Shooting guidance method and device and computer equipment

Abstract

The invention relates to a shooting guidance method and device and computer equipment. The method comprises the following steps: receiving a first preview image from a client; based on the first preview image, determining first scene information, first filter information and recommended posture information through a composition recommendation model; determining posture guidance information through a posture generative adversarial network based on the first scene information, the first filter information and the recommended posture information; and sending the posture guidance information to the client, wherein the posture guidance information is used for the client to display a posture guidance line on a shooting interface so as to guide shooting. By adopting the method, the corresponding posture guidance line can be flexibly displayed based on the preview image for shooting guidance, and the flexibility and applicability of the shooting guidance can be improved.

Description

Technical Field

[0001] This application relates to the field of image technology, and in particular to a shooting guidance method, apparatus, and computer equipment. Background Technology

[0002] During the process of shooting images or videos, static guide lines can be used to guide the user. However, because static guide lines are fixed and unchanging, the flexibility of shooting guidance is relatively low. Summary of the Invention

[0003] Therefore, it is necessary to provide a shooting guidance method, device, and computer equipment to address the aforementioned technical problems.

[0004] In a first aspect, this application provides a shooting guidance method, comprising: receiving a first preview image from a client; determining first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model; determining pose guidance information based on the first scene information, the first filter information, and the recommended pose information using a pose generation adversarial network; and sending the pose guidance information to the client, wherein the pose guidance information is used by the client to display pose guidance lines on the shooting interface to guide shooting.

[0005] In one embodiment, determining the first scene information, the first filter information, and the recommended pose information based on the first preview image using a composition recommendation model includes: determining the first scene information, the first filter information, the recommended pose information, and the composition score of the first preview image based on the first preview image using a composition recommendation model; determining pose guidance information based on the first scene information, the first filter information, and the recommended pose information using a pose generation adversarial network includes: determining pose guidance information based on the first scene information, the first filter information, and the recommended pose information when the composition score is less than a score threshold.

[0006] In one embodiment, the method further includes: determining depth information of the first preview image based on the first preview image; the step of determining first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model includes: determining first scene information, first filter information, and recommended pose information based on the first preview image and the depth information of the first preview image using a composition recommendation model.

[0007] In one embodiment, receiving the first preview image from the client includes: receiving the first preview image from the client and depth information of the first preview image; determining the first scene information, the first filter information, and the recommended pose information based on the first preview image through a composition recommendation model includes: determining the first scene information, the first filter information, and the recommended pose information based on the first preview image and the depth information of the first preview image through a composition recommendation model.

[0008] In one embodiment, the method further includes: determining the subject type in the first preview image; obtaining historical data corresponding to the subject type; and determining the first scene information, the first filter information, and the recommended pose information based on the first preview image using a composition recommendation model, including: determining the first scene information, the first filter information, and the recommended pose information based on the first preview image and the historical data using a composition recommendation model.

[0009] In one embodiment, the method further includes: receiving a second preview image from the client, the second preview image being acquired by the client after displaying the pose guidance lines; determining pose guidance information through a pose generation adversarial network based on the first scene information, the first filter information, and the recommended pose information, including: determining pose guidance information and second scene information through a pose generation adversarial network based on the first scene information, the first filter information, and the recommended pose information; determining second filter information based on the second scene information; determining the similarity between the pose guidance information and the second filter information and the second preview image; and determining the second preview image as the target image if the similarity is greater than or equal to a similarity threshold.

[0010] In one embodiment, the method further includes: adjusting the pose of the second preview image based on the pose guidance information when the similarity is less than the similarity threshold; and applying a filter to the adjusted second preview image based on the second filter information to obtain the target image.

[0011] In one embodiment, the method further includes: sending the target image to the client, the target image being displayed by the client.

[0012] Secondly, this application also provides a shooting guidance device, comprising: a receiving module for receiving a first preview image from a client; a first determining module for determining first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model; a second determining module for determining pose guidance information based on the first scene information, the first filter information, and the recommended pose information using a pose generation adversarial network; and a sending module for sending the pose guidance information to the client, wherein the pose guidance information is used by the client to display pose guidance lines on the shooting interface to guide shooting.

[0013] Thirdly, this application also provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: receiving a first preview image from a client; determining first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model; determining pose guidance information based on the first scene information, the first filter information, and the recommended pose information using a pose generation adversarial network; and sending the pose guidance information to the client, wherein the pose guidance information is used by the client to display pose guidance lines on the shooting interface to guide shooting.

[0014] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0015] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0016] The aforementioned shooting guidance method, apparatus, and computer equipment receive a first preview image from a client. Based on the first preview image, a composition recommendation model determines first scene information, first filter information, and recommended pose information. Based on the first scene information, first filter information, and recommended pose information, a pose generative adversarial network (PGAN) determines pose guidance information and sends it to the client. This pose guidance information is used by the client to display pose guidance lines on the shooting interface to guide shooting. It is evident that corresponding pose guidance information can be generated based on the preview image acquired by the client, allowing the client to display pose guidance lines on the shooting interface to guide image or video shooting. Since different preview images may generate different pose guidance information, shooting guidance can be flexibly displayed based on the preview image, improving the flexibility and applicability of shooting guidance. Furthermore, since the composition recommendation model first determines the scene information, filter information, and recommended pose information corresponding to the preview image, and then the pose generative adversarial network determines the pose guidance information corresponding to the scene information, filter information, and recommended pose information, different information is determined by different models, ensuring the accuracy and reliability of information determination, thereby improving the accuracy and reliability of shooting guidance. Furthermore, using pre-trained models to determine pose guidance information can improve the efficiency of pose guidance information generation, improve shooting guidance efficiency, and thus improve user experience. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of a network architecture provided in an embodiment of this application;

[0019] Figure 2 This is a flowchart illustrating a shooting guidance method provided in an embodiment of this application;

[0020] Figure 3 This is an interactive schematic diagram of a shooting guidance method provided in an embodiment of this application;

[0021] Figure 4 This is an interactive schematic diagram of another shooting guidance method provided in an embodiment of this application;

[0022] Figure 5 This is an interactive schematic diagram of another shooting guidance method provided in the embodiments of this application;

[0023] Figure 6 This is an interactive schematic diagram of a shooting guidance method provided in an embodiment of this application;

[0024] Figure 7 This is an interactive schematic diagram of a shooting guidance method provided in an embodiment of this application;

[0025] Figure 8 This is a schematic diagram of the structure of a shooting guidance device provided in an embodiment of this application;

[0026] Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0028] To better understand the embodiments of this application, the network architecture of this application will be described below.

