Aerosol imaging analysis method for lung ventilation function assessment

Abstract

The present application relates to the technical field of lung function evaluation, in particular to a kind of aerosol image analysis method for lung ventilation function evaluation.First, by changing the concentration of aerosol, obtain lung aerosol image under each concentration;Based on image block level analysis, lung is divided into different functional areas (such as lung lobe, airway partition), and in comparing the particle distribution difference characteristics of different image blocks, quantification deposition index is reflected regional ventilation uniformity;Further combined with density clustering, sub-region is divided, the number change of aerosol particles is analyzed across time points, and metabolic rate index is calculated to represent local metabolic activity;Finally, deposition index and metabolic rate index are fused, and ventilation obstruction evaluation value is generated.This method couples spatial-temporal multi-parameter, effectively reduces the interference of breathing pattern and individual difference, synchronously analyzes ventilation function and metabolic state, provides quantitative basis for lung ventilation disorder analysis, and significantly improves the overall evaluation and clinical decision-making accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of lung function evaluation, and particularly relates to a lung ventilation function evaluation atomization image analysis method. BACKGROUND

[0002] Lung ventilation function is a dynamic index for measuring the process of air entering alveoli and waste gas being discharged from alveoli, and is an important means for diagnosing and monitoring respiratory diseases, and is commonly used in the diagnosis and treatment process of chronic obstructive pulmonary disease (COPD), asthma, pulmonary fibrosis and the like. The lung ventilation function can be mainly evaluated by generating aerosols with specific particle sizes through an atomization device, and then capturing the distribution of the atomized particles in the lungs of a patient, so as to assist a doctor in evaluating the lung ventilation function.

[0003] In the prior art, when the lung ventilation function is evaluated by using the distribution of atomized particles in the lungs, a single concentration of aerosols is usually set, and a static global index (such as an FEV1 / FVC ratio) is used. However, due to the individual differences caused by different inhalation methods and breathing patterns of patients, the atomized tracers with a single concentration are unevenly deposited and distributed in the lungs, and the static global index is difficult to quantitatively evaluate the dynamic deposition and metabolism of the atomized particles in the local lungs of the patient, which reduces the accuracy of the lung ventilation function evaluation. SUMMARY

[0004] In order to solve the technical problem that due to the individual differences caused by different inhalation methods and breathing patterns of patients, the atomized tracers with a single concentration are unevenly deposited and distributed in the lungs, and the static global index is difficult to quantitatively evaluate the dynamic deposition and metabolism of the atomized particles in the local lungs of the patient, which reduces the accuracy of the lung ventilation function evaluation, the purpose of the present application is to provide a lung ventilation function evaluation atomization image analysis method, and the technical solution adopted is as follows:

[0005] The concentration of the aerosols is changed in a descending order in time sequence, and lung atomization images of each patient under different concentrations are obtained;

[0006] Each lung atomization image is regionally segmented to obtain a plurality of image blocks, the particle distribution index of each image block is determined based on the distribution feature difference of the atomized particles between different image blocks in each lung atomization image, and the deposition index of the atomized particles of each image block is determined by analyzing the change trend feature of the particle distribution index in the lung atomization images of the same image block at different time points.

[0007] In the first pulmonary aerosol image, the aerosol particles in each image block are clustered based on the density characteristics of the aerosol particles, so as to divide each image block into a plurality of sub-regions; according to the position distribution of the sub-regions in each image block and the number change characteristics of the aerosol particles in the corresponding regions of the pulmonary aerosol image at different time of the sub-regions, the metabolic rate of the aerosol particles corresponding to each image block is determined;

[0008] The particle deposition index of each image block and the metabolic rate of the aerosol particles are fused, so as to obtain the ventilation obstruction evaluation value of each image block of the patient's lung, which is used for auxiliary evaluation of lung ventilation function.

[0009] Further, the particle distribution index acquisition method comprises:

[0010] In each pulmonary aerosol image, the number difference characteristics of the aerosol particles between all image blocks are analyzed to determine the first distribution factor of each image block;

[0011] Based on the density difference characteristics of the aerosol particles between all image blocks, the second distribution factor of each image block is determined;

[0012] The normalized value of the sum of the first distribution factor and the second distribution factor of each image block is taken as the particle distribution index of each image block.

[0013] Further, the first distribution factor acquisition method comprises:

[0014] In each pulmonary aerosol image, the number average of the aerosol particles of all image blocks is taken as a number reference value, and the normalized value of the difference between the number of the aerosol particles of each image block and the number reference value is taken as the first distribution factor of each image block.

[0015] Further, the second distribution factor acquisition method comprises:

[0016] In each pulmonary aerosol image, the ratio of the number of the aerosol particles to the area of each image block is taken as a density factor;

[0017] The average of the density factors of all image blocks is taken as a density reference value, and the normalized value of the difference between the density factor of each image block and the density reference value is taken as the second distribution factor of each image block.

[0018] Further, the aerosol particle deposition index acquisition method comprises:

[0019] For any one image block, in the lung atomization images at each adjacent two time points, the normalized value of the difference between the particle distribution index at the later time point and the particle distribution index at the former time point is taken as the particle deposition factor of the image block at the adjacent two time points;

[0020] The particle deposition factors of each image block are linearly fitted according to the time sequence, and the normalized value of the slope value of the obtained fitting straight line is taken as the deposition weight of each image block.

