A Method and System for Evaluating the Soothing Efficacy of Cosmetics Based on Deep Learning Models

Abstract

This invention discloses a method and system for evaluating the soothing efficacy of cosmetics based on a deep learning model. The method includes: setting up a control group and a test sample group; after each group is cultured for the same time, zebrafish are fixed with methylcellulose, and side views of the zebrafish in each group are captured under a fluorescence microscope as the original zebrafish images; a binary mask is obtained based on the preprocessed original zebrafish images; the preprocessed original zebrafish images and the corresponding binary masks are used as samples in the training dataset; a semantic segmentation model is trained using the training dataset; the zebrafish images to be detected in each group are preprocessed and input into the trained semantic segmentation model to obtain images containing only the target region in each group, and then the number of fluorescent particles in each group is extracted to evaluate the soothing effect of the cosmetics. This invention combines fluorescence imaging technology and a deep learning model, which can quickly and accurately calculate the number of fluorescent particles in key regions of zebrafish images, thereby realizing the evaluation of the soothing efficacy of cosmetics.

Description

Technical Field

[0001] This invention relates to the field of cosmetic efficacy evaluation technology, and in particular to a method, system, terminal device, computer-readable storage medium, and computer program product for evaluating the soothing efficacy of cosmetics based on a deep learning model. Background Technology

[0002] Traditional safety assessments of cosmetics primarily rely on animal testing, such as with mice, rats, and rabbits. However, animal testing not only presents animal welfare and ethical issues but also faces limitations, such as the long lifespan of animals, the difficulty in producing offspring, and the relative complexity of experimental procedures. Furthermore, animal testing is susceptible to external factors such as physical condition and environment, leading to non-reproducibility and unreliability of results.

[0003] Zebrafish, as a biological model, possesses characteristics such as transparency and a short reproductive cycle, making it an ideal model for studying gene function and drug screening. Early zebrafish imaging technology was mainly used to study embryonic development and kinematics. With the development of fluorescence imaging technology, zebrafish fluorescence imaging technology has also been gradually applied to evaluate the efficacy of cosmetics, offering advantages such as effectiveness, high throughput, and precision. The general procedure for detecting cosmetic efficacy using fluorescence imaging technology is as follows: transferring normally developing 2dpf zebrafish into a six-well plate; incubating the zebrafish under different experimental conditions; fixing the zebrafish after incubation and photographing them under a fluorescence microscope; and counting the number of neutrophils based on the images. Although this method can evaluate cosmetic efficacy by counting neutrophils, it has the following main drawbacks: 1) The experiment generates a large amount of image data, making manual counting a heavy task. Manual counting is extremely dependent on the operator's energy and concentration; prolonged screen viewing can easily lead to visual fatigue, resulting in misjudgments and omissions. Furthermore, due to accumulated fatigue, the accuracy of counting declines with working hours, making it difficult to guarantee the consistency of data between batches; 2) Manual counting has certain subjective limitations. Different operators have difficulty in standardizing the morphological judgment criteria of neutrophils. This subjective difference can lead to large fluctuations in the counting results, resulting in a lack of reliability; 3) Inefficiency. Manual counting can only identify each image and each cell individually, which is extremely inefficient, resulting in a long evaluation cycle for cosmetic efficacy and delaying product development. Summary of the Invention

[0004] To address the shortcomings of the existing technology, this invention provides a method, system, terminal device, computer-readable storage medium, and computer program product for evaluating the soothing efficacy of cosmetics based on a deep learning model. By combining fluorescence imaging technology with a deep learning model, the number of fluorescent particles in key areas of zebrafish images can be extracted quickly and accurately, thereby enabling the evaluation of the soothing efficacy of cosmetics.

[0005] The first objective of this invention is to provide a method for evaluating the soothing efficacy of cosmetics based on a deep learning model.

[0006] The second objective of this invention is to provide a cosmetic soothing efficacy evaluation system based on a deep learning model.

[0007] The third objective of this invention is to provide a terminal device.

[0008] A fourth objective of this invention is to provide a computer-readable storage medium.

[0009] The fifth objective of this invention is to provide a computer program product.

