Title recommendation method and device, electronic equipment and computer storage medium
By extracting the characteristics and target attribute information of the multimedia content to be published to generate the title, the problem of difficulty for users to fill in the title is solved, and the content release efficiency and user experience are improved.
Patent Information
- Application Number
- CN202410981695.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-08
AI Technical Summary
When users create content in a content interaction platform, it is difficult for users to fill in the title, resulting in inefficient content release and affecting the user experience.
By obtaining the characteristics of the multimedia content to be published, candidate titles are determined, and target titles are generated based on target attribute information such as scene, location, time, etc., and recommended to users.
It improves the efficiency of content release, reduces the user's need to create titles, and improves the user experience.
Smart Images

Figure CN118916478A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of interactive content, and in particular to a title recommendation method, device, electronic device and computer storage medium. Background Art
[0002] At present, a variety of interactive contents are displayed in the content interaction platform, and users can perform interactive operations by triggering corresponding interactive contents in the content interaction platform. In the prior art, when users create content on the content interaction platform, they often do not know how to fill in the title and need to spend a long time to conceive the title, which leads to low content creation efficiency when users publish content and affects the content creation experience of users. Summary of the invention
[0003] The embodiments of the present application provide a title recommendation method, device, electronic device and computer storage medium, which can solve the technical problem of low content publishing efficiency caused by not knowing how to fill in the title.
[0004] The present application embodiment provides a title recommendation method, including:
[0005] Acquire multimedia content to be published;
[0006] Extracting features of the multimedia content to be published to obtain content features corresponding to the multimedia content to be published;
[0007] Determining candidate titles of the multimedia content to be published according to the content characteristics;
[0008] Based on the target attribute information of the multimedia content to be published and the candidate titles, a target number of target titles are generated, and the target number of target titles are recommended for the multimedia content to be published, wherein the target attribute information includes at least one of the scene information, location information, time information and facial information of the multimedia content to be published.
[0009] Accordingly, an embodiment of the present application provides a title recommendation device, comprising:
[0010] An acquisition module, used to acquire multimedia content to be published;
[0011] An extraction module, used to extract features of the multimedia content to be published, and obtain content features corresponding to the multimedia content to be published;
[0012] A determination module, used to determine the candidate titles of the multimedia content to be published according to the content features;
[0013] A generation module is used to generate a target number of target titles based on the target attribute information of the multimedia content to be released and the candidate titles, and recommend the target number of target titles for the multimedia content to be released, wherein the target attribute information includes at least one of the scene information, location information, time information and face information of the multimedia content to be released.
[0014] In addition, an embodiment of the present application also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to implement the title recommendation method provided in the embodiment of the present application.
[0015] In addition, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute any one of the title recommendation methods provided in the embodiment of the present application.
[0016] In addition, an embodiment of the present application also provides a computer program product, including a computer program, and when the computer program is executed by a processor, any one of the title recommendation methods provided in the embodiments of the present application is implemented.
[0017] In an embodiment of the present application, multimedia content to be published is obtained; feature extraction is performed on the multimedia content to be published to obtain content features corresponding to the multimedia content to be published; candidate titles of the multimedia content to be published are determined based on the content features; a target number of target titles are generated based on target attribute information of the multimedia content to be published and the candidate titles, and a target number of target titles are recommended for the multimedia content to be published, wherein the target attribute information includes at least one of scene information, location information, time information and face information of the multimedia content to be published, so as to automatically recommend titles of the multimedia content to be published to users based on the content features and the target attribute information, so that users can select titles of the multimedia content to be published from the recommended target titles, without the need for users to create titles of the multimedia content to be published by themselves, thereby improving the publishing efficiency of the multimedia content to be published. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] 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 those skilled in the art, other drawings can be obtained based on these drawings without creative work.
[0019] Figure 1 It is a scene schematic diagram of the title recommendation process provided by an embodiment of the present application;
[0020] Figure 2It is a flowchart of a title recommendation method provided in an embodiment of the present application;
[0021] Figure 3 is a schematic diagram of a multimedia content interface provided by an embodiment of the present application;
[0022] Figure 4 is a schematic diagram of a publishing interface provided in an embodiment of the present application;
[0023] Figure 5 is a schematic diagram of a process for generating a preset title library provided in an embodiment of the present application;
[0024] Figure 6 is a schematic diagram of a process for generating a target title provided by an embodiment of the present application;
[0025] Figure 7 It is another schematic diagram of the publishing interface provided in the embodiment of the present application;
[0026] Figure 8 is a schematic diagram of a location acquisition authorization interface provided in an embodiment of the present application;
[0027] Fig. 9 is a schematic diagram of a location acquisition authorization interface provided in an embodiment of the present application;
[0028] Fig.10 It is another schematic diagram of the publishing interface provided in the embodiment of the present application;
[0029] Fig.11 It is another schematic diagram of the publishing interface provided in the embodiment of the present application;
[0030] Fig.12 It is another schematic diagram of the publishing interface provided in the embodiment of the present application;
[0031] Fig.13 It is another schematic diagram of the publishing interface provided in the embodiment of the present application;
[0032] Fig.14 It is a flowchart of another title recommendation method provided in an embodiment of the present application;
[0033] Fig.15 is a structural diagram of a title recommendation device provided in an embodiment of the present application;
[0034] Fig.16 is a structural schematic diagram of another title recommendation device provided in an embodiment of the present application;
[0035] Fig.17 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0036] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.
[0037] The embodiment of the present application provides a method and device for recommending a title, an electronic device and a computer storage medium. The title recommendation device can be integrated in an electronic device, which can be a server or a terminal.
[0038] Among them, the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, network acceleration services (Content Delivery Network, CDN), as well as big data and artificial intelligence platforms.
[0039] The terminal may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected via wired or wireless communication, and this application does not limit this.
[0040] In addition, the term "plurality" in the embodiments of the present application refers to two or more than two. The terms "first" and "second" in the embodiments of the present application are used to distinguish descriptions and should not be understood as implying relative importance.
[0041] For example, Figure 1 As shown, the terminal obtains the multimedia content to be published, and sends the multimedia content to be published to the server. The server extracts features of the multimedia content to be published, and obtains content features corresponding to the multimedia content to be published; determines candidate titles of the multimedia content to be published according to the content features; generates a target number of target titles according to the target attribute information of the multimedia content to be published and the candidate titles, and recommends a target number of target titles for the multimedia content to be published, wherein the target attribute information includes at least one of scene information, location information, time information, and face information of the multimedia content to be published. The server sends the target number of target titles to the terminal. The terminal displays the target number of target titles on the publishing interface of the multimedia content to be published.
[0042] It should be noted that the description order of the following embodiments is not intended to limit the preferred order of the embodiments.
[0043] In this embodiment, the title recommendation device will be described. To facilitate the description of the title recommendation method of the present application, the title recommendation device will be integrated in a terminal for detailed description, that is, the terminal will be used as the execution subject for detailed description.
[0044] See also Figure 2 , Figure 2 : is a flowchart of a title recommendation method provided by an embodiment of the present application. The title recommendation method may include:
[0045] S201: Acquire multimedia content to be published.
[0046] Among them, the multimedia content to be published refers to the information that the user is about to publish on the client, and its type can be set according to actual conditions. For example, the multimedia content to be published can include at least one of text, audio, image and video, which is not limited in this embodiment of the present application.