[0021] The deposition weight of each image block is multiplied by the mean value of all the particle deposition factors corresponding to each image block, and the normalized value of the obtained product is taken as the atomized particle deposition index of each image block.

[0022] Further, the method for obtaining the sub-regions comprises:

[0023] In the first lung atomization image, the atomized particles in each image block are analyzed by clustering based on the DBSCAN clustering algorithm, and all the cluster clusters are obtained, wherein the minimum sample number and the neighborhood radius are both preset values.

[0024] Each cluster cluster is taken as a sub-region, so as to obtain all the sub-regions corresponding to each image block.

[0025] Further, the method for obtaining the atomized particle metabolic rate comprises:

[0026] In the first lung atomization image, the positional relationship between the sub-regions in each image block is analyzed, so as to determine the metabolic rate confidence of each sub-region.

[0027] The number change characteristics of the atomized particles of the sub-regions in the first lung atomization image and the corresponding regions in other lung atomization images are analyzed, and the atomized particle metabolic factor of each sub-region is determined.

[0028] In each image block of the first lung atomization image, the atomized particle metabolic factor is weighted and averaged by using the metabolic rate confidence of the sub-region, and the normalized value of the obtained weighted result is taken as the atomized particle metabolic rate of each image block.

[0029] Further, the method for obtaining the metabolic rate confidence comprises:

[0030] In the first lung atomization image, in each image block, the largest sub-region is taken as the core region.

[0031] The Euclidean distance between the centroid of each sub-region and the centroid of the core region is normalized, and the normalized value is taken as the metabolic rate confidence of each sub-region.

[0032] Further, the method for obtaining the aerosol particle metabolism factor comprises:

[0033] Each sub-region in each image block in the first lung aerosol image is taken as a contrast region in a corresponding region in other lung aerosol images;

[0034] Each sub-region and all contrast regions corresponding to the sub-region are sorted in time sequence to obtain a sorting sequence, and the sub-region and all contrast regions in the sorting sequence are collectively referred to as target regions;

[0035] In the sorting sequence, the difference between the number of aerosol particles in a previous target region and the number of aerosol particles in a next target region and the time interval between the previous target region and the next target region are taken as a metabolism parameter;

[0036] The average of all metabolism parameters corresponding to each sub-region in each image block in the first lung aerosol image is taken as an aerosol particle metabolism factor of each sub-region.

[0037] Further, the method for obtaining the ventilation obstruction evaluation value comprises:

[0038] In the first lung aerosol image, the product of the aerosol particle deposition index and the aerosol particle metabolism rate corresponding to each image block is taken as a ventilation obstruction factor of each image block after negative correlation mapping and normalization;

[0039] In the first lung aerosol image, the average of the ventilation obstruction factors of all image blocks is taken as an obstruction reference value, and the difference between the ventilation obstruction factor of each image block and the obstruction reference value is taken as a ventilation obstruction evaluation value of each image block after normalization.

[0040] The present application has the following beneficial effects:

[0041] In the present application, the aerosol concentration is designed in descending order, so as to obtain the lung atomization image of the patient at different concentrations, which can be closer to the dynamic fluctuation of aerosol concentration in the natural breathing process, and can correct the influence of the patient's breathing mode (such as deep inspiration, shallow and rapid breathing) on particle deposition, and reduce the evaluation deviation caused by the difference in inhalation method. Under the condition that the patient's lung ventilation function is normal, the distribution of atomized particles will be more uniform, and with the continuous inhalation of aerosol, particle deposition will occur in the ventilation area, so by image block level analysis, the lung is divided into different regions (such as upper and lower lobes, central and peripheral airways), and then the difference characteristics of the distribution of atomized particles between different image blocks and the change trend in the lung atomization image at different concentrations are analyzed, which can quantify the atomized particle deposition index of each image block as an index for evaluating the ventilation function of the lung region corresponding to the image block. Further, the atomized particles converted by radioactive aerosol cannot be exhaled out of the body, but spread to the alveoli and metabolized from the kidneys, and the cell types and local metabolic activity of each region of the lung are different, which affect the distribution of atomized particles, so by analyzing the metabolism of atomized particles in each region of the lung, the ventilation function thereof can be further confirmed. Therefore, in the first lung atomization image, the sub-regions are divided based on the density of atomized particles, and the atomized particle metabolism rate is determined according to the position distribution of the sub-regions and the number change characteristics of the atomized particles in the sub-regions at different times, as another index for measuring the ventilation function. Finally, the two indexes for measuring the ventilation function corresponding to each image block are fused to obtain a ventilation obstruction evaluation value, which assists the doctor in evaluating the lung ventilation function and effectively improves the comprehensiveness and accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, a brief introduction will be given to the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0043] Figure 1 A method flowchart of a lung ventilation function evaluation atomization image analysis method provided by an embodiment of the present application;

[0044] Figure 2 A method flowchart of a particle distribution index acquisition method provided by an embodiment of the present application;

[0045] Figure 3 A method flowchart of an atomized particle metabolism rate acquisition method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the following describes in detail the specific implementation, structure, features and effects of a lung ventilation function evaluation aerosol image analysis method according to the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0048] The following describes in detail the specific scheme of a lung ventilation function evaluation aerosol image analysis method provided by the present application in combination with the accompanying drawings.