[0010] The first objective of this invention can be achieved by adopting the following technical solution:

[0011] A method for evaluating the soothing efficacy of cosmetics based on a deep learning model, the method comprising:

[0012] Set up a normal control group: Place zebrafish in standard dilution water;

[0013] Set up a model control group: Place zebrafish in standard diluted water containing the modeling agent;

[0014] Setting up the test sample group: Zebrafish were placed in standard dilution water containing a modeling agent and a test sample stock solution; the test sample stock solution was a solution prepared from the cosmetic product to be tested;

[0015] Set up a positive control group: Place zebrafish in standard diluted water containing modeling agent and dipotassium glycyrrhizate;

[0016] The zebrafish in each group were normally developed 2dpf neutrophil gene mutant zebrafish;

[0017] After each group was cultured for the same period of time, zebrafish were fixed with methylcellulose and side views of each group of zebrafish were taken under a fluorescence microscope as the original images of zebrafish.

[0018] The original zebrafish image is preprocessed to improve the overall brightness of the image;

[0019] Based on the preprocessed original zebrafish image, a binary mask is obtained; the preprocessed original zebrafish image and the corresponding binary mask are used as samples in the training dataset.

[0020] The semantic segmentation model is trained using the training dataset;

[0021] The zebrafish images to be detected in each group are preprocessed and then input into the trained semantic segmentation model to obtain images containing only the target region in each group; the number of fluorescent particles in each group is extracted based on the images containing only the target region in each group; and the soothing effect of the cosmetic product is evaluated based on the number of fluorescent particles in each group.

[0022] Furthermore, the process of preprocessing the zebrafish images to be detected in each group and inputting them into the trained semantic segmentation model yields images in each group that contain only the target region, including:

[0023] The zebrafish images to be detected in each group are preprocessed and then input into the trained semantic segmentation model, which outputs the corresponding binary mask.

[0024] The binary mask is multiplied pixel-level with the corresponding preprocessed zebrafish image to be detected to obtain a zebrafish image containing only the target region.

[0025] Furthermore, the preprocessing includes:

[0026] Convert the image from RGB space to HSV space;

[0027] Increase the pixel values ​​of the V and S channels of the image to obtain an image with increased brightness;

[0028] Convert the image with increased brightness from HSV space to RGB space;

[0029] The images are either original zebrafish images or zebrafish images to be detected in each group.

[0030] Furthermore, obtaining the binary mask based on the preprocessed original zebrafish image includes:

[0031] The target area is circled in the preprocessed original zebrafish image with a thin line to obtain a preliminary line outline; the target area is two regions on the back of the zebrafish, namely the yolk sac and the corresponding part above the lateral line.

[0032] Based on the initial line outline, morphological operations are used to refine the line outline and remove existing noise or artifacts.

[0033] Finally, the line outline is filled with color and binarized to obtain a binary mask.

[0034] Furthermore, the step of extracting the number of fluorescent particles in each group based on images containing only the target region includes:

[0035] Extract the green channel from the RGB channels of an image containing only the target region and convert it to a grayscale image; filter out pixels in the grayscale image that are larger than a set threshold, and then use the SimpleITK toolkit to treat connected pixels as a fluorescent particle, thereby obtaining the number of fluorescent particles in the image.

[0036] Furthermore, training the semantic segmentation model using the training dataset includes:

[0037] The preprocessed original zebrafish image is input into the semantic segmentation model, which outputs a binary mask.

[0038] The output binary mask is compared with the binary mask corresponding to the preprocessed original zebrafish image. The difference between the two is calculated using the cross-entropy loss function, and backpropagation is performed based on this difference to update and optimize the model parameters.

[0039] Furthermore, the evaluation of the soothing effect of the tested cosmetics based on the number of fluorescent particles in each group includes:

[0040] Using the model control group as a standard, the number of fluorescent particles in other groups is compared. If there are significant differences between the data, the experimental data are considered valid.

[0041] If the experimental data are valid, the soothing effect of the tested cosmetic product is calculated by using the number of fluorescent particles in the model control group, normal control group, and test sample group.

[0042] The second objective of this invention can be achieved by adopting the following technical solution:

[0043] A cosmetic soothing efficacy evaluation system based on a deep learning model, the system comprising:

[0044] The image acquisition module is used to set up control and test sample groups. After culturing for the same period, zebrafish in each group are fixed with methylcellulose, and side views of the zebrafish in each group are captured under a fluorescence microscope as the original images of the zebrafish. The control group includes a normal control group, a model control group, and a positive control group. The normal control group consists of zebrafish placed in standard dilution water, the model control group consists of zebrafish placed in standard dilution water containing a modeling agent, and the positive control group consists of zebrafish placed in standard dilution water containing a modeling agent and dipotassium glycyrrhizate. The test sample group consists of zebrafish placed in standard dilution water containing a modeling agent and a test sample stock solution, wherein the test sample stock solution is a solution prepared from the cosmetic to be tested. The zebrafish in each group are normally developed 2dpf neutrophil gene mutant zebrafish.

