Facial mask preparation process parameter optimization method and system adopting artificial intelligence

Abstract

The invention relates to an artificial intelligence-based mask preparation process parameter optimization method and system, and relates to the field of preparation optimization, and the method comprises the steps: carrying out the light-transmitting image collection of a to-be-impregnated mask cloth, and carrying out the recognition of the thickness of the mask cloth to obtain a first thickness and a second thickness; collecting essence viscosity information to obtain an interval of impregnation process parameters, randomly generating a first impregnation parameter in the interval, and performing impregnation simulation prediction in combination with the viscosity information and the first and second thicknesses to obtain first and second impregnation amounts; performing mask use prediction according to the first and second impregnation amounts in combination with the first and second thickness and viscosity information to obtain a first and second essence dripping rate and a first and second mask cloth deformation rate, performing calculation in combination with the first and second impregnation amounts to obtain a first and second mask score, performing impregnation parameter optimization, and obtaining optimal impregnation parameters for mask impregnation preparation. The technical problems that in the mask preparation process, impregnation parameter setting depends on artificial experience, precise optimization is difficult to achieve, and mask quality and consistency are affected are solved.

Description

Technical Field

[0001] The present invention relates to the field of preparation optimization, and in particular to a method and system for optimizing process parameters of facial mask preparation using artificial intelligence. Background Art

[0002] In the mask preparation process, the impregnation process is a key step to evenly adsorb the essence on the membrane cloth. As an important process parameter in this process, the setting of the impregnation time directly affects the quality of the mask and user experience. The setting of the impregnation time mainly depends on the operator's experience and experimental adjustment. This method is not only inefficient, but also difficult to ensure the quality consistency of each batch of masks. If the impregnation time is too short, the essence cannot be fully adsorbed on the membrane cloth, resulting in uneven distribution of the essence, affecting the skin care effect and user experience of the mask. On the other hand, if the impregnation time is too long, the membrane cloth may cause excessive swelling and deformation, especially for membrane cloths made of certain special materials, such as biofiber membrane cloths. Long-term impregnation will make the membrane cloth soft and easy to deform, thereby affecting the fit and usage experience of the mask. Summary of the invention

[0003] The present invention aims to solve the technical problem that the setting of immersion parameters in the mask preparation process in the prior art relies on manual experience, which is difficult to achieve accurate optimization and affects the quality and consistency of the mask. A mask preparation process parameter optimization method and system using artificial intelligence are provided to solve the problem.

[0004] The technical solution of the present invention to solve the above technical problems is as follows: In a first aspect, the present invention provides a method for optimizing process parameters of a facial mask preparation using artificial intelligence, the method comprising: acquiring a light-transmitting image of a target membrane cloth to be impregnated, identifying the thickness of the membrane cloth, and obtaining a first thickness of a first area and a second thickness of a second area; acquiring viscosity information of an essence liquid, and obtaining an impregnation parameter interval of an impregnation process parameter; randomly generating a first impregnation parameter within the impregnation parameter interval, and performing an impregnation simulation prediction in combination with the viscosity information, the first thickness, and the second thickness, to obtain a first impregnation amount and a second impregnation amount; performing a facial mask usage prediction based on the first impregnation amount and the second impregnation amount, in combination with the first thickness, the second thickness, and the viscosity information, to obtain a first essence dripping rate, a first membrane cloth deformation rate, a second essence dripping rate, and a second membrane cloth deformation rate, and calculating a first mask score and a second mask score in combination with the first impregnation amount and the second impregnation amount, optimizing the impregnation parameters, obtaining optimal impregnation parameters, and performing facial mask impregnation preparation.

[0005] In a second aspect, the present invention provides a facial mask preparation process parameter optimization system using artificial intelligence, the system comprising: an image acquisition module, used to capture a light-transmitting image of a target membrane cloth to be impregnated, identify the membrane cloth thickness, and obtain a first thickness in a first area and a second thickness in a second area; a parameter acquisition module, used to collect viscosity information of an essence liquid, and obtain an impregnation parameter interval of an impregnation process parameter; a simulation prediction module, used to randomly generate a first impregnation parameter within the impregnation parameter interval, and perform an impregnation simulation prediction based on the viscosity information, the first thickness, and the second thickness, to obtain a first impregnation amount and a second impregnation amount; an optimization preparation module, used to perform a facial mask usage prediction based on the first impregnation amount and the second impregnation amount, combined with the first thickness, the second thickness and the viscosity information, to obtain a first essence dripping rate, a first membrane cloth deformation rate, a second essence dripping rate, and a second membrane cloth deformation rate, and to calculate a first mask score and a second mask score based on the first impregnation amount and the second impregnation amount, to optimize the impregnation parameters, to obtain optimal impregnation parameters, and to perform mask impregnation preparation.

[0006] The beneficial effects of the present invention are: by collecting the transparent image of the film cloth to identify the thickness and the viscosity information of the essence, combined with the randomly generated impregnation parameters for simulation prediction, the essence dripping rate and the film cloth deformation rate when the mask is used are predicted, and the mask score is calculated to optimize the impregnation parameters, so as to achieve precise control of the impregnation process and improve the mask quality and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A schematic flow chart of a method for optimizing process parameters for preparing a facial mask using artificial intelligence provided by the present invention.

[0008] Figure 2 A schematic diagram of the structure of a facial mask preparation process parameter optimization system using artificial intelligence provided by the present invention.

