Flexible sensor preparation method, sensor, and facial expression recognition system and method

By adopting the dual-coupled microstructure and electrical double-layer effect of polyvinylidene fluoride and polyvinyl alcohol in facial expression recognition technology, the problems of insufficient signal of a single sensor and interference from the membrane are solved, and the accuracy of facial expression recognition is improved and anti-interference ability, achieving high sensitivity and wide range of facial expression recognition.

CN119063884APending Publication Date: 2024-12-03UNIV OF JINAN
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411079484.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-07
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Traditional facial expression recognition technology uses a single sensor signal, resulting in fewer facial expression features, limited data volume and susceptible to external interference, reducing the recognition accuracy, especially in multimode sensors, membrane interference leads to a decrease in capacitance performance, affecting facial expression recognition. accuracy.

Method used

Polyvinylidene fluoride material is used to prepare porous electrodes by mixed liquid phase separation method, and thin spinning films are prepared in combination with electrospinning method to form a double-coupled microstructure between the electrode layer and the dielectric layer, and polyvinyl alcohol doped with different concentrations The 1-butyl-2,3-dimethylimidazole chloride salt forms an electrical bilayer effect, improving sensor performance and ensuring the accuracy of facial expression recognition.

Benefits of technology

It significantly improves the sensitivity and pressure detection range of the sensor, enhances the accuracy and anti-interference ability of facial expression recognition, and ensures real-time and accurate recognition of facial expressions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119063884A_ABST
    Figure CN119063884A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of facial expression recognition, and provides a flexible sensor preparation method, a sensor, a facial expression recognition system and a facial expression recognition method. In consideration of the promotion effect of a microstructure on capacitive performance, polyvinylidene fluoride is selected as a material, and a porous electrode is prepared by using a mixed liquid phase separation method; a thin spinning film is prepared on the dielectric layer by adopting an electrostatic spinning method, so that a double-coupled microstructure of the electrode layer and the dielectric layer is formed, and the performance of the sensor is remarkably improved. Meanwhile, in order to form an electric double-layer effect and improve the response of capacitance, a dielectric layer of the sensor is selected from polyvinyl alcohol doped with 1-butyl-2, 3-dimethylimidazole chlorine salts with different concentrations; the flexible sensor prepared by the invention combines the advantages of a double-coupled microstructure and an electric double-layer effect, and when information acquisition of facial expression recognition is carried out, ultrahigh sensitivity and an ultra-wide pressure detection range can be ensured, the capacitive performance and the sensor performance can be ensured, and the accuracy of facial expression recognition is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of facial expression recognition, and in particular relates to a method for preparing a flexible sensor, a sensor, a facial expression recognition system and a method. Background Art

[0002] Facial expression recognition plays a vital role in emotion cognition. Traditional facial expression recognition often recognizes facial expressions by collecting single signals such as facial muscle stretching action signals and facial electromyography signals. The facial expression features that can be extracted from a single sensor signal are relatively few, and the amount of sensor signal data is relatively limited. In addition, the collection process is easily interfered by external factors and has poor anti-interference ability, resulting in a decrease in the accuracy of facial expression recognition.

[0003] The inventors have discovered that setting up a multi-mode sensor to simultaneously collect facial capacitive signals and electromyographic signals can solve the problems existing in using a single signal acquisition sensor when recognizing facial expressions; however, in a multi-mode sensor, interference is bound to exist between the membrane layer used to collect capacitive signals and the membrane layer used to collect electromyographic signals, which will affect the capacitive response of the sensor, thereby reducing the capacitive performance and sensor performance, and the accuracy of facial expression recognition cannot be guaranteed. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a flexible sensor preparation method, a sensor, a facial expression recognition system and a method. Considering the promoting effect of microstructure on capacitance performance, polyvinylidene fluoride is selected as the material to prepare a porous electrode using a mixed liquid phase separation method, and the dielectric layer adopts an electrostatic spinning method to prepare a thinner spinning film, forming a microstructure of dual coupling of electrode layer and dielectric layer, which significantly improves the sensor performance; at the same time, in order to form an electric double layer effect to improve the response of the capacitor, the dielectric layer of the sensor uses polyvinyl alcohol doped with different concentrations of 1-butyl-2,3-dimethylimidazolium chloride; the prepared flexible sensor combines the advantages of the dual coupling microstructure and the electric double layer effect, can ensure the capacitance performance and sensor performance, and ensure the accuracy of facial expression recognition.

[0005] In order to achieve the above-mentioned object, in a first aspect, the present invention provides a method for preparing a flexible sensor, which adopts the following technical solution:

[0006] A method for preparing a flexible sensor comprises preparing a dielectric layer, arranging electrodes of a capacitive sensor on both sides of the prepared dielectric layer; and silver-plating the outer side of the capacitive sensor plate to prepare an electrode for collecting electromyographic signals.

