Image comment processing method and device, equipment and storage medium

Abstract

The invention discloses an image comment processing method and device, equipment and a storage medium, and relates to the technical field of Internet. The method comprises the steps of obtaining at least two image comments of a first song; each image comment comprises at least one image; performing classification processing on each image in the at least two image comments to obtain an image type of each image; based on the image type of each image, obtaining a weight coefficient of each image; and based on the weight coefficient of each image, adjusting the sequence of the at least two image comments in the comment area of the first song. According to the method, the display effect of the image comments in the comment area can be optimized.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of Internet technology, and in particular to a method, device, equipment and storage medium for processing image comments. Background Art

[0002] On music platforms, users can post comments on songs or singers. According to the content of the comments, they can be divided into text comments, image comments, voice comments, etc.

[0003] In the related art, most of the image comments in the comment area of ​​a song are sorted according to the user's interactive behavior on each image comment (such as the user's likes, replies, etc.). For example, if the number of likes and replies of the first image comment of the first song is higher than the number of likes and replies of the second image comment, then the first image comment will be displayed in the comment area of ​​the first song in priority compared with the second image comment.

[0004] However, the ordering method of image comments in the related art is relatively simple, and the display effect of songs in the comment area is poor. Summary of the invention

[0005] The embodiment of the present application provides a method, device, equipment and storage medium for processing image comments. The technical solution provided by the embodiment of the present application is as follows:

[0006] According to one aspect of an embodiment of the present application, a method for processing image comments is provided, the method comprising:

[0007] Obtain at least two image reviews of a first song; each of the image reviews contains at least one image;

[0008] Performing classification processing on each image in the at least two image comments to obtain an image type of each image;

[0009] Based on the image type of each of the images, obtaining a weight coefficient of each of the images;

[0010] Based on the weight coefficient of each of the images, the order of the at least two image comments in the comment area of ​​the first song is adjusted.

[0011] According to an aspect provided by an embodiment of the present application, a device for processing image comments is provided, the device comprising:

[0012] A first acquisition module is used to acquire at least two image comments of a first song; each of the image comments contains at least one image;

[0013] A classification processing module, used to perform classification processing on each image in the at least two image comments to obtain an image type of each image;

[0014] A second acquisition module, used for acquiring a weight coefficient of each of the images based on an image type of each of the images;

[0015] The sorting adjustment module is used to adjust the sorting of the at least two image comments in the comment area of ​​the first song based on the weight coefficient of each of the images.

[0016] In some embodiments, the second acquisition module is used to query the weight coefficient of each of the images based on the image type of each of the images from the corresponding relationship between the image type and the weight coefficient.

[0017] In some embodiments, the image review processing device further includes: a construction module, configured to perform a division process on the historical images in the historical image reviews of each song to obtain at least one historical image set; each historical image in the historical image set satisfies a division condition, and each of the division conditions has a corresponding recommendation weight coefficient;

[0018] The construction module is used to perform clustering processing on at least one of the historical image sets to obtain at least one clustered image type corresponding to at least one of the historical image sets;

[0019] The construction module is used to assign the recommendation weight coefficient of the division condition corresponding to at least one of the historical image sets to at least one cluster image type respectively corresponding to at least one of the historical image sets;

[0020] The construction module is used to construct a corresponding relationship between the image type and the weight coefficient based on the recommendation weight coefficient corresponding to each of the clustered image types.

[0021] In some embodiments, the construction module is used to perform segmentation processing on the historical images in the historical image comments of the songs based on the exposure of the historical images in the historical image comments of the songs to obtain the at least one historical image set; the interaction information is used to indicate the user's interaction behavior on the comments;

[0022] The division condition includes the exposure interval to which the exposure amount of the historical image belongs.

[0023] In some embodiments, the construction module is used to obtain the adjustment coefficient corresponding to each of the clustered image types based on the interaction information corresponding to the historical images in each of the clustered image types; the interaction information is used to indicate the user's interaction behavior on the comment;

[0024] The construction module is used to adjust the recommendation weight coefficient corresponding to each of the clustered image types through the adjustment coefficient corresponding to each of the clustered image types to obtain the adjustment weight coefficient of each of the clustered image types;

[0025] The construction module is used to construct a corresponding relationship between the image type and the weight coefficient based on the adjustment weight coefficient corresponding to each of the clustered image types.

[0026] In some embodiments, the construction module is used to average the interaction values ​​in the interaction information corresponding to the historical images in the clustered image type to obtain the average interaction value corresponding to the clustered image type;

[0027] The construction module is used to query the adjustment coefficient corresponding to the numerical interval where the average interaction value is located as the adjustment coefficient corresponding to the cluster image type.

[0028] In some embodiments, the image comment processing device further includes: a third acquisition module, used to acquire feature information of the at least two image comments of the first song before adjusting the order of the at least two image comments in the comment area of ​​the first song based on the weight coefficient of each image; the feature information includes at least one of the following: interaction information, release time information, relevance information, and image quality information, the interaction information is used to indicate the user's interaction behavior on the comment, and the relevance information is used to indicate the relevance degree between the image and the first song;

[0029] The third acquisition module is used to acquire the ranking scores of the at least two image comments of the first song based on the feature information of the at least two image comments;

[0030] The ranking adjustment module is used to adjust the ranking scores of the at least two image comments based on the weight coefficient of each of the images to obtain the adjusted ranking scores of the at least two image comments;

[0031] The ranking adjustment module is used to adjust the ranking of the at least two image comments in the comment area of ​​the first song based on the adjusted ranking scores of the at least two image comments.

[0032] In some embodiments, the image comment processing device further includes: an image description module, configured to perform image description processing on each of the images to obtain description information of each of the images;

[0033] The image description module is used to obtain the correlation information based on the tag information of the first song and the description information of each of the images.

[0034] In some embodiments, the ranking adjustment module is used to obtain a target weight coefficient based on the weight coefficients of the at least two images when the first image comment contains at least two images; the target weight coefficient is the maximum value or average value of the weight coefficients of the at least two images; the first image comment is any one of the at least two image comments;

[0035] The ranking adjustment module is used to adjust the ranking score of the first image comment based on the target weight coefficient to obtain the adjusted ranking score of the first image comment.

[0036] According to one aspect of an embodiment of the present application, a computer device is provided, the computer device comprising a processor and a memory, the memory storing a computer program, the computer program being loaded and executed by the processor to implement the above-mentioned image comment processing method.