[0047] Optionally, the terminal may display a multimedia content interface, which includes published multimedia content and a publishing control, and obtains the multimedia content to be published through the publishing control. Figure 3 As shown, in response to a triggering operation on a publishing control, an album interface is displayed, the album interface including at least one image and / or at least one video, and in response to a selection operation on at least one image and / or at least one video, multimedia content to be published is generated according to the image corresponding to the selection operation and / or the video corresponding to the selection operation, or the terminal may also display images and / or videos, and in response to an operation to publish images and / or videos, multimedia content to be published is generated according to the images and / or videos.
[0048] The method for the terminal to obtain the multimedia content to be published can be set according to actual conditions, and the embodiment of the present application does not limit it here.
[0049] S202: Extract features of the multimedia content to be published to obtain content features corresponding to the multimedia content to be published.
[0050] Among them, the method of feature extraction for the multimedia content to be published can be set according to the actual situation. For example, feature extraction can be performed on the multimedia content to be published through a feature extraction model, or feature extraction can be performed on the multimedia content to be published through a Histogram of Oriented Gradients (HOG), which is not limited in the embodiments of the present application.
[0051] The type of feature extraction model can be set according to actual conditions. For example, the feature extraction model can be a convolutional neural network model or a long short-term memory network model, which is not limited in the embodiments of the present application.
[0052] In some embodiments, when the multimedia content to be published includes images and / or videos, feature extraction may be performed on the images and / or videos in the multimedia content to be published. When the multimedia content to be published includes images and / or videos, if the multimedia content to be published includes multiple images and / or videos, feature extraction may be performed on any image in the multimedia content to be published (for example, feature extraction may be performed on the first object in the multimedia content to be published, and the first object may be an image or a video. When the first object is an image, the first object may be extracted as follows: Figure 4 As shown), extract features from any video in the multimedia content to be released, and / or fuse multiple images and / or videos to obtain a fused object, and extract features from the fused object.
[0053] When performing feature extraction on a video in the multimedia content to be released, feature extraction may be performed on all video frames in the video, or feature extraction may be performed on key frames in the video, which is not limited in the embodiments of the present application.
[0054] S203: Determine candidate titles of the multimedia content to be published according to content features.
[0055] Among them, the method of determining the candidate titles of the multimedia content to be released based on the content characteristics can be set according to the actual situation. For example, the candidate titles of the multimedia content to be released can be determined according to the content characteristics through a preset title library or a large language model. The embodiment of the present application is not limited here.
[0056] When determining candidate titles of the multimedia content to be published according to content features through a large language model, the process of determining candidate titles of the multimedia content to be published according to the content features may be:
[0057] Selecting prompt words that match the content features from a preset prompt word library to obtain target prompt words;
[0058] Generate prompt information according to the multimedia content to be published and the target prompt word;
[0059] Input the prompt information into the large language model to obtain candidate titles.
[0060] When determining candidate titles of the multimedia content to be published according to content characteristics through a preset title library, determining candidate titles of the multimedia content to be published according to the content characteristics includes:
[0061] Determining feature similarity between content features and features corresponding to titles in a preset title library;
[0062] The titles corresponding to the features that meet the feature similarity of the preset similarity condition are determined as candidate titles.
[0063] Among them, the preset similarity conditions can be set according to actual conditions. For example, the titles corresponding to the features with feature similarities greater than the preset similarity threshold can be determined as candidate titles, or the feature similarities can be sorted in order from large to small, and then the titles corresponding to the features with the first preset target number of feature similarities can be determined as candidate titles. The embodiments of the present application are not limited here.
[0064] In an embodiment of the present application, the feature similarity between the content feature and the feature corresponding to the title in the preset title library is determined, and the title corresponding to the feature that meets the preset similarity condition is determined as the candidate title. The candidate title is determined by means of the multimedia content to be published and the preset title library.
[0065] In some embodiments, before determining the feature similarity between the content feature and the feature corresponding to the title in the preset title library, the method further includes:
[0066] Get the title information published historically in the client;
[0067] Performing expansion processing according to the title information to obtain an expanded title, wherein the expansion processing includes keyword replacement processing and / or sentence replacement processing;
[0068] Generate a preset title library based on the expanded titles.
[0069] The title information historically published in the client refers to at least one of the title, topic, and query information used in the multimedia content historically published by other users in the client. Keyword replacement processing refers to replacing words in the title information, and sentence replacement processing refers to generating an expanded title based on the title.
[0070] The method of keyword replacement processing can be set according to actual conditions. For example, the title information can be processed by word segmentation and part-of-speech recognition to obtain nouns, and the nouns in the title information can be replaced with the parent nouns of the nouns to obtain an expanded title. Alternatively, the nouns in the title information can be replaced with synonyms to obtain an expanded title. The embodiments of the present application are not limited here.
[0071] The parent noun of a noun refers to the parent noun of the noun in the target dictionary, and the target dictionary may be, for example, the Chinese version of wordnet. The method for searching for synonyms may be selected according to actual conditions, for example, synonyms may be searched by the word2vec method, which is not limited in the present embodiment.
[0072] The sentence replacement processing method can be set according to actual conditions. For example, the length of the title information can be set, and the title information with the set length can be translated into title information in other languages, and then the title information in other languages can be translated into Chinese title information to obtain the expanded title. Alternatively, the target historical multimedia content and the title of the target historical multimedia content in the historically released multimedia content can be stored in a recall library, and the title information can be expanded according to the recall library through the trained graphic model to obtain the expanded title.
[0073] In an embodiment of the present application, title information historically published in the client is obtained, and expansion processing is performed based on the title information to obtain an expanded title, the expansion processing includes keyword replacement processing and / or sentence replacement processing, and a preset title library is generated based on the expanded title to enrich the titles in the preset title library.
[0074] In some embodiments, before determining the feature similarity between the content feature and the feature corresponding to the title in the preset title library, the method further includes:
[0075] Get the title information published historically in the client;
[0076] Performing filtering and / or deduplication processing on the title information to obtain a processed title, wherein the filtering processing includes rule filtering processing and / or model filtering processing, and the deduplication processing includes character deduplication processing and / or semantic deduplication processing;
[0077] Generate a preset title library based on the processed titles.
[0078] Among them, rule filtering refers to filtering the title information according to preset rules. The preset rules can be set according to actual conditions, for example, the preset rules can be set according to whether it is illegal, category, length, part of speech, content and at least one of keywords, which is not limited in the embodiments of the present application.
[0079] When setting preset rules based on whether there is a violation, the preset rules may be safety disallowance rules, ecological disallowance rules, and tonality disqualification rules. When setting preset rules based on length, the preset rules may be to retain 2 to 12 characters of title information. When setting preset rules based on part of speech, the preset rules may be to filter out title information containing words with specific content, such as brand words, quantifiers, specific locations (except well-known locations), time, and title information of constellations. When setting preset rules based on keywords, the preset rules may be to filter out title information containing preset keywords and icons, such as niche languages or symbols (niche language post-symbols may include pinyin, traditional Chinese, other languages except Chinese and English, except idiomatic words), niche things (niche things such as things, hobbies, or names that only a small number of people know), comparisons, and subjective extreme judgment words (subjective extreme judgment words may be, for example, the most, historically, and must, etc.). When preset rules are set according to the content, the preset rules can be used to filter out descriptive narratives and / or incomplete title information. The descriptive narrative title information generally contains specific complex events, and the incomplete title information may include unclear CP avatars, stuck points, or similarities.