[0029] Figure 1 This is a schematic diagram of a network architecture disclosed in an embodiment of this application. For example... Figure 1 As shown, the network architecture may include a terminal 101 and a server 102. The terminal 101 and the server 102 communicate via a network. This network can be a local area network (LAN), a wide area network (WAN), or a hybrid of both. A data storage system can store the data that the server 102 needs to process. The data storage system can be integrated onto the server 102, or it can be located in the cloud or on another network server.

[0030] A social client is installed or configured on terminal 101. The social client has the function of calling the image acquisition device on terminal 101 to capture images and / or videos. The social client is the client corresponding to server 102. The social client can capture or obtain a preview image through the image acquisition device and send this preview image to server 102 through terminal 101. After receiving the preview image, server 102 can determine posture guidance information based on the preview image and then send the posture guidance information to the social client. After receiving the posture guidance information, the social client can generate posture guidance lines based on the posture guidance information and display the posture guidance lines on the shooting interface to guide shooting. It can then capture another preview image and send it to server 102. After receiving the other preview image, server 102 can determine whether the other preview image meets preset conditions. If it meets the expected conditions, it can determine the other preview image as the final image. If it does not meet the preset conditions, it can adjust the other preview image to obtain the final image and then send the final image to the social client. After receiving the final image, the social client can display the final image so that the user can check the image shooting effect. Image acquisition devices can be cameras, webcams, or other devices with image acquisition capabilities. Hereinafter, social media clients will be referred to simply as clients.

[0031] Terminal 101 can be, but is not limited to, electronic devices equipped with image acquisition devices, such as laptops, smartphones, tablets, IoT devices, and portable wearable devices, and can have a client installed. IoT devices can be smart TVs, smart in-vehicle devices, etc. Portable wearable devices can be smartwatches, smart bracelets, etc.

[0032] Server 102 can be a standalone physical server, a service node in a blockchain system, or a server cluster consisting of multiple physical servers. The server cluster can be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0033] Figure 2 This is a flowchart illustrating a shooting guidance method provided in an embodiment of this application. The shooting guidance method is described from the perspective of the aforementioned server. Figure 2 As shown, the shooting guidance method may include the following steps.

[0034] 201. Receive the first preview image from the client.

[0035] The first preview image can be a preview image captured or acquired by the client, or a compressed preview image captured or acquired by the client. A preview image is a scaled-down, low-resolution, or static image representing or summarizing content that has not been fully loaded or opened to view the full content. The first preview image is a Red-Green-Blue (RGB) image.

[0036] During the process of the client taking pictures or videos, the server can receive the first preview image from the client.

[0037] 202. Based on the first preview image, determine the first scene information, the first filter information, and the recommended pose information through a composition recommendation model.

[0038] The composition recommendation model is a model capable of automatically analyzing image content and providing composition optimization suggestions and scene information. The first scene information is the scene information of the first preview image, used to comprehensively describe and define the semantic and structured data of the visual environment and content presented by the first preview image. The first filter information is the filter information of the first preview image. Filter information is a set of descriptions and parameters for digital processing effects used to change at least one of the visual appearance, style, and atmosphere of an image. Recommended pose information is guiding data provided for a specific object and / or a specific shooting scene regarding at least one of body movements, positional relationships, facial expressions, and compositional arrangements. The specific object can be a person, an animal, a virtual character, or various objects.

[0039] After receiving the first preview image, the first scene information, the first filter information, and the recommended pose information can be determined based on the first preview image through a composition recommendation model.

[0040] When the first preview image is a preview image captured by the client, the first scene information, first filter information, and recommended pose information can be determined based on the first preview image using a composition recommendation model. Specifically, the first preview image can be input into the composition recommendation model, which can then output the first scene information, first filter information, and recommended pose information. Since the client does not compress the preview image, the image quality of the preview image can be guaranteed, thus ensuring the accuracy, reliability, and precision of the determined scene information, filter information, and recommended pose information. Furthermore, because the client does not compress the preview image, the server does not need to decompress the received image, reducing processing time and improving shooting guidance efficiency and user experience.

[0041] If the first preview image is a compressed preview image captured by the client, it can be decompressed to obtain a third preview image. Then, based on this third preview image, a composition recommendation model can determine the first scene information, the first filter information, and the recommended pose information. Specifically, the third preview image can be input into the composition recommendation model, which can then output the first scene information, the first filter information, and the recommended pose information. Because the client compresses the preview image, the amount of data transmitted can be reduced, thereby saving transmission resources and improving transmission efficiency.

[0042] 203. Based on the first scene information, the first filter information, and the recommended pose information, pose generation adversarial network is used to determine pose guidance information.

[0043] Generative adversarial networks (GANs) are networks used to generate realistic, diverse, and controllable pose data for humans, animals, or objects. Pose guidance information is information used to guide the subject in adjusting or achieving a specific pose.

[0044] Based on the first scene information, the first filter information, and the recommended pose information, pose generation adversarial network can determine pose guidance information. That is, the first scene information, the first filter information, and the recommended pose information can be input into the pose generation adversarial network, and the pose generation adversarial network can output pose guidance information.

[0045] 204. Send posture guidance information to the client. The posture guidance information is used by the client to display posture guidance lines on the shooting interface to guide the shooting.

[0046] Pose guidance lines are virtual lines, outlines, or markers superimposed on the shooting interface to visually display the position, direction, or trajectory of the target pose, as well as to adjust the direction. Pose guidance lines can be Augmented Reality (AR) pose guidance lines, Mixed Reality (MR) pose guidance lines, or other types of pose guidance lines.

[0047] Once the pose guidance information for the first preview image is determined, it can be sent to the client. This allows the client to generate pose guidance lines based on the information and display them on the shooting interface to guide the shooting process.

[0048] The aforementioned shooting guidance method receives a first preview image from the client. Based on this image, a composition recommendation model determines first scene information, first filter information, and recommended pose information. Then, based on these information, a pose generative adversarial network (PGAN) determines pose guidance information and sends it to the client. This guidance information is used by the client to display pose guidance lines on the shooting interface to guide shooting. As can be seen, corresponding pose guidance information can be generated based on the preview image acquired by the client, allowing the client to display pose guidance lines on the shooting interface to guide image or video shooting. Since different preview images may generate different pose guidance information, shooting guidance can be flexibly provided based on the preview image, improving the flexibility and applicability of the shooting guidance. Furthermore, since the composition recommendation model first determines the scene information, filter information, and recommended pose information corresponding to the preview image, and then the PGAN determines the corresponding pose guidance information, different information is determined by different models. This ensures the accuracy and reliability of the information determination, thereby improving the accuracy and reliability of the shooting guidance. Furthermore, using pre-trained models to determine pose guidance information can improve the efficiency of pose guidance information generation, improve shooting guidance efficiency, and thus improve user experience.