[0037] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is loaded and executed by a processor to implement the above-mentioned image comment processing method.

[0038] According to one aspect of an embodiment of the present application, a computer program product is provided. The computer program product includes a computer program. The computer program is loaded and executed by a processor to implement the above-mentioned image comment processing method.

[0039] The technical solution provided by the embodiments of the present application includes at least the following beneficial effects:

[0040] In a scenario where a user posts an image comment on a song, a computer device can extract an image in the image comment, obtain an image type of the image through classification processing, and then determine a weight coefficient of the image according to the image type, so that the computer device can adjust the order of the image comment in the comment area of ​​the song; this scheme adjusts the display order of the image comments in the comment area of ​​the song according to the image type of the image in the image comment, which can expand the influencing factors of the order of the image comments in the comment area of ​​the song, improve the accuracy of the order of the song comments, and further optimize the display effect of the image comments in the comment area. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 It is a schematic diagram of a solution implementation environment provided by an exemplary embodiment of the present application;

[0043] Figure 2 is a flowchart of a method for processing image comments provided by an exemplary embodiment of the present application;

[0044] Figure 3 is a schematic diagram of the corresponding relationship between image types and weight coefficients provided by an exemplary embodiment of the present application;

[0045] Figure 4 is a flowchart of a method for processing image comments provided by an exemplary embodiment of the present application;

[0046] Figure 5 is a flow chart of a song review image content ecological optimization method provided by an exemplary embodiment of the present application;

[0047] Figure 6 is a block diagram of an image comment processing device provided by an exemplary embodiment of the present application;

[0048] Figure 7 It is a structural block diagram of a computer device provided by an exemplary embodiment of the present application.

[0049] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application. DETAILED DESCRIPTION

[0050] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0051] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0052] The terms used in this disclosure are for the purpose of describing specific embodiments only and are not intended to limit the disclosure. The singular forms of "a", "said" and "the" used in this disclosure and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0053] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions. For example, the user information and data such as image comments and interactive information involved in this application are obtained with full authorization.

[0054] It should be understood that although the terms first, second, etc. may be used in the present disclosure to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. Depending on the context, the word "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determining".

[0055] Please refer to Figure 1 , which shows a schematic diagram of a solution implementation environment provided by an exemplary embodiment of the present application. Figure 1 As shown, the implementation environment of the solution may include: a terminal 100 and a server 200 .

[0056] The terminal 100 may be an electronic device such as a mobile phone, a tablet computer, a vehicle-mounted terminal (vehicle computer), a wearable device, a personal computer (PC), etc. The terminal 100 may be installed with a client that runs a target application, and the target application may be an application that supports the image comment processing method, which is not limited in this application. In addition, this application does not limit the form of the target application, including but not limited to an application (Application, App), a mini-program, a client, etc. installed in the terminal 100, and may also be in the form of a web page.

[0057] The server 200 may be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and cloud servers that provide basic cloud computing services such as big data and artificial palm image recognition platforms. The server 200 may be the background server of the above-mentioned target application, used to provide background services for the client of the target application.

[0058] The terminal 100 and the server 200 may communicate with each other via a network, such as a wired network or a wireless network.

[0059] In the method for processing image comments provided in the embodiment of the present application, the execution subject of each step may be a computer device, which refers to an electronic device with data calculation, processing and storage capabilities. Figure 1 For example, the method for processing image comments may be executed by the terminal 100, or by the server 200, or by the terminal 100 and the server 200 interacting and cooperating to execute the method, which is not limited in the present application.

[0060] In the embodiment of the present application, the comment area of ​​a song refers to an area in the music platform where users exchange opinions and share feelings. It is a virtual space for interaction between the listeners, publishers, and platforms of the song. Users can view comments, like comments, reply to comments, post comments, etc. in the comment area.

[0061] For example, taking a comment area of ​​a first song as an example, a first user posts a first image comment in the comment area of ​​the first song through a first terminal, and a second user posts a second image comment in the comment area of ​​the first song through a second terminal. The first terminal and the second terminal send the image included in the first image comment and the image included in the second image comment to the server. Accordingly, the server performs classification processing on each image to obtain the image type of each image; then, the server obtains the weight coefficient of each image based on the image type of each image; then, the server adjusts the order of the first image comment and the second image comment in the comment area of ​​the first song based on the weight coefficient of each image.

[0062] Please refer to Figure 2 , which shows a flowchart of a method for processing image comments provided by an exemplary embodiment of the present application. The method can be executed by a computer device, for example, the computer device can be a server of a music application. Figure 2 As shown, the method may include steps 210 to 240.

[0063] Step 210: Obtain at least two image comments of the first song; each image comment contains at least one image.

[0064] In the embodiment of the present application, the image comment refers to a comment containing an image posted by a user in the comment area of ​​a song. Optionally, the image comment may contain text or may not contain text.

[0065] The at least two image comments of the first song may be image comments posted by the same user or by different users. For example, the images in the image comments may express the user's feelings after listening to the song or the user's love for the publisher of the song.

[0066] Step 220: Perform classification processing on each image in the at least two image comments to obtain the image type of each image.

[0067] Optionally, the computer device may extract image features of the image from each image in the at least two image comments by using image recognition technology, and map the image features to a preset category to obtain the image type of the image.

[0068] Exemplarily, the computer device may input each image in the at least two image comments into a pre-trained machine learning model to obtain the image type of each image output by the machine learning model. The machine learning model is trained based on sample images and image types of the sample images. Optionally, the machine learning model may be a classification model.

[0069] Specifically, for example, a computer device inputs a sample image into a machine learning model to be trained, and obtains the predicted image type of each image output by the machine learning model; then, a loss function value is obtained based on the difference between the predicted image type and the image type of the sample image; then, the parameters of the machine learning model are updated according to the loss function value; then, the above process is repeated until the convergence condition of the training is reached to obtain a trained machine learning model.

[0070] Optionally, the above-mentioned image types include at least one of the following: emoticons, singer photos, celebrity photos, film and television photos, mobile phone screenshots, and may also include other types, which are not limited in this application.

[0071] Step 230: Based on the image type of each image, obtain a weight coefficient of each image.

