acquires regional impedance variation parameters that reflect changes in conductivity caused by gas distribution within the lungs. Exemplarily, the impedance imaging unit includes multiple electrodes arranged around the chest wall. The unit generates a regional impedance distribution image of a cross-section of the chest using a reconstruction algorithm by injecting excitation current into electrode pairs and acquiring voltage signals from other electrode pairs.

[0043] It should be noted that the four types of sensing units mentioned above can be used individually or in combination. For example, in an ICU setting, all four sensing units can be deployed simultaneously to achieve the most comprehensive multi-source fusion. In intraoperative anesthesia settings, pressure sensing units and ventilator sensing units can be prioritized. In transport monitoring settings, chest wall movement monitoring units can be deployed, allowing the system to adapt to the needs of different clinical environments.

[0044] When used in combination, the timing synchronization module aligns the signals from each source to ensure that multi-source data are fused on the same time reference. Meanwhile, the specific installation locations, models, and parameters of the aforementioned sensing units are for illustrative purposes only and do not constitute a limitation on the scope of protection of this application. Those skilled in the art can select appropriate combinations of sensing units based on actual clinical needs.

[0045] Example 2: Correspondingly, please refer to Figure 1 As shown, Figure 1 This is a flowchart illustrating the steps of a lung respiratory cloud map and index localization method based on multi-source fusion. The present invention also provides a lung respiratory cloud map and index localization method based on multi-source fusion, utilizing the lung respiratory cloud map and index localization method based on multi-source fusion provided in the first aspect of the present invention. This method includes the following steps: S1. Multiple collection points are set up in different areas of the chest wall of the subject to collect respiratory sound signals from multiple chest wall areas.

[0046] In this step, multiple sampling points are deployed in different areas of the chest wall to collect breath sound signals from multiple chest wall regions. This multi-point deployment achieves spatial coverage sampling of the chest wall surface, providing a spatial resolution basis for subsequent gridding and lung region mapping, enabling medical staff to simultaneously acquire breath sound signals from multiple regions.

[0047] S2. Obtain respiratory mechanics parameters, which include at least one of airway pressure parameters, ventilation parameters, chest wall motion parameters, and regional impedance change parameters.

[0048] In this step, ventilation parameters may include airflow parameters and / or tidal volume parameters, both of which can cover the flow rate and volume of respiratory airflow on the ventilator side. This step introduces mechanical modal data that is physically different from the respiratory sound signal by acquiring respiratory mechanics parameters, including at least one of airway pressure parameters, ventilation parameters, chest wall motion parameters, and regional impedance change parameters. The acquisition of multimodal parameters provides cross-modal constraints for subsequent multi-source fusion inversion. For details on the acquisition of these respiratory mechanics parameters, please refer to Example 1.

[0049] S3. Based on the standardized chest wall grid template, the double lung plane template, and the spatial coordinates of multiple acquisition points set in each chest wall region, determine at least one acquisition point corresponding to each of the different lung regions.

[0050] In this step, the standardized chest wall grid template divides the entire chest wall space, with each grid cell covering a certain chest wall surface area. Collection points are arranged according to the standardized chest wall grid. Each grid cell in the standardized chest wall grid template has a unique location number and a clear correspondence with one or more breath sound collection points. Therefore, the physical location of a breath sound collection point on the chest wall determines which grid cell it falls into. Simultaneously, the projection of the lung anatomy onto the chest wall surface can establish a bi-lung plane template. Therefore, each grid cell has a corresponding projection mapping to a specific part of a lung lobe, allowing direct mapping of the grid cell to its internal lung lobe. Thus, using the known grid positions of the standardized chest wall grid template and the bi-lung plane template, the breath sound signals from the arranged collection points can be assigned to specific lung lobe locations. When the collected breath sound signals are finally output, they display both "grid abnormality" and a "lung location abnormality" message at the lobe level.

[0051] By establishing a mapping relationship between collection points, grid units, and lung lobes, abnormal locations on the chest wall surface and abnormal locations within the lung lobes can be located and output synchronously, solving the problem that abnormal areas cannot be accurately assigned to both the lung lobe and grid levels simultaneously.

[0052] S4. Extract the respiratory sound features of the respiratory sound signal collected at each of the acquisition points.

[0053] In this step, the acquired breath sound signal is first preprocessed through bandpass filtering, noise suppression, heart sound separation, and artifact correction. The feature extraction module then extracts features such as breath sound intensity, breath sound frequency, time-frequency characteristics, and breath phase characteristics from the processed signal, obtaining parameters such as breath sound amplitude envelope, frequency band energy distribution, dominant frequency, spectral centroid, inspiratory-to-expiratory ratio, phase start and end points, and abnormal breath sound events. Therefore, the four types of features extracted in this embodiment comprehensively describe the acoustic properties of breath sounds from three analytical dimensions: time domain, frequency domain, and time-frequency domain. This provides richer input information for subsequent local index calculation and fusion inversion, improving the accuracy of regional inversion.

[0054] S5. Calculate the local breath sound index for each lung region corresponding to each of the acquisition points based on the breath sound characteristics, and invert the local small airway pressure index and local alveolar ventilation index by combining the respiratory mechanics parameters.

