Image-based content identification method and device

Abstract

The invention discloses an image-based content identification method and device. The method comprises the following steps: acquiring an image to be identified, and preprocessing the image to be identified; extracting feature information in the preprocessed image, wherein the feature information comprises edge features, texture features and color features; and based on the feature information, performing image classification by using an image recognition model to generate a corresponding predicted content label, and based on the predicted content label, performing content recognition on the to-be-recognized image and outputting a recognition result. According to the invention, the technical problem of inaccurate identification result of content identification based on images in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to an image-based content recognition method and device. Background Art

[0002] With the rapid development of virtual reality and artificial intelligence technologies, virtual digital humans have been widely used in many fields, such as virtual assistants, virtual entertainment, education, and medical care. Traditional virtual digital human interaction technology is mainly based on voice input for interaction. However, as users' demand for virtual experience increases, single voice interaction can no longer meet the increasingly complex application scenarios. Especially in some image-based application scenarios, such as virtual image design and virtual scene generation, users need to interact with virtual digital humans more intuitively and efficiently through images.

[0003] At present, image-based content recognition technology has made some progress in the field of computer vision, and can identify feature information such as edges, textures, colors, etc. in images, and perform classification and content recognition. However, existing technologies still have certain limitations in terms of image content recognition accuracy and adaptability, especially in the interaction with virtual digital humans. Existing image recognition methods may not be able to accurately understand the image content, resulting in unsatisfactory interaction results.

[0004] Therefore, there is an urgent need for an improved image-based content recognition method that can accurately identify image content by inputting images during the interaction with virtual digital humans, generate corresponding prediction labels, and combine virtual digital human technology to enable users to interact with virtual digital humans naturally, accurately and efficiently through images.

[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0006] The embodiments of the present invention provide a method and device for image-based content recognition, so as to at least solve the technical problem of inaccurate recognition results of image-based content recognition in the prior art.

[0007] According to one aspect of an embodiment of the present invention, there is provided an image-based content recognition method, comprising: acquiring an image to be recognized, and preprocessing the image to be recognized; extracting feature information from the preprocessed image, the feature information comprising edge features, texture features, and color features; based on the feature information, performing image classification using an image recognition model to generate a corresponding predicted content label, and based on the predicted content label, performing content recognition on the image to be recognized and outputting a recognition result.

[0008] According to another aspect of an embodiment of the present invention, there is also provided an image-based content recognition device, including: a preprocessing module, configured to obtain an image to be recognized and preprocess the image to be recognized; an extraction module, configured to extract feature information from the preprocessed image, wherein the feature information includes edge features, texture features and color features; and a recognition module, configured to classify the image based on the feature information using an image recognition model, generate a corresponding category label, and based on the category label, perform content recognition on the image to be recognized and output a recognition result.

[0009] In the embodiment of the present invention, an image to be identified is obtained, and the image to be identified is preprocessed; feature information in the preprocessed image is extracted, and the feature information includes edge features, texture features, and color features; based on the feature information, an image recognition model is used to classify the image, and a corresponding predicted content label is generated, and based on the predicted content label, content recognition is performed on the image to be identified and a recognition result is output. The above solution solves the technical problem of inaccurate recognition results of content recognition based on images in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0011] Figure 1 is a flow chart of a method for image-based content recognition according to an embodiment of the present invention;

[0012] Figure 2 is a flow chart of another image-based content recognition method according to an embodiment of the present invention;

[0013] Figure 3 is a flow chart of a denoising method according to an embodiment of the present invention;

[0014] Figure 4 is a flowchart of a method for training an image recognition model according to an embodiment of the present invention;

[0015] Figure 5 is a schematic structural diagram of an image-based content recognition device according to an embodiment of the present invention;

[0016] Figure 6 A schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0018] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0019] According to an embodiment of the present invention, a method embodiment of an image-based content recognition method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0020] Figure 1 is an image-based content recognition method according to an embodiment of the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0021] Step S102: acquiring an image to be identified and preprocessing the image to be identified.

[0022] Acquire an image to be identified, generate a Gaussian kernel and normalize the Gaussian kernel; pre-process the image to be identified based on the pixel position of the image to be identified, the offset of the pixel position relative to the center of the normalized Gaussian kernel, the horizontal gradient and vertical gradient of the pixel position, the gradient weight, and the noise intensity parameter.

