Ventilator mask seal assist test method and system based on 3D facial data

Abstract

The application discloses a ventilator mask sealing auxiliary test method and system based on 3D face data, relates to the field of mask sealing auxiliary test, and remarkably improves the individualized adaptation effect of the ventilator mask through multi-modal data fusion and intelligent prediction: firstly, digital reconstruction of facial anatomical structure is realized based on high-precision point cloud modeling, so that the initial fit degree of the mask is greatly improved; secondly, through a double closed-loop adjustment mechanism of dynamic action simulation and air pressure leakage rate, the leakage risk caused by actions such as shaking the head and speaking is greatly reduced; finally, combined with a time sequence prediction model, preventive maintenance is realized, the mask failure risk can be early warned, and the treatment interruption rate is significantly reduced, so that the clinical debugging time is greatly reduced compared with the traditional method; through a standardized action set to simulate the night activity state of the patient, combined with a real-time air pressure monitoring system, intermittent leakage problems that cannot be found by traditional static testing can be captured, and the dynamic sealing qualified rate of the mask is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of mask sealing auxiliary testing, and in particular, relates to a ventilator mask sealing auxiliary testing method and system based on 3D facial data. BACKGROUND

[0002] Traditional ventilator masks are designed with standardized sizes, which are difficult to adapt to individual differences in the anatomical structures of different patients, resulting in poor fit, low comfort, and other problems; existing masks are prone to leakage when patients are active (such as turning over and speaking), and lack real-time monitoring and adaptive adjustment mechanisms for air pressure leakage under dynamic actions; mask performance degradation or failure usually relies on manual inspection, which cannot provide early warning for replacement timing, resulting in treatment interruption or waste of consumables. SUMMARY

[0003] To overcome the above technical problems existing in the prior art, the present application proposes a ventilator mask sealing auxiliary testing method and system based on 3D facial data.

[0004] To solve the above technical problems, the present application is realized by the following technical scheme:

[0005] The present application is a ventilator mask sealing auxiliary testing method based on 3D facial data, comprising the following steps:

[0006] S1, collect the current patient's facial feature data, and then generate a high-resolution point cloud model of the current patient according to the facial feature data to obtain a current facial point cloud model;

[0007] S2, construct an initial ventilator mask 3D model and simulate the fit with the current facial point cloud model, adjust the local parameters of the initial ventilator mask 3D model according to the simulation results, and obtain a final ventilator mask 3D model;

[0008] S3, 3D print the final ventilator mask 3D model and perform deformation testing, and then adjust the printing thickness parameter in the 3D printing process according to the test results to obtain a final ventilator mask;

[0009] S4, perform an inflation experiment on the final ventilator mask and continue to adjust the adjusted printing thickness parameter in S3 to obtain a first adjusted ventilator mask;

[0010] S5, make the current patient wear the first adjusted ventilator mask and perform a standard action, then collect the air pressure leakage rate, continue to adjust the local parameters adjusted in S2, and repeat S3, S4, and S5 to obtain a second adjusted ventilator mask;

[0011] S6, predict the air pressure leakage rate at future time based on the air pressure leakage rate collected in S5;

[0012] S7, predicting in advance that the current patient will replace the second adjusted ventilator mask according to the prediction result in S6.

[0013] Preferably, the S1 comprises the following steps:

[0014] S11, setting a plurality of 3D facial feature types to obtain a facial feature type set; collecting the current patient's facial feature data in combination with the facial feature type set to obtain a current patient's facial feature data set;

[0015] S12, setting a facial model resolution threshold; generating a high-resolution point cloud model of the current patient's face with a resolution greater than or equal to the facial model resolution threshold according to the current patient's facial feature data set to obtain a current facial point cloud model;

[0016] By predefining a set of key feature types such as facial contour, nose bridge height, and cheekbone shape, and combining non-contact 3D scanning technology, the patient's facial details can be quickly captured to generate a low-error high-resolution point cloud model. Compared with the traditional fitting adjustment method, the human error is significantly reduced and the adaptation cycle is shortened. The facial point cloud model generation mechanism based on threshold control ensures that the model resolution meets the clinical sealing analysis requirements. High-precision facial modeling can accurately reflect the individual facial curvature differences and avoid the local compression problems caused by traditional standardized masks. At the same time, non-contact scanning has no physical contact risk and is suitable for burn patients or sensitive skin groups.

[0017] Preferably, the S2 comprises the following steps:

[0018] S21, constructing an initial ventilator mask 3D model according to the current facial point cloud model;

[0019] S22, setting the thickness data of the medical-grade silicone buffer layer of the nose pad area corresponding to the initial ventilator mask 3D model, the inner diameter data of the mask air pipe joint, and the edge chamfer radius data of the mask to obtain the initial buffer layer thickness data, the initial joint inner diameter data, and the initial edge chamfer radius data; setting the initial buffer layer thickness data, the initial joint inner diameter data, and the initial edge chamfer radius data to the initial ventilator mask 3D model to obtain a set of ventilator mask 3D models;

[0020] S23, in combination with the facial feature type set and setting a current mask fit threshold set; at the same time, in combination with the facial feature type set, the set of ventilator mask 3D models and each facial feature of the current facial point cloud model are simulated for fit in a virtual environment, and after the simulation is completed, a current fit data set is obtained; adjusting the set of ventilator mask 3D models according to the current fit data to obtain a final ventilator mask 3D model.