[0055] In this step, the index calculation module calculates the local breath sound index for each lung region corresponding to each acquisition point based on the aforementioned breath sound characteristics. This includes the local breath sound intensity index and the local breath sound frequency index. The breath sound characteristics acquired from multiple points on the chest wall are synchronized with airway pressure parameters, airflow, tidal volume, chest wall movement, and / or regional impedance changes over time. A regional grid mapping to the chest wall grid is established based on the spatial location of each acquisition point in the bi-lung plane template and the standardized chest wall grid template. Temporal feature vectors are generated for each region through spatial interpolation and temporal alignment; the local breath sound intensity index and the local breath sound frequency index are calculated based on these temporal feature vectors.

[0056] Then, local small airway pressure and local alveolar ventilation determine the observable breath sound characteristics and respiratory mechanics parameters on the chest wall surface. This embodiment collects multi-source data through steps S1-S4 and can use parameters from at least two different physical mechanisms as constraints for inversion. The local small airway pressure index focuses on the pressure / frequency correlation (time-frequency characteristics + airway pressure + airflow parameters) in the multi-source input, reflecting the mechanical state within the airway. The local alveolar ventilation index focuses on the ventilation efficiency correlation (intensity variation + tidal volume variation + chest wall movement variation + impedance variation) in the multi-source input, reflecting gas exchange function. Therefore, using multi-source data as constraints, the internal parameter values ​​that best approximate the multi-source data output of the forward model can be solved.

[0057] It is understandable that changes in intensity, tidal volume, chest movement, and impedance can be extracted from the already collected breath sound intensity characteristics, tidal volume parameters, chest movement parameters, and regional impedance change parameters.

[0058] Therefore, by limiting the calculation of the local small airway pressure index and the local alveolar ventilation index to be jointly inverted based on at least two types of information, a multi-source constraint mechanism for the inversion step was clarified. The two indices each selected a fusion term most relevant to their physiological meaning, allowing the local small airway pressure index to emphasize the pressure / frequency correlation to reflect the airway mechanical state, and the local alveolar ventilation index to emphasize the ventilation efficiency correlation to reflect gas exchange function. This differentiated fusion source design allows the two indices to be optimally constrained for different internal physiological states, improving the physiological rationality and discriminative power of the inversion results.

[0059] S6. Based on the local breath sound index, the local small airway pressure index, and the local alveolar ventilation index, combined with regional heterogeneity index and respiratory phase delay index, calculate the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index.

[0060] In this step, regional heterogeneity indices are obtained by cross-regional statistical comparison of local indices (local breath sound intensity index, local breath sound frequency index, local small airway pressure index, and local alveolar ventilation index) of each grid cell. These indices can be used to reflect the degree of difference between regions. Respiratory phase delay indices are calculated from the start and end points of the inspiratory / expiratory phases of each region based on the respiratory phase features extracted in the feature extraction step. These indices can be used to quantify the difference in the onset time of inspiratory / expiratory phases in different regions, reflecting the degree of synchronization between regions.

[0061] Therefore, the local breath sound index, local small airway pressure index, and local alveolar ventilation index reflect the local respiratory status of a single grid unit. The two indices reflect the lateral asymmetry and asynchrony between different regions. The index calculation module converts these indices into quantitative scores with uniform dimensions and further calculates the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index, making all three breath sound cloud indices quantifiable numerical indicators. This transforms the originally dispersed regional parameters into a comparable and continuously traceable quantitative index system. For example, a higher total respiratory cloud index indicates more uniform overall ventilation, a higher respiratory cloud heterogeneity sub-index indicates more uneven regional distribution, and a higher respiratory cloud phase delay sub-index indicates a more pronounced degree of regional respiratory asynchrony.

[0062] S7. The respiratory mechanics parameters, local indices, and respiratory cloud indices are mapped to dynamic cloud maps of the two lung regions and chest wall grid positioning maps according to the correspondence between each collection point and the grid unit and the lung region.

[0063] In this step, each local index and each respiratory cloud index is rendered into a dynamic cloud map of the bilateral lung regions and a chest wall grid positioning map, respectively, according to the correspondence between each collection point and the chest wall grid unit and lung region. This achieves dual visualization output of abnormal areas at the lobar level and the chest wall surface level. The dynamic cloud map of the bilateral lung regions uses a bilateral lung planar template as a spatial reference to display the functional status of the lung lobe; the chest wall grid positioning map uses a standardized chest wall grid template as a spatial reference to display the location of abnormalities at the body surface level. The simultaneous refresh of the two maps allows medical staff to not only intuitively see the lung function status, but also accurately compare the location with the patient's body surface for operation.

[0064] S8. Based on the changing trend of continuous time frames, output abnormal prompt information corresponding to the lung region according to the dynamic cloud map of the two lung regions, and / or output the positioning result of the abnormal region on the chest wall surface according to the chest wall grid positioning map.

[0065] Existing regional ventilation monitoring schemes cannot further indicate the risk of pleural cavity lesions based on the spatial distribution characteristics of unilateral abnormalities. This embodiment can identify abnormal areas by analyzing the changes of various local indices and / or respiratory cloud indices across consecutive time frames, and output auxiliary diagnostic prompts for pneumothorax, hemothorax, or hemopneumothorax when unilateral abnormal clusters continuously meet the abnormal conditions, thus achieving early warning of pleural cavity lesions.