[0009] Explanation of the reference numerals: image acquisition module 11 , parameter acquisition module 12 , simulation prediction module 13 , optimization preparation module 14 . DETAILED DESCRIPTION

[0010] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0011] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0012] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0013] Embodiment 1: like Figure 1 As shown, an embodiment of the present invention provides a method for optimizing process parameters of mask preparation using artificial intelligence, the method comprising: S10: collecting a light transmission image of the target membrane cloth to be impregnated, identifying the thickness of the membrane cloth, and obtaining a first thickness of the first area and a second thickness of the second area.

[0014] Exemplarily, a light-transmitting image of the target membrane cloth to be impregnated is collected. This step is usually performed by lighting the back of the membrane cloth, using light to penetrate the membrane cloth and project it onto the image acquisition device, thereby obtaining a light-transmitting image of the membrane cloth. Then, the collected light-transmitting image is used to identify the thickness of the membrane cloth using image processing technology and deep learning algorithms. In this process, the first area (i.e., T area) and the second area (i.e., U area) on the membrane cloth can be accurately distinguished. Since the T area is usually located in the center of the face, such as near the nose, forehead and chin, these areas require more essence to achieve better skin care effects when applying a mask, so the membrane cloth in the T area is designed to be relatively thick to absorb more essence. The U area is located on both sides of the face, such as the cheek area. These areas are relatively flat and require less essence, so the membrane cloth in the U area is designed to be relatively thin. The thickness of the membrane cloth in the T area and the U area can be accurately obtained through light-transmitting image acquisition and thickness recognition technology, providing basic data support for the optimization of subsequent impregnation parameters. For example, if the film cloth in the T zone is not thick enough, the essence may not be fully absorbed, affecting the skin care effect of the mask; if the film cloth in the U zone is too thick, the mask may not fit well when attached, affecting the usage experience.

[0015] S20: collecting viscosity information of the essence and obtaining an immersion parameter range of an immersion process parameter.

[0016] Optionally, the viscosity of the essence directly affects the effect of the dipping process. Therefore, it is necessary to specifically collect the viscosity information of the essence currently being produced for the mask. Usually, the viscosity information is collected by using special equipment such as a viscometer, which can accurately measure the viscosity value of the essence. After obtaining the viscosity information, the dipping parameter interval of the dipping process parameters can be determined in combination with the conventional range of the dipping time. The conventional range of the dipping time is usually 10-30 minutes, but it depends on the material of the membrane cloth and the viscosity of the essence. Among them, the dipping parameter interval is a reasonable dipping time range, which is intended to ensure that the essence can be fully and evenly adsorbed on the membrane cloth, while avoiding swelling and deformation of the membrane cloth due to excessive dipping time. For example, if the viscosity of the essence is high, then the dipping time may need to be appropriately extended to ensure that the essence can fully penetrate into the membrane cloth; conversely, if the viscosity of the essence is low, then the dipping time can be appropriately shortened to avoid excessive swelling of the membrane cloth. By collecting the viscosity information of the essence and determining the dipping parameter interval, accurate data support can be provided for subsequent dipping simulation predictions and mask usage predictions.

[0017] S30: randomly generating a first impregnation parameter within the impregnation parameter range, and performing an impregnation simulation prediction in combination with the viscosity information, the first thickness, and the second thickness to obtain a first impregnation amount and a second impregnation amount.

[0018] Furthermore, after determining the immersion parameter interval, a first immersion parameter is randomly generated within the interval, and this parameter is usually a specific immersion time value. Then, this immersion parameter is combined with the previously collected essence viscosity information, as well as the first thickness of the first area (T area) and the second thickness of the second area (U area) obtained by light transmission image acquisition and thickness recognition technology, and input into the pre-trained immersion simulation prediction model. The immersion simulation prediction model can simulate the adsorption of essence on the membrane cloth under different immersion parameters using a machine learning algorithm. Through simulation prediction, the immersion amount of the first area and the second area after immersion can be obtained, that is, the distribution amount of the essence in the T area and the U area. For example, if the simulation prediction results show that the immersion amount of the T area is insufficient, then it may be necessary to adjust the immersion parameters, such as increasing the immersion time or increasing the viscosity of the essence, to ensure that the T area can absorb enough essence; conversely, if the immersion amount of the U area is too much, resulting in deformation of the membrane cloth, then it is necessary to appropriately reduce the immersion time or reduce the viscosity of the essence. Through this step, precise control of the impregnation process can be achieved, providing key data support for subsequent mask usage prediction and impregnation parameter optimization.

[0019] S40: According to the first impregnation amount and the second impregnation amount, combined with the first thickness, the second thickness and the viscosity information, the mask usage is predicted to obtain the first essence dripping rate, the first film cloth deformation rate, the second essence dripping rate and the second film cloth deformation rate, and the first mask score and the second mask score are calculated in combination with the first impregnation amount and the second impregnation amount, and the impregnation parameters are optimized to obtain the optimal impregnation parameters for mask impregnation preparation.