[0072] For example, the computer device may be pre-set with weight coefficients corresponding to various image types, so that the computing device can determine the weight coefficient of the target image according to the image type of the target image. For example, the weight coefficient corresponding to an image of an emoticon package is 2, and the weight coefficient corresponding to an image of a mobile phone screenshot is -1.

[0073] The weight coefficient is used to indicate the relative importance of the image. The higher the relative importance, the greater the weight coefficient, and the greater the impact on the ranking of the image comments to which the image belongs in the comment area of ​​the first song.

[0074] Exemplarily, the computer device may obtain a reference image of the same or similar image type as the target image, and interactive information (such as likes, comments, etc.) corresponding to the reference image, and set a weight coefficient for the target image. For example, the higher the evaluation value of the interactive information of the reference image, the greater the weight coefficient of the target image. The reference image may be a historical image in the comment area of ​​each song, or a historical image in the comment area of ​​at least one user platform (such as a music platform, a social platform, a video platform, etc.).

[0075] Step 240: Based on the weight coefficient of each image, adjust the ranking of the at least two image comments in the comment area of ​​the first song.

[0076] Optionally, the computer device may sort the image comments in the comment area of ​​the first song in descending order according to the weight coefficients of the images from high to low. That is, the higher the weight coefficient of the image, the higher the display priority of the image in the comment area of ​​the first song, so that the image is ranked higher in the comment area of ​​the first song. Accordingly, when each user views the comment area, the image will be displayed in the comment area with priority.

[0077] Optionally, the computer device can obtain other ranking influencing factors, such as the credibility, activity, image quality, etc. of the user corresponding to the image comment, and comprehensively adjust the ranking of at least two image comments in the comment area of ​​the first song according to the weight coefficient of each image and other ranking influencing factors.

[0078] In summary, the solution provided by the embodiment of the present application is that, in a scenario where a user posts an image comment on a song, a computer device can extract an image from an image comment, obtain an image type of the image through classification processing, and then determine a weight coefficient of the image based on the image type of the image, so that the computer device can adjust the order of the image comment in the comment area of ​​the song; this solution adjusts the display order of image comments in the comment area of ​​a song based on the image type of the image in the image comment, which can expand the influencing factors of the order of image comments in the comment area of ​​a song, improve the accuracy of the order of song comments, and further optimize the display effect of image comments in the comment area.

[0079] Based on the above Figure 2 In the solution in the embodiment shown, in a possible implementation manner, the above step 230 can be implemented as follows:

[0080] Based on the image type of each image, the weight coefficient of each image is queried from the corresponding relationship between the image type and the weight coefficient.

[0081] The correspondence between the image type and the weight coefficient is information pre-established and stored in the computer device. Exemplarily, the computer device creates a database table or file structure for storing the correspondence between the image type and the weight coefficient.

[0082] For example, the corresponding relationship between the above image types and weight coefficients is shown in Table 1.

[0083] Table 1

[0084] Image Type Emoticons Celebrity Pictures Mobile screenshot …… Weight coefficient 2 1.5 -1 ……

[0085] In an embodiment of the present application, by pre-constructing and maintaining the correspondence between image types and weight coefficients, a computer device can directly query the weight coefficient of an image based on the image type from the correspondence between the image type and the weight coefficient, thereby improving the efficiency of obtaining the weight coefficient of the image.

[0086] Based on the solutions in the above-mentioned embodiments of the present application, in a possible implementation scheme, the above-mentioned method for processing image comments further includes steps 201 to 204.

[0087] Step 201: Perform division processing on the historical images in the historical image reviews of each song to obtain at least one historical image set; the historical images in each historical image set meet a division condition, and each division condition has a corresponding recommendation weight coefficient.

[0088] That is, the computer device pre-acquires the historical images in the historical image comments of each song in the music platform; then, the computer device can classify at least one historical image that meets the same division condition among the historical images in the historical image comments of each song into the same historical image set. Therefore, the computer device can obtain one or more historical image sets according to multiple division conditions.

[0089] The computer device pre-stores a variety of division conditions and recommendation weight coefficients corresponding to the division conditions. For example, if division condition a is that the number of likes of the image comments to which the historical image belongs reaches 999+, the recommendation weight coefficient corresponding to the division condition is 1.2; for another example, if division condition b is that the number of clicks on the historical image reaches 9999+, the recommendation weight coefficient corresponding to the division condition is 1.5.

[0090] Step 202: Perform clustering processing on at least one historical image set to obtain at least one clustered image type corresponding to the at least one historical image set.

[0091] The above clustering process can classify historical images of the same or similar type in a historical image set into a cluster, where historical images in the same cluster are similar, while historical images in different clusters are different. Exemplarily, the computer device can perform clustering on at least one historical image set using a K-means clustering algorithm.

[0092] In the embodiment of the present application, each historical image set generally includes multiple historical images, and different historical images may correspond to different types. Therefore, the computer device can obtain at least one clustered image type corresponding to each historical image set by performing clustering processing on each historical image set.

[0093] The clustered image type is consistent with the image type; or the clustered image type is similar to the image type.

[0094] In a possible implementation, after the computer device performs clustering processing on at least one historical image set, manual clustering may be performed to further refine at least one cluster image type corresponding to each historical image set.

[0095] Step 203: assigning a recommended weight coefficient of a division condition corresponding to at least one historical image set to at least one cluster image type corresponding to at least one historical image set.

[0096] For example, the historical image set a corresponding to the division condition a contains A cluster image types. The computer device can assign the recommended weight coefficient of division condition a, 1.2, to the A cluster image types in the historical image set a. Then, the recommended weight coefficient corresponding to the A cluster image types is 1.2. For another example, the historical image set b corresponding to the division condition b contains B cluster image types. The computer device can assign the recommended weight coefficient of division condition b, 1.5, to the B cluster image types in the historical image set b. Then, the recommended weight coefficient corresponding to the B cluster image types is 1.5.

[0097] Step 204: Based on the recommended weight coefficients corresponding to each clustered image type, a corresponding relationship between the image type and the weight coefficient is constructed.

[0098] Optionally, the computing device may determine the recommendation weight coefficient corresponding to a clustered image type as a weight coefficient corresponding to an image type matching the clustered image type.

