Psychological assessment system and method for pregnant and lying-in women based on facial recognition
Through the maternal and maternal psychological evaluation system based on facial recognition, the facial jitter change parameters are extracted using image quality evaluation and deep learning models, and the accuracy and convenience of traditional evaluation methods are solved, real-time daily monitoring and efficient evaluation of maternal and maternal psychological evaluation are achieved.
Patent Information
- Application Number
- CN202510328142.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-19
- Publication Date
- 2025-07-04
AI Technical Summary
The traditional psychological assessment methods of pregnant women are insufficiently accurate, have poor convenience and cannot achieve daily monitoring. Especially in remote areas or areas with scarce medical resources, it is difficult for pregnant women to obtain psychological assessment in a timely manner, resulting in timely attention and intervention in psychological problems.
The maternal and maternal psychological evaluation system based on facial recognition is adopted, and by obtaining simulated video signals, using image quality evaluation model, facial recognition technology and deep learning facial jitter change parameter model, facial jitter change parameters are extracted, and psychological physiological characteristic parameters are generated for evaluation.
Accurate and convenient psychological assessment of pregnant women, real-time daily monitoring can be carried out, reduce interference from human factors, and provide objective evaluation methods, which are easy to apply in any site and improve the accuracy and efficiency of evaluation.
Smart Images
Figure CN120260910A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of data processing, and in particular relates to a system and method for evaluating the psychological status of pregnant women based on facial recognition. Background Art
[0002] At present, the traditional psychological assessment methods for pregnant women have many shortcomings. The assessment tools used are often traditional scales. However, in actual applications, some pregnant women may frequently give false or incorrect answers due to their misunderstanding of the questions or deliberately conceal their true feelings, which makes the assessment results biased or invalid. At the same time, the timeliness of traditional scales is insufficient. The psychological assessment of pregnant women can often only be carried out at a specific time point, such as during a prenatal examination or a postpartum follow-up, and the time interval for conducting another psychological assessment is mostly 3 months to half a year, which makes daily monitoring impossible, resulting in the inability to capture the psychological changes of pregnant women in a timely manner, thereby delaying the best time for attention and intervention.
[0003] In addition, the inability of pregnant women to go to the hospital for psychological evaluation on time due to physical inconvenience is also an issue that cannot be ignored. The physical inconvenience during pregnancy and postpartum period makes travel difficult. Facing a series of tedious processes such as transportation and queuing not only increases the physical burden of pregnant women, but also increases the risk of pregnant women contracting diseases. For pregnant women living in remote areas or areas with relatively scarce medical resources, it is more difficult to go to the hospital for psychological evaluation, which leads to some pregnant women giving up psychological evaluation because they are afraid of trouble or cannot reach the hospital, so that their psychological problems do not receive timely attention and intervention.
[0004] Therefore, the traditional maternal psychological assessment method has obvious deficiencies in accuracy, convenience and daily monitoring. Based on this, how to provide an effective solution to achieve accurate, convenient and daily monitoring of maternal psychological assessment has become a difficult problem that needs to be solved urgently in the existing technology. Summary of the invention
[0005] The purpose of the present invention is to provide a maternal and child psychological assessment system and method based on facial recognition to solve the above-mentioned problems existing in the prior art.
[0006] In order to achieve the above object, the present invention adopts the following technical solutions:
[0007] In a first aspect, the present invention provides a maternal psychological assessment system based on facial recognition, comprising: an acquisition detection module, used to acquire an analog video signal of a maternal to be assessed, receive a preset image quality standard, detect the analog video signal based on an image quality assessment model, and determine whether the analog video signal meets the preset image quality standard;
[0008] The data preprocessing module is communicatively connected to the acquisition and detection module, and is configured to receive the analog video signal that meets the preset image quality standard, and preprocess the analog video signal that meets the preset image quality standard to obtain the basic image data, so as to send the basic image data to the feature annotation module after obtaining the basic image data;
[0009] The feature annotation module is communicatively connected to the data preprocessing module, and is configured to receive the basic image data, perform face recognition and key feature point extraction on the basic image data based on the face recognition technology and the face feature point detection algorithm, obtain the key feature points of the face, and annotate the key feature points in the basic image data to obtain the face image data;
[0010] The parameter extraction module is communicatively connected to the feature annotation module, and is configured to receive the face image data, extract parameters from the face image data based on the face jitter change parameter extraction model of deep learning to obtain the face jitter change parameters, and generate the psychophysiological characteristic parameters according to the face jitter change parameters, so as to perform a psychological assessment on the pregnant woman to be evaluated according to the psychophysiological characteristic parameters.