[0007] An electrospinning method is used to prepare a spinning membrane to form a microstructure of a dual coupling of an electrode layer and a dielectric layer; the dielectric layer is prepared by doping polyvinyl alcohol with different concentrations of 1-butyl-2,3-dimethylimidazolium chloride; the electrode is a porous electrode, and polyvinylidene fluoride is selected as the material and prepared by a mixed liquid phase separation method.

[0008] Furthermore, the preparation of the electrode includes: using a polyvinylidene fluoride film as a substrate, and depositing Ag as an electrode on the polyvinylidene fluoride film by magnetron sputtering.

[0009] Further, a porous membrane is prepared, comprising: adding 5.5 g of N,N-dimethylformamide to polyvinylidene fluoride and stirring at 70° until completely dissolved, then adding 3 g of n-octane, and stirring the mixed solution at 130° until uniformly mixed;

[0010] Then, the spin coating step is performed. Take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it for 15 seconds. After the spin coating is completed, let it stand in the air for 1 minute. Finally, put the glass sheet into water to cool until the film turns white and peels off the glass sheet.

[0011] Furthermore, the Ag thin film was prepared, including: depositing Ag on the porous membrane by magnetron sputtering; after the chamber pressure was pumped to 5×10-3Pa, argon gas was introduced at a flow rate of 20sccm, and the DC power supply was set to 50W at a working gas pressure of 1Pa, and sputtering was performed for 15 minutes.

[0012] Furthermore, the prepared solution was dropped onto a glass slide and spin-coated at a speed of 600 to 1600 rpm for 15 s.

[0013] Furthermore, the preparation of the dielectric layer includes: taking 2g of polyvinyl alcohol particles and adding 18g of deionized water, stirring at 90° and 800 rpm for 8 hours until the polyvinyl alcohol is completely dissolved, to obtain a 10wt% polyvinyl alcohol solution; adding different mass fractions of 1-butyl-2,3-dimethylimidazolium chloride ionic liquid to the prepared polyvinyl alcohol solution, stirring until completely dissolved, to obtain a polyvinyl alcohol / 1-butyl-2,3-dimethylimidazolium chloride precursor solution; electrospinning at a flow rate of 0.5ml / h and an applied voltage of 19kV; the aluminum foil on the stainless steel substrate is used as a receiving base for the electrospun fiber membrane; the collector is kept at a distance of 12 cm from the syringe; and the obtained product is dried at room temperature.

[0014] Furthermore, a sandwich structure sensor was assembled using an electrospun membrane of polyvinyl alcohol / 1-butyl-2,3-dimethylimidazolium chloride as a dielectric layer and a porous membrane prepared by liquid phase separation using polyvinylidene fluoride as an electrode layer.

[0015] In order to achieve the above-mentioned object, in a second aspect, the present invention further provides a flexible sensor, which adopts the following technical solution:

[0016] A sensor is prepared by using the flexible sensor preparation method as described in the first aspect.

[0017] In order to achieve the above object, in a third aspect, the present invention also provides a facial expression recognition system, which adopts the following technical solution:

[0018] A facial expression recognition system comprises a flexible sensor, a data acquisition circuit and a host computer which are connected to each other; the flexible sensor is the sensor as described in the second aspect.

[0019] In order to achieve the above object, in a fourth aspect, the present invention also provides a facial expression recognition system method, which adopts the following technical solution:

[0020] A facial expression recognition method adopts the facial expression recognition system as described in the third aspect, including: using a flexible sensor to collect facial capacitance signals and electromyographic signals; and using a preset neural network model to perform facial expression recognition based on the capacitance signals and electromyographic signals.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] The present invention prepares a capacitive sensor for collecting facial capacitance signals, and an electrode for collecting facial electromyographic signals is arranged on the outer side of the prepared capacitive sensor plate; wherein, considering the promotion effect of the microstructure on the capacitance performance, polyvinylidene fluoride is selected as the material to prepare the porous electrode using the mixed liquid phase separation method, and the dielectric layer adopts the electrostatic spinning method to prepare a thinner spinning film, forming a microstructure of dual coupling of the electrode layer and the dielectric layer, which significantly improves the sensor performance. At the same time, in order to form an electric double layer effect to improve the response of the capacitor, the dielectric layer of the sensor uses polyvinyl alcohol doped with different concentrations of 1-butyl-2,3-dimethylimidazolium chloride; the flexible sensor prepared by the present invention combines the advantages of the dual coupling microstructure and the electric double layer effect, and can ensure ultra-high sensitivity and ultra-wide pressure detection range when collecting information for facial expression recognition, and can ensure the capacitance performance and sensor performance, and ensure the accuracy of facial expression recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0024] Figure 1 It is the signal acquisition and transmission process of the facial expression recognition system based on flexible sensors.

[0025] Figure 2 It is the neural network training and recognition process of the facial expression recognition system based on flexible sensors.