[0021] Preferably, the adjustment of the respirator mask 3D model according to the current fit degree data in S23 to obtain the final respirator mask 3D model comprises the following steps:

[0022] S231, when the current fit degree data set has fit degree data less than the corresponding current mask fit degree threshold value in the current mask fit degree threshold value set, adjust the initial buffer layer thickness data, initial joint inner diameter data and initial edge chamfer radius data until the current fit degree data set has no fit degree data less than the corresponding current mask fit degree threshold value in the current mask fit degree threshold value set, to obtain the final buffer layer thickness data, the final joint inner diameter data, the final edge chamfer radius data and the final respirator mask 3D model; otherwise, no adjustment is needed, and the set respirator mask 3D model is taken as the final respirator mask 3D model;

[0023] Based on the high-precision facial point cloud model, an initial 3D mask model is constructed, and through parameterized design of the nose pad buffer layer thickness, joint inner diameter and edge chamfer radius, combined with virtual fit degree simulation, the structure parameters are dynamically adjusted until the sealing requirement is met. Compared with the traditional try-on method, the sealing qualified rate is greatly improved, and it is especially suitable for patients with complex facial anatomy structure; the medical-grade silicone buffer layer and the self-adaptive chamfer design can reduce the local pressure peak value and reduce the risk of pressure injury in sensitive areas such as the nasal bridge and the malar bone; virtual iterative optimization replaces traditional multiple physical try-on, shortening the adaptation cycle; the parameterized model can be directly connected to 3D printing production, reducing mold development costs.

[0024] Preferably, the S3 comprises the following steps:

[0025] S31, set the current initial 3D printing thickness data; according to the current initial 3D printing thickness data, output the final respirator mask 3D model in the virtual environment in S23 in STL format and perform 3D printing operation to obtain an initial respirator mask;

[0026] S32, set the current deformation test pressure and the current deformation threshold value; perform deformation test on the initial respirator mask under the current deformation test pressure, and after the test is completed, obtain the current deformation data; if the current deformation data is greater than or equal to the current deformation threshold value, adjust the current initial 3D printing thickness data and repeat S31 until the current deformation data is less than the current deformation threshold value, to obtain the current final 3D printing thickness data and the final respirator mask;

[0027] By dynamically adjusting the printing thickness parameter and combining with deformation pressure test closed loop verification, it is ensured that the mask maintains structural integrity under the rated ventilation pressure, avoiding gas leakage problems caused by material deformation in clinical use. Compared with the traditional homogeneous printing scheme, the deformation resistance is greatly improved.

[0028] Preferably, the S4 comprises the following steps:

[0029] S41, set a test air pressure value set and an air pressure detection period; according to the test air pressure value set, the gas with a preset test air pressure value is filled into the final respirator mask for multiple times; after the inflation is completed, the air pressure in the final respirator mask is tested after the air pressure detection period, and a current air pressure experimental data set is obtained;

[0030] S42, set a current air pressure leakage rate threshold value corresponding to each inflation in S41, and obtain a current air pressure leakage rate threshold value set; calculate the air pressure leakage rate data corresponding to each inflation in S41 according to the test air pressure value set and the current air pressure experimental data set, and obtain a current air pressure leakage rate data set;

[0031] S43, when there is leakage rate data greater than or equal to the corresponding air pressure leakage rate threshold value in the current air pressure leakage rate threshold value set in the current air pressure leakage rate data set, return to S32, on the basis that the current deformation data is less than the current deformation threshold, continue to adjust the current final 3D printing thickness data, and repeat S41, S42 and S43 until there is no leakage rate data greater than or equal to the corresponding air pressure leakage rate threshold value in the current air pressure leakage rate threshold value set in the current air pressure leakage rate data set, and a first adjusted respirator mask is obtained; otherwise, it is not necessary to return to S32 for adjustment, and the final respirator mask is taken as the first adjusted respirator mask;

[0032] A test set covering the clinical pressure range is established, the leakage characteristics under different working conditions are captured through periodic air pressure detection, compared with the traditional single-point test, the potential leakage risk can be found in advance, and the product reliability is greatly improved; the 3D printing thickness adjustment is dynamically associated with the deformation data and the leakage rate data, forming a synergistic optimization mechanism of material mechanical properties and sealing performance, which can simultaneously improve the structural strength and air tightness of the mask; the automatic closed loop of "test-analysis-adjustment" is adopted, the printing parameter optimization is automatically triggered when the leakage exceeds the standard, the single iteration time is greatly shortened, the product development cycle is significantly accelerated, and the pass rate of the final product is greatly improved.

[0033] Preferably, the S5 comprises the following steps:

[0034] S51, set a standardized action set and wear the first adjusted respirator mask on the face of the current patient; after wearing, collect the air pressure data in the first adjusted respirator mask at this moment to obtain current wearing initial air pressure data;

[0035] S52, cooperate with the standardized action set to make the current patient perform the standardized action; in the process of performing, real-time air pressure data in the first adjusted respirator mask is collected to obtain a current wearing real-time air pressure data set;

[0036] A current wearing air pressure leakage rate threshold is set; according to the current wearing real-time air pressure data set and the current wearing initial air pressure data, the air pressure leakage rate data corresponding to each current wearing real-time air pressure data is calculated to obtain a current wearing air pressure leakage rate data set;

[0037] S53, when there is air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold in the current wearing air pressure leakage rate data set, return to S23, on the basis that there is no fit degree data less than the corresponding current mask fit degree threshold in the current fit degree data set in S23, continue to adjust the final buffer layer thickness data, the final joint inner diameter data and the final edge chamfer radius data, and repeat S31, S32, S41, S42, S43, S51, S52 and S53 until there is no air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold in the current wearing air pressure leakage rate data set, to obtain a second adjusted respirator mask;

[0038] Otherwise, it is not necessary to return to S23 for adjustment, and the first adjusted respirator mask is taken as the second adjusted respirator mask, and S61 is entered;

[0039] Through the standardized action set, the night activity state of the patient is simulated, and in combination with the real-time air pressure monitoring system, intermittent leakage problems that cannot be found by traditional static tests can be captured; experimental data shows that this method greatly improves the dynamic sealing qualification rate of the mask; an intelligent correlation between the air pressure leakage rate and the structure parameters is established, and when the dynamic leakage is over-standard, the parameter adjustment cycle is automatically triggered. This mechanism greatly shortens the clinical adaptation cycle of the product, and reduces the rework rate; while ensuring that the mechanical performance meets the standard, the wearing comfort is continuously optimized through the fit degree data, and the treatment compliance of the patient is greatly improved.

