Method and system for pediatric lung function image intelligent analysis based on ai and brain-computer technology

Abstract

The application provides a kind of method and system for pediatric lung function image intelligent analysis based on AI and brain-computer technology and computing device, the method comprises: obtaining the brain neurophysiological signal of target personnel, respiratory dynamics image data and physiological sign data, constructs multi-source heterogeneous data set;The multi-source heterogeneous data set is preprocessed, to generate standardized image data;The characteristic extraction is carried out to the standardized image data, and first characteristic vector, second characteristic vector and third characteristic vector are obtained;The first characteristic vector, the second characteristic vector and the third characteristic vector are fused and calculated, to obtain the respiratory function quantitative index of the target personnel;Based on the respiratory function quantitative index, output analysis report.According to the technical scheme of the application, non-invasive intelligent examination and evaluation can be completed without the active cooperation of the subject, the risk of sedation is reduced, the success rate of subject examination and evaluation accuracy are improved, and the technical innovation from passive detection to active evaluation is realized.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence and brain-computer interface fusion technology, specifically to a method, system, and computing device for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer technology. Background Technology

[0002] Pulmonary function testing in children is a crucial step in the diagnosis and treatment of respiratory diseases and health assessment. However, due to the physiological and behavioral characteristics of children, it faces multiple technical bottlenecks in clinical practice. Traditional pulmonary function testing relies on the subject's active cooperation in completing instructed breathing actions (such as forceful inhalation, exhalation, and breath-holding). However, infants and preschool children, whose cognitive and behavioral abilities are not yet mature, often have difficulty understanding or executing the instructions, resulting in poor test compliance, low repeatability, and even test failure or inaccurate results.

[0003] For young children, sedation or anesthesia is often used in clinical practice to assist in testing. However, these methods have problems such as respiratory depression, circulatory risks, and postoperative recovery complications, and require additional medical resources, making it difficult to meet the needs of routine screening and dynamic monitoring. At the same time, while existing imaging and functional assessment methods (such as chest X-rays, lung ultrasound, and pulse oscillation techniques) can provide auxiliary information, their image characteristics are significantly affected by age differences, and they have limitations such as strong subjectivity in image interpretation and easy to miss small lesions.

[0004] Currently, there is an urgent clinical need for a non-invasive, non-cooperative, and intelligent assessment solution that can obtain accurate lung function parameters in real time, in order to overcome the dependence on cooperation in children's lung function testing and provide a reliable basis for the early diagnosis and individualized management of children's respiratory diseases.

[0005] Therefore, a technical solution is needed that can achieve non-invasive intelligent examination and assessment without the active cooperation of the subject, reduce the risk of sedation, improve the success rate of the subject's examination and the accuracy of the assessment, and realize the technological innovation from passive examination to active assessment. Summary of the Invention

[0006] This invention aims to provide a method, system, and computing device for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology. It can achieve non-invasive intelligent examination and assessment without the active cooperation of the subject, reduce the risk of sedation, improve the success rate of subject examination and assessment accuracy, and realize a technological innovation from passive detection to active assessment.

[0007] According to one aspect of the present invention, a method for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology is provided, the method comprising:

[0008] Acquire brain neurophysiological signals, respiratory dynamics imaging data and physiological signs data of target personnel, and construct a multi-source heterogeneous dataset;

[0009] The multi-source heterogeneous dataset is preprocessed to generate standardized image data;

[0010] Feature extraction is performed on the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector;

[0011] The first feature vector, the second feature vector, and the third feature vector are fused and calculated to obtain the respiratory function quantitative index of the target person.

[0012] An analysis report is generated based on the aforementioned quantitative indicators of respiratory function.

[0013] According to some embodiments, brain neurophysiological signals, respiratory dynamics imaging data, and physiological signs data of target personnel are acquired to construct a multi-source heterogeneous dataset, including:

[0014] The brain's neurophysiological signals are collected using a wearable brain-computer interface device;

[0015] The respiratory dynamics imaging data and physiological signs data are collected by external sensing devices.