[0026] Figure 3 It is a preparation flow chart of dual-mode sensor.

[0027] Figure 4 This is a schematic diagram of the three-dimensional structure of the dual-mode sensor. The marks in the figure represent: substrate - polyvinylidene fluoride; electrode - silver; dielectric layer - electrospun membrane of polyvinyl alcohol doped with 1-butyl-2,3-dimethylimidazolium chloride.

[0028] Figure 5 This is a cross-sectional SEM image of the polyvinylidene fluoride porous membrane of the electrode layer in the dual-mode sensor.

[0029] Figure 6 This is the surface SEM image of the polyvinylidene fluoride porous membrane of the electrode layer in the dual-mode sensor.

[0030] Figure 7 It is the SEM image of the electrospun membrane of the interface layer in the dual-mode sensor.

[0031] Figure 8 This is the capacitance response sensitivity of polyvinyl alcohol doped with 1-butyl-2,3-dimethylimidazolium chloride at different concentrations.

[0032] Fig. 9 It is the host computer interface of the facial expression recognition system based on flexible sensors.

[0033] Fig.10 It is the signal waveform of different facial expressions collected.

[0034] Fig.11 It is a structural diagram of a one-dimensional convolutional neural network. DETAILED DESCRIPTION

[0035] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0036] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0037] Facial expression recognition plays a vital role in emotion cognition. Some studies have shown that 60-80% of communication is non-verbal. Non-verbal information includes facial expressions, eye contact, tone of voice, gestures, and body distance. Facial expression is one of the important ways to express emotional state and intention. It is becoming increasingly important to use human emotional expressions such as emotions and moods to improve the relationship between humans and various machine systems. With the rapid development of computer vision and artificial intelligence, face recognition and facial expression recognition technology have made significant progress, but they still face challenges such as privacy security, posture changes, and specific environmental restrictions. In addition, computer vision-based facial expression recognition work mainly focuses on expression recognition of unobstructed parts, so partial facial occlusion has become one of the important obstacles to the development of technology. Lighting changes are also a key issue, because computer vision-based methods need to collect photos for expression training and recognition, and lighting changes directly affect the quality of photos, which in turn affects the accuracy of recognition. At present, a large number of studies are devoted to solving these problems. With the development of science and technology, people have more and more needs for physiological health monitoring, emotional interaction, human-computer interface, psychological research, and psychotherapy based on flexible sensors. For example, detecting the movement and deformation of facial muscles to obtain facial expression information. Identify emotional states such as anxiety, stress, and fatigue; in virtual reality, augmented reality and other technologies, by sensing the user's facial expressions, the system can make corresponding feedback and adjustments to improve the user experience; control devices through facial expressions, such as eye tracking and facial action recognition, to achieve smarter and more convenient operations; help researchers gain a deeper understanding of psychological processes in emotions and cognition. It can be used to monitor changes in patients' facial expressions and evaluate treatment effects; therefore, it is urgent to develop a human facial information detection and intelligent recognition device based on flexible sensors.

[0038] Traditional facial expression recognition often recognizes facial expressions by collecting single signals such as facial muscle stretching action signals and facial electromyography signals; there may be some problems in facial expression recognition by collecting a sensor signal, such as relatively few facial expression features that can be extracted from a single sensor signal; the amount of data of a single sensor signal is also relatively limited; a single sensor signal is easily interfered by external factors during the acquisition process and has poor anti-interference ability. These problems will lead to a decrease in the recognition accuracy of the facial expression system and reduce the robustness of the facial expression recognition system. Different types of sensor signals often have different characteristics. In addition, different types of sensor signals have different responses to the same facial expression action. In order to improve the accuracy of the facial expression recognition system and provide more features that can be used as classification conditions for the facial expression recognition system, we focus on studying the simultaneous use of multiple signals to classify facial expressions.

[0039] As described in the background technology, setting up a multi-mode sensor to simultaneously collect facial capacitive signals and electromyographic signals can solve the problems existing in using a single signal collection sensor when recognizing facial expressions; however, in the multi-mode sensor, there is bound to be interference between the membrane layer used to collect capacitive signals and the membrane layer used to collect electromyographic signals, which will affect the capacitive response of the sensor, thereby reducing the capacitive performance and sensor performance, and the accuracy of facial expression recognition cannot be guaranteed.

[0040] Based on the above problems, in some embodiments, a flexible sensor preparation method is provided, and a flexible sensor prepared by the flexible sensor preparation method is provided; the flexible sensor is a dual-mode sensor, which can simultaneously collect capacitive signals and electromyographic signals, and a facial expression recognition system is constructed with the assistance of 1d-CNN, and two sensors are used to simultaneously collect facial expressions and perform real-time recognition, thereby realizing real-time classification and processing of facial expressions. The proposed dual-mode sensor includes a flexible pressure-capacitive pressure sensor and a dry electrode for electrophysiological signal monitoring.