[0011] In a possible design, the parameter extraction module is further configured to perform spatial feature extraction on the face image data based on the convolutional neural network model, input the face jitter data samples marked with different psychophysiological states into the convolutional neural network model, and update the network parameters in the convolutional neural network model by using the backpropagation algorithm;
[0012] The face jitter change parameter extraction model combines a bidirectional recurrent neural network and a temporal attention mechanism, so as to utilize the face jitter change parameters in the front and back time periods of the face image data to capture the dynamic changes of the face jitter.
[0013] In a possible design, before preprocessing the analog video signal that meets the preset image quality standard, the data preprocessing module is further configured to perform A / D conversion on the analog video signal that meets the preset image quality standard to convert the analog video signal into a digital signal, and process the digital signal to obtain the basic image data, so as to perform recognition and annotation on the basic image data.
[0014] In a possible design, after obtaining the basic image data, the data preprocessing module is further configured to perform denoising processing on the basic image data;
[0015] After obtaining the basic image data, the data preprocessing module is further configured to perform image enhancement operations on the basic image data to enhance the contrast and clarity of the facial features, so that after performing the image enhancement operations on the basic image data, the feature extraction is performed on the basic image data by using the face recognition technology and the face feature point detection algorithm;
[0016] The data preprocessing module, after obtaining the basic image data, is further configured to perform cropping and normalization processing on the basic image data.
[0017] In a possible design, it further includes a data encryption module. The data encryption module is communicatively connected to the acquisition and detection module and the data preprocessing module, and is configured to encrypt the analog video signal that meets the preset image quality standard based on a dynamic encryption algorithm, and transmit the encrypted analog video signal to the data preprocessing module.
[0018] In a possible design, it further includes a correction module. The correction module is communicatively connected to the parameter extraction module, and is configured to establish a dynamic correction model library, and optimize and correct the psychophysiological characteristic parameters according to the dynamic correction models in the dynamic correction model library.
[0019] In a possible design, it further includes a data storage module. The data storage module is communicatively connected to the acquisition and detection module and the feature annotation module, and is configured to store the analog video signal and facial image data of the pregnant and lying-in women to be evaluated.
[0020] In a possible design, it further includes a collection device. The collection device is communicatively connected to the acquisition and detection module, and is configured to collect the analog video signal of the pregnant and lying-in women to be evaluated. Among them, the collection device includes a camera with a resolution greater than or equal to the preset resolution.
[0021] In a possible design, it further includes a user terminal. The user terminal is communicatively connected to the parameter extraction module, and the user terminal is configured to receive the psychophysiological characteristic parameters generated by the parameter extraction module, so as to perform psychological evaluation according to the psychophysiological characteristic parameters.
[0022] In a second aspect, the present invention provides a method for psychological evaluation of pregnant and lying-in women based on face recognition, which is applied to the system for psychological evaluation of pregnant and lying-in women based on face recognition as described in any one of the above. Among them, the method includes:
[0023] Obtain the analog video signal of the pregnant and lying-in women to be evaluated, receive the preset image quality standard, detect the analog video signal based on the image quality evaluation model, and determine whether the analog video signal meets the preset image quality standard;
[0024] Preprocess the analog video signal that meets the preset image quality standard to obtain basic image data;
[0025] Perform face recognition and key feature point extraction on the basic image data by using face recognition technology and face feature point detection algorithm to obtain the key feature points of the face, and annotate the key feature points in the basic image data to obtain facial image data;
[0026] The facial jitter change parameter extraction model based on deep learning extracts parameters from the facial image data to obtain the facial jitter change parameters, generates psychophysiological characteristic parameters according to the facial jitter change parameters, and conducts a psychological assessment of the pregnant and lying-in women to be evaluated based on the psychophysiological characteristic parameters.
[0027] In a possible design, the facial jitter change parameter extraction model based on deep learning is used to extract the facial image data to obtain the facial jitter change parameters, including: extracting spatial feature data from the facial image data based on a convolutional neural network model; extracting the motion features of the facial muscles in the facial image data according to the spatial feature data by using a convolutional neural network model to obtain the facial jitter change parameters.