[0016] According to some embodiments, the multi-source heterogeneous dataset is preprocessed to generate standardized image data, including:

[0017] The multi-source heterogeneous dataset is time-stamp aligned and denoised, and the respiratory dynamics imaging data is normalized to anatomical scale based on the physiological signs data to generate standardized imaging data.

[0018] According to some embodiments, feature extraction is performed on the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector, including:

[0019] The standardized image data is processed by an image encoder based on a convolutional neural network to extract features and generate the first feature vector, which is a lung function feature vector.

[0020] The brain neurophysiological signals are feature extracted using a neural encoder based on a long short-term memory network and / or attention mechanism to generate a second feature vector and a third feature vector. The second feature vector is a respiratory center intention feature vector, and the third feature vector is a respiratory effort feature vector.

[0021] According to some embodiments, the first feature vector, the second feature vector, and the third feature vector are fused and calculated to obtain the respiratory function quantitative index of the target person, including:

[0022] The first feature vector, the second feature vector, and the third feature vector are input into the cross-modal fusion network;

[0023] The first feature vector, the second feature vector, and the third feature vector are aligned and fused with image temporal data and neural temporal data through the cross-modal fusion network to obtain a fused feature vector;

[0024] Based on the fused feature vector, the respiratory function quantitative index of the target person is calculated.

[0025] According to some embodiments, the cross-modal fusion network includes a feature alignment layer, a feature fusion layer, and an index prediction layer, wherein,

[0026] The feature alignment layer uses a time series alignment algorithm to map the image time series data and the neural signal time series data to a unified time coordinate system.

[0027] The feature fusion layer uses an attention mechanism to calculate the weights of each modality feature, and then concatenates and adds the weighted features to obtain the fused feature vector.

[0028] The indicator prediction layer calculates the respiratory function quantitative indicators based on the fused feature vector using a regression algorithm.

[0029] According to some embodiments, the quantitative indicators of respiratory function include ventilation heterogeneity index, small airway function parameters, and respiratory dynamic impedance parameters.

[0030] According to some embodiments, the neurophysiological signals include electroencephalogram (EEG) signals and / or cerebral oxygenation fNIRS signals;

[0031] The respiratory dynamics imaging data includes at least one of the following: pulse oscillation IOS impedance spectrum, lung ultrasound images, chest X-ray images, and chest and abdominal wall motion images.

[0032] The physiological data include at least one of heart rate, blood oxygen saturation, and skin conductance.

[0033] According to another aspect of the present invention, a system for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology is provided, the system comprising:

[0034] The multimodal acquisition unit is used to acquire the target personnel's brain nerve physiological signals, respiratory dynamics imaging data and physiological signs data, and to construct a multi-source heterogeneous dataset;

[0035] The data preprocessing unit is used to preprocess the multi-source heterogeneous dataset to generate standardized image data;

[0036] The feature extraction unit is used to extract features from the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector.

[0037] The intelligent deep fusion analysis unit is used to fuse and calculate the first feature vector, the second feature vector and the third feature vector to obtain the respiratory function quantitative index of the target person.

[0038] The results output unit is used to output an analysis report based on the quantitative indicators of respiratory function.

[0039] According to another aspect of the present invention, a computing device is provided, comprising:

[0040] Processor; and

[0041] A memory that stores a computer program, which, when executed by the processor, implements the method as described in any of the preceding methods.

[0042] According to embodiments of the present invention, the present invention first acquires the target individual's brain neurophysiological signals, respiratory dynamic imaging data, and physiological sign data, constructs a multi-source heterogeneous dataset, preprocesses the multi-source heterogeneous dataset to generate standardized imaging data, extracts features from the standardized imaging data to obtain a first feature vector, a second feature vector, and a third feature vector, and fuses and calculates the first feature vector, the second feature vector, and the third feature vector to obtain the target individual's respiratory function quantitative index, and finally outputs an analysis report based on the respiratory function quantitative index. The present invention enables non-invasive intelligent examination and assessment without the subject's active cooperation, reduces the risk of sedation, improves the success rate of subject examination and assessment accuracy, and achieves a technological innovation from passive detection to active assessment.