[0049] In some embodiments, determining first scene information, first filter information, and recommended pose information based on a first preview image using a composition recommendation model includes: determining the first scene information, first filter information, recommended pose information, and composition score of the first preview image based on the first preview image using a composition recommendation model; and determining pose guidance information based on the first scene information, first filter information, and recommended pose information using a pose generative adversarial network, including: determining pose guidance information based on the first scene information, first filter information, and recommended pose information using a pose generative adversarial network when the composition score of the first preview image is less than a score threshold.

[0050] The composition score of the first preview image is an automated and quantitative evaluation of its composition quality. The composition score of the first preview image is positively correlated with its shooting effect; that is, the higher the composition score, the better the shooting effect.

[0051] Based on the first preview image, a composition recommendation model can determine the first scene information, first filter information, recommended pose information, and the composition score of the first preview image. In other words, the first preview image can be input into the composition recommendation model, which can output the first scene information, first filter information, recommended pose information, and the composition score of the first preview image. Therefore, the composition recommendation model can output not only the first scene information, first filter information, and recommended pose information, but also the composition score of the first preview image.

[0052] Therefore, it can be determined whether the composition score of the first preview image is greater than or equal to a score threshold. If the composition score of the first preview image is greater than or equal to the score threshold, it indicates that the shooting effect of the first preview image is good, and there is no need to provide shooting guidance for the first preview image. The first preview image can be directly identified as the target image, reducing unnecessary processing and interaction processes, thereby saving communication resources, processing resources, and reducing power consumption. If the composition score of the first preview image is less than the score threshold, it indicates that the shooting effect of the first preview image is poor. Based on the first scene information, the first filter information, and the recommended pose information, pose guidance information can be determined through a pose generation adversarial network, and then the pose guidance information can be sent to the client. This allows the client to generate pose guidance lines based on the pose guidance information and display the pose guidance lines on the shooting interface to guide the shooting and capture a better preview image.

[0053] For example, the compositional score of an image can be represented as follows:

[0054] S=α·M_subject+β·M_background+γ·M_rule

[0055] Where S is the image composition score, M_subject is the image subject prominence, M_background is the image background simplicity, M_rule is the image composition rule compliance, α is the weight of the image subject prominence, β is the weight of the image background simplicity, and γ is the weight of the image composition rule compliance. The values ​​of α, β, and γ can be fixed or determined by the composition recommendation model, and the sum of α, β, and γ is 1. Image subject prominence measures the visual attention-grabbing nature of the subject in an image and its distinctness from the background and other elements. Image background simplicity measures the degree to which the background interferes with the subject, or the overall clutter of the image. Image composition rule compliance measures the degree to which the image composition conforms to composition rules.

[0056] In the above embodiments, based on the first preview image, a composition recommendation model is used to determine the first scene information, the first filter information, the recommended pose information, and the composition score of the first preview image. If the composition score of the first preview image is less than a score threshold, pose guidance information is determined using a pose generative adversarial network (PGAN) based on the first scene information, the first filter information, and the recommended pose information. Therefore, when the composition score of the preview image is less than the score threshold, pose guidance information can be determined using a pose generative adversarial network based on the first scene information, the first filter information, and the recommended pose information. This pose guidance information is then sent to the client so that the client can display pose guidance lines on the shooting interface to guide image or video shooting. When the composition score of the preview image is greater than or equal to the score threshold, pose guidance information is not determined, and shooting guidance is not provided. Thus, in cases of poor shooting results, the server can determine pose guidance information so that dynamic shooting guidance can be achieved through interaction between the client and the server, thereby improving the flexibility and applicability of shooting guidance. In cases of good shooting results, shooting guidance is not provided, reducing unnecessary processing and interaction processes, thereby saving communication resources, processing resources, and reducing power consumption. Furthermore, since data processing is performed by the server while information display is performed by the client, the efficiency of dynamic guidance can be improved.

[0057] In some embodiments, the shooting guidance method further includes: determining depth information of the first preview image based on the first preview image; and determining first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model, including: determining the first scene information, first filter information, and recommended pose information based on the first preview image and the depth information of the first preview image using a composition recommendation model.

[0058] The depth information of the first preview image is the straight-line distance from the real-world object to the lens (or observation point) corresponding to each pixel in the first preview image.

[0059] To improve the accuracy, reliability, and precision of the determined scene information, filter information, and recommended pose information, the server can determine the depth information of the first preview image after receiving it. The depth information can be determined using a pre-trained model or other methods. The pre-trained model can be a supervised model, an unsupervised model, or other models with equivalent functionality.

[0060] Then, based on the first preview image and its depth information, the composition recommendation model can determine the first scene information, the first filter information, and the recommended pose information. Specifically, the first preview image and its depth information can be input into the composition recommendation model, which can then output the first scene information, the first filter information, and the recommended pose information.

[0061] In the above embodiments, during the process of using the composition recommendation model to determine the first scene information, the first filter information, and the recommended pose information, not only the first preview image is input, but also the depth information of the first preview image is input. This can improve the accuracy, reliability, and precision of the determined first scene information, first filter information, and recommended pose information, thereby improving the accuracy and reliability of the shooting guidance.

[0062] In some embodiments, receiving a first preview image from a client includes: receiving the first preview image and depth information of the first preview image from the client; determining first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model, including: determining the first scene information, first filter information, and recommended pose information based on the first preview image and depth information of the first preview image using a composition recommendation model.

[0063] The depth information of the first preview image is the straight-line distance from the real-world object to the lens (or observation point) corresponding to each pixel in the first preview image.

[0064] The depth information of the first preview image can be determined by the client and sent to the server along with the first preview image. Accordingly, the server can receive the first preview image and its depth information from the client, and then determine the first scene information, the first filter information, and the recommended pose information based on the first preview image and its depth information using a composition recommendation model.

[0065] In the above embodiments, during the process of determining the first scene information, first filter information, and recommended pose information using the composition recommendation model, not only the first preview image is input, but also the depth information of the first preview image. This improves the accuracy, reliability, and precision of the determined first scene information, first filter information, and recommended pose information, thereby enhancing the accuracy and reliability of the shooting guidance. Furthermore, since the depth information of the first preview image is determined by the client, and the client can use not only the first preview image but also relevant information such as ambient brightness during the determination process, the accuracy, reliability, and precision of the determined depth information of the first preview image can be improved. This, in turn, improves the accuracy, reliability, and precision of the first scene information, first filter information, and recommended pose information determined based on the depth information of the first preview image, further enhancing the accuracy, reliability, and precision of the shooting guidance.

[0066] In some embodiments, the shooting guidance method further includes: determining the subject type in the first preview image; obtaining historical data corresponding to the subject type; and determining first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model, including: determining first scene information, first filter information, and recommended pose information based on the first preview image and historical data using a composition recommendation model.