[0066] In summary, steps S1 to S8 constitute a complete technical chain of "chest wall acquisition—chest wall grid—lung region mapping—feature extraction—parameter inversion—index calculation—dynamic cloud map—anomaly localization—assisted diagnosis," realizing a closed-loop process from acoustic observation of the chest wall surface to estimation of the internal lung lobe status, from local indices to comprehensive indices, and from data calculation to visualization output and clinical alarms. In existing technologies, breath sound display, regional ventilation evaluation, and airway pressure monitoring are fragmented, with each modality of data presented independently, requiring medical staff to make manual judgments by integrating the display interfaces of multiple devices. This application integrates multi-source information in a unified dual-lung plane template and a standardized chest wall grid template, forming a dynamic multi-parameter display scheme that is localizable, quantifiable, and capable of alarms, significantly improving clinical response efficiency and anomaly localization accuracy.

[0067] Further, step S1 includes: S11. Each grid cell of the standardized chest wall grid template is assigned a unique position number.

[0068] S12. Project each of the grid units onto different chest wall regions and set up multiple acquisition points corresponding to each of the grid units.

[0069] S13. An acoustic sensor is set at each of the acquisition points, and the acquisition point where each acoustic sensor is located corresponds to the position number of at least one of the grid cells.

[0070] S14. Acquire the respiratory sound signal of the chest wall region corresponding to the grid unit by means of each of the acoustic sensors.

[0071] In this embodiment, the collection points are arranged according to a standardized grid pattern on the chest wall. For example... Figure 2 and Figure 3 As shown, the standardized chest wall grid template is divided into anterior chest grid area 121, lateral chest grid area 122, and posterior grid area 123. Each grid area contains multiple grid units, each numbered sequentially. For example, the anterior chest grid area contains 7×4=28 grid units, numbered A01 to A28, covering the left / right upper lobes and the anterior lung region. The lateral chest grid area contains 6×4=24 grid units, numbered L01 to L24, covering the right middle lobe and the lateral lung region. The posterior grid area contains 7×4=28 grid units, numbered P01 to P28, covering the left / right lower lobes and the posterior lung region. Each acquisition point located in the anterior chest region (anterior chest of the left lung and anterior chest of the right lung) corresponds to at least one grid cell in the anterior chest grid area; each acquisition point located in the lateral chest region (lateral chest of the left lung and lateral chest of the right lung) corresponds to at least one grid cell in the lateral chest grid area; each acquisition point located in the back region (back of the left lung and back of the right lung) corresponds to at least one grid cell in the back grid area.

[0072] Each grid cell of the standardized chest wall grid template has a unique location number; an acoustic sensor is set at each acquisition point, and the placement of each acoustic sensor corresponds to the location number of at least one grid cell; the respiratory sound signal of the chest wall region where the corresponding grid cell is located is acquired through each acoustic sensor. As a specific implementation of this embodiment, the acoustic sensors are electronic stethoscopes or MEMS microphones. For example, 7 acquisition points are set up in the anterior chest region, corresponding to the subclavian, parasternal, nipple line, and lower lung lobe regions. 5 acquisition points are set up in the lateral chest region, corresponding to the anterior axillary line, mid-axillary line, posterior axillary line, and costophrenic region. 7 acquisition points are set up in the back region, corresponding to the interscapular region, subscapular region, and bilateral lower lung fields. The physical location of each acquisition point is recorded using three-dimensional spatial coordinates for subsequent registration with several grid cells. It should be noted that one acquisition point can correspond to one or more grid cells. For example, acquisition point A1 can correspond to grid cell A01, or simultaneously to grid cells A01 and A02.

[0073] Therefore, by dividing the standardized chest wall grid template into anterior chest, lateral chest, and back grid areas, complete coverage of the entire chest wall surface area is achieved. In particular, the projections of the right and left lower lobes are mainly located on the back and lower lateral chest; the lack of a back grid area would prevent the detection of lower lobe abnormalities. This three-zone division ensures the integrity of the lung lobe projections onto the body surface, thus covering all anatomical locations of the patient's left and right lungs corresponding to the anterior, lateral, and back areas, enabling the system to achieve full lung coverage monitoring capabilities.

[0074] It should be noted that the anterior chest grid area, lateral chest grid area, and back grid area are standardized divisions of the entire chest wall surface area, covering the anatomical locations of the patient's left and right lungs corresponding to the anterior, lateral, and back areas. Data collection points are included within the grid areas according to their physical location. The system can proportionally adjust the standard grid based on the subject's body shape parameters (such as chest circumference, height, and BMI (Body Mass Index)) to accommodate individual differences.

[0075] Therefore, by assigning a unique location number to each grid cell and making the placement of the acoustic sensor correspond to the location number of at least one grid cell, a deterministic location attribution relationship between the acquisition point and the grid cell is established, providing a calculable location basis for the subsequent grid assignment of the index, while also supporting many-to-one attribution between acquisition points and grid cells.