[0049] Please refer to Figure 1 which shows a method flowchart of a lung ventilation function evaluation aerosol image analysis method provided by one embodiment of the present application, and the method includes the following steps:

[0050] Step S1: Change the concentration of aerosol in a time sequence descending order, and obtain the lung aerosol images of each patient at different concentrations.

[0051] Lung ventilation function is a core physiological indicator for measuring the efficiency of air entering and leaving alveoli and gas exchange. Its dynamic evaluation has irreplaceable clinical value for early diagnosis, efficacy monitoring and prognosis of respiratory diseases such as chronic obstructive pulmonary disease (COPD), asthma, pulmonary fibrosis and cystic fibrosis. The distribution of aerosol particles in the lung is usually used to evaluate lung ventilation function. However, due to individual differences in inhalation methods and breathing patterns, a single concentration of aerosol tracer will not be uniformly distributed in the lung. In addition, aerosol particles in the patient's lungs will undergo dynamic deposition and metabolic processes. Therefore, in the embodiments of the present application, by supplying radioactive aerosols of different concentrations, lung aerosol images of patients at different concentrations are obtained. In the subsequent process, by analyzing the distribution characteristics of aerosol particles and the change characteristics between them in the lung aerosol images at different concentrations, more accurate lung ventilation function evaluation can be achieved to assist doctors in making more accurate judgments.

[0052] Specifically, the method for acquiring the lung aerosol image can include: tracer selection 99mTc-DTPA radioactive aerosol, which can be cleared from the alveolar region through trans-epithelial diffusion, and the absorbed dose is discharged through glomerular filtration in the kidney, which is not easy to cause harm to the human body. The aerosol tracer, i.e. radioactive aerosol, is provided in different concentrations, which can produce a gradient response, helping to distinguish between particle deposition caused by individual respiratory patterns and true ventilation abnormalities caused by ventilation obstruction. Using only a single concentration, the particle deposition superposition effect is similar, and it is difficult to distinguish whether the low contrast is due to insufficient tracer dose caused by shallow breathing of the patient or due to pathologically obstructed ventilation that limits gas exchange in the region. By supplying radioactive aerosol with different concentrations, the deposition dynamics of the lung region under concentration change can be observed. The safe dose range of 99mTc-DTPA is 20-30 MBq. In order to analyze the distribution and deposition of aerosol particles in each lung region in the time dynamic image, enhance the contrast, and make aerosol with different concentrations, the total dose should not exceed 30 MBq. The 99mTc-DTPA is aerosolized in descending order of concentration. The aerosolizer is used to convert 99mTc-DTPA solution with different concentrations (in descending order of concentration) into fine particles. The patient is in a supine position, with both arms holding the head, and the probe is as close to the chest as possible. The patient inhales according to the predetermined inhalation protocol (deep inhalation, breath holding for a certain time, and then exhalation according to the doctor's instructions). It is ensured that the tracer can enter each region of the lung as evenly as possible. Then, using single photon emission computed tomography (SPECT), the lungs of the patient after inhaling the tracer are collected, the aerosol timing images of the patient during the breath holding (intermediate time) are obtained, and the collected image data is preprocessed, including but not limited to noise suppression, image enhancement, etc. to facilitate subsequent extraction of the distribution information of the tracer. At this point, the concentration of the aerosol can be changed in descending order in time to obtain the lung aerosol images of the patient under different concentrations.

[0053] It should be noted that the process of preprocessing the image data is a known technology, and the specific process will not be described here; the number of lung aerosol images can be adjusted according to the implementation scenario, but at least one lung aerosol image under each concentration should be ensured, and the concentration of the aerosol can be set to 6 levels, which can be adjusted according to the implementation scenario and is not limited here; the collection and acquisition of patient information data in the embodiment of the present application are authorized by the relevant user, and the process does not violate the relevant laws and regulations and does not violate the public order and good customs.

[0054] Step S2: region segmentation is performed on each lung aerosol image to obtain a plurality of image blocks; in each lung aerosol image, a particle distribution index of each image block is determined based on the distribution feature difference of the aerosol particles between different image blocks; the change trend feature of the particle distribution index of the same image block in the lung aerosol images at different time points is analyzed to determine an aerosol particle deposition index of each image block.

[0055] The core of the lung ventilation function evaluation by analyzing the lung atomization image lies in judging the ventilation function of each region of the lung by using the transport and deposition characteristics of the atomized particles in the respiratory system. First, each lung atomization image can be regionally segmented to obtain a plurality of image blocks.

[0056] Preferably, in an embodiment of the present application, the image block acquisition method comprises:

[0057] Each lung atomization image is regionally segmented based on a pre-trained neural network to obtain a plurality of image blocks, wherein the image blocks in the embodiment of the present application represent different lung regions.