[0080] Model filtering processing refers to classifying the title information through the trained text model to achieve filtering of the title information. The trained text model can be set according to actual conditions, for example, the trained text model can be Bert or ERNIE, which is not limited in the present embodiment of the application.
[0081] Character deduplication refers to changing the case of characters in the title information to lowercase and removing similar characters in the title information. Semantic deduplication refers to removing similar title information. The semantic deduplication method can be processed according to the actual situation. For example, the title information is segmented, and the embedding vector of each title information is counted through the word embedding model to determine the distance between the embedding vectors and the length difference between the title information. Based on the distance and length difference, similar title information is determined and stored in the same title set.
[0082] In an embodiment of the present application, title information historically published in the client is obtained, and the title information is filtered and / or deduplicated to obtain processed titles, the filtering processing includes rule filtering processing and / or model filtering processing, and the deduplication processing includes character deduplication processing and / or semantic deduplication processing, and a preset title library is generated based on the processed titles to reduce the number of titles in the preset title library.
[0083] In some embodiments, after obtaining the processed titles and / or expanded titles, the processed titles and / or expanded titles can be classified according to preset types, and then a preset title library is generated based on the classified titles, so that the title classifications are stored in the preset title library.
[0084] The preset type can be set according to the actual situation. For example, the preset type can be set according to at least one of the scene of the title, the product tone, and the characteristics of the title. When the preset type is set according to the scene of the title, the preset type can be people, food, animals, sports, fashion, scenery, and daily life. When the preset type is set according to the product tone, the preset type can be surreal imagination, inner feelings, perception (perception can include hearing, touch, vision, taste, and smell), time, objective description, and context of things.
[0085] Surreal imagination refers to the imaginary plot, dialogue and / or narration from the second or third person perspective, which is used to stimulate interaction. Inner feelings refer to the description of inner feelings from the first person perspective, self-motivation and / or inner monologue, which are used to stimulate publishing and resonate. Time refers to date, season, and duration, which are used to leave memories of the moment and promote nostalgia. Objective description refers to the concreteness of entities, which meets the basic expression of daily life. Event context refers to the summary vocabulary for videos or multiple pictures, which is used to meet the expression of life fragments.
[0086] When setting preset types according to the characteristics of the title, the preset types can be general titles (general titles are applicable to all scenarios), scene-based titles (scene-based titles are applicable to individual scenarios), fine-grained titles, and ultra-fine-grained titles. Among them, when a title with qualifiers appears, a limiting tag is added to the title. The qualifiers may be, for example, scenic spots, time, materials, colors, shapes, and occupations.
[0087] In an embodiment of the present application, the titles may be classified and stored in a preset title library, so that candidate titles may be generated according to the categories of the titles in the preset title library, thereby improving the diversity of the candidate titles.
[0088] For example, Figure 5 As shown, after obtaining the title information historically published in the client, the title information is expanded to obtain the expanded title, the expanded title is filtered and deduplicated to obtain the processed title, and the processed title is classified and stored in a preset title library.
[0089] S204. Generate a target number of target titles based on target attribute information of the multimedia content to be published and candidate titles, and recommend a target number of target titles for the multimedia content to be published, wherein the target attribute information includes at least one of scene information, location information, time information, and face information of the multimedia content to be published.
[0090] In some embodiments, when the target attribute information includes location information, generating a target number of target titles according to the target attribute information and candidate titles of the multimedia content to be published includes:
[0091] Obtaining location information corresponding to images and / or videos in the media content to be published;
[0092] Adding location information to a target position of a first at least part of a candidate title in the candidate titles to obtain a first candidate title;
[0093] A target number of target titles is determined based on the first candidate title and a second at least part of the candidate titles among the candidate titles.
[0094] The location information corresponding to the image refers to the location information where the image was taken, and the location information corresponding to the video refers to the location information where the video was taken. When the location information obtained is a place name, the location information can be directly added to the target location. When the location information obtained is longitude and latitude, the location information can be converted into a place name first, and then added to the target location.
[0095] The first at least partial candidate title and the second at least partial candidate title constitute a candidate title, and the first at least partial candidate title and the second at least partial candidate title can be set according to actual conditions. For example, the candidate titles may include 5, the first at least partial candidate title may be 4 of the 5 candidate titles, and the second at least partial candidate title may be 1 of the 5 candidate titles. In this case, 4 of the 5 target titles include location information, and 1 target title does not include location information. Alternatively, the first at least partial candidate title may be 3 of the 5 candidate titles, and the second at least partial candidate title may be 2 of the 5 candidate titles. In this case, 3 of the 5 target titles include location information, and 2 target titles do not include location information. The embodiments of the present application are not limited here.
[0096] The target position may be any position in the first at least partial candidate title, for example, the target position may be the front position, the middle position, and the back position in the first at least partial candidate title, and the embodiment of the present application is not limited thereto.
[0097] The target positions of different candidate titles in the first at least part of the candidate titles may be the same or different, and this embodiment of the present application does not limit this.
[0098] In an embodiment of the present application, location information corresponding to images and / or videos in the media content to be published is obtained, the location information is added to the target position of the first at least part of the candidate titles in the candidate titles to obtain the first candidate title, and a target number of target titles is determined based on the first candidate title and the second at least part of the candidate titles in the candidate titles, so that the target number of target titles includes not only target titles containing location information, but also target titles without location information, thereby improving the diversity of the target titles and making the target titles more in line with user needs, thereby improving the publishing rate of multimedia content and further improving user experience.
[0099] In some embodiments, the target attribute information also includes scene information, and obtaining location information corresponding to the image and / or video in the media content to be published includes:
[0100] Matching the scene information with the preset scene information in the scene set;
[0101] If there is preset scene information matching the scene information, then location information corresponding to the image and / or video in the media content to be published is obtained.
[0102] Among them, the preset scene information can be set according to actual conditions. For example, the preset scene information can be scenery and / or food. The food can be, for example, drinks, desserts, Chinese food and / or Western food. The embodiments of the present application are not limited here.
[0103] In an embodiment of the present application, since some scenes do not require location information and some scenes require location information, a scene collection can be set in advance, and then the scene information of the multimedia content to be published is matched with the preset scene information in the scene collection. If there is preset scene information that matches the scene information, the location information corresponding to the image and / or video in the media content to be published is obtained, so that when the scene information of the multimedia content to be published is the preset scene information, the target title recommended for the multimedia content to be published includes the location information, so that the generated target title is more in line with the user's needs.
[0104] In some embodiments, obtaining location information corresponding to an image and / or video in the media content to be published includes:
[0105] Determining whether the text information in the media content to be published includes location information;
[0106] If the text information in the media content to be published does not include location information, then the location information corresponding to the image and / or video in the media content to be published is obtained.
[0107] In an embodiment of the present application, it is determined whether the text information in the media content to be published contains location information. If the text information in the media content to be published does not contain location information, the location information corresponding to the image and / or video in the media content to be published is obtained to avoid inconsistency between the location information in the target title generated based on the location information corresponding to the image and / or video and the location information in the text information.