[0045] An image preprocessing module is used to preprocess the original zebrafish image to improve the overall brightness of the image;

[0046] The training dataset acquisition module is used to obtain the binary mask based on the preprocessed original zebrafish image; the preprocessed original zebrafish image and the corresponding binary mask are used as samples in the training dataset.

[0047] The model training module is used to train the semantic segmentation model using the training dataset;

[0048] The cosmetic evaluation module is used to input the pre-processed zebrafish images to be tested in each group into the trained semantic segmentation model to obtain images containing only the target region in each group; based on the images containing only the target region in each group, the number of fluorescent particles in each group is extracted; based on the number of fluorescent particles in each group, the soothing effect of the cosmetic is evaluated.

[0049] The third objective of this invention can be achieved by adopting the following technical solution:

[0050] A terminal device includes a processor and a memory for storing processor-executable programs. When the processor executes the program stored in the memory, it implements the above-described method for evaluating the soothing efficacy of cosmetics based on a deep learning model.

[0051] The fourth objective of this invention can be achieved by adopting the following technical solution:

[0052] A computer-readable storage medium storing a program that, when executed by a processor, implements the above-described method for evaluating the soothing efficacy of cosmetics based on a deep learning model.

[0053] The fifth objective of this invention can be achieved by adopting the following technical solution:

[0054] A computer program product includes a computer program that, when executed by a processor, implements the above-described method for evaluating the soothing efficacy of cosmetics.

[0055] The present invention has the following advantages over the prior art:

[0056] This invention combines fluorescence imaging technology with a deep learning model. It utilizes zebrafish fluorescence imaging to acquire raw zebrafish image data, ensuring the reliability of the acquired data as it closely reflects the organism's actual reactions. The pre-processed raw zebrafish image data is then used to train a semantic segmentation model. This trained model accurately segments the target region within the raw zebrafish image and automatically counts the neutrophils in that region, thereby evaluating the soothing efficacy of cosmetics. By combining fluorescence imaging technology with a deep learning model, the evaluation process is accelerated, and the accuracy and scientific rigor of evaluating the soothing efficacy of cosmetics are significantly improved, providing a new approach for the development of the beauty industry. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0058] Figure 1 This is a flowchart of the cosmetic soothing efficacy evaluation method based on a deep learning model according to Embodiment 1 of the present invention;

[0059] Figure 2 This is a schematic diagram of the cosmetic soothing efficacy evaluation method based on a deep learning model according to Embodiment 1 of the present invention.

[0060] Figure 3 This is a schematic diagram of the process for obtaining the original image of a fluorescently labeled zebrafish according to Embodiment 1 of the present invention;

[0061] Figure 4 This is the original image of a zebrafish from Embodiment 1 of the present invention;

[0062] Figure 5 This is the preprocessed original image of a zebrafish from Embodiment 1 of the present invention;

[0063] Figure 6 In the middle (a), (b), and (c), respectively, there are multiple preprocessed original zebrafish images of Embodiment 1 of the present invention, a binary mask corresponding to the image in (a), and an image containing only the target region corresponding to the binary mask in (b);

[0064] Figure 7 In the middle (a) and (b), there are multiple images containing only the target region in Embodiment 1 of the present invention, and the number of fluorescent particles extracted corresponding to the image in (a), respectively.

[0065] Figure 8 This is a structural block diagram of the cosmetic soothing efficacy evaluation system based on a deep learning model according to Embodiment 2 of the present invention;

[0066] Figure 9 This is a structural block diagram of the terminal device according to Embodiment 3 of the present invention. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. It should be understood that the specific embodiments described are merely used to explain this application and are not intended to limit this application.

[0068] Example 1:

[0069] like Figure 1 , 2 As shown, this embodiment provides a method for evaluating the soothing efficacy of cosmetics based on a deep learning model, including the following steps:

[0070] S101. Obtain the original image of the fluorescently labeled zebrafish.