[0099] For example, Figure 3As shown, the historical images in the historical image reviews of each song include historical image set a, historical image set b, and other historical image sets. The recommended weight coefficient of the division condition corresponding to historical image set a is 1.2, and the recommended weight coefficient of the division condition corresponding to historical image set b is 1.5. Among them, historical image set a includes cluster image type a1, cluster image type a2, and other cluster images, and historical image set b includes cluster image type b1, cluster image type b2, and other cluster images. Then, the computer device can construct a corresponding relationship between cluster image type a1 and weight coefficient 1.2, a corresponding relationship between cluster image type a2 and weight coefficient 1.2, a corresponding relationship between cluster image type b1 and weight coefficient 1.5, and a corresponding relationship between cluster image type b2 and weight coefficient 1.5.

[0100] Optionally, the computer device may adjust a recommendation weight coefficient corresponding to a cluster image type, and determine the adjusted recommendation weight coefficient as a weight coefficient corresponding to an image type matching the cluster image type.

[0101] The embodiment of the present application shows the process of constructing the correspondence between image types and weight coefficients. The computer device can obtain historical image sets corresponding to multiple division conditions and recommended weight coefficients corresponding to the multiple division conditions based on multiple pre-set division conditions and the recommended weight coefficients corresponding to the multiple division conditions, thereby improving the accuracy of the weight coefficients corresponding to each image type.

[0102] Based on the solutions in the above embodiments of the present application, in a possible implementation scheme, the above step 201 can be implemented as follows:

[0103] Based on the exposure amount of the historical images in the historical image comments of each song, the historical images in the historical image comments of each song are divided to obtain at least one historical image set; the interactive information is used to indicate the interactive behavior of the user on the comment;

[0104] The division condition includes the exposure interval to which the exposure amount of the historical image belongs.

[0105] In the embodiment of the present application, the exposure interval to which the exposure of the above-mentioned historical image belongs is pre-set and stored in the computer device.

[0106] Among them, the exposure of historical images refers to at least one of the number of clicks, downloads, collections, and shares of historical images by users. The number of clicks refers to the number of operations in which users click on and enlarge historical images. The number of downloads refers to the number of times users download and save historical images. The number of collections refers to the number of times users add historical images to their favorites and save them. The number of shares refers to the number of times users share historical images with other users.

[0107] Exemplarily, in the case where the exposure volume of historical images is the number of clicks on historical images by users, division condition 1 is that the click volume interval of the click volume of historical images by users is 0 to 999, division condition 2 is that the click volume interval of the click volume of historical images by users is 1000 to 9999, and division condition 3 is that the click volume interval of the click volume of historical images by users is above 10000; in the case where the exposure volume of historical images is the number of downloads of historical images by users, division condition 1 is that the download volume interval of the download volume of historical images by users is 0 to 99, division condition 2 is that the download volume interval of the download volume of historical images by users is 100 to 999, and division condition 3 is that the download volume interval of the download volume of historical images by users is above 1000.

[0108] The embodiment of the present application shows the basis for performing division processing on the historical images in the historical image comments of each song. The exposure range to which the exposure amount of the historical image belongs can be set in the division condition, so that the computer device performs division processing according to the exposure amount of each historical image, improves the accuracy of dividing the historical images, and improves the correlation between the exposure amount of the historical image and the weight coefficient.

[0109] Based on the solutions in the above-mentioned embodiments of the present application, in a possible implementation scheme, the above-mentioned step 204 can be implemented as steps 204a to 204c.

[0110] Step 204a: based on the interaction information corresponding to the historical images in each cluster image type, obtain the adjustment coefficient corresponding to each cluster image type; the interaction information is used to indicate the user's interaction behavior on the comment.

[0111] The user's interactive behavior on the comment refers to the user's interactive behavior on the comment to which the historical image belongs. Exemplarily, the user's interactive behavior on the comment includes at least one of the following: the user's like, reply / comment, forward / share of the comment, and may also include other interactive behaviors, which are not limited in this application.

[0112] For example, if the comments of each historical image in a certain cluster image type have received a large number of likes from users, the adjustment coefficient corresponding to this cluster image type can be set to a larger value (such as 2). For another example, if the comments of each historical image in a certain cluster image type have received fewer interactions, the adjustment coefficient corresponding to this cluster image type can be set to a smaller value (such as 0.5).

[0113] The corresponding relationship between the numerical interval of the interactive information and the adjustment coefficient can be pre-set and stored in the computer device. That is, the computer device can obtain the interactive information corresponding to the historical images in the clustered image type, and then query the adjustment coefficient corresponding to each clustered image type based on the numerical interval to which the interactive information corresponding to the historical images in the clustered image type belongs and the corresponding relationship between the numerical interval of the interactive information and the adjustment coefficient.

[0114] Exemplarily, the computer device creates a database table or file structure for storing the correspondence between the numerical range of each interactive information and the adjustment coefficient. Specifically, for example, taking the interactive information used to indicate that the user's interactive behavior for a comment is a like, the correspondence between the numerical range of the interactive information and the adjustment coefficient is shown in Table 2.

[0115] Table 2

[0116]

[0117]

[0118] Exemplarily, when the interactive information indicates that the user's interactive behaviors with respect to a comment include multiple types of interactive behaviors such as likes, replies / comments, and forwarding / sharing, the computer device may obtain the adjustment coefficient based on the interactive information of one of the interactive behaviors (such as likes); or, the computer device may use the average value or the sum of the adjustment coefficients corresponding to the interactive information of the multiple interactive behaviors as the adjustment coefficient.

[0119] Step 204b: adjusting the recommendation weight coefficient corresponding to each cluster image type by using the adjustment coefficient corresponding to each cluster image type, and obtaining the adjustment weight coefficient of each cluster image type.

[0120] Exemplarily, the computer device may use the product, average value, or sum of the adjustment coefficient corresponding to each cluster image type and the recommended weight coefficient corresponding to each cluster image type as the adjustment weight coefficient of each cluster image type.

[0121] Step 204c: constructing a corresponding relationship between image types and weight coefficients based on the adjustment weight coefficients corresponding to each clustered image type.

[0122] In an embodiment of the present application, the computing device may determine the adjustment weight coefficient corresponding to a cluster image type as the weight coefficient corresponding to the image type matching the cluster image type.

[0123] In an embodiment of the present application, in the process of constructing a correspondence between image types and weight coefficients, the computer device can adjust the recommended weight coefficient corresponding to the clustered image type based on the user's interactive behavior on the comments to the historical images, thereby improving the correlation between the user's interactive behavior and the weight coefficient.