[0043] According to embodiments of the present invention, by simultaneously acquiring pediatric pulmonary function images and neural signals such as electroencephalography (EEG) and cerebral oxygenation, deep learning is used to extract pulmonary function image features and respiratory-related neural activity features, respectively. These are then fused via a cross-modal network to achieve pulmonary function index prediction, abnormality grading, and quality control assessment. This invention allows for non-invasive pulmonary function assessment without requiring active cooperation from the child, reducing the risk of sedation, and improving the success rate and accuracy of examinations in infants and young children. It is suitable for intelligent screening and diagnosis of pulmonary function in pediatric outpatient clinics, neonatal intensive care units, and primary healthcare institutions.

[0044] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit the invention. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0046] Figure 1 A flowchart illustrating a method for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology, according to an example embodiment, is shown.

[0047] Figure 2 A schematic diagram showing the measurement channel numbers for oxyhemoglobin and deoxyhemoglobin according to an example embodiment is provided.

[0048] Figure 3 A schematic diagram showing the numerical distribution of the characteristic parameters of the six standard frequency bands according to an example embodiment is provided.

[0049] Figure 4 A schematic diagram of the raw voltage waveform of a continuous electroencephalogram (EEG) according to an example embodiment is shown.

[0050] Figure 5 This diagram illustrates the distribution of blood oxygen dynamics in the cerebral cortex during a breathing task according to an example embodiment.

[0051] Figure 6 A schematic diagram illustrating the differences in cortical response under different respiratory states or abnormal conditions according to an example embodiment.

[0052] Figure 7 A schematic diagram of a system for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology, according to an example embodiment, is shown.

[0053] Figure 8 A block diagram of a computing device according to an exemplary embodiment is shown. Detailed Implementation

[0054] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that the invention will be thorough and complete, and the concept of the exemplary embodiments will be fully conveyed to those skilled in the art. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted.

[0055] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a full understanding of embodiments of the invention. However, those skilled in the art will recognize that the technical solutions of the invention can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of the invention.

[0056] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0057] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0058] It should be understood that although the terms first, second, third, etc., may be used herein to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another. Therefore, the first component discussed below may be referred to as the second component without departing from the teachings of the present invention. As used herein, the term "and / or" includes all combinations of any one and more of the associated listed items.

[0059] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of exemplary embodiments, and the modules or processes in the drawings are not necessarily essential for implementing the present invention, and therefore cannot be used to limit the scope of protection of the present invention.

[0060] Pulmonary function testing in children is a crucial basis for the diagnosis and treatment of respiratory diseases (such as asthma, bronchitis, and pulmonary dysplasia) and for health management; its clinical importance is undeniable. However, due to the unique physiological structure and cognitive development of children, traditional testing techniques have long faced multiple insurmountable bottlenecks in clinical practice. Current mainstream pulmonary function tests heavily rely on the active cooperation of the subjects, requiring the completion of standardized mandatory breathing actions (such as forceful inhalation, explosive exhalation, and breath-holding for a specific duration). For infants and preschool children, due to their immature cognitive abilities and difficulty in maintaining sustained attention, they often cannot understand or accurately execute complex instructions, resulting in extremely poor test compliance and low data repeatability. Numerous clinical cases have therefore resulted in test failures or distorted results, seriously affecting the accuracy and reliability of diagnosis.

[0061] To address this challenge, sedation or general anesthesia is often required to complete testing in young children. However, such invasive procedures not only carry potential medical risks such as respiratory depression, circulatory system fluctuations, and postoperative recovery complications, but also require specialized anesthesia teams and monitoring equipment, significantly increasing medical costs and operational complexity, making it difficult to meet the actual needs of large-scale epidemiological screening and long-term dynamic monitoring. Furthermore, while existing auxiliary assessment methods (such as chest X-rays, lung ultrasound, and pulse oscillation techniques) can provide some imaging or functional information, their image characteristics are significantly affected by age differences, and the interpretation process is highly subjective, limiting their ability to identify early, minor lesions or subtle functional changes, easily leading to missed or misdiagnosed cases.

[0062] To this end, this invention proposes a method and system for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology. This system enables non-invasive intelligent examination and assessment without the active cooperation of the subject, reducing the risk of sedation, improving the success rate and accuracy of the examination, and achieving a technological innovation from passive detection to active assessment.

[0063] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention.