[0071] like Figure 3 As shown, the original images of fluorescently labeled zebrafish were obtained through experiments, and the steps are as follows:

[0072] (1) Cosmetic sample confirmation. Confirm the usage method, solubility, stability and other information of the cosmetic sample to be tested.

[0073] (2) Preparation of standard dilution water.

[0074] In one embodiment, accurately weigh 17.2g sodium chloride, 0.76g potassium chloride, 2.9g calcium chloride, and 4.9g magnesium sulfate, dissolve them in distilled or deionized water, and bring the volume to 1000ml. Take 16.67ml of this solution and dilute it again with distilled or deionized water to 1000ml to obtain the standard diluted water. Before use, equilibrate to oxygen saturation with air at (28±2)℃ and store for later use.

[0075] (3) Preparation of stock solution for cosmetic sample to be tested. Prepare a stock solution of a certain concentration of cosmetic sample to be tested, and store it according to the requirements (at room temperature or refrigerated) according to the characteristics of the sample to be tested.

[0076] (4) Sodium dodecyl sulfonate (SLS) was used as a molding agent.

[0077] (5) Positive control. Dipotassium glycyrrhizate was used as the positive control sample.

[0078] (6) Preliminary experiment.

[0079] (6-1) Select 2dpf neutrophil gene mutant zebrafish with normal development.

[0080] (6-2) The experimental groups were divided into normal control group, model control group, positive control group and test sample group.

[0081] Normal control group: zebrafish were placed in standard dilution water.

[0082] Model control group: Zebrafish were placed in standard diluted water containing SLS.

[0083] Test sample group: Zebrafish were placed in standard dilution water containing SLS and test sample stock solution.

[0084] Positive control group: Zebrafish were placed in standard diluted water containing SLS and dipotassium glycyrrhizate.

[0085] Each control group was incubated with zebrafish using a six-well plate, with 15 zebrafish incubated per well.

[0086] (6-3) Wrap the culture vessel with aluminum foil and incubate at (28±2)℃ for 18 hours.

[0087] (6-4) Observe and record the mortality and other toxic effects of zebrafish. The highest concentration at which no malformation, death or other toxic effects of zebrafish were observed was taken as the maximum detection concentration of the sample to be tested.

[0088] (7) Formal experiment.

[0089] (7-1) The experimental groups are the same as the preliminary experiment. The three concentrations in the test sample group are diluted by 2 times, for example, 0.5%, 0.25%, and 0.125%. The highest test concentration shall not be higher than the maximum detection concentration.

[0090] (7-2) Repeat the operations in the preliminary experiment.

[0091] (7-3) Zebrafish were immobilized using methylcellulose and photographed under a fluorescence microscope. Side views of the zebrafish were taken for image analysis. Approximately 500 images from each experimental group were selected to train the segmentation model, enabling it to learn the ability to segment regions of interest. After training, for any new cosmetic sample, only 10 images from each experimental group were selected as the valid experimental data for that group. These 10 images from each group, totaling 40 images, served as the original zebrafish images.

[0092] S102. Preprocess the original zebrafish image.

[0093] The fluorescently labeled zebrafish images taken under a microscope are low in brightness and generally dark, making it difficult to clearly see the zebrafish's outline. In order to clearly show the distinguishable yolk sac and the corresponding dorsal region above the lateral line, this embodiment preprocesses the original zebrafish images.

[0094] Brightness enhancement helps improve the contrast between the foreground and background, making it easier to identify regions of interest (ROIs). By preprocessing the brightness of an image, it can make the image clearer and easier to view, eliminate some image noise, thereby improving image quality and providing better input for subsequent image segmentation tasks.

[0095] This embodiment's brightness processing converts the image from the RGB (Red, Green, Blue) color space to the HSV (Hue, Saturation, Value) color space. The HSV color space separates color information from brightness information. By increasing the value of the V channel, the overall brightness of the image can be significantly improved, enhancing its visibility. Specifically, the V channel matrix of the HSV format image is extracted, and each element is multiplied by a constant 1.3 to increase the value of the V channel matrix. The S channel (saturation channel) controls the intensity or purity of colors in the image. Increasing the S channel value in the same way enhances color saturation, making the image more vibrant and vivid. The original zebrafish image after brightness processing is then converted from HSV to RGB space. Figure 4 and Figure 5 As can be seen, the image after brightness processing can clearly distinguish the zebrafish from the black background, while also clearly identifying the fluorescent cells inside the zebrafish and clearly identifying the edges of the zebrafish's abdomen and back, which facilitates subsequent processing and model training.