[0124] Based on the solutions in the above embodiments of the present application, in a possible implementation solution, the above step 204a can be implemented as follows:

[0125] Taking an average of the interaction values ​​in the interaction information corresponding to the historical images in the clustered image type, to obtain an average interaction value corresponding to the clustered image type;

[0126] The adjustment coefficient corresponding to the numerical range of the average interaction value is queried as the adjustment coefficient corresponding to the cluster image type.

[0127] That is, the computer device may obtain the average value of the interaction values ​​in the interaction information corresponding to the historical images in the clustered image type. Optionally, the above average value may also be a weighted average value, and the interaction values ​​of different interaction information correspond to different weights.

[0128] Exemplarily, the computer device creates a database table or file structure for storing the correspondence between the numerical range of the average interaction value and the adjustment coefficient, so as to quickly query the adjustment coefficient corresponding to the clustered image type according to the average interaction value to ensure the consistency and fairness of the adjustment process.

[0129] In an embodiment of the present application, by pre-setting the correspondence between the numerical range of the average interaction value and the adjustment coefficient, the computer device can obtain the adjustment coefficient corresponding to each cluster image type based on the average interaction value corresponding to each cluster image type, thereby improving the efficiency of obtaining the adjustment coefficient.

[0130] Based on the solutions in the above embodiments of the present application, in a possible implementation scheme, please refer to Figure 4 , which shows a flowchart of a method for processing image comments provided by an exemplary embodiment of the present application. Figure 4 As shown, before the above step 240, the method further includes steps 232 and 234. The above step 240 can be implemented as steps 240a and 240b.

[0131] Step 232: Obtain feature information of at least two image comments of the first song; the feature information includes at least one of the following: interactive information, release time information, relevance information, and image quality information, the interactive information is used to indicate the user's interactive behavior on the comments, and the relevance information is used to indicate the degree of relevance between the image and the first song.

[0132] The above-mentioned interactive information refers to the user's behavioral data on a specific comment, such as likes, replies / comments, forwarding / sharing, etc. The interactive information can reflect the attention and support of the comment by each user.

[0133] The above-mentioned release time information refers to the specific timestamp when the comment is released. The release time information can reflect the freshness and timeliness of the comment.

[0134] The above correlation information can be used to evaluate the degree of association between the image in the image comment and the first song itself.

[0135] For example, the relevance information may indicate the degree of relevance between the image in the image comment and the style of the first song; or another example, the relevance information may indicate the degree of relevance between the image in the image comment and the lyrics of the first song. Using the relevance information as the sorting basis can avoid images that are irrelevant to the first song or contradictory in style from being displayed in the comment area of ​​the first song.

[0136] The above-mentioned image quality information refers to the quality of the image. For example, the image quality information includes resolution, clarity, etc. For example, the computer device can obtain the image quality information of each image in each image review based on the image quality analysis tool.

[0137] Step 234: Based on the feature information of the at least two image comments of the first song, obtain the ranking scores of the at least two image comments.

[0138] Exemplarily, the computer device may store a first calculation formula for obtaining a ranking score according to the feature information. Accordingly, the computer device may substitute the acquired feature information of the image comment into the first calculation formula to obtain the ranking score of the image comment.

[0139] For example, taking the feature information as interactive information, the first calculation formula may be: ranking score = impact factor * interactive value of interactive information, where the impact factor corresponds to the calculation weight of the interactive information.

[0140] Among them, the higher the value corresponding to the interactive information, the higher the ranking score; the closer the time corresponding to the release time information, the higher the ranking score; the higher the value corresponding to the relevance information, the higher the ranking score; the higher the value corresponding to the image quality information, the higher the ranking score.

[0141] Step 240a: Based on the weight coefficient of each image, the ranking scores of at least two image comments are adjusted to obtain adjusted ranking scores of the at least two image comments.

[0142] Optionally, the computer device may store the weight coefficient and feature information of the image to obtain the second calculation formula for the adjusted ranking score. Accordingly, the computer device may substitute the acquired weight coefficient of the image and the feature information of the image comment into the second calculation formula to obtain the adjusted ranking score of the image comment.

[0143] Exemplarily, the computer device may use the product of the weight coefficient of the image and the ranking score of the image comment as the adjusted ranking score of the image comment. For example, taking the feature information as interactive information, the first calculation formula may be: adjusted ranking score = weight coefficient * impact factor * interactive value of interactive information, where the impact factor corresponds to the calculation weight of the interactive information.

[0144] Among them, the larger the weight coefficient of the image, the greater the change in the ranking score of the image comments before and after adjustment.

[0145] Step 240b: Adjust the ranking of the at least two image comments in the comment area of ​​the first song based on the adjusted ranking scores of the at least two image comments.

[0146] Exemplarily, the higher the adjusted ranking score of an image comment, the higher the display priority of the image comment in the comment area of ​​the first song; the lower the adjusted ranking score of an image comment, the lower the display priority of the image comment in the comment area of ​​the first song.

[0147] The embodiment of the present application shows feature information related to the display order of each image comment in the corresponding comment area. The computer device can obtain the initial display order of each image comment in the comment area of ​​the first song based on the feature information, and then optimize the display order of each image comment in the corresponding comment area based on the weight coefficient of the image of each image comment.

[0148] Based on the solutions in the above embodiments of the present application, in a possible implementation scheme, the above image comment processing method further includes:

[0149] Performing image description processing on each image to obtain description information of each image;

[0150] Based on the tag information of the first song and the description information of each image, correlation information is obtained.

[0151] Exemplarily, in the process of obtaining the description information of each image, the computer device can input each image and prompt word into a pre-trained generative language model to obtain the description information of each image output by the generative language model. The generative language model is trained based on sample images, sample prompt words and sample description information of sample images.

[0152] For example, the generative language model includes an encoder and a decoder. The encoder is used to extract image features of an image, and the decoder is used to generate a text description corresponding to the image content.

[0153] Among them, the label information of the above-mentioned first song may include at least one of the following: cheerful, healing, melancholy, sad, etc.

[0154] In an embodiment of the present application, the computer device can perform sentiment analysis on the description information of each image to obtain sentiment information corresponding to the description information of each image; then, obtain correlation information based on the similarity between the sentiment information and the label information of the first song.

[0155] Exemplarily, in the process of performing sentiment analysis on the description information of each image, the computer device can input the label information of the first song and the description information of each image into a pre-trained discriminative language model to obtain the sentiment information corresponding to the description information of each image output by the discriminative language model.

