A method and system for facial skin data analysis

Abstract

This application relates to a facial skin data analysis method and system, belonging to the field of skin data analysis technology. The analysis method includes: collecting current climate data and the user's current skin data; obtaining the user's historical skin data before a preset time period based on a skin database; inputting the current skin data, current climate data, and the historical skin data before the preset time period into a pre-trained skin data prediction model to obtain predicted skin data after the preset time period; and analyzing the predicted skin data to obtain the user's skin change analysis results. This application can help users objectively and clearly understand their own skin changes, thereby facilitating the provision of long-term, reasonable, and personalized skincare advice.

Description

Technical Field

[0001] This application relates to the field of skin data analysis technology, and in particular to a method and system for facial skin data analysis. Background Technology

[0002] With the improvement of people's living standards and the rapid development of medical aesthetics, people are paying more and more attention to the health of their skin. As skin quality is a mirror of the body's physiological health, people are also paying more and more attention to the beauty and care of facial skin, and hope to be able to conveniently and quickly evaluate their skin condition.

[0003] Currently, skin condition is a comprehensive reflection of multiple data points, such as skin moisture content and oil content. Over time, these skin indicators also change, such as a decrease in oil content and a decrease in skin moisture. When caring for their skin, users often judge recent changes based on personal feelings or visual observation. However, this method is often subjective and cannot help users objectively and clearly understand their own skin changes, nor can it provide corresponding long-term and reasonable skincare advice. Summary of the Invention

[0004] To help users objectively and clearly understand changes in their own skin and to provide corresponding long-term and reasonable skincare advice, this application provides a facial skin data analysis method and system.

[0005] In a first aspect, this application provides a facial skin data analysis method, employing the following technical solution:

[0006] A facial skin data analysis method, the analysis method comprising,

[0007] Collect current climate data and the user's current skin data;

[0008] Based on a skin database, retrieve the user's historical skin data from a preset time period.

[0009] The current skin data, current climate data, and historical skin data before a preset time period are input into a pre-trained skin data prediction model to obtain predicted skin data after the preset time period.

[0010] Based on the predicted skin data, the analysis results of the user's skin changes are obtained.

[0011] By adopting the above technical solution, combining current climate data, current user skin data, and historical skin data, and analyzing skin condition trends based on time-series data, the algorithm predicts the user's skin condition after a preset time period, and then calculates the skin change analysis results. This provides users with a reference for tracking skin condition changes over a future period, helping them better understand and improve their skin condition. It also facilitates providing users with personalized skincare advice, including suitable skincare products, skincare methods, and lifestyle adjustment suggestions.

[0012] Optionally, the step of obtaining the user's historical skin data before a preset time period based on a skin database includes:

[0013] Obtain user identity information and verify the user identity information;

[0014] If the verification is successful, the preset time period and target data features input by the user are obtained;

[0015] Based on the preset time period and target feature data, the target data features of the user before the preset time period are obtained from the skin database as historical skin data.

[0016] By adopting the above technical solution, user identity is identified and matched with the user's historical skin data in the skin database. Furthermore, users can set preset time periods and target data features according to their needs, so as to facilitate users to select the data features they are more concerned about based on their own skin type, in order to meet their needs and analyze the user's specific skin problems.

[0017] Optionally, the analysis method further includes a step of training the skin data prediction model, the step including;

[0018] Obtain a sample dataset; wherein the sample dataset includes the user's historical skin data and historical climate data;

[0019] In chronological order, data pairs are constructed based on historical skin data and historical climate data corresponding to multiple preset time periods, and a data combination sequence is established based on the data pairs;

[0020] Multiple training samples are obtained by gradually sliding a sliding time window from the beginning of the data combination sequence to capture data pairs in the data combination sequence in turn; wherein the size of the sliding time window is at least two preset time periods, the step size of each slide is the preset time period, and each training sample includes training input features and training output features. The training input features include the historical skin data sequence within the window, the current skin data sequence, and the corresponding historical climate data sequence. The training output features include the skin data after the next preset time period.

[0021] The multiple training samples are combined into a training sample set, and the training sample set is preprocessed.

[0022] The preprocessed training sample set is divided into a training set and a test set.