[0076] In some embodiments, the bilateral lung planar template is a standardized two-dimensional schematic diagram of the left and right lung contours, marking the positions, boundaries, and contours of five lobes: the left upper lobe, left lower lobe, right upper lobe, right middle lobe, and right lower lobe. The bilateral lung planar template provides spatial definitions of the lung lobes on a standard two-dimensional plane. Further, step S3 includes: S31. Map each of the grid units in the standardized chest wall grid template onto the bilateral lung plane template according to the anatomical projection relationship, so that the multiple grid units correspond to at least one of the lung regions of the left upper lobe, the left lower lobe, the right upper lobe, the right middle lobe, and the right lower lobe.

[0077] In this step, the anatomical projection relationship is based on the known anatomical correspondence between the location of the chest wall surface and the underlying lung lobes. For example, the projection of the 2nd-4th intercostal spaces on the right side of the anterior chest corresponds to the upper lobe of the right lung, and the projection of the 3rd-5th intercostal spaces on the left side of the anterior chest corresponds to the upper lobe of the left lung. The system can adjust the scale of the standard grid according to the body shape parameters of the subject before mapping the chest wall grid units to the upper lobe of the left lung, the lower lobe of the left lung, the upper lobe of the right lung, the middle lobe of the right lung, and the lower lobe of the right lung.

[0078] S32. Determine the mapping relationship between each collection point and each lung region based on the lung region corresponding to each grid unit and the grid unit corresponding to each collection point.

[0079] In this embodiment, the standardized chest wall grid template can serve as a location bridge connecting the acquisition point and the lung lobe. Since the acquisition point falls into a specific grid cell through its physical coordinates (location attribution), and the grid cell simultaneously corresponds to a specific lung lobe through anatomical projection (anatomical projection relationship), the acquisition point is indirectly associated with the lung lobe region. It should be noted that the mapping relationship between the chest wall grid and the anatomical projection of the lung region is based on a pre-defined relationship known in anatomy.

[0080] As an example, suppose grid cell A10 in the anterior chest grid area covers the 3rd-4th intercostal space on the left anterior chest: A breath sound acquisition point (located within the A10 area) is placed within grid cell A10; therefore, the breath sound signal acquired at this acquisition point belongs to grid cell A10. Since grid cell A10 corresponds to the left upper lobe of the lung in anatomical projection, the local index of A10 is also classified as the left upper lobe. In the final output, "A10 grid abnormality" can be displayed on the chest wall grid location map, and "left upper lobe abnormality" can also be indicated at the lobe level.

[0081] Therefore, by mapping grid units to corresponding lobes on a bilateral lung plane template according to anatomical projection relationships, and determining the mapping relationship between acquisition points and lung regions, a calculable mapping channel from the chest wall surface to the internal lung lobes is established. In existing technologies, the correspondence between chest wall auscultation and lung lobes relies on the experience and judgment of medical personnel and cannot be quantified. This application maps grid units to the five lung lobes one-to-one through anatomical projection relationships, enabling the system to automatically locate chest wall abnormalities at the lung lobe level. Simultaneously, this mapping at the grid unit level ensures mapping accuracy; that is, the upper and middle grid units within the same anterior chest grid area can correspond to different lung lobes.

[0082] Further, step S6 includes: S61. The total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index are each preset with a normal reference range; S62. Compare the actual values ​​of the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index with their corresponding normal reference intervals to obtain the deviation of each respiratory cloud index from its corresponding normal reference interval. S63. Calculate the individual score of each of the respiratory cloud indices based on the deviation amount; S64. The individual scores of the total respiratory cloud index, the individual scores of the respiratory cloud heterogeneity index, and the individual scores of the respiratory cloud phase delay index are weighted and summed according to preset weighting coefficients to obtain a comprehensive evaluation value.

[0083] In some embodiments, the overall respiratory cloud index is calculated by cross-lobe weighted aggregation of the local breath sound index, local small airway pressure index, and local alveolar ventilation index of each lung lobe. The respiratory cloud heterogeneity sub-index reflects the spatial non-uniformity of ventilation states among different lung lobes. Its calculation principle is as follows: taking each local index of each lung lobe as a sample set, the dispersion index of this set is calculated using a common statistical method for measuring sample dispersion; the resulting dispersion index is the respiratory cloud heterogeneity sub-index. The respiratory cloud phase delay sub-index is calculated as the proportion of the range of the respiratory phase delay index to the respiratory cycle.

[0084] It should be noted that the calculation of each respiratory cloud index is a relatively mature and general technique, and those skilled in the art can choose an appropriate dispersion measure method according to actual needs. For example, the overall respiratory cloud index can also be calculated using other equivalent aggregation methods (such as simple mean, median aggregation, etc.). Dispersion indicators can also be at least one of the coefficient of variation, standard deviation, range ratio, or Gini coefficient. The respiratory cloud phase delay sub-index can also be calculated using other equivalent phase difference measures (such as root mean square deviation of each lung lobe, delay estimation based on cross-correlation analysis, etc.).

[0085] In this embodiment, the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index are used as quantifiable numerical indicators, each corresponding to a normal reference interval. The degree to which each index deviates from the center of the normal reference interval is normalized to 0-100 to obtain an individual score. The greater the deviation of the indicator from the normal reference interval, the worse the local lung function or ventilation status, and the lower the corresponding quantitative score. Subsequently, the individual scores of each index are weighted and fused using preset weights to obtain a comprehensive evaluation value reflecting the overall respiratory status. This value can be further smoothed temporally using at least one of sliding time window averaging, exponential smoothing, or Kalman filtering.