[0058] It should be noted that the neural network in this embodiment of the present application can adopt a U-Net deep convolutional neural network, and the training process of the neural network is a known technology, which will not be described here. In the embodiment of the present application, it is considered that the regional segmentation results of all lung atomization images are basically the same, that is, the image blocks can be one-to-one corresponding in different lung atomization images.

[0059] In the case that the lung ventilation function of the patient is normal, the atomized particles in each image block of the lung atomization image should be uniformly distributed. However, if there is airway obstruction and ventilation is not smooth, the atomized particles cannot enter the blocked region of the lung with breathing, so the deposition of the atomized particles in this region will be less. Therefore, the difference in the distribution of the atomized particles between different image blocks in each lung atomization image can be compared to determine the particle distribution index of each image block, and the lung ventilation function is preliminarily evaluated.

[0060] Preferably, in an embodiment of the present application, the particle distribution index acquisition method comprises:

[0061] Please refer to Figure 2 which shows a method flowchart of the particle distribution index acquisition method in an embodiment of the present application. The method comprises the following steps:

[0062] Step S201: In each lung atomization image, the number difference characteristics of the atomized particles between all image blocks are analyzed to determine a first distribution factor of each image block.

[0063] In each lung atomization image, the number average of the atomized particles of all image blocks is taken as a number reference value. The number reference value can quantify the average level of the number of atomized particles in all image blocks of the whole lung atomization image, and provides a global reference standard.

[0064] Then, a difference value between the number of atomized particles in each image block and the number reference value is calculated. When the difference value is negative and smaller, it indicates that the number of atomized particles in the image block is less than the average level of the whole, and thus the image block is more likely to be an area with too sparse atomized particles caused by problems in the ventilation function. Conversely, when the difference value is positive and larger, it indicates that the distribution of atomized particles in the image block is more normal, and thus the image block is more likely to be a normal area. Therefore, the difference value is normalized to obtain a first distribution factor of each image block. Since the difference value can be positive or negative, the normalization method can use a function .

[0065] Step S202: In each lung atomization image, a second distribution factor of each image block is determined based on the density difference feature of atomized particles between all image blocks.

[0066] In each lung atomization image, a ratio of the number of atomized particles to the area of each image block is taken as a density factor. The density factor can eliminate the influence of area difference on the number of atomized particles and focus more on the density of atomized particles in a unit area.

[0067] Similarly, a mean value of the density factors of all image blocks is taken as a density reference value, which can be used as a global reference standard for density to quantify the average level of the overall atomized particle density. Then, a difference value between the density factor of each image block and the density reference value is calculated. Since the distribution of atomized particles in the image is more uniform when the patient's lung ventilation function is normal, and the atomized particles cannot enter the blocked area of the lung when the lung ventilation function is abnormal, when the difference value is negative and smaller, it indicates that the number of atomized particles in the image block is less than the average level of the whole, and thus the image block is more likely to be an area with too sparse atomized particles caused by problems in the ventilation function. Conversely, when the difference value is positive and larger, it indicates that the distribution of atomized particles in the image block is more normal, and thus the image block is more likely to be a normal area. Therefore, the difference value is normalized to obtain a second distribution factor of each image block. Since the difference value can be positive or negative, the normalization method can use a function .

[0068] Step S203: In each lung atomization image, the first distribution factor and the second distribution factor of each image block are fused to obtain a particle distribution index of each image block.

[0069] Based on the analysis in the foregoing steps S201 and S202, when the first distribution factor of a certain image block is larger, it indicates that the distribution of the atomized particles in the image block is more normal. Similarly, when the second distribution factor of a certain image block is larger, it also indicates that the distribution of the atomized particles in the image block is more normal. Therefore, the normalized value of the sum of the first distribution factor and the second distribution factor of each image block is taken as the particle distribution index of each image block. At this time, the smaller the particle distribution index of the image block, the more sparse the distribution of the atomized particles in the image block compared with other image blocks, and the more likely the image block has an abnormal ventilation function problem, which needs higher attention. The normalization is a technical means familiar to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization, and the specific normalization method is not limited herein.

[0070] The patient inhales the maximum concentration of aerosol for the first time, and in the normal region of the lung atomization image, the distribution of the atomized particles is more, and in the blocked region, the atomized particles are less. Continuing to reduce the concentration of the inhaled tracer, then in each lung atomization image, under the mutual comparison of the distribution characteristics of the atomized particles between the image blocks, the atomized particles in the normal region will be deposited and superimposed, and in the blocked region, the possibility of the low-concentration aerosol distribution to the region is smaller, and the deposition of the atomized particles may only have a slight increase or remain unchanged, or even due to the metabolism of the aerosol, the amount of the deposited atomized particles may decrease. Therefore, in the embodiment of the present application, continuously reducing the concentration of the radioactive aerosol can improve the contrast between the normal region and the blocked region, and improve the accuracy of the evaluation of the lung ventilation function. Thus, the deposition of the atomized particles in the image block can be used as an index for quantifying the ventilation function.