[0023] The training set is input into a pre-built neural network model for training to obtain a skin data prediction model. The skin data prediction model is then tested based on the test set, and the parameters of the skin data prediction model are corrected to obtain the pre-trained skin data prediction model.

[0024] By employing the above technical solutions, combining historical skin data helps the model understand the changing trends of users' skin conditions, while combining historical climate data helps the model understand the impact of external factors on skin conditions. By using a progressively sliding window, local temporal features can be captured, ensuring that the temporal features of skin and climate data are preserved within a fixed-size time window. This helps the model grasp the patterns of skin condition changes over time, generating a large number of training samples, which is beneficial for the model to learn complex patterns. Furthermore, by flexibly adjusting the prediction duration, the prediction duration of the output target can be adjusted according to needs, enabling predictions of skin conditions for different time ranges in the future. Simultaneously, a skin data prediction model is built and trained using a training set composed of processed training samples, tested using a test set, and optimized and corrected by the model parameters to obtain the optimal skin data prediction model, thereby improving the accuracy of skin data prediction.

[0025] Optionally, the method for testing the skin data prediction model based on the test set includes:

[0026] The test set is input into the skin data prediction model to test the predictive ability of the skin data prediction model; wherein the test indicators include the coefficient of determination, the root mean square error of cross-validation, and the root mean square error of prediction.

[0027] By adopting the above technical solution, the model is tested using test set samples that did not participate in the modeling. The prediction results are statistically compared with conventional measurement values. The performance and stability of the skin data prediction model are evaluated by combining multiple evaluation indicators. This allows the actual predictive ability of the prediction model to be determined, which facilitates parameter correction of the prediction model and further improves the accuracy of the prediction results.

[0028] Optionally, the step of analyzing the predicted skin data to obtain the skin change analysis results includes:

[0029] Extract the current feature value of each target data feature in the current skin data and the predicted feature value of each target data feature in the predicted skin data respectively;

[0030] Calculate the feature difference of each target data feature based on the current feature value and the predicted feature value of each target data feature;

[0031] The skin change analysis value is calculated by weighting the pre-defined weights and feature differences of each target data feature and using the weighted average.

[0032] By adopting the above technical solution, the feature differences of various target data are weighted and averaged to represent the overall degree of skin change. Then, quantitative indicators are used as a reference for users to track the trend of skin condition changes, so as to better understand and improve skin condition.

[0033] Optionally, after extracting the current feature values ​​of each target data feature in the current skin data and the predicted feature values ​​of each target data feature in the predicted skin data, the method further includes:

[0034] Based on the current and predicted feature values ​​of each target data feature, a visualization chart is generated and sent to the user terminal.

[0035] By adopting the above technical solution, in order to improve the user experience, the predicted changes in skin data are displayed to users in the form of visual charts, so as to help users clearly and intuitively understand the changing trends of skin data and then make corresponding skin care strategies.

[0036] Secondly, this application provides a facial skin data analysis system, which adopts the following technical solution:

[0037] A facial skin data analysis system, the analysis system comprising:

[0038] The data acquisition module is used to collect current climate data and the user's current skin data;

[0039] The historical skin data acquisition module is used to acquire the user's historical skin data before a preset time period based on the skin database;

[0040] The predicted skin data generation module is used to input the current skin data, current climate data, and historical skin data before a preset time period into a pre-trained skin data prediction model to obtain predicted skin data after the preset time period.

[0041] The analysis module is used to analyze the predicted skin data to obtain the skin change analysis results of the user.

[0042] By adopting the above technical solution, combining current climate data, current user skin data, and historical skin data, and analyzing skin condition trends based on time-series data, the algorithm predicts the user's skin condition after a preset time period, and then calculates the skin change analysis results. This provides users with a reference for tracking skin condition changes over a future period, helping them better understand and improve their skin condition. It also facilitates providing users with personalized skincare advice, including suitable skincare products, skincare methods, and lifestyle adjustment suggestions.

[0043] Optionally, the analysis system further includes:

[0044] The visualization chart generation module is used to generate visualization charts based on the current and predicted feature values ​​of various target data features and send them to the user terminal.

[0045] By adopting the above technical solution, in order to improve the user experience, the predicted changes in skin data are displayed to users in the form of visual charts, so as to help users clearly and intuitively understand the changing trends of skin data and then make corresponding skin care strategies.