[0086] By setting preset normal reference ranges for three respiratory cloud indices and calculating individual scores based on deviations, and then weighting and fusing them according to preset weights to obtain a comprehensive evaluation value, a quantitative conversion mechanism from multiple indices to a single comprehensive evaluation value is established. In existing technologies, each parameter is presented independently, lacking a unified comprehensive evaluation method. The comprehensive evaluation value of this application allows medical staff to quickly judge the overall respiratory status through a single value, while each individual score retains dimensional information, balancing the simplicity of the overall evaluation with the detail of dimensional diagnosis.

[0087] Further, step S7 includes: S71. The local small airway pressure index, the local alveolar ventilation index, and the local small airway pressure index are located at the positions corresponding to the grid units on the chest wall standardized grid template according to the correspondence between each collection point and each grid unit. S72. The local small airway pressure index, the local alveolar ventilation index, the local small airway pressure index, and the local alveolar ventilation index are located at the positions corresponding to the lung regions on the bilateral lung plane template according to the correspondence between each collection point and each lung region. S73. Render each of the grid cells and each of the local indices on each of the lung regions according to the preset visual coding rules to generate the chest wall grid positioning map and the dynamic cloud map of the bilateral lung regions. S74. Display at least one of the following: the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, the respiratory cloud phase delay sub-index, and the respiratory mechanics parameter, in a preset information area of ​​the dynamic cloud map of the bilateral lung regions and / or the chest wall grid positioning map.

[0088] In this embodiment, a dynamic cloud map is constructed on the left and right lung planar images, and a chest wall grid positioning map is constructed on a standardized chest wall grid template. First, using the acquisition point-grid unit-lung region mapping relationship established in steps S31-S32, grid-level indices are mapped to the lobe level. The preset visual coding rules include: different colors are used to distinguish different parameter types (e.g., the first color layer represents the intensity of local breath sounds, the second color layer represents the frequency of local breath sounds, the third color layer represents the local small airway pressure index, and the fourth color layer represents the local alveolar ventilation index), and color depth, brightness, saturation, or transparency are used to reflect the magnitude or degree of deviation of the corresponding index value. Multiple color layers are superimposed to form a "cloud map" effect.

[0089] In this embodiment, as Figure 4 As shown, respiratory mechanics parameters (such as peak airway pressure and tidal volume) are not applicable to spatial distribution attributes by grid unit or lung lobe as overall or tubular quantities. Therefore, they do not need to be spatially mapped in the dynamic cloud map of the two lung regions. Instead, they can be displayed in numerical, curve or dial form in the side information area of ​​the dynamic cloud map area of ​​the two lung regions, forming a correlation and comparison with the cloud map content.

[0090] The dynamic cloud map of the bilateral lung regions and the chest wall grid positioning map, obtained through multi-source parameter fusion analysis, are updated simultaneously, using a bilateral lung planar template and a standardized chest wall grid template as spatial references, respectively. The dynamic cloud map of the bilateral lung regions visually displays the lung function status, showing which lung region has ventilation problems. The chest wall grid positioning map accurately indicates the location of the collection point (e.g., A10), locating which chest wall region (left upper lobe) corresponds to the abnormality, facilitating medical staff to compare with the actual patient's body surface for operation.

[0091] For example, when the parameters or indices of multiple adjacent grid cells are continuously abnormal, the chest wall grid-lung region mapping and abnormality localization and prompting module highlights the corresponding grid in the chest wall grid localization map and provides prompts for the corresponding lung lobe region and / or potential abnormal regions. In some implementations, if the highlighted grids are mainly concentrated on one side of the chest wall and present as continuous abnormal clusters (e.g., the regions where A14, A15, A18, and A19 are all abnormal), the display terminal further outputs a prompt for unilateral pleural cavity abnormality and marks auxiliary diagnostic information for suspected unilateral pneumothorax, unilateral hemothorax, and / or unilateral hemopneumothorax.

[0092] Therefore, by positioning local indices to corresponding positions on the standardized chest wall grid template and the bilateral lung plane template, two images are generated using preset visual coding rules. The respiratory cloud index and respiratory mechanics parameters are then displayed in a pre-defined information area, ensuring that spatial distribution information and overall parameter information do not interfere with each other. Local indices with spatial distribution attributes are then rendered to their corresponding positions on the images using color coding, while overall parameters without spatial distribution attributes (such as peak airway pressure and tidal volume) are displayed in numerical / curve form in the adjacent information area, improving the information density and readability of the display interface.

[0093] Example 3: Risk Warning and Anomaly Location: Based on the changes of various local indices and / or respiratory cloud indices across consecutive time frames, abnormal regions meeting preset anomaly conditions are identified, and the location results of the abnormal regions on the chest wall surface and / or the corresponding abnormality warning information for the lung regions are output. Specifically, the changes of various local indices and / or respiratory cloud indices between the current time frame and the previous one or more time frames are compared. When the change exceeds a preset anomaly threshold or meets preset anomaly conditions, the grid cell and lung region to which the corresponding acquisition point belongs are identified as abnormal regions.