[0041] Considering the promotion effect of microstructure on capacitance performance, in the preparation method of flexible sensors, polyvinylidene fluoride was selected as the material to prepare porous electrodes using the mixed liquid phase separation method. The dielectric layer was prepared by electrospinning to form a thin spinning membrane, forming a microstructure of dual coupling of electrode layer and dielectric layer, which significantly improved the sensor performance. In order to form an electric double layer effect to improve the response of the capacitor, we used polyvinyl alcohol doped with different concentrations of 1-butyl-2,3-dimethylimidazolium chloride as the dielectric layer of the sensor. The prepared sensor combines the advantages of dual coupling microstructure and electric double layer effect, giving it ultra-high sensitivity and ultra-wide pressure detection range. In addition, the sensor provided is ultra-thin, only 0.25mm, and can fit well on the skin surface. In order to demonstrate the effective application of the prepared sensor in facial expression recognition, the corresponding applications were explored, including the preparation of a facial expression recognition system by combining capacitive signals and electromyographic signals, which can monitor facial expression changes in real time. The system will recognize facial expressions and give feedback. At the same time, it was verified that the strategy of simultaneous collection, training and recognition of two signals can significantly improve the accuracy of facial expression recognition; it well demonstrated the great practicality and potential of the facial expression recognition system based on flexible sensors in facial expression recognition.

[0042] A facial expression recognition system includes a flexible dual-mode sensor, a data acquisition circuit and a host computer. Optionally, the flexible dual-mode sensor includes a wearable pressure-capacitive pressure sensor and a dry electrode bioelectric sensor. The signal acquisition and transmission process of the data acquisition circuit is as follows: the capacitance signal is acquired by the FDC2214 acquisition chip, and the electromyographic signal is acquired by the bioelectric signal acquisition module. After the acquisition of the two sensor signals is completed, they are handed over to the STM32 main control chip for processing, and then sent to the host computer via Bluetooth. After the host computer receives the data, it is handed over to the neural network for real-time recognition. The host computer interface is constructed by the python third-party library PyQt5, and the main interface includes a title, a waveform display interface, an expression recognition result display window, and a button component. The waveform display interface displays the real-time waveforms of the collected capacitance signal and electromyographic signal respectively; the button component window can select the serial port number and baud rate; the expression recognition result display window can display different facial expressions in real time according to the recognition results.

[0043] The preparation of flexible sensors in the facial expression recognition system includes the preparation of a wearable capacitive pressure sensor with a wide range and high sensitivity, and the preparation of a dry electrode bioelectric sensor with a high signal-to-noise ratio and low root mean square noise, including the following steps:

[0044] S1. Preparation of dry electrode bioelectric sensor:

[0045] Optionally, a polyvinylidene fluoride film is used as a substrate, and Ag is deposited on the polyvinylidene fluoride film by magnetron sputtering to form an electrode.

[0046] S2. Preparation of pressure-capacitive pressure sensor:

[0047] S2.1, take 2g of polyvinyl alcohol particles, add 18g of deionized water, stir at 90° and 800 rpm for 8 hours until the polyvinyl alcohol is completely dissolved, to obtain a 10wt% polyvinyl alcohol solution;

[0048] S2.2, adding different mass fractions of 1-butyl-2,3-dimethylimidazolium chloride ionic liquid to the prepared polyvinyl alcohol solution, stirring until completely dissolved, to obtain a polyvinyl alcohol / 1-butyl-2,3-dimethylimidazolium chloride precursor solution;

[0049] S2.3, electrospinning was performed at a flow rate of 0.5 ml / h and an applied voltage of 19 kV. The aluminum foil on the stainless steel substrate served as the receiving base for the electrospun fiber membrane. The collector was kept at a distance of 12 cm from the syringe. The obtained product was dried at room temperature;

[0050] S2.4, add 5.5g N,N-dimethylformamide to polyvinylidene fluoride and stir at 70° until completely dissolved, then add 3g n-octane, and stir the mixed solution at 130° until it is evenly mixed;

[0051] S2.5. Then, perform the spin coating step. Take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it at 600 rpm for 15 seconds. After spin coating, let it stand in the air for 1 minute. Finally, put the glass sheet into water to cool until the film turns white and peels off from the glass sheet.

[0052] S2.6. Assembly of a wearable capacitive pressure sensor with a wide range and high sensitivity: A sandwich-structured capacitive pressure sensor was assembled using an electrospun membrane of polyvinyl alcohol / 1-butyl-2,3-dimethylimidazolium chloride as the dielectric layer and a porous membrane prepared by liquid phase separation of polyvinylidene fluoride as the electrode layer.

[0053] The dual-mode sensor collects two sensor signals, which are simply processed by the main control chip and then sent to the host computer software via Bluetooth. After receiving the data, the host computer software performs simple data format processing and inputs it into the neural network model for facial expression recognition.