[0046] Thirdly, this application provides a computer device, which adopts the following technical solution:

[0047] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in the first aspect.

[0048] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution:

[0049] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as in any of the methods in the first aspect.

[0050] In summary, this application includes at least one of the following beneficial technical effects: combining current climate data, current user skin data, and historical skin data, analyzing skin condition trends based on time-series data, predicting the user's skin condition after a preset time period through an algorithm, and then calculating the skin change analysis results, providing users with a reference for tracking skin condition changes over a future period of time, thereby helping users better understand and improve their skin condition, and also facilitating the provision of personalized skincare advice, including suitable skincare products, skincare methods, and lifestyle adjustment suggestions. Attached Figure Description

[0051] Figure 1 This is a flowchart illustrating a facial skin data analysis method according to one embodiment of this application.

[0052] Figure 2 This is a flowchart illustrating a facial skin data analysis method according to one embodiment of this application.

[0053] Figure 3 This is a flowchart illustrating a facial skin data analysis method according to one embodiment of this application.

[0054] Figure 4 This is a structural block diagram of a facial skin data analysis system according to one embodiment of this application.

[0055] Figure labeling: 101, Data acquisition module; 102, Historical skin data acquisition module; 103, Predictive skin data generation module; 104, Analysis module. Detailed Implementation

[0056] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figure 1-4 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0057] This application discloses a method for facial skin data analysis.

[0058] Reference Figure 1 A facial skin data analysis method, the analysis method includes,

[0059] Step S101: Collect current climate data and user's current skin data;

[0060] The current climate data includes, but is not limited to, environmental factors that may affect the skin, such as temperature and humidity, which can be collected by corresponding sensors; the current skin data includes, but is not limited to, skin moisture content, skin oil content, skin elasticity, etc., which can be collected by skin testing instruments such as bioelectrical impedance analyzers, conductivity meters, sebum meters, and skin elasticity testers.

[0061] Step S102: Based on the skin database, obtain the user's historical skin data before a preset time period;

[0062] The preset time period is the preset time interval required by the user to predict skin data. For example, if it is set to 7 days, then the target data features from 7 days ago need to be obtained as historical skin data in order to predict skin data 7 days later.

[0063] Step S103: Input the current skin data, current climate data and historical skin data before the preset time period into the pre-trained skin data prediction model to obtain the predicted skin data after the preset time period.

[0064] The current skin data, historical skin data, and predicted skin data all include the same category of skin data features, which can be selected by the user according to their own needs.

[0065] Step S104: Based on the predicted skin data, analyze the user's skin change analysis results.

[0066] In the above implementation, by combining current climate data, current user skin data, and historical skin data, and based on time-series data skin condition trend analysis, the algorithm predicts the user's skin condition after a preset time period, and then calculates the skin change analysis results. This serves as a reference for users to track the trend of skin condition changes in the future, helping users to better understand and improve their skin condition. It also facilitates providing users with personalized skin care advice, including suitable skin care products, skin care methods, and lifestyle adjustment suggestions.

[0067] As one implementation of step S102, the step of obtaining the user's historical skin data before a preset time period based on the skin database includes:

[0068] Obtain user identity information and verify the user identity information;

[0069] User identity information includes, but is not limited to, fingerprint information, facial information, iris information, user ID, etc.; user identity information can be verified based on a pre-set identity information database to match corresponding user information data, thereby ensuring the security and accuracy of information matching;

[0070] If the verification passes, the system retrieves the preset time period and target data features entered by the user; if the verification fails, it displays an error message indicating an identity verification failure.

[0071] Among them, the target data features are the data features that users need to pay attention to regarding skin data, such as skin moisture content, skin oil content, skin elasticity, etc.

[0072] Based on a preset time period and target feature data, the target data features of the user before the preset time period are obtained from the skin database as historical skin data.

[0073] The preset time period is the preset time interval required by the user to predict skin data. For example, if it is set to 7 days, then the target data features from 7 days ago need to be obtained as historical skin data in order to predict skin data 7 days later.