[0094] For example, the preset abnormal conditions include at least one of the following situations: The spatial distribution of local indices deviates from the preset normal distribution pattern: When the local indices of multiple grid units show regional clustering anomalies (such as all grid units corresponding to a certain lung lobe having low indices), deviating from the preset uniform distribution pattern, it is judged as abnormal.

[0095] The respiratory phase delay index exceeds the preset normal range: when the difference between the onset time of inspiration / expiration in different regions exceeds the preset threshold, it is judged as abnormal; Deviation between local small airway pressure index and local alveolar ventilation index from the preset matching relationship: Under normal circumstances, small airway pressure and alveolar ventilation should maintain a certain proportional relationship. When this proportional relationship is broken (such as pressure increases but ventilation does not increase), it is judged as pressure / ventilation mismatch. The response of a region to adjustments in respiratory mechanics parameters deviates from the preset expected pattern: When the ventilator parameters are adjusted, the ventilation index of a certain region should improve. If there is no significant improvement in that region, it is judged as an abnormal recruitment response; if the pressure index continues to increase while the ventilation gain is limited, it is judged as an overinflation risk zone. Local alveolar ventilation index consistently below preset ventilation threshold: When the alveolar ventilation index of a certain area is consistently below the preset threshold for multiple consecutive time frames, it is determined to be an area with abnormal ventilation.

[0096] The abnormal areas are marked on the chest wall grid location map at the corresponding grid cell positions of each collection point, and the location results of the abnormal areas on the chest wall surface are output; and / or the abnormal areas are marked on the lung regions corresponding to each collection point on the dynamic cloud map of the bilateral lung regions, and the abnormality prompt information for the corresponding lung regions is output. When the parameters, sub-indices, or total indices of the above-mentioned abnormal areas reach a preset threshold, the risk prompt module generates abnormal prompt information, which includes at least one prompt information generated based on abnormal distribution of regional parameters, respiratory phase delay, pressure / ventilation mismatch, abnormal regional response, and / or abnormal regional ventilation.

[0097] For example, when a region exhibits decreased breath sounds, abnormally increased frequency, elevated local small airway pressure index, and decreased alveolar ventilation index, the system can identify this region as a pressure / ventilation mismatch area or a high-risk area for obstruction; when a region exhibits significantly delayed inspiratory or expiratory phases and a delayed increase in the regional ventilation index, the system can identify this region as a local ventilation delay area; when a region shows a significant improvement in ventilation index and a decrease in pressure index after ventilation parameter adjustments, the system can identify this region as a recruitment response area; when a region experiences a sustained increase in pressure index but limited ventilation gain, the system can identify this region as an overinflation risk area.

[0098] Furthermore, when the local indices and / or respiratory cloud indices of the abnormal area reach preset alarm thresholds, audible and / or visual alarms are triggered. Audible alarms include beeping sounds, voice announcements, or combinations thereof; visual alarms include color changes, icon flashing, area outline highlighting, warning light flashing, or combinations thereof. Alarm levels can be divided into multiple levels based on the severity of the abnormality (e.g., yellow alert, orange warning, red emergency), with different levels corresponding to different audible and visual combinations. For example, the risk warning output interface simultaneously labels the following on the display terminal: chest wall grid location (e.g., "A10 grid abnormality"), corresponding lung lobe label (e.g., "left upper lobe"), abnormality type label (e.g., "pressure / ventilation mismatch"), and alarm level.

[0099] Therefore, by triggering audible and / or visual alarms when the local index and / or respiratory cloud index in the abnormal area reach a preset alarm threshold, a closed-loop mechanism from abnormality identification to clinical alarm is established. The abnormality condition determines "whether it is abnormal," and the alarm threshold determines "whether the abnormality is serious enough to warrant an alarm." This tiered approach avoids alarm fatigue caused by triggering alarms for minor abnormalities, while ensuring that serious abnormalities are detected promptly.

[0100] As a preferred embodiment, when multiple adjacent acquisition points in a unilateral chest wall region continuously meet preset abnormal conditions, auxiliary diagnostic prompts for unilateral pneumothorax, unilateral hemothorax, and / or unilateral hemopneumothorax are output. Specifically, when multiple adjacent acquisition points in the left or right chest region continuously exhibit decreased or absent breath sounds, decreased local alveolar ventilation index, increased respiratory cloud heterogeneity sub-index, and / or unilateral temporal abnormalities in consecutive time frames, the system designates the abnormal grid cluster (multiple adjacent grid units) as a high-risk area for unilateral pleural cavity abnormalities. When the unilateral abnormal grid cluster mainly manifests as decreased breath sounds and decreased ventilation index, the system outputs an auxiliary diagnostic prompt for unilateral pneumothorax; when the unilateral abnormal grid cluster is accompanied by more obvious pressure / ventilation mismatch abnormalities, the system further outputs auxiliary diagnostic prompts for unilateral hemothorax and / or unilateral hemopneumothorax, and triggers a higher-priority audible and visual alarm.

[0101] It should be noted that, for the method embodiments, the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps may be performed in other orders or simultaneously.