[0023] Step S104: extracting feature information from the preprocessed image, wherein the feature information includes edge features, texture features, and color features.

[0024] Step S106, based on the feature information, use the image recognition model to perform image classification, generate corresponding predicted content labels, and based on the predicted content labels, perform content recognition on the image to be recognized and output a recognition result.

[0025] First, the teacher model is used to predict the training data to generate soft labels for the training data. For example, the teacher model is used to encode the training data to obtain embedded features of the training data; the teacher model is used to perform content recognition on the training data and the embedded features to obtain the soft labels for the training data.

[0026] Next, the soft label generated by the teacher model is used to train the student model, and the trained student model is used as the image recognition model. For example, the training data is subjected to data enhancement to obtain the enhanced training data, and the enhanced training data and the soft label are input into the student model to obtain the predicted content label of the enhanced training data output by the student model and the confidence of the predicted content label; based on the predicted content label and the confidence of the predicted content label, the student model is iteratively trained. For example, the following is executed in a loop until the loss value reaches a preset loss threshold: the difference between the predicted content label and the soft label output by the student model is calculated to obtain the cross entropy loss, and a weighted loss calculation is performed based on the cross entropy loss and the confidence to obtain the loss value; based on the loss value, the gradient is calculated using the back propagation algorithm, and the parameters of the student model are updated using the optimization algorithm based on the gradient.

[0027] Figure 2 is another image-based content recognition method according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:

[0028] Step S202: image preprocessing.

[0029] The preprocessing method can be as follows:

[0030] First, the image is normalized to ensure that all pixel values ​​are scaled to the range of [0,1]. This helps eliminate brightness differences between images, thereby reducing interference factors in model training and accelerating its convergence. Subsequently, data augmentation is performed to enrich the diversity of training data. For example, randomly rotating images at different angles; randomly zooming in or out on images to accommodate objects of different sizes; randomly translating images horizontally or vertically to simulate objects in different positions; randomly cropping parts of the image to highlight specific areas; randomly flipping images, either horizontally or vertically, to enhance the model's ability to adapt to changes in direction; and randomly adjusting the brightness, contrast, and saturation of the image to cope with different lighting conditions. These augmentation strategies work together to prevent the model from overfitting and improve its recognition effect in real scenes.

[0031] Next, the image is denoised. First, a Gaussian kernel is generated and normalized. Then, denoising is performed based on the position of each pixel in the image and its corresponding pixel value offset, horizontal and vertical gradients, and noise intensity parameters. The noise intensity parameter is used to adjust the sensitivity of the denoising process to the difference in pixel values. Specifically, the denoising method is as follows: Figure 3 The following steps are shown:

[0032] Step S2022, generating a Gaussian kernel and normalizing the Gaussian kernel.

[0033] Generate a Gaussian kernel and normalize it. Assume that the image has been loaded as a matrix I, where I(i,j) represents the pixel value of the image at position (i,j). First, select the size k of the Gaussian kernel (such as 3x3, 5x5, 7x7, etc.) and the standard deviation σ. A larger Gaussian kernel will increase the blurring effect, but may lose more details; the larger the standard deviation σ, the more significant the blurring effect. Next, generate the Gaussian kernel matrix K based on the Gaussian function.

[0034]

[0035] in, σ is the standard deviation, which determines the width of the Gaussian distribution, (x,y) is the coordinate of the pixel in the image relative to the center of the kernel, and k is the size of the Gaussian kernel.

[0036] For a k×k Gaussian kernel, whose center is (k / / 2, k / / 2), the Gaussian kernel is normalized to ensure that its weighted sum is 1:

[0037]

[0038] Wherein, m represents the offset of the currently processed pixel position in the horizontal direction, and n represents the offset of the currently processed pixel position in the vertical direction.

[0039] Step S2024, perform denoising processing.

[0040] Perform denoising. Multi-scale Gaussian blur captures the features of the image at different scales by applying multiple Gaussian kernels with different standard deviations σ\sigmaσ, thereby enhancing the denoising effect. The improved denoising formula is:

[0041]

[0042] Among them, I(i,j) is the pixel value of the original image at position (i,j), m,n are the offsets of the horizontal and vertical coordinates relative to the center of the Gaussian kernel, tx is the horizontal gradient of the image at each pixel position, ty is the vertical gradient of the image at each pixel position, β is the gradient weight, and N is the noise intensity parameter, which is used to control the sensitivity of the pixel value difference of the image during the denoising process. It is obtained based on the noise intensity η, the pixel value of the image at position (i,j), and the offset value of the pixel value of the image at position (i,j).