[0040] Preferably, the S6 comprises the following steps:

[0041] S61, set a training data proportion and a test data proportion; data division is performed on the current wearing air pressure leakage rate data set by using the training data proportion and the test data proportion to obtain a current wearing air pressure leakage rate training data set, a current wearing air pressure leakage rate test data set and a current wearing air pressure leakage rate to-be-predicted data set.

[0042] S62, construct an initial BP neural network model; train and test the initial BP neural network model by using the current wearing air pressure leakage rate training data set and the current wearing air pressure leakage rate test data set respectively, and obtain a final BP neural network model;

[0043] S63, input the current wearing air pressure leakage rate to be predicted data set into the final BP neural network model to predict the air pressure leakage rate at a future time, and obtain a future wearing air pressure leakage rate data set;

[0044] By using the BP neural network to establish a dynamic leakage rate prediction model, potential leakage risks can be warned in advance, so that the response speed of clinical intervention is greatly improved. Secondly, through the dynamic division mechanism of the training set and the test set, the generalization ability of the model is ensured, and real-time data is reserved for continuous optimization. Finally, the prediction result is intelligently linked with the mask parameter library, and when the predicted leakage rate exceeds the standard, the optimal adjustment scheme is automatically recommended, so that the product iteration cycle is greatly shortened.

[0045] Preferably, the S7 comprises the following steps:

[0046] S71, when there is air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold in the future wearing air pressure leakage rate data set, the time corresponding to the air pressure leakage rate data corresponding to each standard action is recorded respectively to obtain a standard action leakage rate threshold exceeding time set; otherwise, no recording is needed;

[0047] S72, according to the standard action leakage rate threshold exceeding time set, when the current patient wears the second adjusted ventilator mask while performing the standard action, the current patient is prompted to replace the second adjusted ventilator mask in advance;

[0048] By establishing a leakage risk prediction model based on action characteristics and time sequence analysis, a replacement warning can be issued before a specific action causes the ventilator mask leakage rate to exceed the threshold. Secondly, through the action-leakage time mapping database, the mask service life prediction error is greatly reduced. The treatment interruption rate caused by mask leakage of the patient is greatly reduced, and the waste of consumables is also reduced.

[0049] The ventilator mask sealing auxiliary test system based on 3D face data comprises a current patient face point cloud model generation module, an initial ventilator mask 3D model construction module, a ventilator mask 3D model local parameter adjustment module, a current mask 3D printing parameter adjustment module, an inflation experiment printing parameter adjustment module, a standard action local parameter adjustment module, a future wearing air pressure leakage rate prediction module, and a standard action wearing replacement prediction module.

[0050] The present application has the following beneficial effects:

[0051] 1. The present application significantly improves the personalized fitting effect of the respirator mask through multi-modal data fusion and intelligent prediction: first, based on high-precision point cloud modeling, the digital reconstruction of facial anatomical structure is realized, which greatly improves the initial fit of the mask; second, through the double closed-loop adjustment mechanism of dynamic action simulation and air pressure leakage rate, the leakage risk caused by actions such as shaking head and speaking is greatly reduced; finally, combined with the time series prediction model, preventive maintenance is realized, which can early warning of mask failure risk, significantly reduces the treatment interruption rate, and greatly reduces the clinical debugging time compared with traditional methods.

[0052] 2. In the present application, the initial 3D mask model is constructed based on high-precision facial point cloud model, and the virtual fit degree simulation is combined to dynamically adjust the structure parameters until the sealing requirement is met; compared with the traditional try-on method, the sealing qualified rate is greatly improved; the medical grade silicone buffer layer and the self-adaptive chamfer design can reduce the local pressure peak value and reduce the risk of pressure injury in sensitive areas such as nasal bridge and zygomatic bone; virtual iterative optimization replaces traditional physical try-on, shortening the adaptation cycle.

[0053] 3. In the present application, the patient's night activity state is simulated through a standardized action set, combined with a real-time air pressure monitoring system, which can capture intermittent leakage problems that cannot be found by traditional static testing; the dynamic sealing qualified rate of the mask is greatly improved; the intelligent correlation between air pressure leakage rate and structure parameters is established, and when the dynamic leakage is out of standard, the parameter adjustment cycle is automatically triggered; this mechanism significantly shortens the clinical adaptation cycle of the product, while reducing the rework rate; while ensuring that the mechanical performance meets the standards, the wearing comfort is continuously optimized through the fit data.

[0054] Of course, any product implementing the present application does not necessarily need to achieve all the advantages described above. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0056] Figure 1 The flowchart of the respirator mask sealing auxiliary test method based on 3D facial data of the present application;

[0057] Figure 2 The module schematic diagram of the respirator mask sealing auxiliary test system based on 3D facial data of the present application. DETAILED DESCRIPTION

[0058] With reference to the drawings of the embodiments of the application, the technical solutions in the embodiments of the application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the application, but not all the embodiments of the application. Based on the embodiments of the application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the application.

[0059] Embodiment one

[0060] Please refer to Figure 1 The embodiment is a ventilator mask sealing auxiliary test method based on 3D facial data, comprising the following steps:

[0061] S1, acquiring current patient facial feature data, and then generating a high-resolution point cloud model of the current patient according to the facial feature data to obtain a current facial point cloud model;

[0062] The S1 comprises the following steps:

[0063] S11, setting a plurality of 3D facial feature types to obtain a facial feature type set; the facial feature type set comprises facial contour features, nose ridge height features, and cheekbone shape features, etc.; in cooperation with the facial feature type set, a non-contact 3D scanner is used to acquire the current patient facial feature data to obtain a current patient facial feature data set;

[0064] S12, setting a facial model resolution threshold; generating a high-resolution point cloud model of the current patient face with a resolution greater than or equal to the facial model resolution threshold according to the current patient facial feature data set to obtain a current facial point cloud model;