[0102] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0103] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0107] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

Claims

1. A lung respiratory cloud map and index localization method based on multi-source fusion, characterized in that, The methods include: Multiple collection points were set up in different areas of the chest wall of the subject to collect respiratory sound signals from multiple chest wall areas. Acquire respiratory mechanics parameters, which include at least one of airway pressure parameters, ventilation parameters, chest wall movement parameters, and regional impedance change parameters; Based on the standardized chest wall grid template, the double lung plane template, and the spatial coordinates of multiple acquisition points set in each chest wall region, at least one acquisition point corresponding to each of the different lung regions is determined. Extract the respiratory sound features of the respiratory sound signal collected at each of the acquisition points; The local breath sound index for each lung region corresponding to each of the acquisition points is calculated based on the breath sound characteristics, and the local small airway pressure index and local alveolar ventilation index are obtained by inversion with the respiratory mechanics parameters. Based on the local breath sound index, the local small airway pressure index, and the local alveolar ventilation index, combined with regional heterogeneity indicators and respiratory phase delay indicators, the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index are calculated. The respiratory mechanics parameters, local indices, and respiratory cloud indices are mapped to dynamic cloud maps of the two lung regions and chest wall grid positioning maps according to the correspondence between each collection point and the grid unit and the lung region, respectively. Based on the changing trends of continuous time frames, the system outputs abnormality alert information corresponding to the lung regions according to the dynamic cloud map of the bilateral lung regions, and / or outputs the location results of the abnormal regions on the chest wall surface according to the chest wall grid positioning map.

2. The lung respiratory cloud map and index localization method based on multi-source fusion according to claim 1, characterized in that, Multiple sampling points were evenly distributed in different regions of the subject's chest wall to collect breath sound signals from multiple chest wall regions, including: Each grid cell of the standardized chest wall grid template is assigned a unique location number; Each of the aforementioned grid units is projected onto different chest wall regions, and multiple acquisition points corresponding to each of the aforementioned grid units are deployed. An acoustic sensor is set at each of the acquisition points, and the acquisition point where each acoustic sensor is located corresponds to the location number of at least one of the grid cells; The respiratory sound signals of the corresponding chest wall region where the grid unit is located are collected by each of the acoustic sensors.

3. The lung respiratory cloud map and index localization method based on multi-source fusion according to claim 2, characterized in that, The step of assigning a unique location number to each grid cell of the standardized chest wall grid template includes: The standardized chest wall grid template is divided into anterior chest grid area, lateral chest grid area and back grid area, and the multiple grid units contained in each grid area are respectively labeled with the position number in sequence; Based on the divided grid areas, at least one of the following chest wall regions is selected: left anterior chest, left lateral chest, left posterior chest, right anterior chest, right lateral chest, and right posterior chest. The step of projecting each of the grid cells onto different chest wall regions and setting up multiple acquisition points corresponding to each of the grid cells includes: Each of the acquisition points located on the anterior chest of the left and right lungs corresponds to at least one of the grid cells in the anterior chest grid area; Each of the acquisition points located on the left and right lateral chest lungs corresponds to at least one grid cell in the lateral chest grid area; Each of the acquisition points located on the back of the left lung and the back of the right lung corresponds to at least one grid cell in the back grid area.

4. A lung respiratory cloud map and index localization method based on multi-source fusion according to claim 2 or 3, characterized in that, The bilateral lung plane template includes at least one of the following lung regions: left upper lobe, left lower lobe, right upper lobe, right middle lobe, and right lower lobe; The determination of at least one collection point corresponding to each of the different lung regions includes: Each of the grid units in the chest wall standardized grid template is mapped onto the bilateral lung plane template according to the anatomical projection relationship, so that the multiple grid units correspond to at least one of the lung regions of the left upper lobe, the left lower lobe, the right upper lobe, the right middle lobe, and the right lower lobe; Based on the lung region corresponding to each of the grid cells and the grid cell corresponding to each collection point, the mapping relationship between each collection point and each of the lung regions is determined.

5. The lung respiratory cloud map and index localization method based on multi-source fusion according to claim 4, characterized in that, The ventilation parameters include airflow parameters and / or tidal volume parameters; and / or, the respiratory sound characteristics include respiratory sound intensity characteristics, frequency characteristics, time-frequency characteristics, and respiratory phase characteristics; and / or, the local respiratory sound index includes a local respiratory sound intensity index and a local respiratory sound frequency index; wherein... The inversion yields the local small airway pressure index and the local alveolar ventilation index, including: The local small airway pressure index is calculated based on the joint inversion of at least two types of information: the time-frequency characteristics, the respiratory phase characteristics, the airway pressure parameters, the airflow parameters, the tidal volume parameters, the thoracic motion parameters, and the regional impedance change parameters. The local alveolar ventilation index is calculated based on the joint inversion of at least two types of information: the temporal variation of the breath sound intensity characteristics, the respiratory phase characteristics, the temporal variation of the tidal volume parameters, the temporal variation of the chest wall motion parameters, and the temporal variation of the regional impedance parameters.

6. The lung respiratory cloud map and index localization method based on multi-source fusion according to claim 5, characterized in that, The calculation of the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud temporal delay sub-index includes: The total respiratory cloud index, the heterogeneity sub-index of respiratory cloud, and the phase delay sub-index of respiratory cloud are each preset to a normal reference range; The actual values ​​of the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index are compared with their corresponding normal reference intervals to obtain the deviation of each respiratory cloud index from its corresponding normal reference interval. Calculate the individual score for each of the respiratory cloud indices based on the deviation amount; The individual scores of the total respiratory cloud index, the individual scores of the respiratory cloud heterogeneity index, and the individual scores of the respiratory cloud phase delay index are weighted and summed according to preset weighting coefficients to obtain a comprehensive evaluation value.