[0071] However, since the breathing mode (breathing depth, breathing flow rate, etc.) of the patient directly determines the delivery and deposition of the atomized tracer in the lung, although the patient breathes according to the fixed number of seconds under the guidance of the doctor, the differences in breathing habits and lung capacity may lead to misjudgment, mistaking normal ventilation as abnormal, or ignoring mild abnormalities. For example, the patient's breathing habit is shallow, and the patient is unable to inhale the atomized tracer before reaching the breath-holding stage, and the atomized particles may be deposited in advance, resulting in a decrease in the distal alveolar tracer, which is mistaken for a distal ventilation disorder. Therefore, if only the deposition of the atomized particles is used to set a unified deposition standard line for all patients, the possibility of misjudgment is high, and the individual differences of the patients need to be considered. Therefore, in the embodiment of the present application, the distribution comparison and change trend characteristics of the atomized particles in the time sequence lung atomization images of the patient are analyzed to determine the atomized particle deposition index of each image block, thereby improving the accuracy of the judgment of the lung ventilation function.

[0072] Preferably, in an embodiment of the present application, the method for obtaining the atomized particle deposition index comprises:

[0073] For any one image block, in the lung atomization images at each adjacent two time points, the difference between the particle distribution index at the later time point and the particle distribution index at the former time point is calculated. Based on the foregoing analysis, if the image block is a lung ventilation obstruction region, then the difference should be a negative value; otherwise, if the image block is a lung normal region, then the difference should be a positive value. The normalized value of the difference is taken as the particle deposition factor of the image block at the adjacent two time points. The normalization method here can use the function .

[0074] At this point, in all the lung atomization images of each patient, each image block has a particle deposition factor at adjacent two time points.

[0075] The lung normal region continuously inhales aerosol over time, and the particle deposition factor of the same image block between adjacent two time points tends to increase. Therefore, all the particle deposition factors of each image block are linearly fitted in time sequence, and the slope value of the obtained fitting straight line is obtained. If the slope value is positive, it indicates that the particles in the image block have deposited over time, and it is more likely to be a normal region. Conversely, if the slope value is negative, it indicates that the particles in the image block have not deposited over time, and the image block is more likely to be a ventilation obstruction region. The normalized value of the slope value is taken as the deposition weight of each image block. The normalization method here can use the function .

[0076] Based on the foregoing analysis and calculation, the greater the deposition weight of a certain image block, the more obvious the deposition phenomenon of the image block, and the more likely it is a normal region. Moreover, the greater the particle deposition factor, the more likely it is a normal region. Therefore, in all the lung atomization images of each patient, the mean value of all the particle deposition factors of the same image block is multiplied by the corresponding deposition weight, and the normalized value of the obtained product is taken as the atomized particle deposition index of each image block. At this point, the greater the atomized particle deposition index, the more the deposition of atomized particles in the image block conforms to the atomized particle deposition characteristics of the lung normal region, and the more likely it is a normal region. The normalization is a technology known to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization, and the specific normalization method is not limited here.

[0077] It should be noted that the method for obtaining the fitting straight line in the embodiment of the application can use the least square method, which is a known technology, and the specific process is not repeated here.

[0078] At this point, the atomized particle deposition index of each image block can be obtained, which can be used as an index for measuring the lung ventilation function.

[0079] Step S3: In the first pulmonary aerosolization image, the aerosol particles in each image block are clustered based on the density characteristics of the aerosol particles, so as to divide each image block into a plurality of sub-regions; according to the position distribution of the sub-regions in each image block and the number change characteristics of the aerosol particles in the corresponding regions of the sub-regions in the pulmonary aerosolization images at different time, the metabolic rate of the aerosol particles corresponding to each image block is determined.

[0080] The aerosol particles converted by the radioactive aerosol cannot be exhaled out of the body, but spread to the alveoli and are metabolized from the kidneys. The cell types and local metabolic activities of the regions of the lungs are different. In the normal regions of the lungs, the ventilation is good, the oxygen exchange is sufficient, and the energy metabolism of the cells is active. The metabolic speed is faster than that in the ventilation blocked regions. Therefore, by analyzing the metabolic conditions of the regions of the lungs, the ventilation functions of the regions of the lungs can be further confirmed.

[0081] In order to analyze the metabolic conditions of the aerosol particles in each image block, the density characteristics of the aerosol particles in each image block in the pulmonary aerosolization image obtained by inhaling the high-concentration aerosol for the first time can be clustered to obtain a plurality of sub-regions.

[0082] Preferably, in an embodiment of the present application, the method for obtaining the sub-regions comprises:

[0083] In the first pulmonary aerosolization image, the aerosol particles in each image block are clustered and analyzed based on the DBSCAN clustering algorithm to obtain all the clustering clusters, wherein the minimum sample number and the neighborhood radius are both preset values.

[0084] Each clustering cluster is taken as a sub-region, so as to obtain all the sub-regions corresponding to each image block.

[0085] It should be noted that the DBSCAN clustering algorithm is a known technology, and the specific process is not described herein. The minimum sample number is set to 8, and the neighborhood radius is set to 5 pixel values. The specific values can be adjusted according to the implementation scene, and are not limited herein.