[0064] Figure 1 A flowchart illustrating a method for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology, according to an example embodiment, is shown.

[0065] See Figure 1 In S101, brain neurophysiological signals, respiratory dynamics imaging data and physiological signs data of the target personnel are acquired to construct a multi-source heterogeneous dataset.

[0066] According to some embodiments, the brain nerve physiological signals are collected through a wearable brain-computer interface device, and the respiratory dynamics imaging data and physiological signs data are collected through an external sensing device.

[0067] According to some embodiments, brain neurophysiological signals of target personnel are collected through wearable brain-computer interface devices, and respiratory dynamic imaging data and physiological sign data of target personnel are collected through external sensing devices to construct a multi-source heterogeneous dataset.

[0068] According to some embodiments, the neurophysiological signals include electroencephalogram (EEG) signals and / or cerebral oxygenation fNIRS signals. The respiratory dynamic imaging data include at least one of impulse oscillation (IOS) impedance spectrum, lung ultrasound images, chest X-ray images, and chest and abdominal wall motion images. The physiological signs data include at least one of heart rate, blood oxygen saturation, and skin conductance.

[0069] Figure 2A schematic diagram showing the measurement channel numbers for oxyhemoglobin and deoxyhemoglobin according to an example embodiment is provided.

[0070] See Figure 2 , Figure 2 This paper demonstrates an example of neurophysiological signals acquired using a near-infrared brain function signal acquisition device, measured via channels numbered for oxyhemoglobin (HbO) and deoxyhemoglobin (HbR). In this protocol, these channels cover cortical regions related to respiratory control (such as the prefrontal cortex and supplemental motor area). By simultaneously recording changes in HbO and HbR concentrations, quantitative indicators of neurometabolic activity in children at rest or during natural breathing are directly obtained, reflecting the strength of respiratory drive without requiring active exhalation.

[0071] According to some embodiments, this scheme also performs data validity screening based on brain neurophysiological signals. Motion artifacts, crying characteristics, or agitation characteristics in brain neurophysiological signals are monitored in real time. When the detected characteristics exceed a preset threshold, the respiratory dynamics imaging data for the corresponding time period are marked as invalid data and discarded.

[0072] In S103, the multi-source heterogeneous dataset is preprocessed to generate standardized image data.

[0073] According to some embodiments, the multi-source heterogeneous dataset is subjected to timestamp alignment and denoising processing, and the respiratory dynamics imaging data is subjected to anatomical scale normalization processing based on the physiological sign data to generate standardized imaging data.

[0074] According to some embodiments, the multi-source heterogeneous dataset is preprocessed and normalized. The dataset undergoes timestamp alignment and denoising, and based on the age or physiological parameters of the target individuals, the respiratory dynamics imaging data is anatomically normalized to generate standardized imaging data.

[0075] According to some embodiments, anatomical scale normalization processing based on the age or physiological parameters of the target person specifically includes: calling the corresponding anatomical scale correction algorithm according to the age group of the target person (e.g., newborn, infant, preschool and school-age, etc.), and dynamically scaling the spatial resolution of the image according to the height or weight parameters of the target person.

[0076] In S105, feature extraction is performed on the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector.

[0077] According to some embodiments, a convolutional neural network-based image encoder is used to extract features from the standardized image data to generate a first feature vector, which is a lung function feature vector. A neural encoder based on a long short-term memory network and / or attention mechanism is used to extract features from the brain's neurophysiological signals to generate a second feature vector and a third feature vector, where the second feature vector is a respiratory center intention feature vector and the third feature vector is a respiratory effort feature vector.

[0078] According to some embodiments, somatosensory evoked potential features are extracted from brain nerve physiological signals, and blood oxygenation change features of the prefrontal cortex and motor cortex are extracted from brain blood oxygenation signals. Based on the somatosensory evoked potential features and blood oxygenation change features, feature vectors reflecting the passive respiratory response and physiological stress level of the target personnel are quantitatively generated.

[0079] Figure 3 A schematic diagram showing the numerical distribution of the characteristic parameters of the six standard frequency bands according to an example embodiment is provided.