[0065] By predefining a key feature type set of facial contour, nose ridge height, and cheekbone shape, etc., and combining a non-contact 3D scanning technology (such as structured light or laser scanning), the patient facial details can be quickly captured to generate a high-resolution point cloud model with an error ≤0.5mm; compared with the traditional fitting adjustment method, the human error is significantly reduced and the fitting cycle is shortened; the facial point cloud model generation mechanism based on threshold control ensures that the model resolution meets the clinical sealing analysis requirements; the data model support is provided for the subsequent digital pre-fitting to predict the gap distribution of the easy-to-leak areas such as the nose wing and the chin in advance, to guide the mask structure optimization design, and thus the dynamic sealing qualified rate is improved by more than 30%; in addition, the high-precision facial modeling can accurately reflect the individual facial curvature difference, avoiding the local compression problem caused by the traditional standardized mask; at the same time, the non-contact scanning has no physical contact risk, and is suitable for burn patients or sensitive skin groups;

[0066] S2, construct an initial respirator mask 3D model and fit the current facial point cloud model, adjust the local parameters of the initial respirator mask 3D model according to the simulation results, and obtain the final respirator mask 3D model;

[0067] The S2 comprises the following steps:

[0068] S21, constructing an initial respirator mask 3D model according to the current facial point cloud model;

[0069] Specifically, the facial point cloud data corresponding to the current facial point cloud model can be converted into a triangular mesh model (resolution ≥ 0.2mm) through reverse engineering software, and then a NURBS surface reconstruction technology is used to generate an initial mask shell with optimized sealing performance, thereby obtaining the initial respirator mask 3D model;

[0070] S22, setting the nose pad region medical grade silicone buffer layer thickness data, mask air pipe joint inner diameter data and mask edge chamfer radius data corresponding to the initial respirator mask 3D model, obtaining initial buffer layer thickness data, initial joint inner diameter data and initial edge chamfer radius data; using the initial buffer layer thickness data, the initial joint inner diameter data and the initial edge chamfer radius data to set the initial respirator mask 3D model, obtaining the set respirator mask 3D model;

[0071] For example, a 2.5mm thick medical grade silicone buffer layer can be preset; a rotatable 30° air pipe joint (inner diameter 22mm) is designed at the lower part of the mask frame, the internal airflow channel adopts a spiral guide structure to reduce noise (actual measurement ≤ 35dB), and the edge chamfer radius is 0.8mm to improve the wearing comfort;

[0072] S23, cooperating with the set of facial feature types and setting the current mask fit threshold set; at the same time, cooperating with the set of facial feature types, the set respirator mask 3D model and each facial feature of the current facial point cloud model are respectively fit in the virtual environment, after the simulation is completed, the current fit data set is obtained; adjusting the set respirator mask 3D model according to the current fit data, obtaining the final respirator mask 3D model;

[0073] S23, adjusting the set respirator mask 3D model according to the current fit data, obtaining the final respirator mask 3D model, comprising the following steps:

[0074] S231, when the current fit data set has fit data less than the corresponding current mask fit threshold value in the current mask fit threshold set, adjust the initial buffer layer thickness data, initial joint inner diameter data, and initial edge chamfer radius data until the current fit data set does not have fit data less than the corresponding current mask fit threshold value in the current mask fit threshold set, obtaining the final buffer layer thickness data, the final joint inner diameter data, the final edge chamfer radius data, and the final respirator mask 3D model; otherwise, no adjustment is needed, and the set respirator mask 3D model is taken as the final respirator mask 3D model;

[0075] Based on a high-precision facial point cloud model (resolution ≥ 0.2mm), an initial 3D mask model is constructed, and through parameterized design of nose pad buffer layer thickness, joint inner diameter, and edge chamfer radius, combined with virtual fit simulation (leakage gap threshold ≤ 1.5mm), the structure parameters are dynamically adjusted until the sealing requirement is met; compared with the traditional try-on method, the sealing qualified rate is improved by more than 30%, especially suitable for patients with complex facial anatomy; the medical-grade silicone buffer layer and the self-adaptive chamfer design can reduce the local pressure peak value and reduce the risk of pressure injury in sensitive areas such as the nasal bridge and the malar bone; virtual iterative optimization (≤3 times) replaces traditional 5-8 times of physical try-on, shortening the adaptation cycle by more than 60%; the parameterized model can be directly connected to 3D printing production, reducing mold development costs;

[0076] S3, 3D printing the final respirator mask 3D model and performing deformation testing, and then adjusting the printing thickness parameter in the 3D printing process according to the test results to obtain the final respirator mask;

[0077] The S3 includes the following steps:

[0078] S31, set the current initial 3D printing thickness data; according to the current initial 3D printing thickness data, output the final respirator mask 3D model in STL format in the virtual environment in S23 and perform 3D printing operation to obtain the initial respirator mask;

[0079] S32, set the current deformation test pressure and the current deformation threshold value; perform deformation testing on the initial respirator mask under the current deformation test pressure, and after the testing is completed, obtain the current deformation data; if the current deformation data is greater than or equal to the current deformation threshold value, adjust the current initial 3D printing thickness data and repeat S31 until the current deformation data is less than the current deformation threshold value, obtaining the current final 3D printing thickness data and the final respirator mask;

[0080] Exemplary, the initial respirator mask can be placed under a pressure of 15 cmH2O for deformation test deformation, if the deformation data is <0.3 mm, then the deformation test is passed.