[0074] In the above implementation, user identification is performed to match the user's historical skin data in the skin database. Furthermore, users can set preset time periods and target data features according to their needs, thereby allowing users to select data features that they are more concerned about based on their own skin type, in order to meet their needs and analyze the user's specific skin problems.

[0075] Reference Figure 2 As a further implementation of the analysis method, the analysis method also includes a step of training a skin data prediction model, the training step including;

[0076] Step S201: Obtain a sample dataset; wherein, the sample dataset includes the user's historical skin data and historical climate data;

[0077] Step S202: Construct data pairs based on historical skin data and historical climate data corresponding to multiple preset time periods in chronological order, and establish a data combination sequence based on the data pairs;

[0078] Each data pair in the data combination sequence corresponds to a data pair in a preset time period, with the structure being (skin data, historical data).

[0079] Step S203: Using a sliding time window, gradually slide from the beginning of the data combination sequence and capture the data pairs in the data combination sequence in turn to obtain multiple training samples; wherein, the size of the sliding time window is at least two preset time periods, the step size of each slide is a preset time period, and each training sample includes training input features and training output features. The training input features include the historical skin data sequence within the window, the current skin data sequence, and the corresponding historical climate data sequence. The training output features include the skin data after the next preset time period.

[0080] Understandably, by sliding the time window forward step by step, an appropriate number of preset time periods can be captured, and a series of continuous data blocks can be generated, each corresponding to an independent training sample. The size of the sliding time window can be set according to the length of the preset time period. For example, if the preset time period is set to 7 days, the size of the sliding time window can be selected as 2 or more preset time periods. For example, if it is selected as 14 days, the size of the sliding time window is 2 preset time periods.

[0081] Each training sample includes training input features and training output features. The training input features include: the historical skin data sequence within the window (i.e., historical skin data arranged in chronological order at each preset time interval), the current skin data sequence (i.e., the skin data at the end position of the sliding window), and the corresponding historical climate data sequence (containing climate data for all preset time intervals within the window). The training output features are the skin data after the next preset time interval.

[0082] For example, if the preset time period is set to 7 days and the sliding time window is 2 preset time periods, i.e., 14 days, then the training input features include: historical skin data arranged in chronological order every 7 days, skin data at the end of the sliding window, and climate data for all time periods of the window. The training output features are the skin data after the next 7 days.

[0083] Step S204: Combine multiple training samples into a training sample set and perform data preprocessing on the training sample set.

[0084] In order to eliminate the scale differences of various data, data preprocessing is required to improve the convergence speed of the model and improve the predictive performance of the model; data preprocessing includes, but is not limited to, normalization or standardization.

[0085] Step S205: Divide the preprocessed training sample set into a training set and a test set;

[0086] The model is divided into a training set and a test set for training and evaluation.

[0087] Step S206: Input the training set into the pre-built neural network model for training to obtain the skin data prediction model;

[0088] The neural network model can be a recurrent neural network (RNN) or a long short-term memory network (LSTM). These models are good at processing time series data and can capture time dependencies and long-distance dependencies. When building the model, certain parameters such as the number of network layers, the number of neurons, the batch size, the learning rate, and the optimizer can be set. In addition, during the parameter tuning process, different parameter combinations can be tried to follow the trade-off between network performance and model complexity. The model parameters can be optimized by iteratively optimizing the model multiple times, setting a loss function to measure the difference between the predicted value and the true value, and using the optimizer to update the parameters to minimize the loss function.

[0089] Step S207: Test the skin data prediction model based on the test set and correct the parameters of the skin data prediction model to obtain the pre-trained skin data prediction model.

[0090] This involves adjusting the parameters of the skin data prediction model based on the test set, optimizing the model structure or training strategy, until the model achieves the desired performance.

[0091] In the above implementation, combining historical skin data helps the model understand the changing trends of the user's skin condition, and combining historical climate data helps the model understand the impact of external factors on skin condition. By gradually sliding the window, local temporal features can be captured, so that the temporal features of skin and climate data can be preserved within a fixed-size time window. This helps the model grasp the pattern of skin condition changes over time, generating a large number of training samples, which will be beneficial for the model to learn complex patterns. Furthermore, by flexibly adjusting the prediction duration, the prediction duration of the output target can be adjusted according to needs, enabling prediction of skin condition for different time ranges in the future. At the same time, a skin data prediction model is built and trained using a training set composed of processed training samples, tested using a test set, and the optimal skin data prediction model can be obtained by optimizing and correcting the model parameters, thereby improving the accuracy of skin data prediction.