[0096] S103. Based on the preprocessed original zebrafish images, obtain the training dataset.

[0097] To accurately measure the soothing effect of cosmetics, it is necessary to observe the distribution and quantity of transgenic neutrophils in zebrafish. Fluorescent granules in the zebrafish yolk sac and the corresponding dorsal region above the lateral line are crucial for subsequent analysis; therefore, correctly segmenting the yolk sac and the corresponding dorsal region above the lateral line (the target area) in each zebrafish image is essential for accurate particle counting. Binary masks can accurately delineate the target area in the image.

[0098] To obtain the binary mask for each image, engineers first outline the target region in the image with a thin red line, creating an initial line contour. Then, morphological operations such as erosion and dilation are used to further refine the line contour, removing potential noise or artifacts. These operations help smooth edges and fill any small gaps in the line contour. Finally, the line contour is filled with color and binarized to obtain the binary mask. The binary mask is a crucial component of semantic segmentation models, helping to classify pixels in an image as foreground or background. The binary mask serves as the label for the corresponding image.

[0099] Approximately 500 original zebrafish images were obtained from various experimental groups. After preprocessing, the preprocessed original zebrafish images and labels were used as samples in the training dataset.

[0100] S104. Train the semantic segmentation model using the training dataset.

[0101] During the training phase, preprocessed original zebrafish images are used as input data for the semantic segmentation model. The model extracts and learns features from the input data, outputting a binary mask. This binary mask aims to accurately delineate the yolk sac and the corresponding dorsal region above the lateral line in the zebrafish image. The binary mask output by the model is then compared with the binary mask corresponding to the preprocessed original zebrafish image. The difference between the two is calculated using the cross-entropy loss function, and based on this, the model parameters are updated and optimized using the backpropagation algorithm. This iterative training process gradually improves the accuracy of the output binary mask, enabling it to more accurately represent the real-world situation and achieve a more precise definition of the target region.

[0102] S105. After preprocessing, the zebrafish image to be detected is input into the trained semantic segmentation model to obtain an image containing only the target region.

[0103] The original zebrafish image to be detected first undergoes image preprocessing, and then is input into a pre-trained semantic segmentation model with accurate segmentation capabilities. The model outputs a binary mask (target region mask), which focuses on key target regions in the zebrafish image, such as the yolk sac and the corresponding dorsal region above the lateral line. The target region mask is then multiplied pixel-wise with the original image to obtain a zebrafish image containing only the target regions.

[0104] This step combines image preprocessing, deep learning models, and pixel-level masking techniques to achieve efficient and accurate object detection.

[0105] The process of transforming a preprocessed zebrafish image into an image containing only the target region can be found in [reference needed]. Figure 6 .

[0106] S106. Extract the number of fluorescent particles from an image containing only the target region; evaluate the soothing effect of the cosmetic based on the number of fluorescent particles.

[0107] This step automatically extracts the number of fluorescent particles from an image containing only the target area; based on the number of fluorescent particles, the soothing effect of cosmetics can be evaluated quickly and accurately.

[0108] (1) Extract the number of fluorescent particles from an image containing only the target region.

[0109] Since fluorescent images are predominantly composed of green pixels, the green channel is first extracted from the image's RGB channels and converted to grayscale. Then, fluorescent particles corresponding to pixels with intensity exceeding a set threshold are filtered out. Next, the SimpleITK toolkit is used to set a threshold, treating connected components of high-brightness pixel clusters as a single fluorescent particle, thus obtaining the number of fluorescent particles in the image. (See reference...) Figure 7 .

[0110] (2) Evaluate the soothing effect of cosmetics based on the number of fluorescent particles.

[0111] In this embodiment, two cosmetic samples were selected from the test sample group: Sample 16 and Sample 15. Ten images were randomly selected from the zebrafish fluorescence images taken in each group as valid experimental data, and the number of fluorescent particles N extracted from each group is shown in Table 1.

[0112] Table 1. Number of fluorescent particles N extracted from each group

[0113]

[0114] Step 1: Calculate the mean (Mean) and standard error (SE) of neutrophil count for each group.

[0115] Normal control group:

[0116] Data X i The numbers are: 11, 13, 13, 17, 17, 17, 18, 20, 21, 21

[0117] Average value calculation:

[0118] Standard error calculation: First calculate the variance. The formula for calculating variance is: , where n is the number of data points.