[0020] Specifically, after obtaining the first impregnation amount and the second impregnation amount, the mask usage prediction is further performed in combination with the first thickness, the second thickness and the viscosity information of the essence. This step is intended to simulate the situations that may occur during the actual use of the mask, such as the dripping of the essence and the deformation of the membrane cloth. Specifically, if the impregnation time is too long or there is too much essence, the membrane cloth may swell and deform, especially for some membrane cloth materials that are easy to soften, the deformation will be more obvious. At the same time, too much essence may also cause the essence to drip when using the mask, affecting the user experience. Therefore, the essence dripping rate and membrane cloth deformation rate of the first area (T area) and the second area (U area) can be predicted by the mask usage prediction model. For example, if the prediction result shows that the essence dripping rate of the T area is high, then the impregnation parameters may need to be adjusted to reduce the amount of essence in the T area; if the membrane cloth deformation rate of the U area is high, then the impregnation process needs to be optimized to avoid excessive swelling of the membrane cloth. Based on the prediction results, the first impregnation amount and the second impregnation amount are combined to calculate the first mask score and the second mask score. These two scores are comprehensive evaluation indicators that take into account multiple factors such as essence distribution, mask deformation, and user experience. By comparing the mask scores under different impregnation parameters, the impregnation parameters can be optimized to find the optimal impregnation parameter combination for subsequent mask impregnation preparation. This can not only ensure the quality and consistency of the mask, but also improve user experience and satisfaction.

[0021] In a preferred embodiment, a light transmission image of the target membrane cloth to be impregnated is collected, and the membrane cloth thickness is identified to obtain a first thickness of the first area and a second thickness of the second area, including: collecting a light transmission image of the target membrane cloth to be impregnated under preset light transmission parameters to obtain a light transmission image of the membrane cloth; inputting the light transmission image of the membrane cloth into a membrane cloth thickness identification channel, identifying and outputting the first thickness of the first area and the second thickness of the second area, wherein the first area is the T area and the second area is the U area.

[0022] Preferably, in order to achieve accurate optimization of the impregnation parameters, it is necessary to collect a light-transmitting image of the target membrane cloth to be impregnated. The light-transmitting image collection process is carried out under preset light-transmitting parameters to ensure that the collected image can clearly reflect the thickness distribution of the membrane cloth. The preset light-transmitting parameters usually include light source type, light intensity, collection angle, etc. The setting of these parameters needs to be adjusted according to the material and characteristics of the membrane cloth to obtain the best image collection effect.

[0023] After the light-transmitting image of the membrane is collected, it will be input into a specially trained membrane thickness recognition channel for processing. The membrane thickness recognition channel is built based on a deep learning algorithm and can automatically identify and extract the membrane thickness information of the first area (T area) and the second area (U area) in the image. In the first area, due to the prominent facial contour, the membrane needs to be designed to be relatively thick to provide better fit and essence adsorption capacity; in the second area, the facial contour is relatively flat, and the membrane is designed to be relatively thin to reduce unnecessary essence waste and membrane deformation.

[0024] Through the processing of the membrane cloth thickness identification channel, the membrane cloth thickness of the T zone and U zone can be accurately obtained, providing key data support for the subsequent optimization of the dipping parameters. For example, if the identification result shows that the membrane cloth thickness of the T zone is insufficient, then the dipping parameters may need to be adjusted to increase the adsorption of essence in the T zone; if the membrane cloth in the U zone is too thick, the dipping process may need to be optimized to reduce the distribution of essence in the U zone and avoid deformation of the membrane cloth. The implementation of this step not only improves the accuracy and efficiency of mask preparation, but also lays a solid foundation for improving user experience and mask quality.

[0025] In a preferred embodiment, the training steps of the membrane cloth thickness identification channel include: using deep learning to construct a membrane cloth thickness identification channel; collecting a set of sample membrane cloth translucent images according to the detection data log of the membrane cloth, and marking the average thickness of the first area and the second area of ​​the membrane cloth in different sample membrane cloth translucent images to obtain a sample first thickness set and a sample second thickness set; using the sample membrane cloth translucent image set, the sample first thickness set and the sample second thickness set to perform supervised training, verification and testing on the membrane cloth thickness identification channel, and completing the training when the accuracy is greater than or equal to the accuracy threshold.

[0026] In detail, the training of the membrane thickness recognition channel is a deep learning-based process. First, a channel dedicated to membrane thickness recognition is constructed through deep learning technology. This channel can automatically learn and extract feature information from the membrane translucent image. Next, based on the membrane detection data log, a series of sample membrane translucent images are collected. These images cover membranes of different materials, thicknesses, and translucency to ensure the diversity and representativeness of the training data.

[0027] After collecting the sample film cloth light transmission image set, it is necessary to mark the average thickness of the first area (T area) and the second area (U area) of the film cloth in each image. This step is usually completed by manual measurement or high-precision instruments to ensure the accuracy of the marked data. After the marking is completed, the sample first thickness set and the sample second thickness set are obtained. These two sets correspond to the film cloth thickness information of the T area and the U area respectively.

[0028] Finally, the membrane cloth thickness recognition channel is supervised and trained using a set of sample membrane cloth light transmission images, a set of sample first thickness images, and a set of sample second thickness images. During the training process, the channel will continuously learn and adjust its own parameters to minimize the error between the predicted thickness and the actual thickness. At the same time, the performance and accuracy of the channel are evaluated through verification and testing steps. When the accuracy is greater than or equal to the preset accuracy threshold, the channel is considered to have been trained and can be put into practical use. For example, if the accuracy threshold is set to 95%, then when the accuracy of the channel on the test set reaches or exceeds 95%, it can be considered to have sufficient recognition ability to accurately predict the thickness of the membrane cloth.

[0029] In a preferred embodiment, collecting viscosity information of the essence and obtaining the dipping parameter range of the dipping process parameters include: collecting viscosity information of the essence currently being produced for the facial mask; obtaining the dipping parameter range of the dipping process parameters, wherein the dipping process parameters include the dipping time.