[0043]

[0044] This embodiment adopts an improved denoising processing formula, and the pixel value I(i,j) of the original image at position (i,j) is processed in detail, taking into account the offsets m and n relative to the center of the Gaussian kernel, as well as the horizontal gradient tx and vertical gradient ty of each pixel position. By adjusting the gradient weight β and the noise intensity parameter N, the sensitivity to the difference in pixel values ​​during the denoising process can be accurately controlled, thereby better preserving image details, reducing the impact of noise, and improving image quality.

[0045] Step S204: training an image recognition model.

[0046] like Figure 4 As shown, the steps of the method for training an image recognition model include:

[0047] Step S2042: Use the teacher model to predict the training data and generate soft labels for the training data.

[0048] The teacher model is a high-performance deep learning model used to generate soft labels and guide the student model to learn during the knowledge distillation process. Its main function is to extract deep features of training data and perform content recognition on these features to generate information with confidence.

[0049] 1) Model selection and initialization.

[0050] Depending on the image recognition task, select a deep learning architecture (such as ResNet, Inception, or VGG). Then, load the weights of the pre-trained model to leverage its learned features on a large-scale dataset such as ImageNet.

[0051] 2) Prepare the dataset.

[0052] Collect a diverse and representative training dataset. Make sure the dataset contains enough samples to cover all categories. Perform appropriate preprocessing on the data, including normalization and data augmentation (such as random cropping, rotation, flipping, scaling, etc.) to improve the generalization ability of the model.

[0053] 3) Train the model.

[0054] In the specific implementation process of the training model, the back propagation algorithm is first used to optimize the model parameters. Select a deep learning framework (such as TensorFlow or PyTorch), build the required model architecture and initialize the parameters. Next, define the loss function. This embodiment uses the cross entropy loss function to measure the difference between the predicted result and the true label. The calculation formula of the cross entropy loss is:

[0055]

[0056] Among them, Loss represents the loss value, which is used to measure the difference between the model prediction and the true label. The goal is to minimize the loss value to improve the performance of the model. N is the total number of training samples. It is used to normalize the loss to ensure that the loss value is independent of the number of samples. I is the index, indicating the i-th training sample. C is the total number of categories, indicating the number of different categories that the model needs to predict. y ij Represents the indicator function of the true label. When sample i belongs to category j, y ij =1; otherwise, y ij =0, which is used to indicate the category that the model should focus on. is the model's predicted probability that sample i belongs to category j, indicating the model's confidence that the sample belongs to each category. T represents the temperature parameter, which is used to adjust the probability distribution of the model output. When T>1, the probability distribution is smoother; when T<1, the probability distribution is sharper. This allows the model to adaptively explore or converge at different stages. It represents the sum of all categories k and is used to calculate the sum of the predicted probabilities of all categories of sample i by the model. It represents the model's predicted probability that the i-th training sample belongs to the k-th category.

[0057] During the training process, the dataset is divided into a training set and a validation set, and the training data is further divided into mini-batches. In each training cycle (epoch), forward propagation is performed for each mini-batch to calculate the output prediction value of the model. Next, the loss value of the batch is evaluated using the defined loss function. Then, through the back-propagation algorithm, the chain rule is used to calculate the gradient of the loss to the model parameters and update the model parameters. The updated formula is:

[0058]

[0059] Among them, θ new is the model parameter of this time, θ old is the model parameter of the previous time and η is the learning rate. In this way, it can be ensured that the model is optimized in the direction of reducing loss in each iteration.

[0060] During the training process, it is necessary to constantly monitor the performance of the model on the validation set and calculate indicators such as accuracy, recall, and F1-score to evaluate the generalization ability of the model. According to the performance of the validation set, timely adjust hyperparameters such as learning rate and batch size to ensure that the model achieves optimal performance during training. In particular, the learning rate decay strategy can effectively improve the training effect: if the validation performance does not improve within several epochs, reduce the learning rate to help the model converge better.