[0067] The subject type in the first preview image refers to the type of subject in the first preview image. The subject in an image is the visual element or object that carries core information, dominates the visual focus, is most prominent, attracts the most attention, and embodies or conveys the intended meaning. The subject type in the first preview image can be a person, an animal, or an object. If the subject type is a person, it can be further subdivided into different genders and / or different age groups. If the subject type is an animal, it can be further subdivided into different animals. If the subject type is an object, it can be further subdivided into different objects. Historical data refers to historical image data. The historical data corresponding to the subject type is the historical image data corresponding to or related to the subject type.

[0068] The subject type in the first preview image can be determined. This can be achieved using a visual saliency detection algorithm. The first preview image can be input into the visual saliency detection algorithm, which outputs a saliency map. The region with the highest brightness in the saliency map can be identified as the target region. The object type within the target region of the first preview image can then be determined, thus obtaining the subject type in the first preview image.

[0069] The subject type in the first preview image can also be determined through object detection and instance segmentation. First, an object detection model can be used to determine different objects and their confidence scores in the first preview image. The subject type in the first preview image can be determined based on at least one of the following: confidence score, area, and position in the target image. For example, the subject type in the first preview image can be determined based on the object with the highest confidence score. Alternatively, the subject type can be determined based on the object with the largest area. Or, the subject type can be determined based on the object located in the central region of the first preview image. Alternatively, at least two of the following factors—conferencing the confidence score, area, and position in the target image—can be fused to obtain a fusion score for the different objects in the first preview image. The subject type in the first preview image can then be determined based on the object with the highest fusion score.

[0070] The subject type in the first preview image can also be determined through semantic segmentation and attention mechanisms. Pixel-level semantic segmentation can be performed on the first preview image to obtain a semantic segmentation map, which includes objects in different regions. Based on the first preview image, an attention heatmap can be determined, and the subject type in the first preview image can be determined based on the semantic segmentation map and the attention heatmap. The semantic segmentation map is a pixel-level, categorical representation. The attention heatmap is a spatial weight distribution map used to reflect the degree of importance or attention given to different regions of the input image.

[0071] Historical data corresponding to this subject type can be retrieved. This can be done locally, from a dedicated database, or from other servers.

[0072] It can store historical data corresponding to different subject types. It can first determine whether historical data corresponding to the subject type in the first preview image is stored. If such historical data is stored, it can retrieve the historical data corresponding to that subject type. If not, it can retrieve public historical data, which can then be used to determine the historical data corresponding to that subject type. Public historical data consists of historical image data that is widely liked.

[0073] Based on the first preview image and historical data, a composition recommendation model can determine the first scene information, the first filter information, and the recommended pose information. Specifically, the first preview image and historical data can be input into the composition recommendation model, which can then output the first scene information, the first filter information, and the recommended pose information. This recommended pose information is based on poses that match the user's historical preferences.

[0074] In the above embodiments, since the recommended pose information is determined by the composition recommendation model based on the preview image and historical data corresponding to the subject type in the preview image, it can be guaranteed that the determined recommended pose information recommends a pose that conforms to the user's historical preferences. Therefore, it can also be guaranteed that the pose guidance information conforms to the user's historical preferences. It is evident that shooting guidance can be provided based on the user's historical preferences and the current scene. Different preview images and different subject types in the preview images may result in different pose guidance lines, enabling personalized shooting guidance based on the current scene and the user's historical preferences.

[0075] In one embodiment, the shooting guidance method further includes: receiving a second preview image from a client, the second preview image being acquired by the client after displaying pose guidance lines; determining pose guidance information using a pose generation adversarial network based on first scene information, first filter information, and recommended pose information, including: determining pose guidance information and second scene information using a pose generation adversarial network based on the first scene information, first filter information, and recommended pose information; determining second filter information based on the second scene information; determining the similarity between the pose guidance information and the second filter information and the second preview image; and determining the second preview image as the target image if the similarity is greater than or equal to a similarity threshold.

[0076] The second preview image is acquired by the client after the posture guide lines are displayed; that is, the second preview image is determined by the preview image captured or acquired by the client after the posture guide lines are displayed.

[0077] Based on first scene information, first filter information, and recommended pose information, a pose generative adversarial network (PGA) can determine pose guidance information and second scene information. Specifically, the first scene information, first filter information, and recommended pose information can be input into the PGA, which can output not only pose guidance information but also second scene information. The precision of the second scene information is greater than that of the first scene information; that is, the second scene information is determined based on the first scene information, but with higher precision.

[0078] Based on the second scene information, second filter information can be determined. The precision of the second filter information is greater than or higher than that of the first filter information. Then, the similarity between the pose guidance information, the second filter information, and the second preview image can be determined. The pose guidance information can be vectorized to obtain a pose guidance vector, the second filter information can be vectorized to obtain a filter vector, the pose guidance vector and the filter vector can be concatenated to obtain a concatenated vector, the second preview image can be vectorized to obtain an image vector, and the similarity between the concatenated vector and the image vector can be calculated or determined. In the concatenated vector, the pose guidance vector can precede the filter vector, or vice versa.

[0079] Then, it can be determined whether the similarity is greater than or equal to the similarity threshold. If the similarity is greater than or equal to the similarity threshold, it indicates that the pose and composition of the second preview image meet the expected conditions, and the second preview image can be directly identified as the target image.

[0080] In the above embodiment, a second preview image is received from the client. This second preview image is obtained by the client after displaying the pose guidance lines. Based on the first scene information, the first filter information, and the recommended pose information, pose guidance information and second scene information are determined through a pose generative adversarial network. Second filter information is determined based on the second scene information. The similarity between the pose guidance information, the second filter information, and the second preview image is determined. If the similarity is greater than or equal to a similarity threshold, the second preview image is determined as the target image. Therefore, if the preview image obtained under the guidance of the pose guidance lines meets the expected conditions, it can be directly determined as the final image, improving the efficiency of shooting guidance.

[0081] In one embodiment, the shooting guidance method further includes: adjusting the posture of the second preview image based on posture guidance information when the similarity is less than a similarity threshold, and performing filter processing on the adjusted second preview image based on second filter information to obtain the target image.

[0082] If the similarity is less than the similarity threshold, it indicates that the pose and composition of the second preview image do not meet the expected conditions. The pose of the second preview image can be adjusted first based on the pose guidance information, and then filtered based on the second filter information to obtain the target image. Alternatively, the second preview image can be filtered first based on the second filter information, and then the pose of the processed second preview image can be adjusted based on the pose guidance information to obtain the target image. Therefore, when the preview image taken using pose guidance lines does not meet the expected conditions, pose adjustment and filtering can be applied to the preview image to ensure that the pose and composition of the final image meet the requirements.