[0028] Beneficial effects:
[0029] (1) The present invention discloses a psychological assessment system and method for pregnant and lying-in women based on facial recognition, which uses facial recognition technology and facial feature point detection algorithms to extract key feature points of the face, and at the same time uses a facial jitter change parameter model based on deep learning to extract images to obtain the facial jitter change parameters, generates psychophysiological characteristic parameters according to the facial jitter change parameters, and conducts a psychological assessment of the pregnant and lying-in women according to the psychophysiological characteristic parameters, improving the accuracy of the auxiliary assessment, thereby being able to provide accurate auxiliary decision-making basis for the assessment users, providing a brand-new and efficient psychological assessment and monitoring solution for pregnant and lying-in women, effectively meeting the needs of psychological assessment during the special periods of pregnancy and postpartum, and improving the work efficiency of the assessment users.
[0030] (2) The present invention only needs to obtain the analog video signal of the pregnant and lying-in women to be evaluated collected by the acquisition device, detect the basic image data through the image quality assessment model, extract and label the key feature points therein by using facial recognition technology and facial feature point detection algorithms, and use the facial jitter change parameter extraction model based on deep learning to obtain the facial jitter change parameters, generate psychophysiological characteristic parameters according to the facial jitter change parameters, and conduct a psychological assessment of the pregnant and lying-in women to be evaluated based on the psychophysiological characteristic parameters, which can conduct a psychological assessment in real time, complete daily monitoring, reduce the interference of human factors during the assessment process, provide an objective assessment method, and is simple and convenient to operate without the limitation of the site, facilitating application and promotion. Brief Description of the Drawings
[0031] Figure 1 It is a module block diagram of a psychological assessment system for pregnant and lying-in women based on facial recognition provided by an embodiment of the present invention;
[0032] Figure 2The flowchart of a maternal psychological assessment system based on facial recognition provided by an embodiment of the present invention. Detailed implementation manners
[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in combination with the accompanying drawings and the descriptions of the embodiments or the prior art. Obviously, the following descriptions of the structures of the accompanying drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts. It should be noted here that the descriptions of these embodiment modes are used to help understand the present invention, but do not constitute a limitation to the present invention.
[0034] The present invention relates to facial recognition technology. Here, it is solemnly promised that in the process of collecting, processing, and using facial recognition data, the requirements of relevant laws and regulations are strictly followed, and clear authorization has been obtained from all facial recognition data objects to ensure the legality and compliance of data collection behaviors. At the same time, the security and privacy of the data are continuously guaranteed, and the relevant data is only used within the authorized scope, and unauthorized abuse or disclosure is never carried out.
[0035] Embodiment:
[0036] As Figure 1 shown, this embodiment provides a maternal psychological assessment system based on facial recognition, including: an acquisition and detection module, configured to acquire a simulated video signal of a pregnant and lying-in woman to be evaluated, receive a preset image quality standard, detect the simulated video signal based on an image quality evaluation model, determine whether the simulated video signal meets the preset image quality standard, and upload the simulated video signal that meets the preset image quality standard to a data preprocessing module; a data preprocessing module, communicatively connected to the acquisition and detection module, configured to receive the simulated video signal that meets the preset image quality standard, and preprocess the simulated video signal that meets the preset image quality standard to obtain basic image data, so as to send the basic image data to a feature annotation module after obtaining the basic image data; a feature annotation module, communicatively connected to the data preprocessing module, configured to receive the basic image data, perform facial recognition and key feature point extraction on the basic image data based on facial recognition technology and a facial feature point detection algorithm, obtain the key feature points of the face, and label the key feature points in the basic image data to obtain facial image data; a parameter extraction module, communicatively connected to the feature annotation module, configured to receive the facial image data, extract parameters from the facial image data based on a facial jitter change parameter extraction model of deep learning, obtain facial jitter change parameters, and generate psychophysiological characteristic parameters according to the facial jitter change parameters, so as to perform a psychological assessment on the pregnant and lying-in woman to be evaluated according to the psychophysiological characteristic parameters.
[0037] In a possible design, it further includes a collection device, which is communicatively connected to the acquisition and detection module and is used to collect the simulated video signal of the pregnant woman to be evaluated. Among them, the collection device includes, but is not limited to, a camera with a resolution greater than or equal to the preset resolution.