[0119]

[0120]

[0121] Model control group:

[0122] Data X i The numbers are: 31, 32, 36, 37, 39, 39, 40, 41, 43, 44

[0123] Average value calculation:

[0124] Standard error calculation:

[0125] Sample 16:

[0126] Data X i The numbers are: 21, 23, 24, 25, 26, 26, 26, 29, 29, 30

[0127] Average value calculation:

[0128] Standard error calculation:

[0129] Sample 15:

[0130] Data X i The numbers are: 19, 23, 23, 24, 24, 25, 26, 28, 28, 28

[0131] Average value calculation:

[0132] Standard error calculation:

[0133] Positive control group:

[0134] Data X i The numbers are: 15, 15, 16, 16, 19, 20, 22, 22, 26, 28

[0135] Average value calculation:

[0136] Standard error calculation:

[0137] Step Two: Conditions for Experimental Validity

[0138] There was a statistically significant difference in the number of neutrophils between the normal control group and the model control group, and the difference in the mean was greater than twice the within-group standard deviation (SD) of the normal control group, proving that the experiment was effective.

[0139] There was a statistically significant difference in the number of neutrophils between the positive control group and the model control group, and the difference in mean values ​​was greater than twice the within-group standard deviation (SD) of the model control group, proving that the experiment was effective.

[0140] Step 3: Perform statistical analysis

[0141] Based on the validity of the experiment, the following statistical analysis was performed:

[0142] Test whether the data in each group conform to a normal distribution. Use SPSS's statistical analysis function and set the confidence interval to 95%.

[0143] Table 2 shows the test results for whether the data in each group conforms to a normal distribution.

[0144]

[0145] When the sample size is 3 ≤ n ≤ 5000, the results are based on the Shapiro-Wilk (W test). In the normal test results, since the sample size is 10, the results... like Table 2 shows the results. At a confidence interval of 95% (α=0.05), P>0.05, therefore the null hypothesis is not rejected, and the variable is considered to follow a normal distribution. Next, a t-test will be used to compare the model with the control group.

[0146] The t-values ​​for the model control group and other groups are calculated using the following formula:

[0147]

[0148] in, , The average values ​​of the model control group and other groups are given, where n1 and n2 are the number of data points in each group, and s is the average value of the model control group and other groups. p This is the pooled standard deviation. p The calculation formula is:

[0149]

[0150] in, , Let V be the variance of the two groups.

[0151] Taking sample 16 as an example, the pooled standard deviation of the model control group and sample 16 is first calculated:

[0152]

[0153] Then calculate the t value:

[0154]

[0155] Next, calculate the degrees of freedom:

[0156]

[0157] Finally, consult the t-distribution table. When the degrees of freedom d... f When t=18, the t-distribution table shows that the p-value for t=7.55 is <0.05.

[0158] The statistical analysis of the data for sample 15 was similar.

[0159] The calculated p-values ​​for both experimental groups (sample 16 and sample 15) and the model control group were <0.05, demonstrating a significant difference between the sample groups and the model control group. This indicates that the cosmetic samples are effective.

[0160] Step 4: Calculation of Soothing Effect

[0161] Assuming the cosmetic sample is effective, calculate the soothing effect of the sample being tested:

[0162]

[0163] The soothing effects of test sample 16 and test sample 15 were calculated to be 57.48% and 62.62%, respectively.

[0164] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware, and the corresponding program can be stored in a computer-readable storage medium.

[0165] It should be noted that although the method operations of the above embodiments are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the order of execution of the described steps may be changed. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0166] Example 2:

[0167] like Figure 8 As shown, this embodiment provides a cosmetic soothing efficacy evaluation system based on a deep learning model. The system includes an image acquisition module 801, an image preprocessing module 802, a training dataset acquisition module 803, a model training module 804, and a cosmetic evaluation module 805, wherein:

[0168] Image acquisition module 801 is used to set up control group and test sample group. After each group is cultured for the same time, zebrafish are fixed with methylcellulose, and side views of zebrafish in each group are taken under a fluorescence microscope as original images of zebrafish. The control group includes normal control group, model control group and positive control group. The normal control group is zebrafish placed in standard dilution water, the model control group is zebrafish placed in standard dilution water containing modeling agent, and the positive control group is zebrafish placed in standard dilution water containing modeling agent and dipotassium glycyrrhizate. The test sample group is zebrafish placed in standard dilution water containing modeling agent and test sample stock solution. The test sample stock solution is a solution prepared by the cosmetic to be tested. The zebrafish in each group are normally developed 2dpf neutrophil gene mutant zebrafish.