[0083] Pose adjustment of the second preview image may include adjusting the proportion of key points, adjusting the position of key points, or adjusting both the proportion and position of key points. Filter processing may include at least one of the following: color adjustment, contrast adjustment, effect adjustment, and texture adjustment.

[0084] In the above embodiments, when the similarity is less than a similarity threshold, the pose of the second preview image is adjusted based on the pose guidance information, and the adjusted second preview image is processed using a filter based on the second filter information to obtain the target image. It is evident that when the preview image captured under the guidance of the pose guidance line does not meet the expected conditions, pose adjustment and filter processing can be applied to the preview image to ensure that the pose of the final image conforms to the pose corresponding to the pose guidance information, and the composition of the final image conforms to the second filter information, thereby improving the quality and reliability of the final image. Furthermore, since an image meeting the expected conditions can be obtained by processing the preview image, further shooting guidance is unnecessary, reducing the interaction process, thus saving communication resources and improving image capture efficiency.

[0085] In one embodiment, the shooting guidance method further includes: sending a target image to a client for the client to display.

[0086] After obtaining the target image, it can be sent to the client so that the client can display the target image and allow the user to view the shooting effect.

[0087] In the above embodiments, after obtaining the final image through shooting guidance, the final image can be sent to the client for display so that users can view the shooting effect, which can improve the user experience.

[0088] Figure 3 This is an interactive schematic diagram of a shooting guidance method provided in an embodiment of this application. The shooting guidance method is described from the perspectives of the client and the server. Figure 3 As shown, the shooting guidance method may include the following steps.

[0089] 301. The client sends the first preview image to the server.

[0090] Accordingly, the server receives the first preview image from the client.

[0091] When a client invokes an image acquisition device such as a camera or webcam to display a shooting interface, the client can determine the first preview image in real time or periodically through the image acquisition device, and then send the first preview image to the server. The shooting interface can be a photo-taking interface (the interface for capturing images) or a video-recording interface (the interface for capturing videos). The first preview image can be a preview image directly captured by the client through the image acquisition device, or a compressed preview image of a preview image captured by the image acquisition device.

[0092] If the first preview image is a compressed version of the acquired preview image, the client can acquire a third preview image using an image acquisition device, and then compress the third preview image to obtain the first preview image.

[0093] 302. Based on the first preview image, the server determines the first scene information, the first filter information, and the recommended pose information through a composition recommendation model.

[0094] For a detailed description of step 302, please refer to the description of step 202, which will not be repeated here.

[0095] 303. Based on the first scene information, the first filter information, and the recommended pose information, the server determines the pose guidance information through a pose generation adversarial network.

[0096] For a detailed description of step 303, please refer to the description of step 203, which will not be repeated here.

[0097] 304. The server sends posture guidance information to the client.

[0098] Accordingly, the client receives posture guidance information from the server.

[0099] 305. The client generates a posture guidance line based on the posture guidance information.

[0100] After receiving the posture guidance information, the client can use technology to generate posture guidance lines based on the posture guidance information.

[0101] For further detailed description of step 305, please refer to the relevant description of step 204, which will not be repeated here.

[0102] 306. The client displays posture guide lines on the shooting interface to guide the shooting.

[0103] After the client generates the pose guide lines, the pose guide lines can be displayed on the shooting interface to guide the shooting, that is, to guide the user in taking the shot.

[0104] In the above embodiment, the client sends a first preview image to the server; the server receives the first preview image from the client, and based on the first preview image, determines first scene information, first filter information, and recommended pose information through a composition recommendation model. Based on the first scene information, first filter information, and recommended pose information, the server determines pose guidance information through a pose generative adversarial network and sends the pose guidance information to the client; the client generates pose guidance lines based on the pose guidance information and displays the pose guidance lines on the shooting interface to guide shooting. It can be seen that the client can send a preview image to the server, and the server can generate corresponding pose guidance information based on the preview image, so that the client can display pose guidance lines on the shooting interface to guide image or video shooting. Since the preview images captured by the client are different, the pose guidance information generated by the server may be different, and the pose guidance lines displayed by the client will be different. Therefore, the corresponding pose guidance lines can be flexibly displayed based on the preview image for shooting guidance, which can improve the flexibility and applicability of shooting guidance. Furthermore, since the server first uses a composition recommendation model to determine the scene information, filter information, and recommended pose information corresponding to the preview image, and then uses a pose generative adversarial network to determine the pose guidance information corresponding to the scene information, filter information, and recommended pose information, it is evident that different models determine different information, which ensures the accuracy and reliability of information determination, thereby improving the accuracy and reliability of shooting guidance. Furthermore, the server uses a pre-trained model to determine pose guidance information, which improves the efficiency of pose guidance generation, thus improving shooting guidance efficiency and ultimately enhancing the user experience. Moreover, data processing is performed by the server, while information display and image acquisition are performed by the client. Due to the server's high processing power, the efficiency of shooting guidance can be improved.

[0105] Figure 4 This is an interactive schematic diagram of another shooting guidance method provided in an embodiment of this application. The shooting guidance method is described from the perspectives of the client and the server. Figure 4 As shown, the shooting guidance method may include the following steps.

[0106] 401. The client sends the first preview image to the server.

[0107] Accordingly, the server receives the first preview image from the client.

[0108] For a detailed description of step 401, please refer to the description of step 301, which will not be repeated here.

[0109] 402. The server determines the depth information of the first preview image based on the first preview image.

[0110] For a detailed description of step 402, please refer to the relevant description above, and it will not be repeated here.

[0111] 403. Based on the first preview image and its depth information, the server determines the first scene information, the first filter information, and the recommended pose information through a composition recommendation model.

[0112] For a detailed description of step 403, please refer to the relevant description above, and it will not be repeated here.

[0113] 404. Based on the first scene information, the first filter information, and the recommended pose information, pose guidance information is determined through a pose generation adversarial network.

[0114] For a detailed description of step 404, please refer to the description of step 203, which will not be repeated here.

[0115] 405. The server sends posture guidance information to the client.

[0116] Accordingly, the client receives posture guidance information from the server.

[0117] 406. The client generates a posture guidance line based on the posture guidance information.

[0118] For a detailed description of step 406, please refer to the description of step 305, which will not be repeated here.

[0119] 407. The client displays posture guide lines on the shooting interface to guide the shooting.

[0120] For a detailed description of step 407, please refer to the description of step 306, which will not be repeated here.