[0080] See Figure 3 This figure presents the characteristic parameters of six standard frequency bands extracted from EEG signals. Each band corresponds to a specific numerical distribution map. The AARF value is 24.8 for the Delta band (1.5-4Hz), 20.4 for the Theta band (4-7.5Hz), 23.6 for the Alpha band (7.5-14Hz), 19.5 for the Beta-1 band (14-20Hz), 22.5 for the Beta-2 band (20-30Hz), and 18.6 for the Gamma band (30-40Hz). These parameters are adaptive autoregressive features calculated from the original EEG neurophysiological signals, belonging to the first feature vector, used to quantify the energy distribution of different cortical rhythms. In pulmonary function assessment, these feature vector values ​​are used to determine the child's sedation level, attention state, and the intensity of neural oscillations locked with the respiratory rhythm, serving as one of the quantitative indicators of respiratory function, thereby assisting in abnormal grading and quality control judgment.

[0081] Figure 4 A schematic diagram of the raw voltage waveform of a continuous electroencephalogram (EEG) according to an example embodiment is shown.

[0082] See Figure 4The figure shows the raw voltage waveform (in microvolts) of brain neurophysiological signals acquired via continuous EEG. According to the example embodiment, analysis of this waveform reveals multiple steep negative deflections and recovery processes, typically reflecting the changes in the electrical activity of cortical neuronal clusters over time. Using the analysis method in this scheme, this raw signal, without frequency domain conversion, is directly used for signal quality assessment, such as detecting the presence of electromyographic artifacts or poor electrode contact. Simultaneously, temporal features (such as zero-crossing rate, peak amplitude, and waveform slope) are extracted from this waveform and synchronized with the lung function imaging acquisition time, serving as the second feature vector input to a deep cross-modal fusion network for predicting quantitative indicators of respiratory function.

[0083] Figure 5 This diagram illustrates the distribution of blood oxygen dynamics in the cerebral cortex during a breathing task according to an example embodiment.

[0084] See Figure 5 According to the example implementation, the distribution of blood oxygen dynamics in the cerebral cortex during a breathing task was analyzed: blue areas represent decreased blood oxygenation (decreased HbO or increased HbR), and red areas represent increased blood oxygenation (increased HbO). Activation was mainly concentrated in the precentral gyrus and the area near the postcentral gyrus, corresponding to the motor control and sensory feedback cortex related to breathing. The left and right hemispheres showed functionally symmetrical but intensity-differentiated activation patterns, reflecting the unevenness of cortical involvement during respiratory drive. This result is the neural lateral input feature, i.e., the third feature vector, which is used for subsequent cross-modal fusion analysis with the first and second feature vectors.

[0085] In S107, the first feature vector, the second feature vector, and the third feature vector are fused and calculated to obtain the respiratory function quantitative index of the target person.

[0086] According to some embodiments, the first feature vector, the second feature vector, and the third feature vector are input into a cross-modal fusion network. The cross-modal fusion network aligns and fuses the first feature vector, the second feature vector, and the third feature vector with image temporal data and neural temporal data to obtain a fused feature vector. Based on the fused feature vector, the respiratory function quantification index of the target person is calculated.

[0087] According to some embodiments, the cross-modal fusion network includes a feature alignment layer, a feature fusion layer, and an indicator prediction layer. The feature alignment layer uses a time-series alignment algorithm to map the image time-series data and the neural signal time-series data to a unified time coordinate system. The feature fusion layer uses an attention mechanism to calculate the weights of each modality feature, and then concatenates and adds the weighted features to obtain the fused feature vector. The indicator prediction layer calculates the respiratory function quantification index based on the fused feature vector using a regression algorithm.

[0088] According to some embodiments, the index prediction layer predicts quantitative values ​​related to ventilation distribution, airway resistance, and lung volume—i.e., quantitative indicators of respiratory function—based on the spliced ​​fused feature vector and using a regression algorithm. These quantitative indicators include the ventilation heterogeneity index, small airway function parameters, and respiratory dynamic impedance parameters.

[0089] In S109, an analysis report is output based on the aforementioned respiratory function quantification indicators.