[0054] In the following embodiments, a facial expression recognition system is provided, including a signal acquisition board, a flexible dual-mode sensor, and a host computer.

[0055] like Figure 1 As shown in the figure, the entire facial expression recognition system may include three parts: front-end signal acquisition and transmission circuit, signal analysis and interaction, and application scenarios of facial expression recognition. The front-end signal acquisition and transmission circuit includes a capacitive acquisition module, an electromyography signal acquisition module, an MCU microcontroller, and a Bluetooth wireless data transmission module, which are responsible for real-time acquisition of facial expressions. After simple processing by the microcontroller chip, the signals are sent to the host computer via Bluetooth. The signal analysis and interaction part includes our independently developed host computer interface and a trained one-dimensional convolutional neural network model. The process of real-time signal acquisition, training, and classification is shown in the figure. Figure 2 As shown in the figure, specifically: first, the capacitance and electromyographic signals of different expressions are collected, and the 8 expressions are used as 8 classification labels to distinguish the 8 facial expressions. The collected data are divided into training data sets and test data sets to input the neural network for training, and the trained model is imported into the host computer interface we designed. When receiving facial expression data, our network model will automatically analyze the data and give a correct expression recognition result based on the classification label.

[0056] In some embodiments, a method for preparing a wearable dual-mode sensor with a wide range and high sensitivity, such as Figure 3 As shown, the following steps are included:

[0057] Preparation of PVA / [BMMI][Cl] electrospinning membrane: Take 2g PVA particles and add 18g deionized water, stir at 90°800 rpm for 8 hours until PVA is completely dissolved to obtain a 10wt% PVA solution. Add different mass fractions of [BMMI][Cl] ionic liquid to the prepared PVA solution and stir until completely dissolved to obtain a PVA / [BMMI][Cl] precursor solution. Electrospinning is performed at a flow rate of 0.5ml / h and an applied voltage of 19kV. The aluminum foil on the stainless steel substrate serves as the receiving base for the electrospun fiber membrane. The collector is kept at a distance of 12 cm from the syringe. The obtained product is dried at room temperature.

[0058] Preparation of PVDF porous membrane: First, take 1.5g PVDF and add 5.5g DMF and stir at 70° until completely dissolved, then add 3g n-octane and stir the mixed solution at 130° until it is evenly mixed. Then, perform the spin coating step, take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it at 600 rpm for 15s, after spin coating, let it stand in the air for 1min, and finally put the glass sheet into water to cool until the film turns white and peels off from the glass sheet.

[0059] Preparation of Ag film: First, the PVDF film prepared above was selected, and Ag was deposited on the PVDF film by magnetron sputtering. After the chamber pressure was pumped to 5×10-3Pa, argon gas was introduced at a flow rate of 20sccm, and the DC power supply was set to 50W at a working gas pressure of 1Pa, and the silver film was prepared by sputtering for 15 minutes.

[0060] In some embodiments, a method for preparing a wearable dual-mode sensor with a wide range and high sensitivity, such as Figure 3 As shown, the following steps are included:

[0061] Preparation of PVA / [BMMI][Cl] electrospinning membrane: Take 2g PVA particles and add 18g deionized water, stir at 90°800 rpm for 8 hours until PVA is completely dissolved to obtain a 10wt% PVA solution. Add different mass fractions of [BMMI][Cl] ionic liquid to the prepared PVA solution and stir until completely dissolved to obtain a PVA / [BMMI][Cl] precursor solution. Electrospinning is performed at a flow rate of 0.5ml / h and an applied voltage of 19kV. The aluminum foil on the stainless steel substrate serves as the receiving base for the electrospun fiber membrane. The collector is kept at a distance of 12 cm from the syringe. The obtained product is dried at room temperature.

[0062] Preparation of PVDF porous membrane: First, take 1.5g PVDF and add 5.5g DMF and stir at 70° until completely dissolved, then add 3g n-octane and stir the mixed solution at 130° until it is evenly mixed. Then, perform the spin coating step, take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it at 800 rpm for 15s, after spin coating, let it stand in the air for 1min, and finally put the glass sheet into water to cool until the film turns white and peels off from the glass sheet.

[0063] Preparation of Ag film: First, the PVDF film prepared above was selected, and Ag was deposited on the PVDF film by magnetron sputtering. After the chamber pressure was pumped to 5×10-3Pa, argon gas was introduced at a flow rate of 20sccm, and the DC power supply was set to 50W at a working gas pressure of 1Pa, and the silver film was prepared by sputtering for 15 minutes.