[0121] In the above embodiments, the client sends a first preview image to the server; the server receives the first preview image from the client, determines the depth information of the first preview image based on the first preview image, and determines first scene information, first filter information, and recommended pose information through a composition recommendation model based on the first preview image and its depth information. Based on the first scene information, first filter information, and recommended pose information, the server determines pose guidance information through a pose generative adversarial network and sends the pose guidance information to the client; the client generates pose guidance lines based on the pose guidance information and displays the pose guidance lines on the shooting interface to guide shooting. It can be seen that the client can send a preview image to the server, and the server can generate depth information of the preview image based on the preview image, and can generate corresponding pose guidance information based on the preview image and its depth information. This allows the client to display pose guidance lines on the shooting interface to guide image or video shooting. Since different preview images are captured by the client, the pose guidance information generated by the server may be different, and the pose guidance lines displayed by the client may also be different. Therefore, the server can flexibly display corresponding pose guidance lines based on the preview image for shooting guidance, improving the flexibility and applicability of shooting guidance. Furthermore, since the server first uses a composition recommendation model to determine the scene information, filter information, and recommended pose information corresponding to the preview image, and then uses a pose generative adversarial network to determine the pose guidance information corresponding to the scene information, filter information, and recommended pose information, it is evident that different models determine different information, which ensures the accuracy and reliability of information determination, thereby improving the accuracy and reliability of shooting guidance. Further, in the process of using the composition recommendation model to determine the first scene information, first filter information, and recommended pose information, the server inputs not only the first preview image but also the depth information of the first preview image, which improves the accuracy, reliability, and precision of the determined first scene information, first filter information, and recommended pose information, thereby improving the accuracy and reliability of shooting guidance. Furthermore, the server uses a pre-trained model to determine pose guidance information, which improves the generation efficiency of pose guidance information, improves shooting guidance efficiency, and thus improves user experience. Furthermore, data processing is performed by the server, while information display and image acquisition are performed by the client. Due to the server's high processing power, the efficiency of shooting guidance can be improved.

[0122] Figure 5 This is an interactive schematic diagram of another shooting guidance method provided in this application embodiment. The shooting guidance method is described from the perspectives of the client and the server. Figure 5 As shown, the shooting guidance method may include the following steps.

[0123] 501. The client determines the depth information of the first preview image based on the first preview image.

[0124] The client can determine the first preview image through an image acquisition device, and then determine the depth information of the first preview image based on it. The depth information of the first preview image can be determined using a pre-trained model, or other methods. The pre-trained model can be a supervised model, an unsupervised model, or other models with equivalent functionality. The depth information of the first preview image can be determined solely from the first preview image, or it can be determined from the first preview image along with other information, such as ambient brightness.

[0125] For further details on step 501, please refer to the relevant descriptions above, and they will not be repeated here.

[0126] 502. The client sends the first preview image and its depth information to the server.

[0127] Accordingly, the server receives the first preview image and its depth information from the client.

[0128] After the client determines the depth information of the first preview image, it can send the first preview image and its depth information to the server. Correspondingly, the server can receive the first preview image and its depth information from the client.

[0129] 503. Based on the first preview image and its depth information, the server determines the first scene information, the first filter information, and the recommended pose information through a composition recommendation model.

[0130] For a detailed description of step 503, please refer to the relevant description above, and it will not be repeated here.

[0131] 504. Based on the first scene information, the first filter information, and the recommended pose information, pose generation adversarial network is used to determine pose guidance information.

[0132] For a detailed description of step 504, please refer to the description of step 203, which will not be repeated here.

[0133] 505. The server sends posture guidance information to the client.

[0134] Accordingly, the client receives posture guidance information from the server.

[0135] 506. The client generates a posture guidance line based on the posture guidance information.

[0136] For a detailed description of step 506, please refer to the description of step 305, which will not be repeated here.

[0137] 507. The client displays posture guide lines on the shooting interface to guide the shooting.

[0138] For a detailed description of step 507, please refer to the description of step 306, which will not be repeated here.

[0139] In the above embodiments, the client determines the depth information of the first preview image based on the first preview image and sends the first preview image and its depth information to the server. The server receives the first preview image and its depth information from the client, and based on the first preview image and its depth information, determines the first scene information, the first filter information, and the recommended pose information through a composition recommendation model. Based on the first scene information, the first filter information, and the recommended pose information, the server determines pose guidance information through a pose generative adversarial network and sends the pose guidance information to the client. The client generates pose guidance lines based on the pose guidance information and displays the pose guidance lines on the shooting interface to guide shooting. It can be seen that the client can send the preview image and its depth information to the server, and the server can generate corresponding pose guidance information based on the preview image and its depth information. This allows the client to display pose guidance lines on the shooting interface to guide image or video shooting. Since the preview images captured by the client are different, the pose guidance information generated by the server may be different, and the pose guidance lines displayed by the client will also be different. Therefore, the server can flexibly display corresponding pose guidance lines based on the preview image for shooting guidance, which can improve the flexibility and applicability of shooting guidance. Furthermore, since the depth information of the first preview image is determined by the client, and the client can use not only the first preview image but also relevant information such as ambient brightness during the determination process, the accuracy, reliability, and precision of the determined depth information of the first preview image can be improved. This, in turn, improves the accuracy, reliability, and precision of the first scene information, first filter information, and recommended pose information determined based on the depth information of the first preview image, thereby enhancing the accuracy, reliability, and precision of the shooting guidance. Further, since the server first determines the scene information, filter information, and recommended pose information corresponding to the preview image using a composition recommendation model, and then determines the pose guidance information corresponding to the scene information, filter information, and recommended pose information using a pose generative adversarial network, it is evident that different information is determined by different models, ensuring the accuracy and reliability of the information determination, thus improving the accuracy and reliability of the shooting guidance. Furthermore, in the process of using the composition recommendation model to determine the first scene information, first filter information, and recommended pose information, the server inputs not only the first preview image but also its depth information, further improving the accuracy, reliability, and precision of the determined first scene information, first filter information, and recommended pose information, thereby enhancing the accuracy and reliability of the shooting guidance. Furthermore, the server uses a pre-trained model to determine pose guidance information, which improves the efficiency of pose guidance generation and shooting guidance, thereby enhancing the user experience. Additionally, data processing is performed by the server, while information display and image acquisition are handled by the client. Due to the server's high processing power, the efficiency of shooting guidance is further improved.

[0140] Figure 6 This is an interactive schematic diagram of another shooting guidance method provided in this application embodiment. The shooting guidance method is described from the perspectives of the client and the server. Figure 6 As shown, the shooting guidance method may include the following steps.

[0141] 601. The client sends the first preview image to the server.

[0142] Accordingly, the server receives the first preview image from the client.

[0143] For a detailed description of step 601, please refer to the description of step 301, which will not be repeated here.

[0144] 602. The server determines the subject type in the first preview image.

[0145] 603. The server retrieves historical data corresponding to the subject type.

[0146] 604. Based on the first preview image and historical data, the server determines the first scene information, the first filter information, and the recommended pose information through a composition recommendation model.

[0147] For a detailed description of steps 602-604, please refer to the relevant descriptions above, and they will not be repeated here.