[0086] The metabolism of the aerosol particles can be specifically manifested as the reduction of the number of the particles. However, in the embodiment of the present application, the radioactive aerosol is continuously inhaled, so that there are new aerosol particles in the upper trachea, and the new aerosol particles are less deposited in the lower region close to the alveoli, so that the lower region is less affected. The metabolic condition calculated by relying on the number of the aerosol particles is more accurate. Therefore, the position distribution of the sub-regions in each image block of the pulmonary aerosolization image can provide the confidence of the metabolic calculation, so that the metabolic rate of the aerosol particles corresponding to each image block is determined by combining the position distribution of the sub-regions with the number change characteristics of the aerosol particles in the corresponding regions of the sub-regions in different pulmonary aerosolization images.

[0087] Preferably, in one embodiment of the present application, the method for obtaining the metabolic rate of the aerosolized particles comprises the following steps:

[0088] Referring to Figure 3 Fig. 3 shows a flow chart of the method for obtaining the metabolic rate of the aerosolized particles in one embodiment of the present application, which comprises the following steps:

[0089] Step S301: In the first lung aerosolization image, the positional relationship between the sub-regions in each image block is analyzed to determine the metabolic rate confidence of each sub-region.

[0090] Since the distribution of the aerosolized particles is related to ventilation, the upper end of the trachea will first contact the aerosol inhaled by the patient, so in the lung aerosolization image, the distribution of the aerosolized particles will decrease from the upper end to the lower end of the trachea. Therefore, in the first lung aerosolization image, the largest sub-region in each image block is taken as the core region, which is the region with the densest distribution of the aerosolized particles, i.e., the region closest to the upper end.

[0091] Then in each image block of the first lung aerosolization image, the Euclidean distance between the centroid of each sub-region and the centroid of the core region is calculated. Based on the foregoing analysis, the closer to the alveolar region at the lower end, the less new aerosolized particles are deposited, and the higher the confidence is. Therefore, the greater the Euclidean distance, the higher the confidence. Thus, the value of the Euclidean distance after normalization is taken as the metabolic rate confidence of each sub-region. The normalization is a technique well known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, and the specific normalization method is not limited herein.

[0092] It should be noted that in this embodiment of the present application, when calculating the Euclidean distance, the lower left corner of the first lung aerosolization image can be taken as the origin, and the horizontal right direction and the vertical upward direction can be taken as the positive directions of the coordinate axes to construct a coordinate system for calculating the Euclidean distance.

[0093] Step S302: The number variation characteristics of the aerosolized particles in the sub-regions in the first lung aerosolization image and the corresponding regions in other lung aerosolization images are analyzed to determine the metabolic factor of the aerosolized particles in each sub-region.

[0094] The metabolic condition of the aerosolized particles can be characterized by the number characteristics of the aerosolized particles. Each sub-region in each image block in the first lung aerosolization image can be taken as a comparison region in the corresponding region in other lung aerosolization images.

[0095] Then each sub-region and all the comparison regions corresponding thereto are sorted in time sequence to obtain a sorting sequence, and the sub-regions and the comparison regions in the sorting sequence are collectively referred to as target regions, which correspond to the same region at different time points.

[0096] In the sorting sequence, in each adjacent two target regions, the difference value of the number of atomized particles in the former target region and the number of atomized particles in the latter target region is taken as the ratio of the time interval, as a metabolic parameter. The greater the difference value of the number of atomized particles, the greater the degree of reduction of the number of atomized particles in the adjacent two target regions. The ratio of the time interval corresponding to the adjacent two target regions can quantify the rate of reduction of the number, and the metabolic parameter is obtained. The greater the metabolic parameter, the faster the metabolic rate.

[0097] Based on the foregoing process, a metabolic parameter can be calculated between each adjacent two target regions. Finally, the average value of all metabolic parameters corresponding to each sub-region in each image block in the first lung atomization image is taken as the atomized particle metabolism factor of each sub-region. It is easy to understand that the greater the atomized particle metabolism factor, the faster the atomized particle metabolism rate of the sub-region, and the higher the probability of normal ventilation function.

[0098] Step S303: In the first lung atomization image, the metabolic rate confidence of the sub-region in each image block and the atomized particle metabolism factor are fused to obtain the atomized particle metabolism rate of each image block.

[0099] Based on the foregoing analysis, the greater the metabolic rate confidence of the sub-region, the higher the confidence of the calculation of the metabolic degree, the greater the atomized particle metabolism factor of the sub-region, and the faster the atomized particle metabolism rate. Therefore, in each image block, the metabolic rate confidence of the sub-region is weighted to obtain the atomized particle metabolism factor, and the normalized value of the obtained weighted result is taken as the atomized particle metabolism rate of each image block. At this time, by multiplying the metabolic rate confidence of the sub-region and the atomized particle metabolism rate in each image block, and normalizing the average level of the products corresponding to all sub-regions, the atomized particle metabolism rate is obtained. The greater the value, the faster the metabolism rate of the particulate matter of the corresponding image block, and the higher the possibility of being a normal ventilation function region. The normalization is a technical means familiar to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization. The specific normalization method is not limited here.

[0100] At this point, the atomized particle metabolism rate of each image block in the first lung atomization image can be obtained as another indicator for evaluating the lung ventilation function of the corresponding region.

[0101] Step S4: The particle deposition indicator and the atomized particle metabolism rate of each image block are fused to obtain the ventilation obstruction evaluation value of each image block of the patient's lung, which is used for auxiliary evaluation of lung ventilation function.