[0081] By dynamically adjusting the printing thickness parameter and combining the deformation pressure test closed loop verification, it is ensured that the mask maintains structural integrity under the rated ventilation pressure (such as 15-30 cmH2O), avoiding the problem of gas leakage caused by material deformation in clinical use. Compared with the traditional homogeneous printing scheme, the anti-deformation ability is improved by more than 40%; the STL model is linked with the physical test data for feedback, a quantitative relationship database of printing parameters-deformation performance is established, providing a process benchmark for subsequent batch production, so that the product qualified rate is stably above 98%;

[0082] S4, performing an inflation experiment on the final respirator mask and continuing to adjust the printing thickness parameter adjusted in S3 to obtain a first adjusted respirator mask;

[0083] The S4 includes the following steps:

[0084] S41, setting a test air pressure value set and an air pressure detection period; according to the test air pressure value set, the air pressure in the final respirator mask is filled with gas with a preset test air pressure value multiple times; after the inflation is completed, the air pressure in the final respirator mask is tested after the air pressure detection period, and a current air pressure experiment data set is obtained;

[0085] S42, setting a gas pressure leakage rate threshold value corresponding to each inflation in S41 to obtain a current gas pressure leakage rate threshold value set; calculating the gas pressure leakage rate data corresponding to each inflation in S41 according to the test air pressure value set and the current air pressure experiment data set to obtain a current gas pressure leakage rate data set; wherein, the calculation of the gas pressure leakage rate is the prior art;

[0086] S43, when there is leakage rate data greater than or equal to the corresponding gas pressure leakage rate threshold value in the current gas pressure leakage rate threshold value set in the current gas pressure leakage rate data set, return to S32, on the basis that the current deformation data is less than the current deformation threshold, continue to adjust the current final 3D printing thickness data, and repeat S41, S42 and S43 until there is no leakage rate data greater than or equal to the corresponding gas pressure leakage rate threshold value in the current gas pressure leakage rate threshold value set in the current gas pressure leakage rate data set, to obtain a first adjusted respirator mask; otherwise, no return to S32 for adjustment, the final respirator mask is taken as the first adjusted respirator mask;

[0087] Exemplary, as follows:

[0088] The ventilator mask is sequentially filled with 4 cmH2O, 8 cmH2O, and 12 cmH2O standard pressure gases, and after each inflation, a 5-second stabilization time is maintained, and pressure decay data is collected through a high-precision pressure sensor (±0.2% FS);

[0089] The leakage rate threshold is set to 0.8 L / min (low pressure), 1.2 L / min (medium pressure), and 1.5 L / min (high pressure), respectively, and when a leakage rate of 1.3 L / min is detected at a pressure of 8 cmH2O, the system automatically triggers the 3D printing thickness to be adjusted from 3.8 mm to 4.2 mm, and the structural deformation value is verified to be always below the safety threshold of 0.15 mm through the deformation sensor;

[0090] The optimized mask is tested for 3 iterations, and finally realizes that the leakage rate of all pressure segments is below the threshold standard (actual measurement data: 4 cmH2O / 0.5 L / min, 8 cmH2O / 1.1 L / min, 12 cmH2O / 1.4 L / min), and the test process is completed through a 4-channel parallel system, and the single-cycle time consumption is 2 minutes and 45 seconds;

[0091] A test set covering the clinical pressure range is established, and the leakage characteristics under different working conditions are captured through periodic air pressure detection, which can detect 90% of potential leakage risks in advance compared with traditional single-point testing, and the product reliability is improved by 40%; the 3D printing thickness adjustment is dynamically associated with the deformation data and the leakage rate data to form a collaborative optimization mechanism of material mechanics performance and sealing performance, which can simultaneously improve the structural strength and air tightness of the mask by 35%; the "test-analysis-adjustment" automatic closed loop is adopted, and when the leakage exceeds the standard, the printing parameter optimization is automatically triggered, the single iteration time is shortened to 1 / 3 of the traditional method, the product development cycle is significantly accelerated, and the final product pass rate can reach more than 98%;

[0092] S5, making the current patient wear the first adjusted ventilator mask and performing a standard action, and collecting air pressure leakage rate to continue adjusting the local parameters adjusted in S2, and repeating S3, S4, and S5 to obtain a second adjusted ventilator mask;

[0093] S5 includes the following steps:

[0094] S51, setting a standardized action set and wearing the first adjusted ventilator mask on the face of the current patient; the standardized action set includes shaking the head and speaking, etc.; after wearing, the air pressure data in the first adjusted ventilator mask at this moment is collected to obtain the current wearing initial air pressure data;

[0095] S52, cooperating with the standardized action set to make the current patient perform the standardized action; during the execution, the air pressure data in the first adjusted ventilator mask is collected in real time to obtain a current wearing real-time air pressure data set;

[0096] setting a current wearing air pressure leakage rate threshold value; calculating air pressure leakage rate data corresponding to each current wearing real-time air pressure data according to the current wearing real-time air pressure data set and the current wearing initial air pressure data, to obtain a current wearing air pressure leakage rate data set;

[0097] S53, when there is air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold value in the current wearing air pressure leakage rate data set, returning to S23, on the basis that there is no fit degree data less than the corresponding current mask fit degree threshold value in the current fit degree data set, continuing to adjust the final buffer layer thickness data, the final joint inner diameter data and the final edge chamfer radius data, and repeating S31, S32, S41, S42, S43, S51, S52 and S53 until there is no air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold value in the current wearing air pressure leakage rate data set, to obtain a second adjusted respirator mask;

[0098] Otherwise, no adjustment is needed to return to S23, the first adjusted respirator mask is taken as the second adjusted respirator mask, and S61 is entered;

[0099] For example, as follows:

[0100] The first adjusted respirator mask (initial pressure 8 cmH2O) is worn on the patient's face, and static sealing data is collected through the built-in pressure sensor (accuracy ±0.25 cmH2O);

[0101] The patient performs the following standardized actions in turn: shaking the head left and right by 45 degrees (2 times / second), reading a standard text (sound intensity 65 dB), and simulating a sleep turning-over action, and pressure fluctuation data is recorded synchronously;

[0102] The test data shows that the head shaking action causes the pressure to drop to 6.8 cmH2O (leakage rate 1.8 L / min), which exceeds the preset threshold value 1.5 L / min; at this time, the parameter optimization is automatically triggered: the buffer layer thickness is increased from 5 mm to 6 mm, the joint inner diameter is adjusted from 22 mm to 20 mm, and the edge chamfer radius is corrected from 2 mm to 2.5 mm;