[0038] Specifically, when using the collection device to collect the simulated video signal of the pregnant woman to be evaluated, the system will give shooting prompts, such as reminders of shooting angle, distance, and remaining shooting duration, to ensure sufficient shooting duration to ensure that sufficient and effective simulated video signals are collected.
[0039] Furthermore, the acquisition and detection module receives the preset image quality standard of the evaluation user, and uses the image quality evaluation model to detect the simulated video signal. The preset image quality standard includes, but is not limited to, resolution, lighting conditions, and focus degree. When the quality of the simulated video signal reaches the preset image quality standard, the simulated video signal is uploaded to the data preprocessing module and the hospital cloud server.
[0040] When the simulated video signal is uploaded to the hospital cloud server, network slicing technology is used to isolate public network data from medical data, ensuring the performance guarantee of the SLA (Service-Level Agreement) transmitted to the hospital cloud server.
[0041] In a possible design, it further includes a data encryption module, which is communicatively connected to the acquisition and detection module and the data preprocessing module, and is used to encrypt the simulated video signal that reaches the preset image quality standard based on a dynamic encryption algorithm, and transmit the encrypted simulated video signal to the data preprocessing module, so as to improve the security of the transmission of the simulated video signal and prevent unauthorized access, theft, tampering, or leakage of data. As an optional implementation method, preferably, the steps of encrypting the facial video data that reaches the preset image quality standard based on the dynamic encryption algorithm include: when the data starts to be uploaded, an automatic timestamp is generated, where the timestamp is the start time of data upload; a random number generator is used to generate a random number as the key; according to the timestamp and the random key, the facial video data that reaches the preset image quality standard is encrypted using an asymmetric encryption algorithm.
[0042] Specifically, this system uses a key storage server and sets strict access permissions. Only authorized personnel can access it after multi-factor authentication, which includes but is not limited to password authentication and dynamic verification code authentication. Regular security checks and vulnerability repairs are performed on the key storage server to prevent illegal intrusion to obtain keys. Compared with the symmetric key encryption method, the asymmetric encryption algorithm has higher security. Different keys are used for encryption and decryption, and one key cannot be derived from the other, which can ensure that data is not stolen or tampered with during transmission. A digital signature can be created during the encryption process to ensure the integrity and authenticity of the data.
[0043] In a possible design, the data preprocessing module receives the encrypted analog video signal uploaded by the acquisition detection module, decrypts it using the key in the key storage server, performs A / D conversion (Analog-to-Digital Conversion) on the decrypted analog digital signal to convert the analog video signal into a digital signal, and processes the digital signal to obtain basic image data. Among them, the basic image data can be continuous frame image data, which is a prior art and its principle will not be elaborated here.
[0044] In a possible design, after obtaining the basic image data, the data preprocessing module is further configured to perform denoising processing on the basic image data to eliminate the influence of interference factors on the processing of the basic image data after denoising.
[0045] After obtaining the basic image data, the data preprocessing module is further configured to perform image enhancement operations on the basic image data to enhance the contrast and clarity of facial features, so that after the image enhancement operation on the basic image data, feature extraction is performed on the basic image data using face recognition technology and facial feature point detection algorithms.
[0046] After obtaining the basic image data, the data preprocessing module is further configured to perform cropping and normalization processing on the basic image data to eliminate the differences between different basic image data after cropping and normalization processing.
[0047] Specifically, the denoising process includes, but is not limited to, removing the shadows caused by light and the noise generated by the acquisition device to improve the clarity and quality of the basic image data; performing image enhancement operations on the basic image data to enhance the contrast and clarity of facial features, making the key facial feature points more prominent, such as eyes, nose, and mouth. At the same time, a loss function aligned with the human face is introduced to measure the difference between the prediction result and the basic image data, and face tracking technology is added to number the human faces in each frame of image data; cropping is to adjust the basic image data to a unified size and perform normalization processing to normalize the pixel values of the basic image data to a certain range, usually between 0 and 1.
[0048] In a possible design, the parameter extraction module is also used to extract spatial features from the facial image data based on a convolutional neural network model, and set convolutional kernels of specific sizes and shapes according to the characteristics of facial feature changes during pregnancy. According to the spatial feature data, the convolutional kernels are used to capture the motion features of facial muscles in the facial image data to obtain facial jitter change parameters.