[0061] In addition, an early termination strategy can be implemented to prevent overfitting. For example, a threshold is set to stop training when the validation loss does not improve over multiple epochs. This strategy ensures that the model performs evenly on the training set and validation set. Finally, the weights of the model with the best performance are saved as the basis for subsequent use. Through the above steps, it is ensured that the teacher model can learn through effective back-propagation, and ultimately achieve high accuracy and good generalization ability.

[0062] 4) Generate soft labels.

[0063] After the teacher model is trained, it can be used to predict the input training data. The output of the teacher model is a probability distribution vector, and the length of each vector is N (the number of categories). For each training sample, the soft label generated by the teacher model is expressed as: p = [p1, p2, ..., pN], where pi represents the probability that the sample belongs to category i. The soft label not only provides the classification result of the sample, but also reflects the relative confidence of the sample in each category.

[0064] Step S2044, use the soft labels generated by the teacher model to train the student model, and use the trained student model as the image recognition model.

[0065] The student model can be a similar architecture to the teacher model, but is usually lightweight by reducing the number of layers, width, or depth. The initial parameters of the student model can be randomly generated or use certain pre-trained parameters (such as pre-training on a small dataset) to improve training efficiency.

[0066] When training the student model, the difference between the predicted content label and the soft label output by the student model is calculated to obtain the cross entropy loss, and a weighted loss calculation is performed based on the cross entropy loss and the confidence to obtain the loss value; based on the loss value, a back propagation algorithm is used to calculate the gradient, and based on the gradient, the parameters of the student model are updated using an optimization algorithm.

[0067] Specifically, the steps include:

[0068] 1. Calculate the difference between the predicted content label and the soft label.

[0069] This application uses an improved cross entropy loss function. The improved function combines the temperature parameter T to adjust the probability distribution of the output, so that the model can better explore the characteristics of each category in the early stage of training:

[0070]

[0071] Among them, Loss cross-entropy Represents the cross entropy loss, which is used to measure the difference between the student model output and the teacher model soft label. ik is the soft label of the teacher model, indicating the true probability that sample i belongs to category j. ik is the predicted probability output by the student model, indicating the student model's confidence in predicting that sample i belongs to category k. K is the total number of categories, indicating the number of different categories that the model needs to predict. S is the temperature parameter, which is used to adjust the smoothness of the output probability distribution. β ij It is the predicted probability of the student model that sample i belongs to category j, reflecting the prediction confidence of the student model for each category on this sample.

[0072] The embodiment of the present application can effectively measure the difference between the probability distribution of the model output and the true label through the above-mentioned cross entropy loss function, thereby prompting the model to learn more accurate classification decisions. Secondly, the cross entropy loss function has a larger penalty for low-probability events and provides a relatively smooth gradient, which makes the optimization process more stable and can avoid the problem of local optimality. In addition, the cross entropy loss function can often accelerate the convergence of the model because it can provide a strong gradient signal when the prediction is close to the true label.

[0073] 2) Weighted loss calculation based on cross entropy loss and confidence

[0074] When calculating the weighted loss, not only can the confidence ci be used, but also a dynamic weighting mechanism based on sample importance can be combined. By introducing a sample importance score wi, which is calculated based on the error rate or confidence of the student model for the sample, the loss function can be modified as follows:

[0075] Loss weighted =c i ·w i ·Loss cross-entropy

[0076] Here, wi can be calculated by evaluating the performance of the sample in past training, so that the model pays more attention to samples that perform poorly in learning.

[0077] 3) Calculate the gradient.

[0078] When calculating the gradient, in addition to using the back-propagation algorithm to calculate the gradient of the loss, a gradient clipping strategy can also be introduced to prevent gradient explosion during training. By setting the threshold ∈, the gradient calculation process can be adjusted:

[0079]

[0080] in, It represents the gradient of the weighted loss function with respect to the model parameter θ, guiding the direction and magnitude of parameter update. is the norm of the gradient, which is used to detect the gradient size to achieve gradient clipping. ∈ is the threshold of gradient clipping. When the norm of the gradient exceeds the threshold, adjustment is made to prevent gradient explosion.

[0081] 4) Update the parameters of the student model.

[0082] In the parameter update step, combined with the dynamic learning rate adjustment mechanism, the learning rate can be adaptively adjusted according to the change in loss of each batch. For example, when the loss value fails to decrease significantly in several consecutive training cycles, the learning rate can be reduced to promote a more robust convergence of the model:

[0083]

[0084] Among them, η new is the updated learning rate, which determines the magnitude of parameter updates. base Is the basic learning rate, the initial learning rate value set. Loss prev is the loss value of the previous training cycle. current is the loss value of the current training cycle.