[0103] After 3 iterations of adjustment, the dynamic leakage rate is stabilized below 1.2 L / min (actual measurement data: head shaking 1.1 L / min, speaking 0.9 L / min, and turning over 1.0 L / min);

[0104] By simulating the patient's night activity state through a standardized action set (head shaking / speaking, etc.) combined with a real-time air pressure monitoring system, intermittent leakage problems that cannot be detected by traditional static tests can be captured; experimental data show that this method improves the dynamic sealing qualification rate of the mask by 42%; an intelligent correlation between air pressure leakage rate and structural parameters (cushion / joint / chamfer) is established, and when the dynamic leakage exceeds the standard, the parameter adjustment cycle is automatically triggered. This mechanism shortens the product clinical fitting cycle by 60% and reduces the rework rate by 35%; while ensuring that the mechanical performance meets the standard (leakage rate < threshold), the fit data is continuously optimized for wearing comfort, and the patient's treatment compliance is improved by 55%;

[0105] S6, predicting the air pressure leakage rate at a future time based on the air pressure leakage rate collected in S5;

[0106] The S6 includes the following steps:

[0107] S61, setting a training data ratio and a test data ratio; using the training data ratio and the test data ratio to divide the current wearing air pressure leakage rate data set, to obtain a current wearing air pressure leakage rate training data set, a current wearing air pressure leakage rate test data set, and a current wearing air pressure leakage rate to be predicted data set;

[0108] S62, constructing an initial BP neural network model; using the current wearing air pressure leakage rate training data set and the current wearing air pressure leakage rate test data set to train and test the initial BP neural network model respectively, to obtain a final BP neural network model; wherein the training error threshold and the test accuracy threshold involved in the training and testing can be set according to actual requirements, which will not be described here;

[0109] S63, inputting the current wearing air pressure leakage rate to be predicted data set into the final BP neural network model to predict the air pressure leakage rate at a future time, to obtain a future wearing air pressure leakage rate data set;

[0110] By using a BP neural network to establish a dynamic leakage rate prediction model, potential leakage risks can be warned in advance, and the response speed of clinical intervention is improved by 60% (test data shows that the prediction accuracy is 92%); secondly, through the dynamic division mechanism of the training set and the test set (typical ratio 7:3), the model generalization ability (test set error <5%) is ensured, and 15% of real-time data is reserved for continuous optimization; finally, the prediction result is intelligently linked with the mask parameter library, and when the predicted leakage rate exceeds the standard, the optimal adjustment scheme is automatically recommended, which shortens the product iteration cycle by 40%;

[0111] S7, predicting in advance that the current patient replaces the second adjusted ventilator mask according to the prediction result in S6;

[0112] The S7 comprises the following steps:

[0113] S71, when the air pressure leakage rate data in the future wearing data set is greater than or equal to the current wearing air pressure leakage rate threshold, the time corresponding to the air pressure leakage rate data corresponding to each standard action is recorded respectively, and a standard action leakage rate threshold exceeding time set is obtained; otherwise, no recording is required;

[0114] S72, according to the standard action leakage rate threshold exceeding time set, the current patient is prompted to replace the second adjusted respirator mask when the current patient wears the second adjusted respirator mask while performing a standard action;

[0115] By establishing a leakage risk prediction model based on action characteristics and time sequence analysis, a replacement warning can be issued 30 seconds before a specific action (such as turning over / speaking) causes the respirator mask leakage rate to exceed the threshold (clinical tests show that the warning accuracy rate reaches 89%); secondly, through the action-leakage time mapping database, the mask service life prediction error is reduced from ± 15 days to ± 3 days; the treatment interruption rate caused by mask leakage is reduced by 67%, and the waste of consumables is reduced by 37%;

[0116] Through multi-modal data fusion and intelligent prediction, the individualized adaptation effect of the respirator mask is significantly improved: first, based on high-precision point cloud modeling, the digital reconstruction of facial anatomical structure is realized, which improves the initial fit of the mask by more than 40%; secondly, through dynamic action simulation and air pressure leakage rate double closed-loop adjustment mechanism, the leakage risk caused by actions such as shaking head and speaking is reduced by 67%; finally, combined with the time sequence prediction model, preventive maintenance is realized, which can early warn the mask failure risk and reduce the treatment interruption rate to below 3%. The whole system forms a "modeling-optimization-prediction" full-process intelligent closed loop, which reduces the clinical debugging time by 60% compared with the traditional method.

[0117] Embodiment two

[0118] Please refer to Figure 2 The embodiment discloses a respirator mask sealing auxiliary test system based on 3D facial data, which can realize the method of the above-mentioned embodiment, comprising a current patient facial point cloud model generation module, an initial respirator mask 3D model construction module, a respirator mask 3D model local parameter adjustment module, a current mask 3D printing parameter adjustment module, an inflation experiment printing parameter adjustment module, a standard action local parameter adjustment module, a future wearing air pressure leakage rate prediction module, and a standard action wearing replacement prediction module.

[0119] The current patient facial point cloud model generation module collects current patient facial feature data, and then generates a high-resolution point cloud model of the current patient according to the facial feature data to obtain a current facial point cloud model.