[0064] In some embodiments, a method for preparing a wearable dual-mode sensor with a wide range and high sensitivity, such as Figure 3 As shown, the following steps are included:

[0065] Preparation of PVA / [BMMI][Cl] electrospinning membrane: Take 2g PVA particles and add 18g deionized water, stir at 90°800 rpm for 8 hours until PVA is completely dissolved to obtain a 10wt% PVA solution. Add different mass fractions of [BMMI][Cl] ionic liquid to the prepared PVA solution and stir until completely dissolved to obtain a PVA / [BMMI][Cl] precursor solution. Electrospinning is performed at a flow rate of 0.5ml / h and an applied voltage of 19kV. The aluminum foil on the stainless steel substrate serves as the receiving base for the electrospun fiber membrane. The collector is kept at a distance of 12 cm from the syringe. The obtained product is dried at room temperature.

[0066] Preparation of PVDF porous membrane: First, take 1.5g PVDF and add 5.5g DMF and stir at 70° until completely dissolved, then add 3g n-octane and stir the mixed solution at 130° until it is evenly mixed. Then, perform the spin coating step, take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it at 1000 rpm for 15s, after spin coating, let it stand in the air for 1min, and finally put the glass sheet into water to cool until the film turns white and peels off from the glass sheet.

[0067] Preparation of Ag film: First, the PVDF film prepared above was selected, and Ag was deposited on the PVDF film by magnetron sputtering. After the chamber pressure was pumped to 5×10-3Pa, argon gas was introduced at a flow rate of 20sccm, and the DC power supply was set to 50W at a working gas pressure of 1Pa, and the silver film was prepared by sputtering for 15 minutes.

[0068] In some embodiments, a method for preparing a wearable dual-mode sensor with a wide range and high sensitivity, such as Figure 3 As shown, the following steps are included:

[0069] Preparation of PVA / [BMMI][Cl] electrospinning membrane: Take 2g PVA particles and add 18g deionized water, stir at 90°800 rpm for 8 hours until PVA is completely dissolved to obtain a 10wt% PVA solution. Add different mass fractions of [BMMI][Cl] ionic liquid to the prepared PVA solution and stir until completely dissolved to obtain a PVA / [BMMI][Cl] precursor solution. Electrospinning is performed at a flow rate of 0.5ml / h and an applied voltage of 19kV. The aluminum foil on the stainless steel substrate serves as the receiving base for the electrospun fiber membrane. The collector is kept at a distance of 12 cm from the syringe. The obtained product is dried at room temperature.

[0070] Preparation of PVDF porous membrane: First, take 1.5g PVDF and add 5.5g DMF and stir at 70° until completely dissolved, then add 3g n-octane and stir the mixed solution at 130° until it is evenly mixed. Then, perform the spin coating step, take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it at 1200 rpm for 15s, after spin coating, let it stand in the air for 1min, and finally put the glass sheet into water to cool until the film turns white and peels off from the glass sheet.

[0071] Preparation of Ag film: First, the PVDF film prepared above was selected, and Ag was deposited on the PVDF film by magnetron sputtering. After the chamber pressure was pumped to 5×10-3Pa, argon gas was introduced at a flow rate of 20sccm, and the DC power supply was set to 50W at a working gas pressure of 1Pa, and the silver film was prepared by sputtering for 15 minutes.

[0072] In some embodiments, a method for preparing a wearable dual-mode sensor with a wide range and high sensitivity, such as Figure 3 As shown, the following steps are included:

[0073] Preparation of PVA / [BMMI][Cl] electrospinning membrane: Take 2g PVA particles and add 18g deionized water, stir at 90°800 rpm for 8 hours until PVA is completely dissolved to obtain a 10wt% PVA solution. Add different mass fractions of [BMMI][Cl] ionic liquid to the prepared PVA solution and stir until completely dissolved to obtain a PVA / [BMMI][Cl] precursor solution. Electrospinning is performed at a flow rate of 0.5ml / h and an applied voltage of 19kV. The aluminum foil on the stainless steel substrate serves as the receiving base for the electrospun fiber membrane. The collector is kept at a distance of 12 cm from the syringe. The obtained product is dried at room temperature.

[0074] Preparation of PVDF porous membrane: First, take 1.5g PVDF and add 5.5g DMF and stir at 70° until completely dissolved, then add 3g n-octane and stir the mixed solution at 130° until it is evenly mixed. Then, perform the spin coating step, take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it at 1400 rpm for 15s, after spin coating, let it stand in the air for 1min, and finally put the glass sheet into water to cool until the film turns white and peels off from the glass sheet.

[0075] Preparation of Ag film: First, the PVDF film prepared above was selected, and Ag was deposited on the PVDF film by magnetron sputtering. After the chamber pressure was pumped to 5×10-3Pa, argon gas was introduced at a flow rate of 20sccm, and the DC power supply was set to 50W at a working gas pressure of 1Pa, and the silver film was prepared by sputtering for 15 minutes.