[0156] The embodiment of the present application illustrates a process for acquiring correlation information. The computer device may first acquire the description information of each image through image description processing; and then acquire the correlation information of the image comments to which each image belongs through the correlation between the tag information of the first song and the description information of each image, thereby improving the accuracy of the correlation information.

[0157] In a possible implementation, the present application can pre-train a multimodal large model, which can be implemented as the reasoning function of the above-mentioned classification model and generative language model. In other words, the multimodal large model can output the image type and image description information of the image based on the input image and prompt words.

[0158] Based on the solutions in the above-mentioned embodiments of the present application, in a possible implementation scheme, the above-mentioned step 240a can be implemented as step 240a1 and step 240a2.

[0159] Step 240a1: When the first image comment includes at least two images, obtain a target weight coefficient based on the weight coefficients of each of the at least two images; the target weight coefficient is the maximum value or average value of the weight coefficients of each of the at least two images; the first image comment is any one of the at least two image comments.

[0160] The computer device may use the maximum value or average value of the weight coefficients of at least two images as the target weight coefficient according to the actual situation. Selecting the maximum value of the weight coefficients of at least two images as the target weight coefficient emphasizes the influence of the maximum weight coefficient on the image comments, which is suitable for highlighting; selecting the average value of the weight coefficients of at least two images as the target weight coefficient avoids excessive influence of the maximum weight coefficient on the image comments, which is suitable for balancing the influence of the weight coefficients of each image on the image comments.

[0161] Exemplarily, when the target weight coefficient is an average value of the weight coefficients of at least two images, the average value may be a weighted average value of the weight coefficients of at least two images.

[0162] Step 240a2: adjusting the ranking score of the first image comment based on the target weight coefficient to obtain an adjusted ranking score of the first image comment.

[0163] In the embodiment of the present application, the computer device adjusts the ranking score of the first image comment according to the target weight coefficient according to the preset adjustment rule to obtain the adjusted ranking score of the first image comment.

[0164] The embodiments of the present application can improve the diversity of schemes for determining weight coefficients when an image comment contains multiple images; specifically, if an image comment contains multiple images, the computer device can adjust the ranking score of the image comment based on the target weight coefficient obtained by the maximum value of the weight coefficients of each image, so as to improve the influence of the image corresponding to the maximum value of the weight coefficient in the image comment on the ranking of the image comments; or, the computer device can adjust the ranking score of the image comment based on the target weight coefficient obtained by the average value of the weight coefficients of each image, so as to improve the influence of each image in the image comment on the ranking of the image comments.

[0165] Based on the schemes shown in the above embodiments, the present application proposes a method for optimizing the ecological content of song review images, which suppresses low-quality image comments and improves the priority of high-quality image comments through multimodal image quality evaluation and image classification processing. Specifically, the method can first cluster images of interest to users, find images that are of interest to users and highly interactive, and then use a multimodal image classification algorithm to output labels for the images to improve the display priority of images of interest and high interactivity. At the same time, the method also combines information such as the relevance of images and songs in image comments, image quality, and the amount of interaction of image comments for comprehensive sorting, and gives priority to displaying high-quality images of interest to users, thereby optimizing the content ecology of review images.

[0166] That is to say, this method analyzes, understands and clusters the images with high likes, replies and downloads in song comments through algorithm content, and searches for universal and highly interactive image types. The model is trained for images of highly interactive image types to improve the accuracy of image classification of highly interactive image types, so as to increase the display weight of highly interactive image types in the song comment area. At the same time, this method can also analyze the features of images that users are not interested in (such as images of low quality or mobile phone screenshots), and through algorithm evaluation and identification, reduce the display weight of images that users are not interested in in the song comment area.

[0167] Please refer to Figure 5 , which shows a flow chart of a song review image content ecological optimization method provided by an exemplary embodiment of the present application. Figure 5 As shown, first, the computer device establishes a corresponding relationship 510 between image types and weight coefficients by dividing, clustering, assigning, and other processes on the historical images in the historical image reviews of each song.

[0168] For example, the computer device may pre-acquire historical images in historical image reviews of n songs, and perform division processing on the historical images in the historical image reviews of the n songs to obtain m historical image sets. The historical images in each historical image set satisfy a division condition, and each division condition has a corresponding recommendation weight coefficient. Afterwards, the computer device may perform clustering processing on the m historical image sets to obtain at least one cluster image type corresponding to the m historical image sets; afterward, the computer device assigns the recommendation weight coefficient of the division condition corresponding to the m historical image sets to at least one cluster image type corresponding to the m historical image sets; afterward, the computer device may construct a corresponding relationship 510 between image types and weight coefficients based on the recommendation weight coefficients corresponding to each cluster image type.

[0169] Secondly, taking the comment area of ​​the first song including image comment A, image comment B, image comment C, image comment D and other image comments as an example, the computer device can perform classification processing on each image in image comment A, image comment B, image comment C, image comment D and other image comments to obtain the image type 501 of each image in image comment A, image comment B, image comment C, image comment D and other image comments; thereafter, the computer device can query the weight coefficient 502 of each image in image comment A, image comment B, image comment C, image comment D and other image comments from the correspondence 510 between the image type and the weight coefficient based on the image type 501 of each image.

[0170] In addition, the computer device can obtain feature information 503 of image comment A, image comment B, image comment C, image comment D and other image comments; thereafter, the computer device obtains ranking scores 504 of image comment A, image comment B, image comment C, image comment D and other image comments based on the feature information 503 of image comment A, image comment B, image comment C, image comment D and other image comments.

[0171] The characteristic information includes interactive information, release time information, relevance information, and image quality information. The interactive information is used to indicate the user's interactive behavior on the comment (such as liking, replying, downloading the image, etc.), and the relevance information is used to indicate the relevance between the image and the first song.

[0172] Then, the computer device adjusts the ranking scores 504 of image comment A, image comment B, image comment C, image comment D and other image comments based on the weight coefficient 502 of each image in image comment A, image comment B, image comment C, image comment D and other image comments to obtain adjusted ranking scores 505 of image comment A, image comment B, image comment C, image comment D and other image comments.

[0173] Finally, the computer device adjusts the ranking of image comment A, image comment B, image comment C, image comment D and other image comments in the comment area of ​​the first song based on the adjusted ranking scores 505 of image comment A, image comment B, image comment C, image comment D and other image comments.