[0120] The initial respirator mask 3D model construction module constructs an initial respirator mask 3D model according to the current face point cloud model;

[0121] The respirator mask 3D model local parameter adjustment module simulates the fitting degree of the initial respirator mask 3D model and the current face point cloud model, and then adjusts the local parameters of the initial respirator mask 3D model according to the simulation results until the fitting degree data reaches the preset threshold, thereby obtaining a final respirator mask 3D model;

[0122] The current mask 3D printing parameter adjustment module performs 3D printing on the final respirator mask 3D model and performs deformation testing, and then adjusts the printing thickness parameter in the 3D printing process according to the test results until the test results meet the requirements, thereby obtaining a final respirator mask;

[0123] The inflation experiment printing parameter adjustment module performs an inflation experiment on the final respirator mask, and continues to adjust the adjusted printing thickness parameter in S3 according to the air pressure leakage rate in the inflation experiment until the air pressure leakage rate in the inflation experiment meets the preset threshold, thereby obtaining a first adjusted respirator mask;

[0124] The standard action local parameter adjustment module makes the current patient wear the first adjusted respirator mask and perform a standard action, continues to adjust the adjusted local parameter in S2 according to the air pressure leakage rate of the first adjusted respirator mask in the execution process, and repeats S3, S4 and S5 until the air pressure leakage rate of the first adjusted respirator mask in the execution process meets the preset threshold, thereby obtaining a second adjusted respirator mask;

[0125] The future wearing air pressure leakage rate prediction module predicts the air pressure leakage rate at a future time based on the air pressure leakage rate of the first adjusted respirator mask in the execution process in S5, thereby obtaining a future wearing air pressure leakage rate data set;

[0126] The standard action wearing replacement prediction module predicts in advance that the current patient will replace the second adjusted respirator mask according to the future wearing air pressure leakage rate data set.

[0127] In the description of the present specification, the description of the terms "one embodiment", "example", "specific example" and the like means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0128] The preferred embodiments of the application disclosed above are only to facilitate the elucidation of the application. The preferred embodiments do not describe all the details of the application and limit the application to the specific embodiments described. Obviously, many modifications and variations can be made in light of the teachings above. The description is chosen and described in order to provide the best illustration of the application principles and their practical application, so that those skilled in the art can well understand and utilize the application.

Claims

1. A method for ventilator mask seal assist testing based on 3D facial data, characterized by, Comprise the following steps: S1, collect the current patient facial feature data, and generate a high-resolution point cloud model of the current patient according to the facial feature data, to obtain the current facial point cloud model; S2, construct an initial respirator mask 3D model and simulate the fit degree with the current facial point cloud model, adjust the local parameters of the initial respirator mask 3D model according to the simulation result, and obtain the final respirator mask 3D model; S3, 3D printing of the final respirator mask 3D model and deformation test, and according to the test result, the printing thickness parameter in the 3D printing process is adjusted, and the final respirator mask is obtained; S4, the final respirator mask is inflated and the printing thickness parameter adjusted in S3 is continuously adjusted, and the first adjusted respirator mask is obtained; S5, make the current patient wear the first adjusted respirator mask and perform the standard action, then collect the air pressure leakage rate, adjust the local parameters adjusted in S2, and repeat S3, S4 and S5 until the air pressure leakage rate of the first adjusted respirator mask in the execution process meets the preset threshold, and the second adjusted respirator mask is obtained; S6, predict the air pressure leakage rate at future time based on the air pressure leakage rate collected in S5; S7, according to the prediction result in S6, predict the current patient to replace the second adjusted respirator mask in advance.

2. The 3D facial data based ventilator mask seal assist test method of claim 1, wherein, The S1 comprises the following steps: S11, set a plurality of 3D facial feature types to obtain a facial feature type set; cooperate with the facial feature type set to collect the current patient facial feature data to obtain the current patient facial feature data set; S12, set a facial model resolution threshold; generate a high-resolution point cloud model of the current patient's face with a resolution greater than or equal to the facial model resolution threshold according to the current patient facial feature data set, to obtain the current facial point cloud model.

3. The 3D facial data based respirator mask seal assist test method of claim 2, wherein, The S2 comprises the following steps: S21, construct an initial respirator mask 3D model according to the current facial point cloud model; S22, set the nose pad area medical grade silicone buffer layer thickness data, mask air pipe joint inner diameter data and mask edge chamfer radius data corresponding to the initial respirator mask 3D model, to obtain the initial buffer layer thickness data, initial joint inner diameter data and initial edge chamfer radius data; set the initial respirator mask 3D model using the initial buffer layer thickness data, initial joint inner diameter data and initial edge chamfer radius data, to obtain the set respirator mask 3D model; S23, cooperate with the facial feature type set and set the current mask fit degree threshold set; at the same time, cooperate with the facial feature type set to simulate the fit degree of the set respirator mask 3D model and each facial feature of the current facial point cloud model in a virtual environment, and after the simulation is completed, the current fit degree data set is obtained; adjust the set respirator mask 3D model according to the current fit degree data to obtain the final respirator mask 3D model.

4. The 3D facial data based ventilator mask seal assist test method of claim 3, wherein, The S23 adjusts the set respirator mask 3D model according to the current fit degree data to obtain the final respirator mask 3D model, which comprises the following steps: S231, when the current fit data set has fit data less than the corresponding current mask fit threshold value in the current mask fit threshold set, adjusting the initial buffer layer thickness data, initial joint inner diameter data, and initial edge chamfer radius data until the current fit data set does not have fit data less than the corresponding current mask fit threshold value in the current mask fit threshold set, obtaining the final buffer layer thickness data, final joint inner diameter data, final edge chamfer radius data, and final respirator mask 3D model; otherwise, no adjustment is needed, and the set respirator mask 3D model is taken as the final respirator mask 3D model.

5. The 3D facial data based ventilator mask seal assist test method of claim 4, wherein, The S3 includes the following steps: S31, outputting the final respirator mask 3D model in STL format in the virtual environment in S23 and performing 3D printing operation to obtain an initial respirator mask; S32, setting a current deformation test pressure and a current deformation threshold value; performing deformation test on the initial respirator mask under the current deformation test pressure, and obtaining current deformation data after the test is completed; if the current deformation data is greater than or equal to the current deformation threshold value, adjusting the current initial 3D printing thickness data and repeating S31 until the current deformation data is less than the current deformation threshold value, obtaining the current final 3D printing thickness data and the final respirator mask.