[0090] According to some embodiments, after obtaining the quantitative indicators of respiratory function, the corresponding report template is automatically matched based on the age group in the target person's physiological sign data, such as a newborn template, an infant template, and a preschool template. The quantitative indicators of respiratory function are mapped to preset fields in the report template, and the calculated values ​​are automatically compared with the standard values ​​for the same age group to generate difference values ​​(such as percentage deviation).

[0091] According to some embodiments, the analysis report includes graphical elements to aid visualization. For example, heatmap overlay: the ventilation heterogeneity index is converted into color codes and overlaid on the target individual's chest image (chest X-ray or ultrasound) to generate a fused image with lesion highlights. Temporal waveform plot: the time series of neural signals aligned with the cross-modal fusion network is plotted side-by-side with the respiratory dynamics time series, demonstrating the synchronicity or lag between neural drive and pulmonary response.

[0092] According to some embodiments, the analysis report supports real-time preview of the electronic version of the analysis report on the local display terminal. The electronic version of the analysis report is automatically pushed to the doctor's workstation through the hospital's intranet interface, and is also uploaded to the cloud storage for later access.

[0093] According to some implementations, the electronic analysis report includes an interactive index. When viewing a certain abnormal indicator in the report, clicking on the indicator can directly jump to retrieve the corresponding original data fragment, enabling rapid tracing of the chain of evidence.

[0094] Figure 6 A schematic diagram illustrating the differences in cortical response under different respiratory states or abnormal conditions according to an example embodiment.

[0095] See Figure 6According to the example embodiment, through fusion calculation, the relationship between respiratory state and brain neurophysiological signals can be analyzed. Cortical responses differ under different respiratory states or abnormal conditions. As shown in the figure, the left region exhibits widespread elevated blood oxygenation (red), indicating that this hemisphere bears the main load during respiratory regulation or compensation. The right region, on the other hand, is dominated by decreased blood oxygenation (blue), showing inhibition of neural activity or reduced resource allocation. Finally, the analysis shows that asymmetric activation patterns can serve as important neural markers for identifying abnormal respiratory patterns and can be used in the analysis report for grading of abnormal lung function and assessing functional compensation.

[0096] The following describes a specific implementation. First, the child wears an fNIRS / EEG headband to collect physiological signals while in a quiet state (natural breathing or sleep). Simultaneously, multi-source medical imaging data, such as IOS respiratory images, lung ultrasound, or chest X-rays, as well as brain function and physiological signals, are acquired. Next, AI algorithms are used to preprocess the collected data, including signal denoising, multimodal data registration, and age-based normalization, to eliminate individual differences and noise interference. Subsequently, a bimodal coding network is used to extract deep features from the physiological signals and imaging data, and these are jointly analyzed using a fusion prediction model to output lung function assessment results and quality control confidence levels. The system automatically generates a pediatric-specific report containing assessment indicators, functional classifications, abnormality alerts, and clinical recommendations. Finally, the report is reviewed and confirmed by a professional physician, who then conducts appropriate clinical interventions, achieving non-invasive, precise, and intelligent assessment and management of pediatric lung function.

[0097] The present invention, through processing, feature extraction, and fusion calculation and analysis of the collected brain nerve physiological signals, respiratory dynamic imaging data and physiological signs data, establishes for the first time a complete causal chain model of "brain respiratory center activity - respiratory muscle effort - chest wall movement - lung ventilation imaging - lung function indicators".

[0098] This invention incorporates growth and development priors, dynamically adjusting feature weights and criteria according to age group (newborn / infant / preschool / school-aged). It does not rely on the subject's active breathing; instead, it determines breathing intention through neural signals and, combined with passive image acquisition, achieves non-invasive, passive, real-time, and accurate detection and assessment of lesions in the natural state of infants / young children.

[0099] Figure 7 A schematic diagram of a system for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology, according to an example embodiment, is shown.

[0100] See Figure 7 The system for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology includes a multimodal acquisition unit 01, a data preprocessing unit 02, a feature extraction unit 03, an intelligent deep fusion analysis unit 04, and a result output unit 05.