[0148] 605. Based on the first scene information, the first filter information, and the recommended pose information, the server determines pose guidance information through a pose generation adversarial network.

[0149] For a detailed description of step 605, please refer to the description of step 203, which will not be repeated here.

[0150] 606. The server sends posture guidance information to the client.

[0151] Accordingly, the client receives posture guidance information from the server.

[0152] 607. The client generates a posture guidance line based on the posture guidance information.

[0153] For a detailed description of step 607, please refer to the description of step 305, which will not be repeated here.

[0154] 608. The client displays posture guide lines on the shooting interface to guide the shooting.

[0155] For a detailed description of step 608, please refer to the description of step 306, which will not be repeated here.

[0156] In the above embodiment, the client sends a first preview image to the server; the server receives the first preview image from the client, determines the subject type in the first preview image, obtains historical data corresponding to the subject type, and, based on the first preview image and historical data, determines first scene information, first filter information, and recommended pose information through a composition recommendation model; based on the first scene information, first filter information, and recommended pose information, determines pose guidance information through a pose generation adversarial network, and sends the pose guidance information to the client; the client generates pose guidance lines based on the pose guidance information and displays the pose guidance lines on the shooting interface to guide shooting. As can be seen, the client can send a preview image to the server. The server can generate corresponding pose guidance information based on the preview image and historical data corresponding to the subject types in the preview image. The client can then display pose guidance lines on the shooting interface to guide image or video shooting. Since the recommended pose information is determined by the composition recommendation model based on the preview image and historical data corresponding to the subject types in the preview image, it can be guaranteed that the recommended pose information matches the user's historical preferences. This ensures that the pose guidance information also matches the user's historical preferences, allowing for shooting guidance based on the user's historical preferences and the current scene. Different preview images and different subject types may result in different pose guidance lines, enabling personalized shooting guidance based on the current scene and the user's historical preferences. Furthermore, since the server first uses the composition recommendation model to determine the scene information, filter information, and recommended pose information corresponding to the preview image, and then uses the pose generative adversarial network to determine the corresponding pose guidance information, different information is determined by different models. This ensures the accuracy and reliability of the information determination, thereby improving the accuracy and reliability of the shooting guidance. Furthermore, the server uses a pre-trained model to determine pose guidance information, which improves the efficiency of pose guidance generation and shooting guidance, thereby enhancing the user experience. Additionally, data processing is performed by the server, while information display and image acquisition are handled by the client. Due to the server's high processing power, the efficiency of shooting guidance is further improved.

[0157] Figure 7 This is an interactive schematic diagram of another shooting guidance method provided in the embodiments of this application. For example... Figure 7 As shown, the shooting guidance method may include the following steps.

[0158] 701. The client acquires the third preview image, compresses the third preview image to obtain the first preview image, and determines the depth information of the third preview image based on the third preview image.

[0159] When a client invokes an image acquisition device such as a camera or webcam to display a shooting interface, the client can acquire a third preview image through the image acquisition device. This third preview image can then be compressed to obtain a first preview image, and its depth information can be determined based on the first preview image. The shooting interface can be a photo capture interface (for taking pictures) or a video recording interface (for recording videos).

[0160] 702. The client sends the depth information of the first and third preview images to the server.

[0161] Accordingly, the server receives depth information from the first and third preview images from the client.

[0162] 703. The server decompresses the first preview image to obtain the third preview image.

[0163] After receiving the depth information of the first and third preview images from the client, the server can decompress the first preview image to obtain the third preview image.

[0164] 704. The server determines the subject type in the third preview image and obtains the historical data corresponding to the subject type in the third preview image.

[0165] 705. Based on the third preview image, its depth information, and historical data, the server determines the first scene information, the first filter information, the recommended pose information, and the composition score of the third preview image through a composition recommendation model.

[0166] The server can determine the first scene information, the first filter information, the recommended pose information, and the composition score of the third preview image based on the third preview image, its depth information, and historical data through a composition recommendation model. In other words, the third preview image, its depth information, and historical data can be input into the composition recommendation model, and the composition recommendation model can output the first scene information, the first filter information, the recommended pose information, and the composition score of the third preview image.

[0167] 706. When the composition score is less than the score threshold, the server determines the pose guidance information and the second scene information through a pose generation adversarial network based on the first scene information, the first filter information and the recommended pose information.

[0168] It can be determined whether the composition score of the third preview image is greater than or equal to a score threshold. If the composition score of the third preview image is greater than or equal to the score threshold, it indicates that the shooting effect of the third preview image is good, and no shooting guidance is needed. The third preview image can be directly identified as the target image. If the composition score of the third preview image is less than the score threshold, it indicates that the shooting effect of the third preview image is poor. Based on the first scene information, the first filter information, and the recommended pose information, pose generation adversarial network can be used to determine pose guidance information and second scene information.

[0169] 707. The server sends posture guidance information to the client.

[0170] Accordingly, the client receives posture guidance information from the server.

[0171] 708. The client generates a posture guide line based on the posture guidance information and displays the posture guide line on the shooting interface to guide the shooting.

[0172] 709. The client acquires the fourth preview image and compresses it to obtain the second preview image.

[0173] After the client displays the posture guide lines, it can capture a fourth preview image, which can then be compressed to obtain a second preview image. The client can capture the fourth preview image after a preset time has elapsed since the posture guide lines were displayed.

[0174] 710. The client sends a second preview image to the server.

[0175] Accordingly, the server receives a second preview image from the client.

[0176] 711. The server decompresses the second preview image to obtain the fourth preview image, determines the second filter information based on the second scene information, determines the pose guidance information and the similarity between the second filter information and the fourth preview image.

[0177] 712. If the similarity is greater than or equal to the similarity threshold, the server will identify the fourth preview image as the target image.

[0178] 713. When the similarity is less than the similarity threshold, the server adjusts the pose of the fourth preview image based on the pose guidance information, and performs filter processing on the adjusted fourth preview image based on the second filter information to obtain the target image.

[0179] 714. The server sends the target image to the client.

[0180] Accordingly, the client receives the target image from the server.

[0181] 715. The client displays the target image.

[0182] After receiving the target image, the client can display it. The client can display the target image on the shooting interface or on other interfaces.

[0183] For further detailed descriptions of steps 701-715, please refer to the relevant descriptions above, and they will not be repeated here.

[0184] It should be understood that some steps in the above embodiments may be omitted, and some steps may be combined. For example, the client-side compression processing step and the server-side decompression processing step may be omitted.

[0185] It should be understood that the same or corresponding content in the different embodiments described above can be referenced to each other.

[0186] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0187] It should be understood that the same or similar information in the different embodiments described above can be referenced to each other.

[0188] Based on the same inventive concept, this application also provides a shooting guidance device for implementing the shooting guidance method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more shooting guidance device embodiments provided below can be found in the limitations of the shooting guidance method described above, and will not be repeated here.