[0108] In some embodiments, after determining whether the text information in the to-be-published media content includes location information, the method further includes:
[0109] If the text information in the media content to be published includes location information, determining whether the scene information corresponding to the image and / or video in the media content to be published is the target scene information;
[0110] If the scene information corresponding to the image and / or video in the media content to be published is the target scene information, the location information corresponding to the image and / or video in the media content to be published is obtained.
[0111] The target scene information may be a preset scene information in the scene set. In an embodiment of the present application, when the text information in the media content to be published includes location information and the scene information corresponding to the image and / or video in the media content to be published is the target scene information, it indicates that the user may want to add the location information corresponding to the image and / or video in the target title. Therefore, the location information corresponding to the image and / or video in the media content to be published is obtained, so that the target title generated based on the location information corresponding to the image and / or video better meets the needs of the user.
[0112] In some embodiments, adding location information to a target position of a first at least part of a candidate title in a candidate title to obtain a first candidate title includes:
[0113] Determine whether the length of the location information meets a preset length condition;
[0114] If the length of the location information meets the preset length condition, the location information is added to the target position of the first at least part of the candidate title in the candidate title to obtain the first candidate title.
[0115] Among them, the preset length condition can be set according to actual conditions. For example, location information with less than 4 characters can be determined as location information that meets the preset length condition. The embodiment of the present application does not limit this.
[0116] In an embodiment of the present application, it is determined whether the length of the location information meets the preset length condition; if the length of the location information meets the preset length condition, the location information is added to the target position of the first at least part of the candidate title in the candidate title to obtain the first candidate title, and the length of the target title generated based on the first candidate title is limited to simplify the target title so that the target title better meets the needs of users.
[0117] In some embodiments, adding location information to a target position of a first at least part of a candidate title in a candidate title to obtain a first candidate title includes:
[0118] Adding location information to a target position of a first at least part of the candidate titles in the candidate titles to obtain an initial title;
[0119] The first candidate title is determined based on the candidate titles corresponding to the first initial title and the second initial title. The first initial title is the initial title corresponding to a length that meets the first preset length condition, and the second initial title is the initial title corresponding to a length that does not meet the first preset length condition.
[0120] Among them, the first preset length condition can be set according to actual conditions. For example, an initial title with less than 20 characters can be determined as the initial title corresponding to the length that meets the first preset length condition. The embodiment of the present application is not limited here.
[0121] In an embodiment of the present application, location information is added to the target position of the first at least part of the candidate title in the candidate title to obtain an initial title, and the first candidate title is determined based on the candidate titles corresponding to the first initial title and the second initial title. The first initial title is the initial title corresponding to the length that meets the first preset length condition, and the second initial title is the initial title corresponding to the length that does not meet the first preset length condition. When the length of the initial title does not meet the first preset length condition, the candidate title corresponding to the initial title is used as the target title to simplify the target title so that the target title better meets the needs of users.
[0122] In some embodiments, adding location information to a target position of a first at least part of a candidate title in a candidate title to obtain a first candidate title includes:
[0123] Determine whether the length of the location information meets a preset length condition;
[0124] If the length of the location information meets the preset length condition, then adding the location information to the target position of the first at least part of the candidate title in the candidate title to obtain an initial title;
[0125] The first candidate title is determined based on the candidate titles corresponding to the first initial title and the second initial title. The first initial title is the initial title corresponding to a length that meets the first preset length condition, and the second initial title is the initial title corresponding to a length that does not meet the first preset length condition.
[0126] In the embodiment of the present application, the length of the target title is limited by the length of the location information and the length of the initial title, and the target title is simplified so that the target title better meets the needs of users.
[0127] In some embodiments, obtaining location information corresponding to a video in the media content to be published includes:
[0128] Acquire location information corresponding to a portion of frame images of a video in the media content to be published, where the portion of frame images includes a cover image and / or a preset number of frame images of the video in the media content to be published;
[0129] The location information corresponding to the partial frame images is determined as the location information corresponding to the video in the media content to be published.
[0130] Among them, the first preset number of frame images can be set according to actual conditions. For example, the first preset number of frame images can be the first frame image or the first 3-4 frame images, which is not limited in the embodiment of the present application.
[0131] In an embodiment of the present application, since the location where the video was shot may include multiple locations, location information corresponding to partial frame images of the video in the media content to be released is obtained, and the partial frame images include a cover image of the video in the media content to be released and / or a previous preset number of frame images. The location information corresponding to the partial frame images is determined as the location information corresponding to the video in the media content to be released, so as to avoid errors in determining the location information corresponding to the video.
[0132] It should be noted that when some frame images include the cover image of the video in the media content to be released and the previous preset number of frame images, the location information corresponding to the cover image can be obtained first, and when the location information of the cover image cannot be obtained, the location information corresponding to the previous preset number of frame images can be obtained.
[0133] In some embodiments, the target attribute information includes time information, and generating a target number of target titles according to the target attribute information and candidate titles of the multimedia content to be published includes:
[0134] Matching the time information with the candidate time information in the candidate title, the time information including at least one of shooting time information, editing time information and historical publishing time information of the image and / or video in the multimedia content to be published;
[0135] A target number of target titles is determined based on the candidate titles corresponding to the candidate time information that matches the time information.
[0136] When the time information is the time information of a video, the time information may be the time information corresponding to a portion of the frame images of the video, where the portion of the frame images includes a cover image of the video in the media content to be released and / or a preset number of frame images.
[0137] The time information may be specific time information, and the candidate time information may be a certain time period. For example, the candidate time information may be at least one of season information, summer and winter vacation time information, Spring Festival information, day of the week information, and morning, noon, and evening information.
[0138] The terminal first determines whether the time information matches the season information, then determines whether the time information matches the winter and summer vacation time information, then determines whether the time information matches the Spring Festival information, then determines whether the time information matches the day of the week information, and finally determines whether the time information matches the morning, noon and evening information.
[0139] The time information matches the candidate time information, which means that the time corresponding to the time information is within the time period corresponding to the candidate time information.
[0140] In an embodiment of the present application, time information is matched with candidate time information in candidate titles, where the time information includes at least one of shooting time information, editing time information, and historical release time information of images and / or videos in the multimedia content to be released. A target number of target titles is determined based on the candidate titles corresponding to the candidate time information that matches the time information, to avoid conflicts between the time in the target title and the time of the images and / or videos in the multimedia content to be released, so that the target title is more in line with user needs, thereby improving the release rate of multimedia content and further improving user experience.
[0141] In some embodiments, generating a target number of target titles according to target attribute information of the multimedia content to be published and candidate titles includes:
[0142] generating target titles according to target attribute information of the multimedia content to be published and candidate titles, wherein the number of target titles is greater than the target number;
[0143] Determine the target title corresponding to the maximum feature similarity among the feature similarities corresponding to the target titles greater than the target number as the first target title;
[0144] According to a random rule, a first target number of second target titles are selected from the remaining target titles, and the remaining target titles are target titles other than the first target title among the target titles greater than the target number;
[0145] A target number of target titles is determined based on the first target title and the second target title.
[0146] The maximum feature similarity among the feature similarities corresponding to the target titles greater than the target number refers to the maximum feature similarity among the feature similarities between the candidate titles and the content features corresponding to the target titles greater than the target number. The target number is one more than the first target number.