[0174] Exemplarily, the computer device can obtain the comment area 520 for the first song based on the adjusted ranking scores 505 of image comments A, image comments B, image comments C, image comments D and other image comments; the computer device can obtain the comment area 530 for the first song based on the ranking scores 504 of image comments A, image comments B, image comments C, image comments D and other image comments.

[0175] in, Figure 5 The execution method of each step in is the same as the execution method of the above-mentioned embodiments of the present application, and will not be repeated here.

[0176] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0177] Please refer to Figure 6, which shows a block diagram of an image comment processing device provided by an exemplary embodiment of the present application. The device has the functions of implementing the above examples, and the functions can be implemented by hardware, or by hardware executing corresponding software. The device can be the server 200 introduced above, or it can be set in the server 200. Figure 6 As shown, the device may include a first acquisition module 601 , a classification processing module 602 , a second acquisition module 603 and a sorting adjustment module 604 .

[0178] The first acquisition module 601 is used to acquire at least two image comments of the first song; each image comment contains at least one image;

[0179] A classification processing module 602 is used to perform classification processing on each image in at least two image comments to obtain an image type of each image;

[0180] A second acquisition module 603 is used to acquire a weight coefficient of each image based on the image type of each image;

[0181] The ranking adjustment module 604 is used to adjust the ranking of at least two image comments in the comment area of ​​the first song based on the weight coefficient of each image.

[0182] In some embodiments, the second acquisition module 603 is used to query the weight coefficient of each image based on the image type of each image from the corresponding relationship between the image type and the weight coefficient.

[0183] In some embodiments, the image review processing device further includes: a construction module, configured to perform a division process on the historical images in the historical image reviews of each song to obtain at least one historical image set; each historical image in the historical image set satisfies a division condition, and each division condition has a corresponding recommendation weight coefficient;

[0184] A construction module, configured to perform clustering processing on at least one historical image set to obtain at least one clustered image type corresponding to the at least one historical image set;

[0185] A construction module, used to assign a recommended weight coefficient of a division condition corresponding to at least one historical image set to at least one cluster image type respectively corresponding to at least one historical image set;

[0186] The construction module is used to construct a corresponding relationship between image types and weight coefficients based on the recommended weight coefficients corresponding to each clustered image type.

[0187] In some embodiments, a construction module is used to perform a segmentation process on the historical images in the historical image comments of each song based on the exposure amount of the historical images in the historical image comments of each song to obtain at least one historical image set; the interaction information is used to indicate the user's interactive behavior on the comment;

[0188] The division condition includes the exposure interval to which the exposure amount of the historical image belongs.

[0189] In some embodiments, a construction module is used to obtain an adjustment coefficient corresponding to each cluster image type based on interaction information corresponding to historical images in each cluster image type; the interaction information is used to indicate the user's interaction behavior with respect to the comment;

[0190] A construction module is used to adjust the recommendation weight coefficient corresponding to each cluster image type through the adjustment coefficient corresponding to each cluster image type to obtain the adjustment weight coefficient of each cluster image type;

[0191] The construction module is used to construct a corresponding relationship between the image type and the weight coefficient based on the adjustment weight coefficient corresponding to each cluster image type.

[0192] In some embodiments, a construction module is used to average the interaction values ​​in the interaction information corresponding to the historical images in the clustered image type to obtain an average interaction value corresponding to the clustered image type;

[0193] A module is constructed to query the adjustment coefficient corresponding to the numerical interval where the average interaction value is located as the adjustment coefficient corresponding to the cluster image type.

[0194] In some embodiments, the image comment processing device further includes: a third acquisition module, which is used to acquire feature information of at least two image comments of the first song before adjusting the order of the at least two image comments in the comment area of ​​the first song based on the weight coefficient of each image; the feature information includes at least one of the following: interaction information, release time information, relevance information, and image quality information, the interaction information is used to indicate the user's interaction behavior on the comment, and the relevance information is used to indicate the relevance degree between the image and the first song;

[0195] A third acquisition module, configured to acquire ranking scores of at least two image comments based on feature information of at least two image comments of the first song;

[0196] A ranking adjustment module 604, configured to adjust the ranking scores of at least two image comments based on the weight coefficients of the images to obtain adjusted ranking scores of the at least two image comments;

[0197] The ranking adjustment module 604 is used to adjust the ranking of the at least two image comments in the comment area of ​​the first song based on the adjusted ranking scores of the at least two image comments.

[0198] In some embodiments, the image comment processing device further includes: an image description module, configured to perform image description processing on each image to obtain description information of each image;

[0199] The image description module is used to obtain correlation information based on the tag information of the first song and the description information of each image.

[0200] In some embodiments, the ranking adjustment module 604 is used to obtain a target weight coefficient based on the weight coefficients of the at least two images when the first image comment includes at least two images; the target weight coefficient is a maximum value or an average value of the weight coefficients of the at least two images; the first image comment is any one of the at least two image comments;

[0201] The ranking adjustment module 604 is used to adjust the ranking score of the first image comment based on the target weight coefficient to obtain the adjusted ranking score of the first image comment.

[0202] It should be noted that the device provided in the above embodiment, when implementing its functions, is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0203] Please refer to Figure 7 , which shows a block diagram of a computer device 700 provided by an exemplary embodiment of the present application. The computer device 700 may be Figure 1 The terminal 100 in the implementation environment shown may also be Figure 1 The server 200 in the illustrated implementation environment is used to implement the image comment processing method provided in the above embodiment. Specifically:

[0204] Typically, the computer device 700 includes a processor 710 and a memory 720 .

[0205] The processor 710 may include one or more processing cores, such as a 4-core processor, a 7-core processor, etc. The processor 710 may be implemented in at least one hardware form of digital signal processing (DSP), field programmable gate array (FPGA), and programmable logic array (PLA). The processor 710 may also include a main processor and a coprocessor. The main processor is a processor for processing data in an awake state, also known as a central processing unit (CPU); the coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 710 may be integrated with a graphics processing unit (GPU), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 710 may also include an artificial intelligence (AI) processor, which is used to process computing operations related to machine learning.

[0206] The memory 720 may include one or more computer-readable storage media, which may be non-transitory. The memory 720 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 720 is used to store a computer program, which is configured to be executed by one or more processors to implement the above-mentioned image comment processing method.