[0092] In one embodiment of this application, the pre-built neural network model can be an LSTM neural network model; a multi-layer LSTM network can be defined, with the first layer responsible for extracting low-level features from the time series data, and subsequent layers responsible for learning the high-level representation of these features when predicting changes in skin condition. In the final output layer, a fully connected layer can be used to obtain the prediction result, and an activation function (such as ReLU) can be used to remove negative values.

[0093] As one implementation method for using the LSTM neural network model, the following steps are included before dividing the preprocessed training sample set into training and test sets:

[0094] Transform the training input features into a three-dimensional array (a, b, c), where a represents the number of training samples, b represents the number of preset time periods captured by the sliding window, and c represents the number of features within each preset time period (including skin data and climate data).

[0095] The training output features are converted into a two-dimensional array (a, d), where a represents the number of training samples and d represents the number of features of the skin data after the next preset time period.

[0096] In the above implementation, through this data format adjustment step, the training input features and training output features in the training set and test set will conform to the data format required by the LSTM neural network model, and then they can be input into the LSTM neural network model for training and testing.

[0097] As one implementation of step S207, the method for testing the skin data prediction model based on the test set includes:

[0098] The test set is input into the skin data prediction model to test the predictive ability of the skin data prediction model; the test indicators include the coefficient of determination, the root mean square error of cross-validation, and the root mean square error of prediction.

[0099] In the above implementation, the model is tested using test set samples that are not involved in modeling, and the prediction results are statistically compared with conventional measured values ​​to understand the model's generalization performance and avoid overfitting. When the model performs well on the validation set, the final performance of the model is evaluated using the test set. By combining multiple evaluation indicators to evaluate the performance and stability of the skin data prediction model, the actual predictive ability of the prediction model can be determined, which facilitates parameter correction of the prediction model and further improves the accuracy of the prediction results.

[0100] Reference Figure 3 As one implementation of step S104, the step of analyzing and obtaining skin change analysis results based on predicted skin data includes:

[0101] Step S301: Extract the current feature values ​​of each target data feature in the current skin data and the predicted feature values ​​of each target data feature in the predicted skin data respectively;

[0102] Among them, the target data features in skin data may include skin moisture content, skin oil content, skin elasticity, etc.

[0103] Step S302: Calculate the feature difference of each target data feature based on the current feature value and the predicted feature value of each target data feature;

[0104] Among them, the feature difference of the target data features is the difference between the predicted feature value and the current feature value;

[0105] Step S303: Calculate the skin change analysis value as the skin change analysis result by performing a weighted average based on the preset weights and feature differences of each target data feature.

[0106] The preset weights of each target data feature can be allocated according to the importance of different target data features or the user's attention to the target data feature. For example, changes in skin moisture content may be more important than changes in skin oil content or skin elasticity, so the weight of skin moisture content can be set to 0.5, the weight of skin oil content to 0.3, and the weight of skin elasticity to 0.2.

[0107] Understandably, since different users may have different data metrics to focus on, the preset weights of various target data features can be adjusted according to individual skin type and needs; for example, users with oily skin may pay more attention to the skin's oil content, while users with dry skin may be more sensitive to the skin's moisture content.

[0108] In one embodiment of this application, the current skin data and the predicted skin data have three characteristic values: skin moisture content, skin oil content, and skin elasticity. The skin change analysis value can be calculated using the following weighted average calculation formula:

[0109] Skin change analysis value = (weight of skin moisture content * characteristic difference of skin moisture content) + (weight of skin oil content * characteristic difference of skin oil content) + (weight of skin elasticity * characteristic difference of skin elasticity).

[0110] In the above implementation, the feature differences of each target data feature are weighted and averaged to represent the overall degree of skin change. In turn, quantitative indicators are used as a reference for users to track the trend of skin condition changes, so as to better understand and improve skin condition.

[0111] As a further implementation of the analysis method, after extracting the current feature values ​​of each target data feature in the current skin data and the predicted feature values ​​of each target data feature in the predicted skin data, the method further includes:

[0112] Based on the current and predicted feature values ​​of each target data feature, a visualization chart is generated and sent to the user terminal.