[0049] Input the facial jitter data samples labeled with different psychophysiological states into the convolutional neural network model, and at the same time use the backpropagation algorithm to update the network parameters in the convolutional neural network model; when the parameter extraction module extracts parameters from the facial image data based on the deep learning-based facial jitter change parameter extraction model, the facial jitter change parameter extraction model combines a bidirectional recurrent neural network and a temporal attention mechanism to utilize the facial jitter change parameters in the front and back time periods of the facial image data to accurately capture the dynamic changes of facial jitter.
[0050] Specifically, the spatial features of the facial image data include facial contours and facial features. Since facial swelling may occur during pregnancy, convolutional kernels more adaptable to facial changes are designed in the convolutional neural network model to accurately capture the motion features of facial muscles; the bidirectional recurrent neural network can make full use of the facial jitter change parameters in the front and back time periods of the facial image data sequence to accurately capture the dynamic changes of facial jitter, and the temporal attention mechanism focuses on the facial features in the time periods closely related to the physiological and psychological changes during pregnancy, including but not limited to the period of frequent fetal movements or the stage of prenatal anxiety.
[0051] Furthermore, a residual network structure is introduced into the hidden layer of the recurrent time network model to solve the problem of gradient disappearance caused by network deepening and better learn the change patterns of facial jitter features at different stages of pregnancy and childbirth. Each residual block is optimized according to the physiological change rules during pregnancy. For example, compared with the early pregnancy, the increased physical burden in the late pregnancy may lead to changes in the amplitude and frequency of facial jitter.
[0052] In a possible design, it also includes a correction module and a communication connection parameter extraction module, which are used to establish a dynamic correction model library and optimize and correct the psychophysiological characteristic parameters according to the dynamic correction models in the dynamic correction model library.
[0053] Specifically, pregnant women can be divided into adolescent pregnant women, pregnant women of appropriate age, elderly pregnant women and extremely elderly pregnant women according to their age. Because pregnant women of different age groups have different physiological functions and psychological response patterns, corresponding dynamic correction models are established for them respectively, and the parameters of the psychological and physiological characteristics of each age group are dynamically adjusted and corrected. The dynamic correction model compares the parameters input by the evaluation user with the data generated in real time, and continuously optimizes the correction parameters through the difference between the two to ensure that the generated psychological and physiological characteristics are more in line with the actual individual situation.
[0054] In a possible design, it also includes a data storage module, a communication connection acquisition detection module and a feature labeling module, which are used to store analog video signals and facial image data of the pregnant woman to be evaluated. An optimized data storage algorithm is introduced, and a large amount of data can be stored quickly and safely using a large-capacity, high-speed storage medium.
[0055] Furthermore, it also includes a data interface module, which adopts communication protocols and data transmission technology to realize data exchange between the system and external devices or networks, ensure seamless transmission and efficient sharing of data, can be easily integrated with different types of external devices and network environments, has good compatibility and scalability, and enhances the flexibility and application scope of the system.
[0056] In summary, this embodiment discloses a psychological assessment system for pregnant and lying-in women based on facial recognition. The key feature points of the face are extracted by using facial recognition technology and facial feature point detection algorithms. At the same time, an image is extracted by using a facial jitter change parameter model based on deep learning to obtain the facial jitter change parameters. Psychological and physiological characteristic parameters are generated according to the facial jitter change parameters, and the psychological assessment of pregnant and lying-in women is carried out according to the psychological and physiological characteristic parameters, improving the accuracy of auxiliary assessment. Thus, an accurate auxiliary decision-making basis can be provided for the assessment user. By adopting a non-invasive video acquisition method, the pregnant and lying-in women can complete the psychological assessment by facing the camera in a natural state, reducing their burden. A new and efficient psychological assessment and monitoring solution is provided for pregnant and lying-in women, effectively meeting the needs of psychological assessment during the special periods of pregnancy and postpartum, improving the work efficiency of the assessment user. Only the simulated video signal of the pregnant and lying-in woman to be assessed collected by the acquisition device needs to be obtained, the basic image data is detected by the image quality assessment model, the key feature points are extracted and marked by using facial recognition technology and facial feature point detection algorithms, the facial jitter change parameters are obtained by using a facial jitter change parameter extraction model based on deep learning, psychological and physiological characteristic parameters are generated according to the facial jitter change parameters, and the psychological assessment of the pregnant and lying-in woman to be assessed is carried out based on the psychological and physiological characteristic parameters. The psychological assessment can be carried out in real time to complete the daily monitoring, reducing the interference of human factors during the assessment process, providing an objective assessment method, and being simple and convenient to operate without the limitation of the venue, facilitating application and promotion.