[0147] For example, the target number is 5, the first target number is 4, and the target titles greater than the target number include target title t1, target title t2, target title t3, target title t4, target title t5, target title t6, and target title t7. The feature similarity between the candidate title and the content feature corresponding to target title t1 is 0.9, the feature similarity between the candidate title and the content feature corresponding to target title t2 is 0.89, the feature similarity between the candidate title and the content feature corresponding to target title t3 is 0.83, the feature similarity between the candidate title and the content feature corresponding to target title t4 is 0.84, and the feature similarity between the candidate title and the content feature corresponding to target title t5 is 0.81. The feature similarity between the candidate title and content features corresponding to target title t4 is 0.87, the feature similarity between the candidate title and content features corresponding to target title t7 is 0.88, and the maximum feature similarity is 0.9. Therefore, the first target title is target title t1, and the remaining target titles are target title t2, target title t3, target title t4, target title t5, target title t6 and target title t7. Target title t2, target title t3, target title t5 and target title t7 are randomly selected from the remaining target titles as the second target title. At this time, the five target titles are target title t1, target title t2, target title t3, target title t5 and target title t7.
[0148] In an embodiment of the present application, based on the target attribute information and candidate titles of the multimedia content to be published, a target number of target titles greater than the target number can be generated, and then a target number of target titles are screened out from the target titles greater than the target number based on feature similarity and random rules, so that not only the relevance of the target titles and the multimedia content to be published can be guaranteed, but also the diversity of the target titles can be guaranteed, so that the target titles are more in line with user needs, the publishing rate of the multimedia content to be published is increased, and the user experience is improved.
[0149] In some embodiments, selecting a first target number of second target titles from the remaining target titles according to a random rule includes:
[0150] Determine the preset type corresponding to each remaining target title;
[0151] According to a random rule, a first target number of second target titles are screened out from the remaining target titles corresponding to the preset type matching the target preset type.
[0152] Among them, the target preset type can be set according to actual conditions. For example, when the preset types include people, food, animals, sports, fashion, scenery, daily life, surreal imagination, inner feelings, perception, time, objective description, context of things, general titles, general titles for scenes, fine-grained titles and ultra-fine-grained titles, the target preset type can be a fine-grained title, an objective description title and a general title, and the embodiments of the present application are not limited here.
[0153] The number of target titles corresponding to each target preset type can be set according to actual conditions. For example, the first target number is 4, and the target preset types are fine-grained titles, objective description titles, and general titles. Then, a second target title corresponding to a fine-grained title, two second target titles corresponding to customer description titles, and a second target title corresponding to a general title can be screened out from the remaining target titles. This embodiment of the present application is not limited here.
[0154] When there are multiple remaining target titles corresponding to a preset type that matches the target preset type, a second target number of target titles can be randomly selected from the remaining target titles corresponding to the preset type that matches the target preset type as the second target title, or the target titles with the first second target number of feature similarities among the remaining target titles corresponding to the preset type that matches the target preset type can be used as the second target title. The embodiments of the present application are not limited here.
[0155] For example, the number of second targets is 1, and the preset type that matches the preset target type is a large general title. The target titles corresponding to the large general titles in the remaining target titles include target title t2, target title t3, and target title t4. It is necessary to screen out a target title from the target titles corresponding to the large general titles as the second target title. The feature similarity between the candidate title corresponding to target title t2 and the content feature is 0.89, the feature similarity between the candidate title corresponding to target title t3 and the content feature is 0.83, and the feature similarity between the candidate title corresponding to target title t4 and the content feature is 0.84. When randomly selecting one from the target titles corresponding to the large general titles in the remaining target titles as the second target title, any one of target title t2, target title t3, and target title t4 can be used as the second target title. When the target title with the first feature similarity among the target titles corresponding to the large general titles in the remaining target titles is used as the second target title, since the feature similarity of target title t2 is the largest, target title t2 is used as the second target title.
[0156] In an embodiment of the present application, a preset type corresponding to each remaining target title is determined, and according to a random rule, a first target number of second target titles are screened out from the remaining target titles corresponding to the preset type that matches the target preset type, so that the second target titles are more in line with user needs, further improving the publication rate of the multimedia content to be published and improving the user experience.
[0157] In some embodiments, when the target attribute information includes scene information, generating a target number of target titles according to the target attribute information and candidate titles of the multimedia content to be published includes:
[0158] According to the scene information, a target number of target titles are screened out from the candidate titles.
[0159] Among them, according to the scene information, a target number of target titles of the first preset type can be screened out from the candidate titles. For example, when the scene information is a screenshot or an emoticon package, the target title is a candidate title corresponding to the general title. Alternatively, according to the scene information, the first target number of candidate titles can be screened out from the candidate titles to obtain the target number of target titles. For example, when the scene information is a landscape, it is determined to screen out the target titles from the candidate titles based on the feature similarity. At this time, the candidate titles corresponding to the first target number of feature similarities can be determined as the target titles.
[0160] In some embodiments, when the target attribute information includes facial information, generating a target number of target titles according to the target attribute information and candidate titles of the multimedia content to be published includes:
[0161] Determine user information based on facial information;
[0162] Determine the matching degree between the user information and the candidate titles;
[0163] Sort the matching degrees from large to small, and determine the candidate titles corresponding to the first target number of matching degrees as the target number of target titles.
[0164] Among them, the user information can be set according to actual conditions. For example, the user information may include at least one of gender and age, which is not limited in the embodiments of the present application.
[0165] In some embodiments, when the target attribute information includes location information, time information, scene information, and face information, generating a target number of target titles according to the target attribute information and candidate titles of the multimedia content to be published includes:
[0166] Filter the target title from the candidate titles according to the time information, scene information and face information;
[0167] Generate a target number of target titles based on the target title and location information.
[0168] The process of selecting the target title from the candidate titles according to the time information, the scene information and the face information may be as follows:
[0169] Matching the time information with the candidate time information in the candidate title;
[0170] Filtering candidate titles corresponding to candidate time information matching the time information from the candidate titles to obtain a third target number of target titles;
[0171] According to the scenario information, a fourth target number of target titles is screened out from the third target number of target titles, and the third target number is greater than the fourth target number;
[0172] Determining user information based on the facial information, and determining a degree of matching between the user information and a third target number of target titles;
[0173] Sort the matching degrees from large to small, and determine the target titles corresponding to the first five target quantity matching degrees as the target titles of the fifth target quantity. The fifth target quantity is equal to the target quantity, or the fifth target quantity is greater than the target quantity and the fifth target quantity is less than the fourth target quantity.
[0174] When the fifth target quantity is equal to the target quantity, the process of generating the target title of the target quantity according to the target title and the location information may be:
[0175] Adding location information to the first at least part of the target titles in the fifth target number of target titles to obtain the target title with the added location information;
[0176] The target title of the added location information and the second at least part of the target titles in the target title of the fifth target quantity are determined as the target titles of the target quantity.
[0177] For example, Figure 6 As shown, the number of targets is 5. When the multimedia content to be published includes images and / or videos, feature extraction is performed on the images and / or videos to obtain content features. According to the content features, candidate titles are screened from the preset title library. Then, according to the time information, scene information, face information, location information and candidate titles, 5 target titles are generated. The 5 target titles can be as follows Figure 7 As shown (it should be understood that Figure 7 If the 5 target titles are not displayed, you can slide left and right to display the 5 target titles).