[0113] The required visualization charts can be created using visualization tools or programming libraries. The types of visualization charts can include line charts, bar charts, etc., to intuitively compare the differences between the current feature values ​​and the predicted feature values. In addition, the user terminal can be the user's smart mobile terminal such as mobile phone, tablet computer, etc.

[0114] In the above implementation, in order to improve the user experience, the predicted changes in skin data are displayed to users in the form of visual charts, so as to help users clearly and intuitively understand the changing trends of skin data and make corresponding skin care strategies.

[0115] This application also discloses a facial skin data analysis system.

[0116] Reference Figure 4 A facial skin data analysis system, the analysis system includes:

[0117] Data acquisition module 101 is used to collect current climate data and the user's current skin data;

[0118] The historical skin data acquisition module 102 is used to acquire the user's historical skin data before a preset time period based on the skin database;

[0119] The skin data prediction generation module 103 is used to input current skin data, current climate data and historical skin data before a preset time period into a pre-trained skin data prediction model to obtain predicted skin data after the preset time period.

[0120] Analysis module 104 is used to analyze the predicted skin data to obtain the user's skin change analysis results.

[0121] In the above implementation, by combining current climate data, current user skin data, and historical skin data, and based on time-series data skin condition trend analysis, the algorithm predicts the user's skin condition after a preset time period, and then calculates the skin change analysis results. This serves as a reference for users to track the trend of skin condition changes in the future, helping users to better understand and improve their skin condition. It also facilitates providing users with personalized skin care advice, including suitable skin care products, skin care methods, and lifestyle adjustment suggestions.

[0122] As a further implementation of the facial skin data analysis system, the analysis system also includes a skin data prediction model training module, which is configured as follows:

[0123] Obtain a sample dataset; the sample dataset includes users' historical skin data and historical climate data;

[0124] Data pairs are constructed based on historical skin data and historical climate data corresponding to multiple preset time periods, in chronological order, and a data combination sequence is established based on the data pairs;

[0125] A sliding time window is used to gradually slide backward from the beginning of the data combination sequence and capture data pairs in the data combination sequence in turn to obtain multiple training samples. The size of the sliding time window is at least two preset time periods, and the step size of each slide is a preset time period. Each training sample includes training input features and training output features. The training input features include the historical skin data sequence within the window, the current skin data sequence, and the corresponding historical climate data sequence. The training output features include the skin data after the next preset time period.

[0126] Multiple training samples are combined into a training sample set, and the training sample set is preprocessed.

[0127] The preprocessed training sample set is divided into a training set and a test set.

[0128] The training set is input into a pre-built neural network model for training to obtain a skin data prediction model. The skin data prediction model is then tested based on a test set, and the parameters of the skin data prediction model are corrected to obtain a pre-trained skin data prediction model.

[0129] In the above embodiments, a skin data prediction model is established and trained using a training set composed of processed training samples, tested using a test set, and the optimal skin data prediction model is obtained by optimizing and correcting the model parameters, thereby improving the accuracy of skin data prediction.

[0130] As a further implementation of the facial skin data analysis system, the analysis system also includes:

[0131] The visualization chart generation module is used to generate visualization charts based on the current and predicted feature values ​​of various target data features and send them to the user terminal.

[0132] In the above implementation, in order to improve the user experience, the predicted changes in skin data are displayed to users in the form of visual charts, so as to help users clearly and intuitively understand the changing trends of skin data and make corresponding skin care strategies.

[0133] The facial skin data analysis system of this application embodiment can implement any of the above-described facial skin data analysis methods, and the specific working process of each module in the facial skin data analysis system can be referred to the corresponding process in the above-described method embodiments.

[0134] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for example, the division of a certain module is merely a logical functional division, and in actual implementation there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0135] This application also discloses a computer device.

[0136] Computer equipment, including memory, processor, and computer program stored in memory and executable on the processor, wherein the processor executes the computer program to implement the facial skin data analysis method described above.

[0137] This application also discloses a computer-readable storage medium.

[0138] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as in any of the facial skin data analysis methods described above.