[0207] Those skilled in the art will understand that Figure 7 The structure shown in the figure does not constitute a limitation on the computer device 700, and the computer device 700 may include more or less components than shown in the figure, or combine some components, or adopt a different arrangement of components.

[0208] In an exemplary embodiment, a computer-readable storage medium is also provided, in which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned image comment processing method is implemented. Optionally, the computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a solid state drive (SSD) or an optical disk, etc. Among them, the random access memory may include a resistance random access memory (ReRAM) and a dynamic random access memory (DRAM).

[0209] In an exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program, the computer program being stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the above-mentioned image comment processing method.

[0210] It should be understood that the "multiple" mentioned in this article refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. In addition, the step numbers described in this article only illustrate a possible execution sequence between the steps. In some other embodiments, the above steps may not be executed in the order of the numbers, such as two steps with different numbers are executed at the same time, or two steps with different numbers are executed in the opposite order to the diagram. The embodiments of the present application are not limited to this.

[0211] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for processing image comments, characterized in that: The method comprises: Obtain at least two image reviews of a first song; each of the image reviews contains at least one image; Performing classification processing on each image in the at least two image comments to obtain an image type of each image; Based on the image type of each of the images, obtaining a weight coefficient of each of the images; Based on the weight coefficient of each of the images, the order of the at least two image comments in the comment area of ​​the first song is adjusted.

2. The method according to claim 1, characterized in that: The obtaining of a weight coefficient of each image based on the image type of each image includes: Based on the image type of each of the images, the weight coefficient of each of the images is queried from the corresponding relationship between the image type and the weight coefficient.

3. The method according to claim 2, characterized in that The method further comprises: Performing a division process on the historical images in the historical image reviews of each song to obtain at least one historical image set; each historical image in the historical image set satisfies a division condition, and each of the division conditions has a corresponding recommendation weight coefficient; Performing clustering processing on at least one of the historical image sets to obtain at least one clustered image type corresponding to at least one of the historical image sets; Assigning the recommendation weight coefficient of the division condition corresponding to at least one of the historical image sets to at least one cluster image type respectively corresponding to at least one of the historical image sets; Based on the recommendation weight coefficients corresponding to the respective clustered image types, a corresponding relationship between the image types and the weight coefficients is constructed.

4. The method according to claim 3, characterized in that: The performing of division processing on the historical images in the historical image reviews of each song to obtain at least one historical image set includes: Based on the exposure amount of the historical images in the historical image comments of the songs, a division process is performed on the historical images in the historical image comments of the songs to obtain the at least one historical image set; the interaction information is used to indicate the user's interaction behavior on the comments; The division condition includes the exposure interval to which the exposure amount of the historical image belongs.

5. The method according to claim 3, characterized in that: The constructing a corresponding relationship between the image type and the weight coefficient based on the recommendation weight coefficient corresponding to each of the clustered image types includes: Based on the interaction information corresponding to the historical images in each of the clustered image types, an adjustment coefficient corresponding to each of the clustered image types is obtained; the interaction information is used to indicate the user's interaction behavior with respect to the comment; By adjusting the adjustment coefficient corresponding to each of the clustered image types, the recommendation weight coefficient corresponding to each of the clustered image types is adjusted to obtain the adjustment weight coefficient of each of the clustered image types; Based on the adjustment weight coefficients corresponding to the respective clustered image types, a corresponding relationship between the image types and the weight coefficients is constructed.

6. The method according to claim 5, characterized in that The interactive information corresponding to the historical images in each of the cluster image types is obtained to obtain the adjustment coefficient corresponding to each of the cluster image types, including: Taking an average value of the interaction values ​​in the interaction information corresponding to the historical images in the clustered image type to obtain an average interaction value corresponding to the clustered image type; The adjustment coefficient corresponding to the value interval where the average interaction value is located is queried as the adjustment coefficient corresponding to the cluster image type.

7. The method according to any one of claims 1 to 6, characterized in that: Before adjusting the ranking of at least two image comments in the comment area of ​​the first song based on the weight coefficients of the images, the method further includes: Acquire characteristic information of the at least two image comments of the first song; the characteristic information includes at least one of the following: interaction information, release time information, relevance information, and image quality information, the interaction information is used to indicate the user's interaction behavior on the comments, and the relevance information is used to indicate the degree of relevance between the image and the first song; Based on the feature information of the at least two image comments of the first song, obtaining ranking scores of the at least two image comments; The step of adjusting the order of the at least two image comments in the comment area of ​​the first song based on the weight coefficient of each of the images includes: Adjusting the ranking scores of the at least two image comments based on the weight coefficients of the images to obtain adjusted ranking scores of the at least two image comments; Based on the adjusted ranking scores of the at least two image comments, the rankings of the at least two image comments in the comment area of ​​the first song are adjusted.

8. The method according to claim 7, characterized in that The method further comprises: Performing image description processing on each of the images to obtain description information of each of the images; The correlation information is obtained based on the tag information of the first song and the description information of each of the images.

9. The method according to claim 7, characterized in that: The step of adjusting the ranking scores of the at least two image comments based on the weight coefficients of the images to obtain the adjusted ranking scores of the at least two image comments includes: In the case where the first image comment includes at least two of the images, a target weight coefficient is obtained based on the weight coefficients of the at least two images; the target weight coefficient is a maximum value or an average value of the weight coefficients of the at least two images; and the first image comment is any one of the at least two image comments; The ranking score of the first image comment is adjusted based on the target weight coefficient to obtain an adjusted ranking score of the first image comment.

10. A device for processing image comments, characterized in that: The device comprises: A first acquisition module is used to acquire at least two image comments of a first song; each of the image comments contains at least one image; A classification processing module, used to perform classification processing on each image in the at least two image comments to obtain an image type of each image; A second acquisition module, used for acquiring a weight coefficient of each of the images based on an image type of each of the images; The sorting adjustment module is used to adjust the sorting of the at least two image comments in the comment area of ​​the first song based on the weight coefficient of each of the images.

11. A computer device, characterized in that: The computer device comprises a processor and a memory, wherein a computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement the method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to be executed by a processor to implement the method according to any one of claims 1 to 9.

13. A computer program product, characterized in that The computer program product comprises a computer program, which is loaded and executed by a processor to implement the method according to any one of claims 1 to 9.

Images