[0030] Specifically, viscosity, as a key physical property of the essence, directly affects its adsorption and penetration effect on the membrane cloth. Therefore, it is necessary to accurately collect the viscosity information of the essence currently being produced for the mask, which is usually done through professional equipment such as a viscometer to ensure the accuracy and reliability of the measurement results. Based on the collected viscosity information and the understanding of the mask production process, it is necessary to obtain the dipping parameter range of the dipping process parameters. Among the dipping process parameters, the dipping time determines the residence time of the essence on the membrane cloth, which in turn affects the final quality of the mask. The determination of the dipping parameter range is obtained by comprehensively considering multiple factors such as the viscosity of the essence, the material of the membrane cloth, and the desired mask effect. For example, for an essence with a higher viscosity, the dipping time may need to be appropriately extended to ensure that it can fully penetrate into the membrane cloth; while for an essence with a lower viscosity, the dipping time can be relatively shortened to avoid excessive swelling of the membrane cloth. By scientifically and rationally determining the dipping parameter range, powerful process guidance can be provided for subsequent mask production.

[0031] In a preferred embodiment, a first impregnation parameter is randomly generated within the impregnation parameter interval, and an impregnation simulation prediction is performed in combination with the viscosity information, the first thickness and the second thickness to obtain the first impregnation amount and the second impregnation amount, including: randomly generating a first impregnation parameter within the impregnation parameter interval; inputting the first impregnation parameter, the viscosity information, the first thickness and the second thickness into a pre-trained impregnation amount prediction channel, and predicting and outputting the first impregnation amount and the second impregnation amount, wherein the impregnation amount prediction channel uses sample impregnation parameters, sample viscosity information, sample first thickness and sample second thickness as input features, and uses the sample first impregnation amount and the sample second impregnation amount as output features for training.

[0032] Furthermore, after the impregnation parameter interval is determined, a first impregnation parameter is randomly generated within this interval. This parameter represents a specific condition in the impregnation process, such as a random value of the impregnation time, which will play a key role in subsequent simulation predictions.

[0033] Subsequently, the generated first impregnation parameter, together with the previously collected serum viscosity information, and the first thickness of the first area (T area) and the second thickness of the second area (U area) obtained by membrane thickness recognition, are input into the pre-trained impregnation amount prediction channel. This impregnation amount prediction channel is a machine learning-based model that establishes a complex relationship between impregnation parameters, viscosity information, membrane thickness, and impregnation amount by learning a large amount of sample data.

[0034] During the training process, the impregnation amount prediction channel uses sample impregnation parameters, sample viscosity information, sample first thickness and sample second thickness as input features, and sample first impregnation amount and sample second impregnation amount as output features. Through this training method, the model can learn the change rules of the essence impregnation amount in the T zone and U zone on the membrane cloth under different conditions.

[0035] When new data (i.e., the generated first impregnation parameter, viscosity information, first thickness, and second thickness) is input, the impregnation amount prediction channel can predict and output the first impregnation amount and the second impregnation amount according to the previously learned rules, for example, 11ml and 6ml respectively. These two impregnation amounts represent the adsorption amount of the essence in the T zone and U zone, respectively, and are important indicators for evaluating the impregnation effect. For example, if the prediction result shows that the impregnation amount in the T zone is insufficient, it may be necessary to adjust the impregnation parameters or optimize the essence formula to improve the adsorption effect in the T zone.

[0036] In a preferred embodiment, according to the first impregnation amount and the second impregnation amount, combined with the first thickness, the second thickness and the viscosity information, the mask usage prediction is performed to obtain the first essence dripping rate, the first film cloth deformation rate, the second essence dripping rate and the second film cloth deformation rate, including: according to the historical preparation data log of the first area, collecting the sample first impregnation amount set, the first thickness set, the sample viscosity information set, the sample first essence dripping rate set, and the sample first film cloth deformation rate set as the first usage training data of the first area, wherein the first film cloth deformation rate includes the proportion of the film cloth size deformation in the first area after impregnation; based on machine learning, constructing the first usage prediction branch of the first area, using the first usage training data for supervised training, verification and testing Test until convergence; according to the historical preparation data log of the second area, collect the sample second impregnation amount set, the second thickness set, the sample viscosity information set, the sample second essence dripping rate set, and the sample second membrane cloth deformation rate set as the second usage training data for the second area; based on machine learning, construct the second usage prediction branch for the second area, and use the second usage training data for supervised training, verification and testing until convergence; input the first impregnation amount, the first thickness and the viscosity information into the first usage prediction branch, and predict the output to obtain the first essence dripping rate and the first membrane cloth deformation rate; input the second impregnation amount, the second thickness and the viscosity information into the second usage prediction branch, and predict the output to obtain the second essence dripping rate and the second membrane cloth deformation rate.