[0178] When the fifth target quantity is greater than the target quantity, the process of generating the target title of the target quantity according to the target title and the location information may be:
[0179] Determine the target title corresponding to the maximum feature similarity among the feature similarities corresponding to the target titles of the fifth target number as the first target title;
[0180] According to the random rule, the second target titles of the first target quantity are screened out from the remaining target titles, the remaining target titles are the target titles of the fifth target quantity except the first target title, and the sum of the number of the first target title and the second target title is the target quantity;
[0181] Adding location information to the first or at least a portion of the target title in the first target title and the second target title to obtain the target title after the location information is added;
[0182] The target title to which the location information is added and the second at least partial target title are determined as the target number of target titles.
[0183] In some embodiments, recommending a target number of target titles for the multimedia content to be published includes:
[0184] Displaying the target titles of the multimedia content to be published and the target quantity via the publishing interface;
[0185] In response to a selection operation on a target title, displaying the target title corresponding to the selection operation at a title position of the multimedia content to be published;
[0186] In response to a publishing operation on the multimedia content to be published, a target title corresponding to the selection operation is published as a title of the multimedia content to be published together with the multimedia content to be published.
[0187] After acquiring the multimedia content to be published, the terminal may display a publishing interface, display the multimedia content to be published on the publishing interface, generate a target number of target titles, and then display the target number of target titles on the publishing interface.
[0188] Optionally, when the publishing interface is displayed, if the client does not have the authority to obtain the location of the terminal, a location acquisition authorization interface is displayed, and the location of the terminal is obtained in response to a confirmation operation on the location acquisition authorization interface.
[0189] For example, the publishing interface can be Figure 4 As shown, the location acquisition authorization interface can be as follows Figure 8 As shown, in response to a trigger operation of "Allow once" or "Allow when using App" in the location acquisition authorization interface, the location of the terminal is acquired.
[0190] Optionally, when the publishing interface is displayed, if the client does not have permission to obtain the location information of pictures and / or videos, a location acquisition authorization interface for obtaining pictures and / or videos is displayed, and in response to a confirmation operation on the location acquisition authorization interface, the location information corresponding to the pictures and / or videos of the multimedia content to be published is obtained.
[0191] For example, the publishing interface can be Figure 4 As shown, the displayed location acquisition authorization interface can be as follows Fig. 9 As shown, in response to a click operation of "Agree" on the location acquisition authorization interface, location information corresponding to the picture and / or video of the multimedia content to be published is acquired.
[0192] It should be noted that, in response to a denial operation on the location acquisition authorization interface, a target title can still be generated, but the target title does not include location information.
[0193] In some embodiments, the location acquisition authorization interface may also include a close prompt message, which is used to cancel the location information acquisition permission. Fig. 9 As shown, the shutdown prompt information may include "You can turn off this feature in Settings-Privacy Settings-More".
[0194] In some embodiments, after the publishing interface is displayed, since it takes time to generate the target title, when the target title is not displayed on the publishing interface, "Smart title generating" may be displayed at the title position of the multimedia content to be published, for example, Figure 4 As shown, when the target title is displayed on the publishing interface, "Smart title has been generated" can be displayed at the title position of the multimedia content to be published, for example, Figure 7 shown.
[0195] In some embodiments, when the target title is not displayed on the publishing interface, if a title creation operation is detected, the process of generating the target title is stopped, that is, the target title is neither generated nor displayed.
[0196] In some embodiments, when the generation of the target title fails, a retry control may be displayed on the publishing interface so that the target title is generated in response to a triggering operation on the retry control. Fig.10 shown.
[0197] In some embodiments, after the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published in response to the selection operation on the target title, the method further includes:
[0198] In response to a trigger operation on a target title at the title position, displaying the target title in an editing state;
[0199] In response to an editing operation on a target title in an editing state, editing the target title in the editing state to obtain an edited target title;
[0200] Displays the edited target title in the title position.
[0201] The editing operation may include at least one of a deletion operation, a modification operation and a creation operation.
[0202] In some embodiments, after the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published in response to the selection operation on the target title, the method further includes:
[0203] Hide other target titles in the publishing interface;
[0204] In response to a trigger operation on a target title at a title position, displaying the target title in an editing state, including:
[0205] In response to a trigger operation on a target title at a title position, the target title in an editing state is displayed, and other target titles are displayed.
[0206] At this time, the terminal may replace the target title at the title position with another target title, or modify the target title.
[0207] For example, after displaying the target title of the target quantity on the publishing interface, the publishing interface can Figure 7 As shown, the target title is "20° Spring PLOG~". In response to the selection operation of the target title, "20° Spring PLOG~" is displayed at the title position of the multimedia content to be published and other target titles are hidden. At this time, the publishing interface can be as shown in FIG. Fig.11 As shown, in response to the triggering operation of the target title at the title position, after the target title in the editing state is displayed, the publishing interface can be as shown in FIG. Fig.12 shown.
[0208] In an embodiment of the present application, in response to a selection operation on a target title, the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published, and other target titles are hidden in the publishing interface. In response to a trigger operation on the target title at the title position, the target title in an editing state is displayed, and other target titles are displayed, so that not only the publishing interface can be simplified, but also the target title can be edited.
[0209] In some embodiments, the multimedia content to be published includes multiple objects, the objects include images and / or videos, the target number of target titles is a title determined according to the first object of the multiple objects, and after the multimedia content to be published and the target number of target titles are displayed via the publishing interface, the following further includes:
[0210] In response to a replacement operation on a first object, obtaining the first object after replacement;
[0211] According to the first object after replacement, the target title of the target quantity is updated;
[0212] Updated goal title to show goal quantity via the publish interface.
[0213] The first object is the first object. The process of updating the target number of target titles according to the first object after replacement can be specifically referred to the process of generating the target number of target titles according to the first object, and this embodiment will not be repeated here.
[0214] In the embodiment of the present application, the target title can be updated according to the real-time changes of the first object, so that the target title better meets the needs of the user.
[0215] In some embodiments, in response to a replacement operation on the first object, it is determined whether a selection operation on the target title has been responded to, and the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published; if a selection operation on the target title has been responded to, and the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published, the target number of target titles is not updated.
[0216] In some embodiments, the publishing interface may also include a recommended title deletion control, such as Fig.13 In response to the triggering operation of the recommended title deletion control, the target number of target titles are deleted, and "No more recommended titles after closing, you can use them again after re-downloading" is displayed.
[0217] In some embodiments, the publishing interface can also display location recommendation information. In this case, the publishing interface, for example, Figure 7 shown.
[0218] In some embodiments, for the simplicity of the publishing interface, when the target number of target titles are displayed on the publishing interface, the location recommendation information is not displayed, and when the target number of target titles are not displayed on the publishing interface, the location recommendation information is displayed.
[0219] In some embodiments, in response to a text filling operation, location recommendation information is displayed and the text information is displayed.
[0220] As can be seen from the above, in the embodiment of the present application, multimedia content to be published is obtained; feature extraction is performed on the multimedia content to be published to obtain content features corresponding to the multimedia content to be published; candidate titles for the multimedia content to be published are determined based on the content features; a target number of target titles are generated based on the target attribute information of the multimedia content to be published and the candidate titles, and a target number of target titles are recommended for the multimedia content to be published, wherein the target attribute information includes at least one of scene information, location information, time information and facial information of the multimedia content to be published, so as to automatically recommend titles of the multimedia content to be published to users based on the content features and the target attribute information, so that users can select titles of the multimedia content to be published from the recommended target titles, without the need for users to create titles of the multimedia content to be published by themselves, thereby improving the publishing efficiency of the multimedia content to be published.