[0101] According to some embodiments, the multimodal acquisition unit 01 is used to acquire the target person's brain neurophysiological signals, respiratory dynamics imaging data, and physiological signs data to construct a multi-source heterogeneous dataset. The data preprocessing unit 02 is used to preprocess the multi-source heterogeneous dataset to generate standardized image data. The feature extraction unit 03 is used to extract features from the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector. The intelligent deep fusion analysis unit 04 is used to fuse and calculate the first feature vector, the second feature vector, and the third feature vector to obtain the target person's respiratory function quantitative index. The result output unit 05 is used to output an analysis report based on the respiratory function quantitative index.

[0102] According to some embodiments, the system of the present invention collects brain neurophysiological signals, respiratory dynamic imaging data and physiological signs data of target personnel through a bedside portable device (NICU / pediatric outpatient clinic), inputs the data into an artificial intelligence medical detection and analysis system, performs data preprocessing, feature extraction and fusion calculation analysis, and realizes non-invasive, passive, real-time and accurate generation of analysis reports, and quickly and conveniently assists in assessing the physical condition of the subject.

[0103] According to some embodiments, the system of the present invention interfaces with the hospital's PACS / EMR system to digitally manage and store analysis reports. Simultaneously, the system connects to a cloud-based SaaS platform to upload analysis reports to cloud storage for later retrieval.

[0104] The system of this invention acquires lung function images and neural signals such as electroencephalogram (EEG) and cerebral blood oxygenation of subjects simultaneously. It uses deep learning to extract lung function image features and respiratory-related neural activity features, and realizes index prediction, abnormality classification and quality control judgment through a cross-modal fusion network.

[0105] This invention effectively solves the problems of extremely low voluntary cooperation, high risks associated with sedation / anesthesia, and poor compliance during children's examinations. It can complete non-invasive lung function assessment without the need for children's active cooperation, reducing the risk of sedation and improving the success rate and accuracy of examinations for infants and young children. It is suitable for intelligent lung function screening in pediatric outpatient clinics, neonatal intensive care units, and primary healthcare institutions.

[0106] Figure 8 A block diagram of a computing device according to an exemplary embodiment is shown.

[0107] like Figure 8 As shown, the computing device 30 includes a processor 12 and a memory 14. The computing device 30 may also include a bus 22, a network interface card 16, and an I / O interface 18. The processor 12, memory 14, network interface card 16, and I / O interface 18 can communicate with each other via the bus 22.

[0108] Processor 12 may include one or more general-purpose CPUs (Central Processing Units), microprocessors, or application-specific integrated circuits, for executing relevant program instructions. According to some embodiments, computing device 30 may also include a high-performance display adapter (GPU) 20 for accelerating processor 12.

[0109] Memory 14 may include a machine system readable medium in the form of volatile memory, such as random access memory (RAM), read-only memory (ROM), and / or cache memory. Memory 14 is used to store one or more programs containing instructions, as well as data. Processor 12 may read the instructions stored in memory 14 to perform the methods described above according to embodiments of the present invention.

[0110] The computing device 30 can also communicate with one or more networks via a network interface card 16. The network interface card is used for data processing or external communication. The network interface card includes a root system-on-a-chip (SoC) and multiple interfaces, through which the SoC performs data communication. The SoC includes a processor and a memory, on which a computer program is stored. When the processor runs the computer program stored in the memory, it implements the method according to an embodiment of the present invention.

[0111] Bus 22 can include address bus, data bus, control bus, etc. Bus 22 provides a path for exchanging information between components.

[0112] It should be noted that, in specific implementations, the computing device 30 may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the device described above may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0113] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), network storage devices, cloud storage devices, or any type of medium or device suitable for storing instructions and / or data.

[0114] This invention also provides a computer program product comprising a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments.

[0115] Those skilled in the art will clearly understand that the technical solutions of the present invention can be implemented by means of software and / or hardware. In this specification, "unit" and "module" refer to software and / or hardware capable of independently performing or cooperating with other components to perform a specific function, wherein the hardware may be, for example, a field-programmable gate array (FPGA), an integrated circuit, etc.

[0116] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0117] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0118] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0119] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0120] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0121] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention.

[0122] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0123] Exemplary embodiments of the present invention have been specifically shown and described above. It should be understood that the present invention is not limited to the detailed structures, arrangements, or implementations described herein; rather, the present invention is intended to cover various modifications and equivalent arrangements contained within the spirit and scope of the appended provisions.