[0037] Optionally, in order to accurately predict the use effect of the mask, especially the essence dripping rate and the membrane cloth deformation rate, a prediction method based on machine learning is adopted. First, for the first area (T zone) of the mask, a series of sample data are collected according to the historical preparation data log, including the sample first impregnation amount set, the first thickness set, the sample viscosity information set, and the corresponding sample first essence dripping rate set and the sample first membrane cloth deformation rate set. These sample data constitute the first usage training data of the first area, wherein the first membrane cloth deformation rate specifically refers to the proportion of the membrane cloth size deformation in the T zone after immersion. For example, after deformation, the area of ​​the T zone membrane cloth after flattening is 11 square centimeters. Under normal circumstances, when the T zone membrane cloth is not deformed, the area of ​​the T zone membrane cloth after flattening is 10 square centimeters. Then the first membrane cloth deformation rate is (11-10) / 10=0.1. The first membrane cloth deformation rate is an important indicator for evaluating the fit and comfort of the mask. The first essence dripping rate of the sample is, for example, the probability of essence dripping during use. For example, if 3 out of 10 facial masks having the same first impregnation volume, first thickness, and viscosity information have essence dripping during use, then the first essence dripping rate of the sample is 0.3.

[0038] Then, based on these rich training data, a machine learning algorithm is used to build a first usage prediction branch specifically for the T zone. This prediction branch gradually learns the complex relationship between the impregnation amount, film thickness, viscosity information, essence dripping rate, and film deformation rate through continuous supervised training, verification, and testing until the model converges, that is, the prediction performance reaches a stable state.

[0039] Similarly, a similar approach was used for the second area (U area) of the mask. Corresponding sample data was collected based on the historical preparation data log of the U area, and the second usage prediction branch for the second area was constructed. This branch was also rigorously trained, validated, and tested to ensure its prediction accuracy.

[0040] In actual applications, when new first impregnation amount, first thickness and viscosity information are obtained, these data are input into the first usage prediction branch that has been trained, and the first essence dripping rate and the first membrane cloth deformation rate of the T zone can be predicted and output. Similarly, the second impregnation amount, second thickness and viscosity information are input into the second usage prediction branch to obtain the second essence dripping rate and the second membrane cloth deformation rate of the U zone. For example, if the prediction result shows that the essence dripping rate of the T zone is high, it may be necessary to adjust the impregnation parameters or optimize the membrane cloth material to reduce the waste of essence; if the membrane cloth deformation rate of the U zone is large, it is necessary to consider improving the production process of the membrane cloth or selecting a more suitable membrane cloth material to improve the overall quality of the mask.

[0041] In a preferred embodiment, the first impregnation amount and the second impregnation amount are combined to calculate the first mask score and the second mask score, and the impregnation parameters are optimized to obtain the optimal impregnation parameters, and the mask impregnation preparation is performed, including: according to the first impregnation amount, the first essence dripping rate and the first membrane cloth deformation rate, based on the first area preparation function, the first mask score is calculated to obtain the second mask score based on the second area preparation function according to the second impregnation amount, the second essence dripping rate and the second membrane cloth deformation rate; according to the first mask score and the second mask score, a weighted calculation is performed to obtain the mask preparation score; continue to randomly generate impregnation parameters, calculate the mask preparation score, and optimize the impregnation parameters until convergence, obtain the optimal impregnation parameters with the largest mask preparation score, and perform mask impregnation preparation.

[0042] In detail, in order to evaluate the quality of the mask under different dipping parameters and optimize the dipping parameters to obtain the best preparation effect, the first mask score is calculated based on the first dipping amount, the first essence dripping rate and the first membrane deformation rate using a specially designed first zone preparation function. This score comprehensively reflects the essence adsorption amount, dripping and membrane deformation degree of the T zone after dipping, and is an important indicator for evaluating the quality of the T zone mask.

[0043] Similarly, for the U zone, the second mask score is calculated based on the second impregnation amount, the second essence dripping rate and the second film cloth deformation rate through the second zone preparation function. This score also combines the key indicators of the U zone and provides a quantitative basis for evaluating the quality of the U zone mask.

[0044] Next, in order to comprehensively evaluate the preparation effect of the entire mask, a weighted calculation is performed based on the first mask score and the second mask score to obtain the mask preparation score. This score takes into account the mask quality of the T zone and U zone and is a comprehensive indicator for evaluating the preparation effect of the entire mask.

[0045] In order to find the optimal impregnation parameters, a random search method is used to continuously generate new impregnation parameters and calculate the corresponding mask preparation scores. By comparing the mask preparation scores under different impregnation parameters, the optimal solution can be gradually approached. This process continues until the mask preparation score converges, that is, it no longer changes significantly with the change of the impregnation parameters, and the optimal impregnation parameters with the largest mask preparation score can be determined.

[0046] Finally, using the optimal impregnation parameters for mask impregnation preparation can ensure that the mask can achieve the best essence absorption, the least essence dripping, and the smallest deformation of the mask cloth in both the T zone and the U zone, thereby providing the best user experience. For example, if it is found through optimization that at a certain impregnation time, the essence absorption in the T zone is the largest and the dripping rate is the lowest, while the deformation rate of the mask cloth in the U zone is also the smallest, then this impregnation time is the optimal impregnation parameter and will be used in the subsequent mask preparation process.

[0047] In a preferred embodiment, the first region preparation function is as follows: ;in, Rate the first mask, , and is the weight, is the first dipping amount, The preset first impregnation volume for the first area is, for example, 10 ml, is the first essence dripping rate, is the deformation rate of the first membrane cloth.