[0076] In some embodiments, a method for preparing a wearable dual-mode sensor with a wide range and high sensitivity, such as Figure 3 As shown, the following steps are included:

[0077] Preparation of PVA / [BMMI][Cl] electrospinning membrane: Take 2g PVA particles and add 18g deionized water, stir at 90°800 rpm for 8 hours until PVA is completely dissolved to obtain a 10wt% PVA solution. Add different mass fractions of [BMMI][Cl] ionic liquid to the prepared PVA solution and stir until completely dissolved to obtain a PVA / [BMMI][Cl] precursor solution. Electrospinning is performed at a flow rate of 0.5ml / h and an applied voltage of 19kV. The aluminum foil on the stainless steel substrate serves as the receiving base for the electrospun fiber membrane. The collector is kept at a distance of 12 cm from the syringe. The obtained product is dried at room temperature.

[0078] Preparation of PVDF porous membrane: First, take 1.5g PVDF and add 5.5g DMF and stir at 70° until completely dissolved, then add 3g n-octane and stir the mixed solution at 130° until it is evenly mixed. Then, perform the spin coating step, take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it at 1600 rpm for 15s, after spin coating, let it stand in the air for 1min, and finally put the glass sheet into water to cool until the film turns white and peels off from the glass sheet.

[0079] Preparation of Ag film: First, the PVDF film prepared above was selected, and Ag was deposited on the PVDF film by magnetron sputtering. After the chamber pressure was pumped to 5×10-3Pa, argon gas was introduced at a flow rate of 20sccm, and the DC power supply was set to 50W at a working gas pressure of 1Pa, and the silver film was prepared by sputtering for 15 minutes.

[0080] Different from the traditional method of using electromyographic signals alone to identify facial expressions, facial electromyographic signals and capacitive signals are collected at the same time, and the characteristics of the two signals are used to jointly improve the accuracy of facial expression recognition. The signal acquisition system (size 3cm*7cm) includes a capacitance digital conversion module (CDC), an electromyographic signal test module (EMG), a main control chip (MCU) and a Bluetooth module, etc., which can simultaneously collect capacitance signals and electromyographic signals and transmit the collected signals to the host computer through Bluetooth wireless for data judgment and expression recognition. Specifically, after the capacitance signal is collected by the acquisition chip, it is sent to the main control chip STM32 through I2C. At the same time, after the electromyographic signal is collected by the acquisition chip, it undergoes a series of amplification and filtering operations and then the data is handed over to STM32. The main control chip processes the data and sends it to the host computer through Bluetooth. After the raw data is processed, it is sent to a one-dimensional convolutional neural network for classification and judgment. Different expressions can be identified according to the classification results. The system interface will display different pictures, realizing real-time data collection, data classification and judgment, and human-computer interaction. Most traditional facial expression recognition methods rely on cameras to capture images, and then upload the captured images to the trained network model for classification and judgment, so as to identify the correct expression. Facial expression recognition methods based on computer vision are easily affected by different backgrounds, lighting and head postures. Wearable flexible sensors can capture subtle facial muscle movements, identify changes in facial expressions, and are robust to posture, occlusion and lighting changes in different experimental environments. We can recognize facial expressions well under different lighting conditions and different occlusion conditions. We designed the recognition of 8 facial expressions, including smile, surprise, contempt, confusion, sadness, anger, blinking left and right. At the same time, it was verified that facial occlusion and lighting changes had no effect on the accuracy of our expression recognition when wearing a mask and at night, proving that our facial expression recognition system can be used in different environments. At the same time, it was proved that the system has high wearability. We used a self-made acquisition system to separately collect capacitive signals and electromyographic signals under different expressions. As can be seen from the figure, the data acquisition is stable, which verifies the practicality of our facial expression acquisition system.

[0081] In terms of data acquisition system integration, the present invention provides a set of small and convenient data acquisition circuit boards with a size of 3cm*7cm. The acquisition board integrates capacitive signal acquisition, electromyographic signal acquisition, microcontroller circuit and Bluetooth wireless data transmission; compared with bulky facial expression acquisition instruments, the circuit is simple and light.

[0082] In terms of sensor performance, the pressure-capacitive pressure sensor prepared by the present invention has a wide range and high sensitivity, and the capacitance is defined as C=εA / 4kπd (where ε is the effective dielectric constant, d is the distance between the two electrodes, k is the electrostatic constant, and A is the effective area of ​​the electrode). In principle, the change in capacitance is controlled by the variables ε, A and d. However, A remains unchanged under external pressure. Therefore, ε and d determine the change in capacitance. When pressure is applied, both ε and d of the prepared sensor change to a certain extent. First, for the prepared sensor, since the microstructures of the electrode layer and the dielectric layer are densely distributed and uniform, they are easily compressed, resulting in a significant decrease in the d value. In addition, according to the Lichterecker mixing rule, the effective dielectric constant of the dielectric layer can be defined as ε α= V air εαair+V c εαc, where V air and V c are the volume fractions of air and dielectric material (V air +V c =1), ε air and ε c are the dielectric constants of air and dielectric layer materials, respectively. Due to the double-coupled microstructure between the electrode layer and the dielectric layer, a large number of air gaps are formed, and the air is squeezed out under small pressure, resulting in ε α Significant increase.