[0221] According to the following Fig.14 , the title recommendation method provided by this application is further explained.
[0222] See also Fig.14 , Fig.14 : is a flowchart of a title recommendation method provided by an embodiment of the present application. The title recommendation method may include:
[0223] S1401. In response to a to-be-published instruction for multimedia content to be published, recommend a target number of target titles for the multimedia content to be published, where the target titles are titles determined based on target attribute information of the multimedia content to be published, where the target attribute information includes at least one of scene information, location information, time information, and facial information of the multimedia content to be published.
[0224] The instruction to be published refers to an instruction for generating a target title and displaying a publishing interface. The method for obtaining the instruction to be published can be set according to actual conditions, and the embodiment of the present application does not limit it here.
[0225] For example, when the terminal acquires images and / or videos, in response to an operation to be published on the images and / or videos, it generates multimedia content to be published based on the images and / or videos, and generates instructions to be published (exemplarily, the operation to be published is a trigger operation on a publishing control, which automatically displays the publishing control when the images and / or videos are acquired, and in response to the trigger operation on the publishing control, generates multimedia content to be published based on the images and / or videos, and generates instructions to be published). For another example, the terminal can display a multimedia content interface, which includes published multimedia content and a publishing control, select images and / or videos through the publishing control, generate multimedia content to be published based on the images and / or videos, and generate instructions to be published.
[0226] S1402: Display the multimedia content to be published and the target titles of the target quantity via the publishing interface.
[0227] The publishing interface can be, for example, Figure 7 shown.
[0228] S1403: In response to a selection operation on a target title, display the target title corresponding to the selection operation at a title position of the multimedia content to be published.
[0229] After the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published, the publishing interface can be as follows: Fig.11 shown.
[0230] In some embodiments, after the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published in response to the selection operation on the target title, the method further includes:
[0231] Hide other target titles, where other target titles are target titles other than the target title corresponding to the selection operation among the target titles of the target quantity;
[0232] In response to a trigger operation on a target title at a title position, displaying the target title in an editing state and displaying other target titles;
[0233] In response to an editing operation on a target title in an editing state, editing the target title in the editing state to obtain an edited target title;
[0234] Displays the edited target title in the title position.
[0235] For example, in response to a trigger operation on a target title at a title position, the target title in the edited state and other target titles displayed may be displayed as follows: Fig.12 shown.
[0236] In this embodiment, in response to a selection operation on a target title, the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published, and other target titles are hidden; in response to a trigger operation on the target title at the title position, the target title in an editing state is displayed, and other target titles are displayed; in response to an editing operation on the target title in an editing state, the target title in an editing state is edited to obtain an edited target title; the edited target title is displayed at the title position, so that not only the publishing interface can be simplified, but also the target title can be modified or replaced.
[0237] S1404: In response to the publishing operation on the media content to be published, publish the target title corresponding to the selection operation as the title of the multimedia content to be published together with the multimedia content to be published.
[0238] The target title corresponding to the selection operation is published as the title of the multimedia content to be published together with the multimedia content to be published, which means that the target title corresponding to the selection operation is published together with the multimedia content to be published as the title of the multimedia content to be published.
[0239] In an embodiment of the present application, in response to a to-be-published instruction for multimedia content to be published, a target number of target titles are recommended for the multimedia content to be published, the target title being a title determined according to target attribute information of the multimedia content to be published, the target attribute information including at least one of scene information, location information, time information and face information of the multimedia content to be published, the multimedia content to be published and the target number of target titles are displayed via a publishing interface, in response to a selection operation on the target title, the target title corresponding to the selection operation is displayed at the title position of the multimedia content to be published, in response to a publishing operation on the media content to be published, the target title corresponding to the selection operation is published as the title of the multimedia content to be published together with the multimedia content to be published, so as to realize automatic recommendation of the title of the multimedia content to be published for the user according to the content characteristics and the target attribute information, so that the user can select the title of the multimedia content to be published from the recommended target titles, without the need for the user to create the title of the multimedia content to be published by himself, thereby improving the publishing efficiency of the multimedia content to be published.
[0240] The explanation of terms, other implementation methods and corresponding beneficial effects of this embodiment can refer to the above-mentioned title recommendation method embodiment, and this embodiment will not be repeated here.
[0241] In order to facilitate better implementation of the title recommendation method provided in the embodiment of the present application, the embodiment of the present application also provides a device based on the above title recommendation method. The meanings of the terms are the same as those in the above title recommendation method, and the specific implementation details can refer to the description in the method embodiment.
[0242] For example, Fig.15 As shown, the title recommendation device may include:
[0243] The acquisition module 1501 is used to acquire multimedia content to be published.
[0244] The extraction module 1502 is used to extract features of the multimedia content to be published, and obtain content features corresponding to the multimedia content to be published.
[0245] The determination module 1503 is used to determine the candidate titles of the multimedia content to be published according to the content characteristics.
[0246] The generation module 1504 is used to generate a target number of target titles and recommend a target number of target titles for the multimedia content to be released based on the target attribute information and candidate titles of the multimedia content to be released, wherein the target attribute information includes at least one of scene information, location information, time information and face information of the multimedia content to be released.
[0247] Optionally, the target attribute information includes location information, and the generating module 1504 is specifically used to execute:
[0248] Obtaining location information corresponding to images and / or videos in the media content to be published;
[0249] Adding location information to a target position of a first at least part of a candidate title in the candidate titles to obtain a first candidate title;
[0250] A target number of target titles is determined based on the first candidate title and a second at least part of the candidate titles among the candidate titles.
[0251] Optionally, the target attribute information also includes scene information, and the generating module 1504 is specifically used to execute:
[0252] Matching the scene information with the preset scene information in the scene set;
[0253] If there is preset scene information matching the scene information, then location information corresponding to the image and / or video in the media content to be published is obtained.
[0254] Optionally, the generating module 1504 is specifically configured to execute:
[0255] Determining whether the text information in the media content to be published includes location information;
[0256] If the text information in the media content to be published does not include location information, then the location information corresponding to the image and / or video in the media content to be published is obtained.
[0257] Optionally, the generating module 1504 is further configured to execute:
[0258] If the text information in the media content to be published includes location information, determining whether the scene information corresponding to the image and / or video in the media content to be published is the target scene information;
[0259] If the scene information corresponding to the image and / or video in the media content to be published is the target scene information, the location information corresponding to the image and / or video in the media content to be published is obtained.
[0260] Optionally, the generating module 1504 is specifically configured to execute:
[0261] Determine whether the length of the location information meets a preset length condition;
[0262] If the length of the location information meets the preset length condition, the location information is added to the target position of the first at least part of the candidate title in the candidate title to obtain the first candidate title.
[0263] Optionally, the generating module 1504 is specifically configured to execute:
[0264] Adding location information to a target position of a first at least part of the candidate titles in the candidate titles to obtain an initial title;
[0265] The first candidate title is determined based on the candidate titles corresponding to the first initial title and the second initial title. The first initial title is the initial title corresponding to a length that meets the first preset length condition, and the second initial title is the initial title corresponding to a length that does not meet the first preset length condition.