[0048] Specifically, in order to quantitatively evaluate the mask quality of the first zone (T zone), a specially designed first zone preparation function was used. The specific calculation formula is: This function comprehensively considers key indicators such as the first impregnation amount, the first essence dripping rate, and the first mask cloth deformation rate, and obtains the first mask score through weighted calculation. Specifically, the calculation formula of the first mask score includes three main parts: the first impregnation amount ( ) and the preset first dipping amount ( ), which reflects the closeness between the actual impregnation amount and the expected impregnation amount; the first essence drip rate ( ), which measures the dripping of the essence during the immersion process. The lower the dripping rate, the better the essence adsorption effect; and the deformation rate of the first membrane cloth ( ), which indicates the size deformation ratio of the membrane cloth in the T zone after impregnation. The smaller the deformation rate, the better the membrane cloth can maintain its original shape, providing better fit and comfort.

[0049] Optionally, the second area preparation function can replace the preset first impregnation amount in the first area preparation function with the preset second impregnation amount of the second area, for example, 5 ml, and other parameters are the same, namely, including the second membrane cloth deformation rate, the second essence dripping rate and the second impregnation amount, to calculate the second mask score.

[0050] In this formula, , and are weight coefficients, which represent the relative importance of the first impregnation amount, the first essence dripping rate, and the first membrane deformation rate in the scoring. By adjusting these weight coefficients, the priority of different indicators can be flexibly set according to actual needs. For example, if it is considered that maintaining the membrane shape in the T zone is more important than reducing essence dripping, then it can be appropriately increased. The value of makes the deformation rate of the membrane cloth have a greater impact on the score. For example, they are 0.3, 0.4 and 0.3 respectively.

[0051] Taking the actual preparation process as an example, assuming that a preset first dipping amount is set , and collected the actual impregnation amount , the first essence drip rate and the first membrane deformation rate Substituting these data into the first region preparation function, the first mask score can be calculated By comparing the The value can be used to evaluate the effect of different preparation conditions on the quality of T-zone masks, thereby optimizing the impregnation parameters and improving the overall quality of the mask.

[0052] The method for optimizing the process parameters of facial mask preparation using artificial intelligence provided by the embodiment of the present invention has at least the following technical effects: 1. The membrane thickness recognition channel constructed through deep learning can accurately identify the membrane thickness of different areas of the mask (T zone and U zone), which not only improves the accuracy of membrane thickness measurement, but also provides key data support for the optimization of subsequent impregnation process parameters. Regional analysis makes the mask preparation more in line with the skin needs of different areas of the face, improving the use effect and comfort of the mask.

[0053] 2. Using the pre-trained impregnation amount prediction channel, combined with randomly generated impregnation parameters, viscosity information, and film cloth thickness data, the impregnation simulation prediction is performed to obtain the first impregnation amount and the second impregnation amount, which can predict the adsorption amount of essence under different impregnation conditions in advance, providing a scientific basis for optimizing the impregnation process parameters. At the same time, the usage prediction branch constructed by machine learning can accurately predict the essence dripping rate and film cloth deformation rate during the use of the mask, further improving the accuracy and reliability of mask preparation.

[0054] 3. By constructing the preparation functions of the first and second regions, combined with key indicators such as impregnation amount, essence dripping rate and membrane cloth deformation rate, the first mask score and the second mask score are calculated, and the mask preparation score is obtained by weighted calculation. This comprehensive scoring system can comprehensively evaluate the preparation quality of the mask and provide a quantitative basis for the optimization of the impregnation parameters. By continuously randomly generating impregnation parameters and calculating the mask preparation score until the optimal impregnation parameters are obtained, the intelligent optimization of the mask preparation process parameters is achieved, which significantly improves the preparation efficiency and quality of the mask.

[0055] Embodiment 2: like Figure 2 As shown, based on the same inventive concept of a method for optimizing process parameters of a facial mask preparation using artificial intelligence provided in Example 1, an embodiment of the present invention further provides a system for optimizing process parameters of a facial mask preparation using artificial intelligence, the system comprising: The image acquisition module 11 is used to acquire a light transmission image of the target membrane cloth to be impregnated, identify the thickness of the membrane cloth, and obtain a first thickness of the first area and a second thickness of the second area.

[0056] The parameter acquisition module 12 is used to collect the viscosity information of the essence and obtain the dipping parameter range of the dipping process parameters.

[0057] The simulation prediction module 13 is used to randomly generate a first impregnation parameter within the impregnation parameter range, and perform an impregnation simulation prediction in combination with the viscosity information, the first thickness and the second thickness to obtain a first impregnation amount and a second impregnation amount.

[0058] The optimization preparation module 14 is used to predict the use of the mask according to the first impregnation amount and the second impregnation amount, combined with the first thickness, the second thickness and the viscosity information, to obtain the first essence dripping rate, the first film cloth deformation rate, the second essence dripping rate and the second film cloth deformation rate, and to calculate the first mask score and the second mask score in combination with the first impregnation amount and the second impregnation amount, to optimize the impregnation parameters, to obtain the optimal impregnation parameters, and to prepare the mask by impregnation.

[0059] Furthermore, the image acquisition module 11 is also used to perform the following steps: The target membrane cloth to be impregnated is subjected to light transmission image acquisition under preset light transmission parameters to obtain a light transmission image of the membrane cloth; the light transmission image of the membrane cloth is input into a membrane cloth thickness identification channel, and the identification output obtains a first thickness of the first area and a second thickness of the second area, wherein the first area is the T area and the second area is the U area.

[0060] Furthermore, the image acquisition module 11 is also used to perform the following steps: Deep learning is used to construct a membrane cloth thickness recognition channel; according to the detection data log of the membrane cloth, a set of sample membrane cloth translucent images is collected, and the average thickness of the first area and the second area of ​​the membrane cloth in different sample membrane cloth translucent images is marked to obtain a sample first thickness set and a sample second thickness set; the sample membrane cloth translucent image set, the sample first thickness set and the sample second thickness set are used to perform supervised training, verification and testing on the membrane cloth thickness recognition channel, and the training is completed when the accuracy is greater than or equal to the accuracy threshold.