[0189] In one exemplary embodiment, Figure 8 This is a schematic diagram of a shooting guidance device provided in an embodiment of this application. Figure 8 As shown, the shooting guidance device may include:

[0190] Receiver module 801 is used to receive the first preview image from the client;

[0191] The first determining module 802 is used to determine first scene information, first filter information and recommended pose information based on the first preview image and through a composition recommendation model.

[0192] The second determining module 803 is used to determine pose guidance information based on the first scene information, the first filter information and the recommended pose information through a pose generation adversarial network.

[0193] The sending module 804 is used to send posture guidance information to the client. The posture guidance information is used by the client to display posture guidance lines on the shooting interface to guide the shooting.

[0194] In some embodiments, the first determining module 802 is specifically used to determine first scene information, first filter information, recommended pose information and composition score of the first preview image based on the first preview image through a composition recommendation model; the second determining module 803 is specifically used to determine pose guidance information based on the first scene information, first filter information and recommended pose information through a pose generation adversarial network when the composition score of the first preview image is less than a score threshold.

[0195] In some embodiments, the shooting guidance device further includes: a third determining module, configured to determine the depth information of the first preview image based on the first preview image; and a first determining module 802, specifically configured to determine the first scene information, the first filter information, and the recommended pose information through a composition recommendation model based on the first preview image and the depth information of the first preview image.

[0196] In some embodiments, the receiving module 801 is specifically used to receive a first preview image and depth information of the first preview image from the client; the first determining module 802 is specifically used to determine first scene information, first filter information and recommended pose information based on the first preview image and depth information of the first preview image through a composition recommendation model.

[0197] In some embodiments, the shooting guidance device further includes: a fourth determining module, configured to determine the subject type in the first preview image; an acquisition module, configured to acquire historical data corresponding to the subject type in the first preview image; and a first determining module 802, specifically configured to determine first scene information, first filter information, and recommended pose information based on the first preview image and historical data through a composition recommendation model.

[0198] In some embodiments, the receiving module 801 is further configured to receive a second preview image from a client, the second preview image being acquired by the client after displaying the pose guidance lines; the second determining module 803 is specifically configured to determine pose guidance information and second scene information through a pose generation adversarial network based on first scene information, first filter information, and recommended pose information; the shooting guidance device further includes: a fifth determining module, configured to determine second filter information based on the second scene information; a sixth determining module, configured to determine the similarity between the pose guidance information and the second filter information and the second preview image; and a seventh determining module, configured to determine the second preview image as the target image if the similarity is greater than or equal to a similarity threshold.

[0199] In some embodiments, the shooting guidance device further includes: an adjustment module, configured to adjust the posture of the second preview image based on posture guidance information when the similarity is less than a similarity threshold; and a processing module, configured to perform filter processing on the adjusted second preview image based on second filter information to obtain a target image.

[0200] In some embodiments, the sending module 804 is further configured to send a target image to a client, the target image being displayed by the client.

[0201] Each module in the aforementioned shooting guidance device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0202] In one exemplary embodiment, a computer device is provided, the internal structure of which can be as shown in the figure. Figure 9As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements a shooting guidance method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0203] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0204] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0205] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0206] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0207] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0208] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0209] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A shooting guidance method, characterized in that, include: Receive the first preview image from the client; Based on the first preview image, the first scene information, the first filter information, and the recommended pose information are determined by the composition recommendation model. Based on the first scene information, the first filter information, and the recommended pose information, pose guidance information is determined through a pose generation adversarial network. The posture guidance information is sent to the client, and the posture guidance information is used by the client to display posture guidance lines on the shooting interface to guide the shooting.

2. The method according to claim 1, characterized in that, The step of determining the first scene information, the first filter information, and the recommended pose information based on the first preview image using a composition recommendation model includes: Based on the first preview image, the first scene information, the first filter information, the recommended pose information, and the composition score of the first preview image are determined by the composition recommendation model. The step of determining pose guidance information through a pose generation adversarial network based on the first scene information, the first filter information, and the recommended pose information includes: If the composition score is less than the score threshold, pose guidance information is determined by a pose generation adversarial network based on the first scene information, the first filter information, and the recommended pose information.

3. The method according to claim 1, characterized in that, The method further includes: Based on the first preview image, determine the depth information of the first preview image; The step of determining the first scene information, the first filter information, and the recommended pose information based on the first preview image using a composition recommendation model includes: Based on the first preview image and its depth information, the first scene information, the first filter information, and the recommended pose information are determined by a composition recommendation model.

4. The method according to claim 1, characterized in that, Receiving the first preview image from the client includes: Receive a first preview image and depth information of the first preview image from the client; The step of determining the first scene information, the first filter information, and the recommended pose information based on the first preview image using a composition recommendation model includes: Based on the first preview image and its depth information, the first scene information, the first filter information, and the recommended pose information are determined by a composition recommendation model.

5. The method according to claim 1, characterized in that, The method further includes: Determine the subject type in the first preview image; Obtain the historical data corresponding to the subject type; The step of determining the first scene information, the first filter information, and the recommended pose information based on the first preview image using a composition recommendation model includes: Based on the first preview image and the historical data, the first scene information, the first filter information, and the recommended pose information are determined by the composition recommendation model.

6. The method according to claim 1, characterized in that, The method further includes: Receive a second preview image from the client, the second preview image being acquired by the client after displaying the posture guideline; The step of determining pose guidance information through a pose generation adversarial network based on the first scene information, the first filter information, and the recommended pose information includes: Based on the first scene information, the first filter information, and the recommended pose information, pose guidance information and second scene information are determined by a pose generation adversarial network. Based on the second scene information, determine the second filter information; Determine the similarity between the posture guidance information and the second filter information, and the second preview image; If the similarity is greater than or equal to the similarity threshold, the second preview image is determined as the target image.

7. The method according to claim 6, characterized in that, The method further includes: If the similarity is less than the similarity threshold, the pose of the second preview image is adjusted based on the pose guidance information; Based on the second filter information, the adjusted second preview image is processed with a filter to obtain the target image.

8. The method according to claim 6 or 7, characterized in that, The method further includes: The target image is sent to the client, and the target image is used by the client to display the target image.

9. A shooting guidance device, characterized in that, include: The receiving module is used to receive the first preview image from the client; The first determining module is used to determine first scene information, first filter information, and recommended pose information based on the first preview image using a composition recommendation model. The second determining module is used to determine posture guidance information based on the first scene information, the first filter information and the recommended posture information through a posture generation adversarial network. The sending module is used to send the posture guidance information to the client, and the posture guidance information is used by the client to display posture guidance lines on the shooting interface to guide the shooting.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-8.

Images