Claims

1. A method for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology, characterized in that, include: Acquire brain neurophysiological signals, respiratory dynamics imaging data and physiological signs data of target personnel, and construct a multi-source heterogeneous dataset; The multi-source heterogeneous dataset is preprocessed to generate standardized image data; Feature extraction is performed on the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector; The first feature vector, the second feature vector, and the third feature vector are fused and calculated to obtain the respiratory function quantitative index of the target person. An analysis report is generated based on the aforementioned quantitative indicators of respiratory function.

2. The method according to claim 1, characterized in that, Acquire neurophysiological signals, respiratory dynamics imaging data, and physiological signs data of the target personnel to construct a multi-source heterogeneous dataset, including: The brain's neurophysiological signals are collected using a wearable brain-computer interface device; The respiratory dynamics imaging data and physiological signs data are collected by external sensing devices.

3. The method according to claim 1, characterized in that, The multi-source heterogeneous dataset is preprocessed to generate standardized image data, including: The multi-source heterogeneous dataset is time-stamp aligned and denoised, and the respiratory dynamics imaging data is normalized to anatomical scale based on the physiological signs data to generate standardized imaging data.

4. The method according to claim 1, characterized in that, Feature extraction is performed on the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector, including: The standardized image data is processed by an image encoder based on a convolutional neural network to extract features and generate the first feature vector, which is a lung function feature vector. The brain neurophysiological signals are feature extracted using a neural encoder based on a long short-term memory network and / or attention mechanism to generate a second feature vector and a third feature vector. The second feature vector is a respiratory center intention feature vector, and the third feature vector is a respiratory effort feature vector.

5. The method according to claim 4, characterized in that, The first feature vector, the second feature vector, and the third feature vector are fused and calculated to obtain the respiratory function quantitative index of the target person, including: The first feature vector, the second feature vector, and the third feature vector are input into the cross-modal fusion network; The first feature vector, the second feature vector, and the third feature vector are aligned and fused with image temporal data and neural temporal data through the cross-modal fusion network to obtain a fused feature vector; Based on the fused feature vector, the respiratory function quantitative index of the target person is calculated.

6. The method according to claim 5, characterized in that, The cross-modal fusion network includes a feature alignment layer, a feature fusion layer, and an index prediction layer. in, The feature alignment layer uses a time series alignment algorithm to map the image time series data and the neural signal time series data to a unified time coordinate system. The feature fusion layer uses an attention mechanism to calculate the weights of each modality feature, and then concatenates and adds the weighted features to obtain the fused feature vector. The indicator prediction layer calculates the respiratory function quantitative indicators based on the fused feature vector using a regression algorithm.

7. The method according to claim 6, characterized in that, The quantitative indicators of respiratory function include the ventilation heterogeneity index, small airway function parameters, and respiratory dynamic impedance parameters.

8. The method according to claim 1, characterized in that, The neurophysiological signals include electroencephalogram (EEG) signals and / or cerebral oxygenation fNIRS signals; The respiratory dynamics imaging data includes at least one of the following: pulse oscillation IOS impedance spectrum, lung ultrasound images, chest X-ray images, and chest and abdominal wall motion images. The physiological data include at least one of heart rate, blood oxygen saturation, and skin conductance.

9. A system for intelligent analysis of pediatric pulmonary function images based on AI and brain-computer interface technology, characterized in that, include: The multimodal acquisition unit is used to acquire the target personnel's brain nerve physiological signals, respiratory dynamics imaging data and physiological signs data, and to construct a multi-source heterogeneous dataset; The data preprocessing unit is used to preprocess the multi-source heterogeneous dataset to generate standardized image data; The feature extraction unit is used to extract features from the standardized image data to obtain a first feature vector, a second feature vector, and a third feature vector. The intelligent deep fusion analysis unit is used to fuse and calculate the first feature vector, the second feature vector and the third feature vector to obtain the respiratory function quantitative index of the target person. The results output unit is used to output an analysis report based on the quantitative indicators of respiratory function.

10. A computing device, characterized in that, include: processor; as well as A memory storing a computer program that, when executed by the processor, implements the method as described in any one of claims 1-8.

Images