[0061] Furthermore, the parameter acquisition module 12 is further configured to perform the following steps: The viscosity information of the essence liquid currently being produced for the facial mask is collected; and the dipping parameter range of the dipping process parameters is obtained, wherein the dipping process parameters include the dipping time.

[0062] Furthermore, the simulation prediction module 13 is also used to perform the following steps: A first impregnation parameter is randomly generated within the impregnation parameter interval; the first impregnation parameter, viscosity information, first thickness and second thickness are input into a pre-trained impregnation amount prediction channel, and the first impregnation amount and the second impregnation amount are predicted and output, wherein the impregnation amount prediction channel uses the sample impregnation parameter, sample viscosity information, sample first thickness and sample second thickness as input features, and uses the sample first impregnation amount and the sample second impregnation amount as output features for training.

[0063] Furthermore, the optimization preparation module 14 is also used to perform the following steps: According to the historical preparation data log of the first area, a sample first impregnation amount set, a first thickness set, a sample viscosity information set, a sample first essence dripping rate set, and a sample first membrane cloth deformation rate set are collected as the first usage training data of the first area, wherein the first membrane cloth deformation rate includes the proportion of the membrane cloth size deformation in the first area after impregnation; based on machine learning, a first usage prediction branch of the first area is constructed, and the first usage training data is used for supervised training, verification and testing until convergence; according to the historical preparation data log of the second area, a sample second impregnation amount set, a second thickness set, a sample viscosity information set, a sample second essence dripping rate set, and a sample second membrane cloth deformation rate set are collected as the second usage training data of the second area; based on machine learning, a second usage prediction branch of the second area is constructed, and the second usage training data is used for supervised training, verification and testing until convergence; the first impregnation amount, first thickness and viscosity information are input into the first usage prediction branch, and the first essence dripping rate and the first membrane cloth deformation rate are obtained by prediction output; the second impregnation amount, second thickness and viscosity information are input into the second usage prediction branch, and the second essence dripping rate and the second membrane cloth deformation rate are obtained by prediction output.

[0064] Furthermore, the optimization preparation module 14 is also used to perform the following steps: According to the first impregnation amount, the first essence dripping rate and the first membrane cloth deformation rate, based on the first area preparation function, the first mask score is calculated; according to the second impregnation amount, the second essence dripping rate and the second membrane cloth deformation rate, based on the second area preparation function, the second mask score is calculated; according to the first mask score and the second mask score, a weighted calculation is performed to obtain the mask preparation score; continue to randomly generate impregnation parameters, calculate the mask preparation score, and optimize the impregnation parameters until convergence, obtain the optimal impregnation parameters with the largest mask preparation score, and perform mask impregnation preparation.

[0065] Furthermore, the optimization preparation module 14 also includes: The first region preparation function is as follows: ;in, Rate the first mask, , and is the weight, is the first dipping amount, a preset first impregnation amount for the first area, is the first essence dripping rate, is the deformation rate of the first membrane cloth.

[0066] Through the above-mentioned detailed description of a method for optimizing process parameters of facial mask preparation using artificial intelligence, those skilled in the art can clearly know that a system for optimizing process parameters of facial mask preparation using artificial intelligence is provided in this embodiment. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0067] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing process parameters of facial mask preparation using artificial intelligence, characterized in that: The method comprises: Capturing a light transmission image of the target membrane cloth to be impregnated, identifying the membrane cloth thickness, and obtaining a first thickness of the first area and a second thickness of the second area; Collecting viscosity information of the essence and obtaining an immersion parameter range of an immersion process parameter; Randomly generate a first dipping parameter within the dipping parameter range, and perform dipping simulation prediction in combination with the viscosity information, the first thickness, and the second thickness to obtain a first dipping amount and a second dipping amount; According to the first impregnation amount and the second impregnation amount, combined with the first thickness, the second thickness and the viscosity information, the mask usage is predicted, and the first essence dripping rate, the first film cloth deformation rate, the second essence dripping rate and the second film cloth deformation rate are obtained. In combination with the first impregnation amount and the second impregnation amount, the first mask score and the second mask score are calculated to obtain the impregnation parameters, and the impregnation parameters are optimized to obtain the optimal impregnation parameters for mask impregnation preparation.

2. The method for optimizing process parameters of facial mask preparation using artificial intelligence according to claim 1, characterized in that: The target membrane cloth to be impregnated is imaged through light transmission, and the membrane cloth thickness is identified to obtain a first thickness of a first area and a second thickness of a second area, including: Capturing a light transmission image of the target membrane cloth to be impregnated under preset light transmission parameters to obtain a light transmission image of the membrane cloth; The light-transmitting image of the membrane cloth is input into a membrane cloth thickness identification channel, and the identification output obtains a first thickness of a first area and a second thickness of a second area, wherein the first area is a T area and the second area is a U area.