[0169] Image preprocessing module 802 is used to preprocess the original zebrafish image to improve the overall brightness of the image;

[0170] The training dataset acquisition module 803 is used to obtain a binary mask based on the preprocessed original zebrafish image; the preprocessed original zebrafish image and the corresponding binary mask are used as samples in the training dataset.

[0171] Model training module 804 is used to train the semantic segmentation model using the training dataset;

[0172] The cosmetic evaluation module 805 is used to input the zebrafish images to be tested in each group into the trained semantic segmentation model after preprocessing, so as to obtain images containing only the target region in each group; extract the number of fluorescent particles in each group based on the images containing only the target region in each group; and evaluate the soothing effect of the cosmetic to be tested based on the number of fluorescent particles in each group.

[0173] The specific implementation of each module in this embodiment can be found in Embodiment 1 above, and will not be repeated here. It should be noted that the system provided in this embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure can be divided into different functional modules to complete all or part of the functions described above.

[0174] Example 3:

[0175] This embodiment provides a terminal device, which can be a computer, such as... Figure 9 As shown, the processor 902, memory, input device 903, display 904, and network interface 905 are connected via system bus 901. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium 906 and an internal memory 907. The non-volatile storage medium 906 stores the operating system, computer programs, and database. The internal memory 907 provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. When the processor 902 executes the computer programs stored in the memory, it implements the cosmetic soothing efficacy evaluation method based on a deep learning model described in Embodiment 1 above.

[0176] Example 4:

[0177] This embodiment provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the cosmetic soothing efficacy evaluation method based on a deep learning model described in Embodiment 1 above.

[0178] It should be noted that the computer-readable storage medium in this embodiment can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0179] Example 5:

[0180] This embodiment provides a computer program product, including a computer program that, when executed by a processor, implements the cosmetic soothing efficacy evaluation method of Embodiment 1 described above.

[0181] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope disclosed in the present invention, based on the technical solution and inventive concept of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. A method for evaluating the soothing efficacy of cosmetics based on a deep learning model, characterized in that, The method includes: Set up a normal control group: Place zebrafish in standard dilution water; Set up a model control group: Place zebrafish in standard diluted water containing the modeling agent; Setting up the test sample group: Zebrafish were placed in standard dilution water containing a modeling agent and a test sample stock solution; the test sample stock solution was a solution prepared from the cosmetic product to be tested; Set up a positive control group: Place zebrafish in standard diluted water containing modeling agent and dipotassium glycyrrhizate; The zebrafish in each group were normally developed 2dpf neutrophil gene mutant zebrafish; After each group was cultured for the same period of time, zebrafish were fixed with methylcellulose and side views of each group of zebrafish were taken under a fluorescence microscope as the original images of zebrafish. The original zebrafish image is preprocessed to improve the overall brightness of the image; Based on the preprocessed original zebrafish image, a binary mask is obtained; the preprocessed original zebrafish image and the corresponding binary mask are used as samples in the training dataset. The semantic segmentation model is trained using the training dataset; The zebrafish images to be detected in each group are preprocessed and then input into the trained semantic segmentation model to obtain images containing only the target region in each group; the number of fluorescent particles in each group is extracted based on the images containing only the target region in each group; the soothing effect of the cosmetic product to be tested is evaluated based on the number of fluorescent particles in each group. The step of obtaining a binary mask based on the preprocessed original zebrafish image includes: The target area is circled in the preprocessed original zebrafish image with a thin line to obtain a preliminary line outline; the target area is two regions on the back of the zebrafish, namely the yolk sac and the corresponding part above the lateral line. Based on the initial line outline, morphological operations are used to refine the line outline and remove existing noise or artifacts. Finally, the line outline is filled with color and binarized to obtain a binary mask; The step of training the semantic segmentation model using the training dataset includes: The preprocessed original zebrafish image is input into the semantic segmentation model, which outputs a binary mask. The output binary mask is compared with the binary mask corresponding to the preprocessed original zebrafish image. The difference between the two is calculated using the cross-entropy loss function, and backpropagation is performed based on this difference to update and optimize the model parameters.