[0139] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device; the program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0140] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0141] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A method for analyzing facial skin data, characterized in that: The analytical methods include, Collect current climate data and the user's current skin data; Based on a skin database, retrieve the user's historical skin data from a preset time period. The current skin data, current climate data, and historical skin data before a preset time period are input into a pre-trained skin data prediction model to obtain predicted skin data after the preset time period. Based on the predicted skin data, the user's skin change analysis results are obtained; The analysis method further includes a step of training the skin data prediction model, the step including; Obtain a sample dataset; wherein the sample dataset includes the user's historical skin data and historical climate data; In chronological order, data pairs are constructed based on historical skin data and historical climate data corresponding to multiple preset time periods, and a data combination sequence is established based on the data pairs; Multiple training samples are obtained by gradually sliding a sliding time window from the beginning of the data combination sequence to capture data pairs in the data combination sequence in turn; wherein the size of the sliding time window is at least two preset time periods, the step size of each slide is the preset time period, and each training sample includes training input features and training output features. The training input features include the historical skin data sequence within the window, the current skin data sequence, and the corresponding historical climate data sequence. The training output features include the skin data after the next preset time period. The multiple training samples are combined into a training sample set, and the training sample set is preprocessed. The preprocessed training sample set is divided into a training set and a test set. The training set is input into a pre-built neural network model for training to obtain a skin data prediction model. The skin data prediction model is then tested based on the test set, and the parameters of the skin data prediction model are corrected to obtain the pre-trained skin data prediction model. The pre-built neural network model uses an LSTM neural network model, and before dividing the preprocessed training sample set into training and test sets, the following steps are also included: The training input features are transformed into a three-dimensional array (a, b, c), where a represents the number of training samples, b represents the number of preset time periods captured by the sliding window, and c represents the number of features in each preset time period, including skin data and climate data. The training output features are converted into a two-dimensional array (a, d), where a represents the number of training samples and d represents the number of features of the skin data after the next preset time period.

2. The facial skin data analysis method according to claim 1, characterized in that: The step of obtaining the user's historical skin data before a preset time period based on the skin database includes: Obtain user identity information and verify the user identity information; If the verification is successful, the preset time period and target data features input by the user are obtained; Based on the preset time period and target feature data, the target data features of the user before the preset time period are obtained from the skin database as historical skin data.

3. The facial skin data analysis method according to claim 1, characterized in that, The method for testing the skin data prediction model based on the test set includes: The test set is input into the skin data prediction model to test the predictive ability of the skin data prediction model; wherein the test indicators include the coefficient of determination, the root mean square error of cross-validation, and the root mean square error of prediction.

4. The facial skin data analysis method according to claim 2, characterized in that, The steps for analyzing skin change results based on the predicted skin data include: Extract the current feature value of each target data feature in the current skin data and the predicted feature value of each target data feature in the predicted skin data respectively; Calculate the feature difference of each target data feature based on the current feature value and the predicted feature value of each target data feature; The skin change analysis value is calculated by weighting the data characteristics of each target with preset weights and feature differences, and then taking the weighted average.

5. The facial skin data analysis method according to claim 4, characterized in that, After extracting the current feature values ​​of each target data feature in the current skin data and the predicted feature values ​​of each target data feature in the predicted skin data, the method further includes: Based on the current and predicted feature values ​​of each target data feature, a visualization chart is generated and sent to the user terminal.

6. A facial skin data analysis system, characterized in that, A facial skin data analysis system for performing any one of claims 1 to 5, the analysis system comprising: The data acquisition module (101) is used to collect current climate data and the user's current skin data; The historical skin data acquisition module (102) is used to acquire the user's historical skin data before a preset time period based on the skin database; The skin data prediction generation module (103) is used to input the current skin data, the current climate data and the historical skin data before the preset time period into the pre-trained skin data prediction model to obtain the predicted skin data after the preset time period. The analysis module (104) is used to analyze the predicted skin data to obtain the skin change analysis results of the user.

7. A facial skin data analysis system according to claim 6, characterized in that, The analysis system also includes: The visualization chart generation module is used to generate visualization charts based on the current and predicted feature values ​​of various target data features and send them to the user terminal.

8. A computer device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that: The computer program is stored that can be loaded by a processor and executed according to any one of claims 1 to 5.

Images