[0085] Finally, the updated learning rate is used to update the parameters:

[0086]

[0087] Step S206, using an image recognition model to recognize the intent in the image.

[0088] Input the image into the trained image recognition model and output the intent corresponding to the image.

[0089] The present application also provides an image-based content recognition device, such as Figure 5 As shown, it includes: a preprocessing module 52, configured to obtain an image to be identified and preprocess the image to be identified; an extraction module 54, configured to extract feature information from the preprocessed image, the feature information including edge features, texture features and color features; a recognition module 56, configured to classify the image based on the feature information using an image recognition model, generate a corresponding category label, and based on the category label, perform content recognition on the image to be identified and output a recognition result.

[0090] It should be noted that the image-based content recognition device provided in the above 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 of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the image-based content recognition device provided in the above embodiment and the image-based content recognition method embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0091] Figure 6 FIG. 1 shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present disclosure. It should be noted that: Figure 6 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0092] like Figure 6 As shown, the electronic device includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage part 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for system operation are also stored. The CPU 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0093] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage section 1008 as needed.

[0094] In particular, according to an embodiment of the present disclosure, the process described below with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the central processing unit (CPU) 1001, various functions defined in the method and apparatus of the present application are executed. In some embodiments, the electronic device may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0095] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for content recognition based on an image, characterized in that: include: Acquire an image to be identified, and preprocess the image to be identified; Extracting feature information from the preprocessed image, wherein the feature information includes edge features, texture features, and color features; Based on the feature information, an image recognition model is used to perform image classification, a corresponding predicted content label is generated, and based on the predicted content label, content recognition is performed on the image to be recognized and a recognition result is output.

2. The method according to claim 1, characterized in that: The image recognition model is obtained through the following training, including: Using the teacher model to predict the training data, and generating soft labels for the training data; The soft labels generated by the teacher model are used to train the student model, and the trained student model is used as the image recognition model.

3. The method according to claim 2, characterized in that The teacher model is used to predict the training data to generate soft labels for the training data, including: Encoding the training data through the teacher model to obtain embedded features of the training data; The training data and the embedded features are subjected to content recognition by the teacher model to obtain the soft labels of the training data.

4. The method according to claim 2, characterized in that: Using the soft labels generated by the teacher model to train a student model, the method further comprises: Performing data enhancement on the training data to obtain enhanced training data, inputting the enhanced training data and the soft label into the student model, and obtaining a predicted content label of the enhanced training data and a confidence level of the predicted content label output by the student model; The student model is iteratively trained based on the predicted content label and the confidence of the predicted content label.

5. The method according to claim 4, characterized in that Based on the predicted content label and the confidence of the predicted content label, the student model is iteratively trained, including: looping the following until the loss value reaches a preset loss threshold: Calculating the difference between the predicted content label output by the student model and the soft label to obtain a cross entropy loss, and performing a weighted loss calculation based on the cross entropy loss and the confidence to obtain the loss value; Based on the loss value, a back-propagation algorithm is used to calculate the gradient, and based on the gradient, an optimization algorithm is used to update the parameters of the student model.

6. The method according to claim 1, characterized in that Preprocessing the image to be identified includes: generating a Gaussian kernel and normalizing the Gaussian kernel; The image to be identified is preprocessed based on the pixel position of the image to be identified, the offset of the pixel position relative to the normalized Gaussian kernel center, the horizontal gradient and vertical gradient of the pixel position, the gradient weight, and the noise intensity parameter.

7. An image-based content recognition device, characterized in that: include: A preprocessing module is configured to obtain an image to be identified and preprocess the image to be identified; An extraction module is configured to extract feature information from the preprocessed image, wherein the feature information includes edge features, texture features, and color features; The recognition module is configured to classify the image based on the feature information using an image recognition model to generate a corresponding category label, and based on the category label, perform content recognition on the image to be recognized and output a recognition result.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 6.

9. A computer device, characterized in that: include: Memory and processor, The memory stores a computer program; The processor is used to execute the computer program stored in the memory, and when the computer program is run, the processor is enabled to execute the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images