2. The method for evaluating the soothing efficacy of cosmetics according to claim 1, characterized in that, The process involves preprocessing the zebrafish images to be detected in each group and then inputting them into a trained semantic segmentation model to obtain images in each group that contain only the target region, including: The zebrafish images to be detected in each group are preprocessed and then input into the trained semantic segmentation model, which outputs the corresponding binary mask. The binary mask is multiplied pixel-level with the corresponding preprocessed zebrafish image to be detected to obtain a zebrafish image containing only the target region.

3. The method for evaluating the soothing efficacy of cosmetics according to any one of claims 1 to 2, characterized in that, The preprocessing includes: Convert the image from RGB space to HSV space; Increase the pixel values ​​of the V and S channels of the image to obtain an image with increased brightness; Convert the image with increased brightness from HSV space to RGB space; The images are either original zebrafish images or zebrafish images to be detected in each group.

4. The method for evaluating the soothing efficacy of cosmetics according to any one of claims 1 to 2, characterized in that, The step of extracting the number of fluorescent particles in each group based on images containing only the target region includes: Extract the green channel from the RGB channels of an image containing only the target region and convert it to a grayscale image; filter out pixels in the grayscale image that are larger than a set threshold, and then use the SimpleITK toolkit to treat connected pixels as a fluorescent particle, thereby obtaining the number of fluorescent particles in the image.

5. The method for evaluating the soothing efficacy of cosmetics according to any one of claims 1 to 2, characterized in that, The evaluation of the soothing effect of the tested cosmetics based on the number of fluorescent particles in each group includes: Using the model control group as a standard, the number of fluorescent particles in other groups is compared. If there are significant differences between the data, the experimental data are considered valid. If the experimental data are valid, the soothing effect of the tested cosmetic can be calculated by using the number of fluorescent particles in the model control group, normal control group, and test sample group.

6. A cosmetic soothing efficacy evaluation system based on a deep learning model, characterized in that, The system includes: The image acquisition module is used to set up control and test sample groups. After culturing for the same period, zebrafish in each group are fixed with methylcellulose, and side views of the zebrafish in each group are captured under a fluorescence microscope as the original images of the zebrafish. The control group includes a normal control group, a model control group, and a positive control group. The normal control group consists of zebrafish placed in standard dilution water, the model control group consists of zebrafish placed in standard dilution water containing a modeling agent, and the positive control group consists of zebrafish placed in standard dilution water containing a modeling agent and dipotassium glycyrrhizate. The test sample group consists of zebrafish placed in standard dilution water containing a modeling agent and a test sample stock solution, wherein the test sample stock solution is a solution prepared from the cosmetic to be tested. The zebrafish in each group are normally developed 2dpf neutrophil gene mutant zebrafish. An image preprocessing module is used to preprocess the original zebrafish image to improve the overall brightness of the image; The training dataset acquisition module is used to obtain the binary mask based on the preprocessed original zebrafish image; the preprocessed original zebrafish image and the corresponding binary mask are used as samples in the training dataset. The model training module is used to train the semantic segmentation model using the training dataset; The cosmetic evaluation module is used to input the zebrafish images to be tested in each group into the trained semantic segmentation model after preprocessing, so as to obtain images containing only the target region in each group; extract the number of fluorescent particles in each group based on the images containing only the target region in each group; and evaluate the soothing effect of the cosmetic based on the number of fluorescent particles in each group. The step of obtaining a binary mask based on the preprocessed original zebrafish image includes: The target area is circled in the preprocessed original zebrafish image with a thin line to obtain a preliminary line outline; the target area is two regions on the back of the zebrafish, namely the yolk sac and the corresponding part above the lateral line. Based on the initial line outline, morphological operations are used to refine the line outline and remove existing noise or artifacts. Finally, the line outline is filled with color and binarized to obtain a binary mask; The step of training the semantic segmentation model using the training dataset includes: The preprocessed original zebrafish image is input into the semantic segmentation model, which outputs a binary mask. The output binary mask is compared with the binary mask corresponding to the preprocessed original zebrafish image. The difference between the two is calculated using the cross-entropy loss function, and backpropagation is performed based on this difference to update and optimize the model parameters.

7. A terminal device, comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the cosmetic soothing efficacy evaluation method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the cosmetic soothing efficacy evaluation method according to any one of claims 1 to 5.

Images