[0083] In addition, a neural network model for facial expression recognition is provided. Optionally, the model consists of two convolutional pooling layers and three fully connected layers, which can distinguish eight expressions. After the data is uploaded to the host computer, it is simply processed and then directly input into the neural network for recognition and judgment.

[0084] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. A method for preparing a flexible sensor, characterized in that: Prepare a capacitive sensor, and set an electrode for collecting facial electromyographic signals on the outer side of the prepared capacitive sensor plate; An electrospinning method is used to prepare a spinning membrane to form a microstructure of a dual coupling of an electrode layer and a dielectric layer; the dielectric layer is prepared by doping polyvinyl alcohol with different concentrations of 1-butyl-2,3-dimethylimidazolium chloride; the electrode is a porous electrode, and polyvinylidene fluoride is selected as the material and prepared by a mixed liquid phase separation method.

2. The method for preparing a flexible sensor according to claim 1, characterized in that: The preparation of the electrode comprises: using a polyvinylidene fluoride film as a substrate, and depositing Ag as an electrode on the polyvinylidene fluoride film by magnetron sputtering.

3. The method for preparing a flexible sensor according to claim 2, characterized in that: The porous membrane is prepared, comprising: adding 5.5 g of N,N-dimethylformamide to polyvinylidene fluoride and stirring at 70° until the mixture is completely dissolved, then adding 3 g of n-octane, and stirring the mixed solution at 130° until the mixture is uniformly mixed; Then, the spin coating step is performed. Take a glass sheet, heat the glass sheet to 130°, then drop the prepared solution on the glass sheet and spin coat it for 15 seconds. After the spin coating is completed, let it stand in the air for 1 minute. Finally, put the glass sheet into water to cool until the film turns white and peels off the glass sheet.

4. The method for preparing a flexible sensor according to claim 3, characterized in that: The preparation of the Ag film includes: depositing Ag on the porous film by magnetron sputtering; pumping the chamber pressure to 5×10-3Pa, introducing argon gas at a flow rate of 20sccm, setting the DC power supply to 50W at a working gas pressure of 1Pa, and sputtering for 15 minutes.

5. The method for preparing a flexible sensor according to claim 3, characterized in that: The prepared solution was dropped onto a glass slide and spin-coated at a speed of 600 to 1600 rpm for 15 s.

6. The method for preparing a flexible sensor according to claim 1, characterized in that: The preparation of the dielectric layer includes: taking 2g of polyvinyl alcohol particles and adding 18g of deionized water, stirring at 90° and 800 rpm for 8 hours until the polyvinyl alcohol is completely dissolved, to obtain a 10wt% polyvinyl alcohol solution; adding different mass fractions of 1-butyl-2,3-dimethylimidazolium chloride ionic liquid to the prepared polyvinyl alcohol solution, stirring until it is completely dissolved, to obtain a polyvinyl alcohol / 1-butyl-2,3-dimethylimidazolium chloride precursor solution; electrospinning at a flow rate of 0.5ml / h and an applied voltage of 19kV; the aluminum foil on the stainless steel substrate is used as a receiving base for the electrospun fiber membrane; the collector and the syringe are kept at a distance of 12cm; and the obtained product is dried at room temperature.

7. The method for preparing a flexible sensor according to claim 1, characterized in that: A sandwich-structured sensor was assembled using an electrospun membrane of polyvinyl alcohol / 1-butyl-2,3-dimethylimidazolium chloride as a dielectric layer and a porous membrane prepared by liquid phase separation of polyvinylidene fluoride as an electrode layer.

8. A sensor, characterized in that The flexible sensor is prepared by the method for preparing the flexible sensor as described in any one of claims 1 to 7.

9. A facial expression recognition system, characterized in that: It comprises a flexible sensor, a data acquisition circuit and a host computer which are interconnected; the flexible sensor adopts the sensor as claimed in claim 8.

10. A facial expression recognition method, characterized in that: The facial expression recognition system as claimed in claim 9 is adopted, comprising: using flexible sensors to collect facial capacitance signals and electromyographic signals; and using a preset neural network model to perform facial expression recognition based on the capacitance signals and electromyographic signals.

Citation Information

Cited By

  • Two-dimensional honeycomb-shaped and three-dimensional spinous process-shaped double-coupled microstructure nanofiber membrane as well as preparation method and application thereof

    CN119956557A

  • Two-dimensional honeycomb and three-dimensional spine-shaped double-coupling microstructure nanofiber membrane, and preparation method and application thereof

    CN119956557B