3. The method for optimizing process parameters of facial mask preparation using artificial intelligence according to claim 2, characterized in that: The training steps of the membrane thickness recognition channel include: Use deep learning to build a membrane thickness recognition channel; According to the detection data log of the membrane cloth, a set of light transmission images of the sample membrane cloth is collected, and the average thickness of the first area and the second area of ​​the membrane cloth in the light transmission images of different sample membrane cloths is marked to obtain a sample first thickness set and a sample second thickness set; The sample film cloth light transmission image set, the sample first thickness set and the sample second thickness set are used to perform supervised training, verification and testing on the film cloth thickness identification channel, and the training is completed when the accuracy is greater than or equal to the accuracy threshold.

4. The method for optimizing process parameters of facial mask preparation using artificial intelligence according to claim 1, characterized in that: Collect the viscosity information of the essence and obtain the dipping parameter range of the dipping process parameters, including: Collect the viscosity information of the essence currently being produced for the mask; An immersion parameter interval of an immersion process parameter is obtained, wherein the immersion process parameter includes an immersion time.

5. The method for optimizing process parameters of facial mask preparation using artificial intelligence according to claim 1, characterized in that: Randomly generating a first impregnation parameter within the impregnation parameter interval, combining the viscosity information, the first thickness and the second thickness, performing an impregnation simulation prediction, and obtaining a first impregnation amount and a second impregnation amount, including: randomly generating a first dipping parameter within the dipping parameter interval; The first impregnation parameter, viscosity information, first thickness and second thickness are input into a pre-trained impregnation amount prediction channel to predict and output the first impregnation amount and the second impregnation amount, wherein the impregnation amount prediction channel uses the sample impregnation parameter, sample viscosity information, sample first thickness and sample second thickness as input features, and uses the sample first impregnation amount and sample second impregnation amount as output features for training.

6. The method for optimizing process parameters of facial mask preparation using artificial intelligence according to claim 1, characterized in that: According to the first impregnation amount and the second impregnation amount, combined with the first thickness, the second thickness and the viscosity information, the mask usage prediction is performed to obtain the first essence dripping rate, the first film cloth deformation rate, the second essence dripping rate and the second film cloth deformation rate, including: According to the historical preparation data log of the first area, a sample first impregnation amount set, a first thickness set, a sample viscosity information set, a sample first essence dripping rate set, and a sample first membrane cloth deformation rate set are collected as the first usage training data of the first area, wherein the first membrane cloth deformation rate includes the proportion of the membrane cloth size deformation in the first area after impregnation; Based on machine learning, construct a first usage prediction branch for the first region, and use the first usage training data for supervised training, verification, and testing until convergence; According to the historical preparation data log of the second area, a second impregnation amount set of samples, a second thickness set, a viscosity information set of samples, a second essence dripping rate set of samples, and a second membrane cloth deformation rate set of samples are collected as second usage training data of the second area; Based on machine learning, construct a second usage prediction branch for the second region, and use the second usage training data for supervised training, verification, and testing until convergence; Inputting the first impregnation amount, the first thickness and the viscosity information into the first usage prediction branch, and predicting and outputting a first essence dripping rate and a first membrane cloth deformation rate; The second impregnation amount, the second thickness and the viscosity information are input into the second usage prediction branch, and the prediction output obtains the second essence dripping rate and the second membrane cloth deformation rate.

7. The method for optimizing process parameters of facial mask preparation using artificial intelligence according to claim 1, characterized in that: Combining the first impregnation amount and the second impregnation amount, calculating a first mask score and a second mask score, optimizing the impregnation parameters, obtaining the optimal impregnation parameters, and performing mask impregnation preparation, including: According to the first impregnation amount, the first essence dripping rate and the first membrane cloth deformation rate, based on the first area preparation function, a first mask score is calculated; according to the second impregnation amount, the second essence dripping rate and the second membrane cloth deformation rate, based on the second area preparation function, a second mask score is calculated; According to the first mask score and the second mask score, weighted calculation is performed to obtain a mask preparation score; Continue to randomly generate dipping parameters, calculate the mask preparation score, and optimize the dipping parameters until convergence, obtain the optimal dipping parameters with the largest mask preparation score, and perform mask dipping preparation.

8. The method for optimizing process parameters of facial mask preparation using artificial intelligence according to claim 7, characterized in that: The first region preparation function is as follows: ； in, Rate the first mask, , and is the weight, is the first dipping amount, a preset first impregnation amount for the first area, is the first essence dripping rate, is the deformation rate of the first membrane cloth.

9. A facial mask preparation process parameter optimization system using artificial intelligence, characterized in that: A method for optimizing process parameters of facial mask preparation using artificial intelligence according to any one of claims 1 to 8, the system comprising: An image acquisition module is used to acquire a light-transmitting image of the target membrane cloth to be impregnated, identify the thickness of the membrane cloth, and obtain a first thickness of the first area and a second thickness of the second area; A parameter acquisition module, used to collect viscosity information of the essence and obtain an immersion parameter range of an immersion process parameter; A simulation prediction module, used for randomly generating a first impregnation parameter within the impregnation parameter interval, and performing an impregnation simulation prediction in combination with the viscosity information, the first thickness and the second thickness to obtain a first impregnation amount and a second impregnation amount; The optimization preparation module is used to predict the use of the mask according to the first impregnation amount and the second impregnation amount, combined with the first thickness, the second thickness and the viscosity information, to obtain the first essence dripping rate, the first film cloth deformation rate, the second essence dripping rate and the second film cloth deformation rate, and to calculate the first mask score and the second mask score in combination with the first impregnation amount and the second impregnation amount, to optimize the impregnation parameters, to obtain the optimal impregnation parameters, and to prepare the mask by impregnation.

Images