6. The 3D facial data-based ventilator mask seal assist test method of claim 5, wherein, The S4 includes the following steps: S41, setting a test air pressure value set and an air pressure detection period; according to the test air pressure value set, filling the gas with a preset test air pressure value into the final respirator mask multiple times; after the inflation is completed, testing the air pressure in the final respirator mask after the air pressure detection period to obtain a current air pressure experiment data set; S42, setting a current air pressure leakage rate threshold value corresponding to each inflation in S41 to obtain a current air pressure leakage rate threshold value set; calculating the air pressure leakage rate data corresponding to each inflation in S41 according to the test air pressure value set and the current air pressure experiment data set to obtain a current air pressure leakage rate data set; S43, when there is leakage rate data greater than or equal to the corresponding air pressure leakage rate threshold value in the current air pressure leakage rate threshold value set in the current air pressure leakage rate data set, returning to S32 to continue adjusting the current final 3D printing thickness data, and repeating S41, S42, and S43 until there is no leakage rate data greater than or equal to the corresponding air pressure leakage rate threshold value in the current air pressure leakage rate threshold value set in the current air pressure leakage rate data set, obtaining a first adjusted respirator mask; otherwise, no return to S32 is needed for adjustment, and the final respirator mask is taken as the first adjusted respirator mask.

7. The 3D facial data-based ventilator mask seal assist test method of claim 6, wherein, The S5 includes the following steps: S51, setting a standardized action set and wearing the first adjusted respirator mask on the face of the current patient; after the wearing is completed, collecting the air pressure data in the first adjusted respirator mask at this moment to obtain current wearing initial air pressure data; S52, cooperate with the standardized action set, and make the current patient perform the standardized action; in the execution process, real-time acquisition of the air pressure data in the first adjusted respirator mask is performed to obtain a current wearing real-time air pressure data set; a current wearing air pressure leakage rate threshold value is set; according to the current wearing real-time air pressure data set and the current wearing initial air pressure data, air pressure leakage rate data corresponding to each current wearing real-time air pressure data is calculated to obtain a current wearing air pressure leakage rate data set; S53, when there is air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold value in the current wearing air pressure leakage rate data set, returning to S23 to continue adjusting the final buffer layer thickness data, the final joint inner diameter data and the final edge chamfer radius data, and repeating S31, S32, S41, S42, S43, S51, S52 and S53 until there is no air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold value in the current wearing air pressure leakage rate data set, to obtain a second adjusted respirator mask; otherwise, without returning to S23 for adjustment, the first adjusted respirator mask is taken as the second adjusted respirator mask, and S6 is entered.

8. The 3D facial data-based ventilator mask seal assist test method of claim 7, wherein, The S6 includes the following steps: S61, data division is performed on the current wearing air pressure leakage rate data set to obtain a current wearing air pressure leakage rate training data set, a current wearing air pressure leakage rate test data set and a current wearing air pressure leakage rate to-be-predicted data set; S62, an initial BP neural network model is constructed; the initial BP neural network model is trained and tested by using the current wearing air pressure leakage rate training data set and the current wearing air pressure leakage rate test data set respectively to obtain a final BP neural network model; S63, the current wearing air pressure leakage rate to-be-predicted data set is input into the final BP neural network model to predict the air pressure leakage rate at a future time, and a future wearing air pressure leakage rate data set is obtained.

9. The 3D facial data-based ventilator mask seal assist test method of claim 8, wherein, The S7 includes the following steps: S71, when there is air pressure leakage rate data greater than or equal to the current wearing air pressure leakage rate threshold value in the future wearing air pressure leakage rate data set, the time corresponding to the air pressure leakage rate data corresponding to each standard action is recorded respectively to obtain a standard action leakage rate threshold value time set; otherwise, no recording is needed; S72, according to the standard action leakage rate threshold value time set, when the current patient performs the standard action while wearing the second adjusted respirator mask, it is predicted in advance that the current patient will replace the second adjusted respirator mask.

10. A system for implementing a 3D facial data based ventilator mask seal assist test method as claimed in any one of claims 1-9, characterized by: The current patient face point cloud model generation module, the initial respirator mask 3D model construction module, the respirator mask 3D model local parameter adjustment module, the current mask 3D printing parameter adjustment module, the inflation experiment printing parameter adjustment module, the standard action local parameter adjustment module, the future wearing air pressure leakage rate prediction module and the standard action wearing replacement prediction module are included. The current patient face point cloud model generation module acquires current patient facial feature data, and then generates a high-resolution point cloud model of the current patient according to the facial feature data to obtain a current face point cloud model. The initial respirator mask 3D model construction module constructs an initial respirator mask 3D model according to the current facial point cloud model; The respirator mask 3D model local parameter adjustment module simulates the fitting degree of the initial respirator mask 3D model and the current facial point cloud model, and then adjusts the local parameters of the initial respirator mask 3D model according to the simulation results until the fitting degree data reaches a preset threshold, thereby obtaining a final respirator mask 3D model; The current mask 3D printing parameter adjustment module performs 3D printing on the final respirator mask 3D model and performs deformation testing, and then adjusts the printing thickness parameter in the 3D printing process according to the test results until the test results meet the requirements, thereby obtaining a final respirator mask; The inflation experiment printing parameter adjustment module performs an inflation experiment on the final respirator mask, and continues to adjust the adjusted printing thickness parameter in S3 according to the air pressure leakage rate in the inflation experiment until the air pressure leakage rate in the inflation experiment meets a preset threshold, thereby obtaining a first adjusted respirator mask; The standard action local parameter adjustment module makes the current patient wear the first adjusted respirator mask and perform a standard action, continues to adjust the adjusted local parameter in S2 according to the air pressure leakage rate of the first adjusted respirator mask in the execution process, and repeats S3, S4 and S5 until the air pressure leakage rate of the first adjusted respirator mask in the execution process meets a preset threshold, thereby obtaining a second adjusted respirator mask; The future wearing air pressure leakage rate prediction module predicts the air pressure leakage rate at a future time based on the air pressure leakage rate of the first adjusted respirator mask in the execution process in S5, thereby obtaining a future wearing air pressure leakage rate data set; The standard action wearing replacement prediction module predicts in advance that the current patient will replace the second adjusted respirator mask according to the future wearing air pressure leakage rate data set.

Images