7. The lung respiratory cloud map and index localization method based on multi-source fusion according to claim 5, characterized in that, The process of mapping the respiratory mechanics parameters, local indices, and respiratory cloud indices to grid cells and lung regions according to the correspondence between each collection point and the lung region, respectively, into a dynamic cloud map of the bilateral lung regions and a grid positioning map of the chest wall, includes: The local small airway pressure index, the local alveolar ventilation index, the local small airway pressure index, and the local alveolar ventilation index are located at the positions corresponding to the grid units on the standardized chest wall grid template according to the correspondence between each collection point and each grid unit; and according to the correspondence between each collection point and each lung region, they are located at the positions corresponding to each lung region on the bilateral lung plane template. Each of the grid cells and each of the local indices on each of the lung regions are rendered using preset visual encoding rules to generate the chest wall grid positioning map and the dynamic cloud map of the bilateral lung regions. The respiratory cloud total index, the respiratory cloud heterogeneity sub-index, the respiratory cloud phase delay sub-index, and at least one of the respiratory mechanics parameters are displayed in association with a preset information area in the dynamic cloud map of the bilateral lung regions and / or the chest wall grid positioning map.

8. The lung respiratory cloud map and index localization method based on multi-source fusion according to claim 1, characterized in that, The method of outputting abnormality alerts for the lung regions based on the change trends of continuous time frames, according to the dynamic cloud map of the bilateral lung regions, and / or outputting the location results of the abnormal regions on the chest wall surface according to the chest wall grid positioning map, includes: Based on the changes of the respiratory mechanics parameters, the local indices, and the respiratory cloud indices between consecutive time frames, the grid cells and lung regions corresponding to the collection points that meet the preset abnormal conditions are identified as abnormal regions. The abnormal areas are marked on the grid cells corresponding to each of the collection points on the chest wall grid positioning map, and the positioning results of the abnormal areas on the chest wall surface are output. And / or, mark the abnormal areas on the lung regions corresponding to each of the collection points in the dynamic cloud map of the bilateral lung regions, and output the abnormal prompt information corresponding to the lung regions; When the grid cells corresponding to multiple adjacent acquisition points in a unilateral chest wall region continuously meet the preset abnormal conditions, auxiliary diagnostic prompts for unilateral pneumothorax, unilateral hemothorax, and / or unilateral hemopneumothorax are output. Specifically, when each of the local indices and / or the respiratory cloud indices in the abnormal region reaches a preset alarm threshold, an audible alarm and / or a visual alarm are triggered.

9. A lung respiratory cloud map and index localization system based on multi-source fusion, characterized in that, The system includes: The breath sound acquisition module is used to set up multiple acquisition points in different areas of the chest wall of the subject and acquire breath sound signals from multiple chest wall areas. A respiratory mechanics parameter acquisition module is used to acquire respiratory mechanics parameters, which include at least one of airway pressure parameters, ventilation parameters, chest wall motion parameters, and regional impedance change parameters. The lung region mapping module is used to determine at least one of the acquisition points corresponding to different lung regions based on the standardized chest wall grid template, the dual lung plane template, and the spatial coordinates of multiple acquisition points set in each chest wall region. The feature extraction module is used to extract the respiratory sound features of the respiratory sound signal collected at each of the acquisition points; The region inversion module is used to calculate the local breath sound index of each lung region corresponding to each of the acquisition points based on the breath sound characteristics, and to invert the local small airway pressure index and local alveolar ventilation index by combining the respiratory mechanics parameters. The index calculation module is used to calculate the total respiratory cloud index, the respiratory cloud heterogeneity sub-index, and the respiratory cloud phase delay sub-index based on the local breath sound index, the local small airway pressure index, and the local alveolar ventilation index, combined with regional heterogeneity indicators and respiratory phase delay indicators. The dynamic cloud map generation module is used to map the respiratory mechanics parameters, local indices, and respiratory cloud indices into dynamic cloud maps of the two lung regions and chest wall grid positioning maps respectively, according to the correspondence between each collection point and the grid unit and the lung region. The anomaly localization and alert module is used to output anomaly alert information corresponding to the lung region based on the change trend of the continuous time frame and the dynamic cloud map of the bilateral lung regions, and / or output the localization result of the abnormal region on the chest wall surface based on the chest wall grid localization map.

10. A lung respiratory cloud map and index localization system based on multi-source fusion according to claim 9, characterized in that, The respiratory sound acquisition module includes multiple acoustic sensors disposed in various regions of the chest wall; The respiratory mechanics parameter acquisition module includes at least one of the following sensing units: The pressure sensing unit includes a fiber optic pressure sensor and / or a MEMS pressure sensor disposed on the endotracheal tube, for acquiring the airway pressure parameters. A ventilator sensing unit is installed on the ventilator tubing to acquire the ventilation parameters; A thoracic motion monitoring unit, mounted on the chest wall, is used to acquire the thoracic motion parameters; An impedance imaging unit, mounted on the chest wall, is used to acquire impedance change parameters of the region.

Images