[0057] As Figure 2 shown, this embodiment provides a psychological assessment method for pregnant and lying-in women based on facial recognition, which is applied to the psychological assessment system for pregnant and lying-in women based on facial recognition in the above embodiment. Among them, the method may but is not limited to include the following steps:
[0058] S1. Obtain the simulated video signal of the pregnant and lying-in woman to be assessed, receive the preset image quality standard, detect the simulated video signal based on the image quality assessment model, and judge whether the simulated video signal meets the preset image quality standard;
[0059] S2. Preprocess the simulated video signal that meets the preset image quality standard to obtain the basic image data;
[0060] S3. Perform facial recognition and key feature point extraction on the basic image data by using facial recognition technology and facial feature point detection algorithms to obtain the key feature points of the face, and mark the key feature points in the basic image data to obtain the facial image data;
[0061] S4. Extract parameters from the facial image data using a facial jitter change parameter extraction model based on deep learning to obtain facial jitter change parameters, generate psychophysiological characteristic parameters according to the facial jitter change parameters, and perform a psychological assessment on the pregnant and lying-in women to be evaluated based on the psychophysiological characteristic parameters.
[0062] In a possible design, use a facial jitter change parameter extraction model based on deep learning to extract the facial image data to obtain facial jitter change parameters, including: extracting spatial feature data from the facial image data based on a convolutional neural network model; extracting the motion features of facial muscles in the facial image data using a convolutional neural network model according to the spatial feature data to obtain facial jitter change parameters.
[0063] In summary, a method for psychological assessment of pregnant and lying-in women based on facial recognition provided in this embodiment includes: obtaining a simulated video signal of the pregnant and lying-in women to be evaluated, receiving a preset image quality standard, detecting the simulated video signal based on an image quality assessment model to determine whether the simulated video signal meets the preset image quality standard; preprocessing the simulated video signal that meets the preset image quality standard to obtain basic image data; performing facial recognition and key feature point extraction on the basic image data using a facial recognition technology and a facial feature point detection algorithm to obtain key feature points of the face, and marking the key feature points in the basic image data to obtain facial image data; extracting parameters from the facial image data using a facial jitter change parameter extraction model based on deep learning to obtain facial jitter change parameters, generating psychophysiological characteristic parameters according to the facial jitter change parameters, and performing a psychological assessment on the pregnant and lying-in women to be evaluated based on the psychophysiological characteristic parameters. This improves the accuracy of auxiliary assessment, thereby being able to provide accurate auxiliary decision-making basis for the evaluating user, providing a brand-new and efficient psychological assessment and monitoring solution for pregnant and lying-in women, effectively meeting the needs of psychological assessment during the special periods of pregnancy and postpartum, improving the work efficiency of the evaluating user, and being able to perform psychological assessment in real time to complete daily monitoring, reducing the interference of human factors during the assessment process, providing an objective assessment method, and being simple and convenient to operate without the limitation of the venue, facilitating application and promotion.
[0064] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A maternal psychological assessment system based on facial recognition, characterized in that Including: An acquisition and detection module, configured to acquire a simulated video signal of a pregnant woman to be evaluated, receive a preset image quality standard, detect the simulated video signal based on an image quality evaluation model, and determine whether the simulated video signal meets the preset image quality standard; A data preprocessing module, communicatively connected to the acquisition and detection module, configured to receive the simulated video signal that meets the preset image quality standard, preprocess the simulated video signal that meets the preset image quality standard to obtain basic image data, and send the basic image data to a feature annotation module; A feature annotation module, communicatively connected to the data preprocessing module, configured to receive the basic image data, perform face recognition and key feature point extraction on the basic image data based on face recognition technology and a face feature point detection algorithm, obtain key feature points of the face, and annotate the key feature points in the basic image data to obtain face image data; A parameter extraction module, communicatively connected to the feature annotation module, configured to receive the face image data, extract parameters from the face image data based on a face jitter change parameter extraction model of deep learning to obtain face jitter change parameters, and generate psychophysiological characteristic parameters according to the face jitter change parameters, so as to perform a psychological evaluation on the pregnant woman to be evaluated according to the psychophysiological characteristic parameters.