[0102] Based on the foregoing steps, two indexes of measuring the lung ventilation function of the image block in the corresponding region of the lung can be obtained. Here, the two indexes can be fused to obtain a ventilation obstruction evaluation value of each image block of the lung of the patient, which is used to assist the doctor in auxiliary evaluation of the lung ventilation function.

[0103] Preferably, in an embodiment of the present application, the method for obtaining the ventilation obstruction evaluation value comprises:

[0104] When the deposition index of the aerosolized particles corresponding to the image block is larger, it indicates that the deposition of the aerosolized particles in the image block is more in line with the deposition characteristics of the aerosolized particles in the normal region of the lung, and it is more likely to be a normal region. When the metabolism rate of the aerosolized particles corresponding to the image block is larger, it indicates that the metabolism rate of the particles in the corresponding image block is faster, and the possibility of being a normal region of the lung ventilation function is higher. Therefore, the aerosolized particle deposition index and the aerosolized particle metabolism rate are negatively correlated with the ventilation obstruction evaluation value. Therefore, in the first lung aerosolization image, the value obtained by negatively correlating and normalizing the product of the aerosolized particle deposition index and the aerosolized particle metabolism rate corresponding to each image block is taken as the ventilation obstruction factor of each image block. The larger the ventilation obstruction factor, the higher the possibility of lung ventilation function damage of the image block, and the greater the degree of attention required by the doctor. The negatively correlating and normalizing here can adopt the formula , wherein, represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0105] In the first lung aerosolization image, the mean value of the ventilation obstruction factors of all the image blocks is taken as the obstruction reference value, and the value obtained by normalizing the difference between the ventilation obstruction factor of each image block and the obstruction reference value is taken as the ventilation obstruction evaluation value of each image block. The larger the difference, the higher the degree of the ventilation obstruction factor exceeding the average level, and the larger the ventilation obstruction evaluation value, that is, the higher the possibility of abnormality of the lung region corresponding to the image block. The normalization here adopts the function .

[0106] At this point, the ventilation obstruction evaluation value of each image block in the first lung aerosolization image can be obtained. The larger the value, the higher the possibility of abnormality of the lung region corresponding to the image block. Therefore, all the image blocks are arranged in descending order according to the ventilation obstruction evaluation value to obtain an evaluation sequence. In the evaluation sequence, the earlier the image block, the greater the possibility of abnormality, and the higher the priority of attention required by the doctor. The lung regions are displayed to the doctor in the order in the evaluation sequence, thereby assisting the doctor in more accurate lung ventilation function evaluation.

[0107] In summary, in the embodiment of the present application, the aerosol concentration is designed in descending order, so as to obtain the lung atomization image of the patient under different concentrations, which can be closer to the dynamic fluctuation of aerosol concentration in the natural breathing process, and can correct the influence of the patient's breathing mode (such as deep inspiration, shallow and rapid breathing) on particle deposition, and reduce the evaluation deviation caused by the difference in inhalation mode. Under the condition that the patient's lung ventilation function is normal, the distribution of atomized particles will be more uniform, and with the continuous inhalation of aerosol, particle deposition will occur in the ventilation area, so by image block level analysis, the lung is divided into different regions (such as upper and lower lobes, central and peripheral airways), and then the distribution difference characteristics of atomized particles between different image blocks and the change trend in the lung atomization image under different concentrations are analyzed, which can quantify the atomized particle deposition index of each image block as an index for evaluating the ventilation function of the lung region corresponding to the image block. Further, the atomized particles converted by radioactive aerosol cannot be exhaled out of the body, but spread to the alveoli and metabolized from the kidneys, and the cell types and local metabolic activity of each region of the lung are different, which affect the distribution of atomized particles, so by analyzing the metabolism of atomized particles in each region of the lung, the ventilation function thereof can be further confirmed. Therefore, in the first lung atomization image, the sub-regions are divided based on the density of atomized particles, and the atomized particle metabolism rate is determined according to the position distribution of the sub-regions and the number change characteristics of the atomized particles in the sub-regions at different times, as another index for measuring the ventilation function. Finally, the two indexes for measuring the ventilation function corresponding to each image block are fused to obtain the ventilation obstruction evaluation value, which assists the doctor in evaluating the lung ventilation function, and effectively improves the comprehensiveness and accuracy.

[0108] It should be noted that the above-mentioned order of the embodiments is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0109] Each embodiment in the specification is described in a progressive manner, and the same and similar parts between each embodiment can be referred to each other, and each embodiment mainly describes the difference from other embodiments.