2. The maternal psychological assessment system based on facial recognition according to claim 1, characterized in that, The parameter extraction module is further configured to perform spatial feature extraction on the face image data based on a convolutional neural network model, input face jitter data samples labeled under different psychophysiological states into the convolutional neural network model, and update network parameters in the convolutional neural network model by using a backpropagation algorithm; The face jitter change parameter extraction model adopts a bidirectional recurrent neural network and a temporal attention mechanism, so as to capture the dynamic changes of face jitter by using face jitter change parameters in the front and back time periods of the face image data.
3. The maternal psychological assessment system based on facial recognition according to claim 1, characterized in that, Before preprocessing the simulated video signal that meets the preset image quality standard, the data preprocessing module is further configured to perform A / D conversion on the simulated video signal that meets the preset image quality standard to convert the simulated video signal into a digital signal, process the digital signal to obtain basic image data, so as to perform recognition and annotation on the basic image data.
4. The maternal psychological assessment system based on facial recognition according to claim 1, wherein After obtaining the basic image data, the data preprocessing module is further configured to perform denoising processing on the basic image data; After obtaining the basic image data, the data preprocessing module is further configured to perform an image enhancement operation on the basic image data to enhance the contrast and clarity of facial features, so that after performing the image enhancement operation on the basic image data, feature extraction is performed on the basic image data by using face recognition technology and a face feature point detection algorithm; After obtaining the basic image data, the data preprocessing module is further configured to perform cropping and normalization processing on the basic image data.
5. The maternal psychological assessment system based on facial recognition according to claim 1, characterized in that It further includes a data encryption module, which is communicatively connected to the acquisition and detection module and the data preprocessing module, and is configured to encrypt the simulated video signal that meets the preset image quality standard based on a dynamic encryption algorithm, and transmit the encrypted simulated video signal to the data preprocessing module.
6. The maternal psychological assessment system based on facial recognition according to claim 1, characterized in that, It further includes a calibration module, which is communicatively connected to the parameter extraction module and is used to establish a dynamic calibration model library and optimize and calibrate the psychophysiological characteristic parameters according to the dynamic calibration models in the dynamic calibration model library.
7. The maternal psychological assessment system based on facial recognition according to claim 1, characterized in that It further includes a data storage module, which is communicatively connected to the acquisition and detection module and the feature annotation module and is used to store the simulated video signals and facial image data of the pregnant and lying-in women to be evaluated.
8. A maternal psychological assessment system based on facial recognition according to claim 1, characterized in that, It further includes an acquisition device, which is communicatively connected to the acquisition and detection module and is used to acquire the simulated video signals of the pregnant and lying-in women to be evaluated. The acquisition device includes a camera with a resolution greater than or equal to a preset resolution.
9. A method for psychological assessment of pregnant and lying-in women based on facial recognition, characterized in that, It is applied to the facial recognition-based psychological assessment system for pregnant and lying-in women according to any one of claims 1 to 8, wherein the method includes: Acquire the simulated video signals of the pregnant and lying-in women to be evaluated, receive a preset image quality standard, detect the simulated video signals based on an image quality assessment model, and determine whether the simulated video signals meet the preset image quality standard; Preprocess the simulated video signals that meet the preset image quality standard to obtain basic image data; Use a face recognition technology and a facial feature point detection algorithm to perform face recognition and key feature point extraction on the basic image data to obtain the key feature points of the face, and annotate the key feature points in the basic image data to obtain facial image data; Extract parameters from the facial image data based on a deep learning-based facial jitter change parameter extraction model to obtain facial jitter change parameters, generate psychophysiological characteristic parameters according to the facial jitter change parameters, and perform a psychological assessment on the pregnant and lying-in women to be evaluated based on the psychophysiological characteristic parameters.
10. The method for psychological assessment of pregnant and lying-in women based on facial recognition according to claim 9, characterized in that, Extract parameters from the facial image data using a deep learning-based facial jitter change parameter extraction model to obtain facial jitter change parameters, including: perform spatial feature extraction on the facial image data based on a convolutional neural network model to obtain spatial feature data; use the convolutional neural network model to extract the motion features of the facial muscles in the facial image data according to the spatial feature data to obtain facial jitter change parameters.