Claims

1. A method of aerosolized image analysis for pulmonary ventilation function assessment, characterized in that, The method comprises: sequentially change the concentration of the aerosol in a descending order of time, and obtain the lung atomization images of each patient under different concentrations; regionally segment each lung atomization image to obtain a plurality of image blocks; in each lung atomization image, determine a particle distribution index of each image block based on the distribution feature difference of the atomized particles between different image blocks; analyze the change trend feature of the particle distribution index of the same image block in the lung atomization images at different time points, and determine an atomized particle deposition index of each image block; in the first lung atomization image, cluster the atomized particles in each image block based on the density feature of the atomized particles, thereby dividing each image block into a plurality of sub-regions; and determine the atomized particle metabolic rate of each image block according to the position distribution of the sub-regions in each image block and the number change feature of the atomized particles in the corresponding regions of the sub-regions in the lung atomization images at different time points; fuse the particle deposition index and the atomized particle metabolic rate of each image block to obtain the ventilation obstruction evaluation value of each image block of the patient's lung, which is used for auxiliary evaluation of the lung ventilation function; the method for obtaining the ventilation obstruction evaluation value comprises: in the first lung atomization image, the product of the atomized particle deposition index and the atomized particle metabolic rate of each image block is negatively correlated and normalized, and the value after the negative correlation and normalization is taken as the ventilation obstruction factor of each image block; in the first lung atomization image, the mean value of the ventilation obstruction factors of all image blocks is taken as the obstruction reference value, and the difference between the ventilation obstruction factor of each image block and the obstruction reference value is normalized, and the value after the normalization is taken as the ventilation obstruction evaluation value of each image block.

2. The method of claim 1, wherein the method further comprises: the method for obtaining the particle distribution index comprises: in each lung atomization image, analyze the number difference feature of the atomized particles between all image blocks to determine a first distribution factor of each image block; based on the density difference feature of the atomized particles between all image blocks, determine a second distribution factor of each image block; the sum of the first distribution factor and the second distribution factor of each image block is normalized, and the value after the normalization is taken as the particle distribution index of each image block.

3. The method of claim 2, wherein the method further comprises: the method for obtaining the first distribution factor comprises: in each lung atomization image, the number mean value of the atomized particles of all image blocks is taken as a number reference value, and the difference between the number of the atomized particles of each image block and the number reference value is normalized, and the value after the normalization is taken as the first distribution factor of each image block.

4. The method of claim 2, wherein the method further comprises: determining a lung volume of the subject based on the lung image data. the method for obtaining the second distribution factor comprises: in each lung atomization image, the ratio of the number of the atomized particles to the area of each image block is taken as a density factor; the mean value of the density factors of all image blocks is taken as a density reference value, and the difference between the density factor of each image block and the density reference value is normalized, and the value after the normalization is taken as the second distribution factor of each image block.

5. The method of claim 1, wherein the method further comprises: determining a lung volume of the subject based on the lung image data. the method for obtaining the atomized particle deposition index comprises: For any one image block, in the lung atomization images at each two adjacent time points, the normalized value of the difference between the particle distribution index at the later time point and the particle distribution index at the former time point is taken as the particle deposition factor of the image block at the two adjacent time points; The particle deposition factors of each image block are linearly fitted according to the time sequence, and the normalized value of the slope value of the obtained fitting straight line is taken as the deposition weight of each image block; The deposition weight of each image block is multiplied by the mean value of all the particle deposition factors corresponding to each image block, and the normalized value of the obtained product is taken as the atomized particle deposition index of each image block.

6. The method of claim 1, wherein the method further comprises: determining a lung volume of the subject based on the lung image data. The method for obtaining the sub-regions comprises: In the first lung atomization image, the atomized particles in each image block are analyzed by clustering based on the DBSCAN clustering algorithm, and all the clustering clusters are obtained, wherein the minimum sample number and the neighborhood radius are both preset values; Each clustering cluster is taken as a sub-region, thereby obtaining all the sub-regions corresponding to each image block.

7. The method of claim 1, wherein the method is a method of aerosolized image analysis for pulmonary ventilation function assessment. The method for obtaining the atomized particle metabolic rate comprises: In the first lung atomization image, the positional relationship between the sub-regions in each image block is analyzed, thereby determining the metabolic rate confidence of each sub-region; The number change characteristics of the atomized particles of the sub-regions in the first lung atomization image and the corresponding regions in other lung atomization images are analyzed, thereby determining the atomized particle metabolic factor of each sub-region; In each image block of the first lung atomization image, the atomized particle metabolic factor is weighted and averaged by using the metabolic rate confidence of the sub-region, and the normalized value of the obtained weighted result is taken as the atomized particle metabolic rate of each image block.

8. The method of claim 7, wherein the method further comprises: determining a lung volume of the patient based on the lung image data. The method for obtaining the metabolic rate confidence comprises: In the first lung atomization image, the largest sub-region in each image block is taken as the core region; The Euclidean distance between the centroid of each sub-region and the centroid of the core region is normalized, and the normalized value is taken as the metabolic rate confidence of each sub-region.

9. The method of claim 7, wherein the method further comprises: determining a lung volume of the patient based on the lung image data. The method for obtaining the atomized particle metabolic factor comprises: Each sub-region in each image block of the first lung atomization image is taken as a comparison region in other lung atomization images; Each sub-region and all the comparison regions corresponding thereto are sorted in time sequence to obtain a sorting sequence, and the sub-regions and the comparison regions in the sorting sequence are collectively referred to as target regions; In each two adjacent target regions in the sorting sequence, the ratio of the difference between the number of atomized particles in the former target region and the number of atomized particles in the latter target region to the time interval is taken as a metabolic parameter; The mean value of all the metabolic parameters corresponding to each sub-region in each image block of the first lung atomization image is taken as the atomized particle metabolic factor of each sub-region.

Images