[0266] Optionally, the generating module 1504 is specifically configured to execute:
[0267] Acquire location information corresponding to a portion of frame images of a video in the media content to be published, where the portion of frame images includes a cover image and / or a preset number of frame images of the video in the media content to be published;
[0268] The location information corresponding to the partial frame images is determined as the location information corresponding to the video in the media content to be published.
[0269] Optionally, the target attribute information includes time information, and the generating module 1504 is specifically used to execute:
[0270] Matching the time information with the candidate time information in the candidate title, the time information including at least one of shooting time information, editing time information and historical publishing time information of the image and / or video in the multimedia content to be published;
[0271] A target number of target titles is determined based on the candidate titles corresponding to the candidate time information that matches the time information.
[0272] Optionally, the generating module 1504 is specifically configured to execute:
[0273] generating target titles according to target attribute information of the multimedia content to be published and candidate titles, wherein the number of target titles is greater than the target number;
[0274] Determine the target title corresponding to the maximum feature similarity among the feature similarities corresponding to the target titles greater than the target number as the first target title;
[0275] According to a random rule, a first target number of second target titles are selected from the remaining target titles, and the remaining target titles are target titles other than the first target title among the target titles greater than the target number;
[0276] A target number of target titles is determined based on the first target title and the second target title.
[0277] Optionally, the generating module 1504 is specifically configured to execute:
[0278] Determine the preset type corresponding to each remaining target title;
[0279] According to a random rule, a first target number of second target titles are screened out from the remaining target titles corresponding to the preset type matching the target preset type.
[0280] Optionally, the determination module 1503 is specifically configured to execute:
[0281] Determining feature similarity between content features and features corresponding to titles in a preset title library;
[0282] The titles corresponding to the features that meet the feature similarity of the preset similarity condition are determined as candidate titles.
[0283] Optionally, the acquisition module 1501 is further configured to execute:
[0284] Get the title information published historically in the client;
[0285] Performing expansion processing according to the title information to obtain an expanded title, wherein the expansion processing includes keyword replacement processing and / or sentence replacement processing;
[0286] Generate a preset title library based on the expanded titles.
[0287] Optionally, the acquisition module 1501 is further configured to execute:
[0288] Get the title information published historically in the client;
[0289] Performing filtering and / or deduplication processing on the title information to obtain a processed title, wherein the filtering processing includes rule filtering processing and / or model filtering processing, and the deduplication processing includes character deduplication processing and / or semantic deduplication processing;
[0290] Generate a preset title library based on the processed titles.
[0291] Optionally, the generating module 1504 is specifically configured to execute:
[0292] Displaying the target titles of the multimedia content to be published and the target quantity via the publishing interface;
[0293] In response to a selection operation on a target title, displaying the target title corresponding to the selection operation at a title position of the multimedia content to be published;
[0294] In response to a publishing operation on the multimedia content to be published, a target title corresponding to the selection operation is published as a title of the multimedia content to be published together with the multimedia content to be published.
[0295] Optionally, the multimedia content to be published includes multiple objects, the objects include images and / or videos, the target number of target titles is a title determined according to the first object in the multiple objects, and the generation module 1504 is further used to execute:
[0296] In response to a replacement operation on a first object, obtaining the first object after replacement;
[0297] Update the target title of the target quantity according to the first object after replacement;
[0298] Updated goal title to show goal quantity via the publish interface.
[0299] Optionally, the generating module 1504 is further configured to execute:
[0300] Hide other target titles, where other target titles are target titles other than the target title corresponding to the selection operation among the target titles of the target quantity;
[0301] In response to a trigger operation on a target title at a title position, displaying the target title in an editing state and displaying other target titles;
[0302] In response to an editing operation on a target title in an editing state, editing the target title in the editing state to obtain an edited target title;
[0303] Displays the edited target title in the title position.
[0304] In specific implementation, the above modules can be implemented as independent entities, or can be arbitrarily combined and implemented as the same or several entities. The specific implementation methods and corresponding beneficial effects of the above modules can be found in the previous method embodiments, which will not be repeated here.
[0305] In order to facilitate better implementation of the title recommendation method provided in the embodiment of the present application, the embodiment of the present application also provides a device based on the above title recommendation method. The meanings of the terms are the same as those in the above title recommendation method, and the specific implementation details can refer to the description in the method embodiment.
[0306] For example, Fig.16 As shown, the title recommendation device may include:
[0307] The first response module 1601 is used to respond to a to-be-published instruction for multimedia content to be published, and recommend a target number of target titles for the multimedia content to be published, where the target titles are titles determined based on target attribute information of the multimedia content to be published, and the target attribute information includes at least one of scene information, location information, time information, and facial information of the multimedia content to be published.
[0308] The interface display module 1602 is used to display the target titles of the multimedia content to be published and the target quantity via the publishing interface.
[0309] The second response module 1603 is used to respond to the selection operation on the target title and display the target title corresponding to the selection operation at the title position of the multimedia content to be published.
[0310] The third response module 1604 is used to respond to the publishing operation of the to-be-published media content and publish the target title corresponding to the selection operation as the title of the to-be-published multimedia content together with the to-be-published multimedia content.
[0311] Optionally, the second response module 1603 is further configured to execute:
[0312] Hide other target titles, where other target titles are target titles other than the target title corresponding to the selection operation among the target titles of the target quantity;
[0313] In response to a trigger operation on a target title at a title position, displaying the target title in an editing state and displaying other target titles;
[0314] In response to an editing operation on a target title in an editing state, editing the target title in the editing state to obtain an edited target title;
[0315] Displays the edited target title in the title position.
[0316] In specific implementation, the above modules can be implemented as independent entities, or can be arbitrarily combined and implemented as the same or several entities. The specific implementation methods and corresponding beneficial effects of the above modules can be found in the previous method embodiments, which will not be repeated here.
[0317] The present application also provides an electronic device, which may be a server or a terminal, etc. Fig.17 As shown, it shows a schematic diagram of the structure of the electronic device involved in the embodiment of the present application, specifically:
[0318] The electronic device may include components such as a processor 1701 with one or more processing cores, a memory 1702 with one or more computer-readable storage media, a power supply 1703, and an input unit 1704. Those skilled in the art will appreciate that Fig.17 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.
[0319] The processor 1701 is the control center of the electronic device, which uses various interfaces and lines to connect various parts of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing computer programs and / or modules stored in the memory 1702, and calling data stored in the memory 1702. Optionally, the processor 1701 may include one or more processing cores; preferably, the processor 1701 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 1701.
[0320] The memory 1702 can be used to store computer programs and modules. The processor 1701 executes various functional applications and data processing by running the computer programs and modules stored in the memory 1702. The memory 1702 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, a computer program required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 1702 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 1702 may also include a memory controller to provide the processor 1701 with access to the memory 1702.
[0321] The electronic device also includes a power supply 1703 for supplying power to each component. Preferably, the power supply 1703 can be logically connected to the processor 1701 through a power management system, so as to manage charging, discharging, and power consumption through the power management system. The power supply 1703 can also include one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components.
[0322] The electronic device may further include an input unit 1704, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.
[0323] Although not shown, the electronic device may further include a display unit, etc., which will not be described in detail herein. Specifically in this embodiment, the processor 1701 in the electronic device will load the executable files corresponding to the processes of one or more computer programs into the memory 1702 according to the following instructions, and the processor 1701 will run the computer programs stored in the memory 1702, thereby realizing various functions, such as:
[0324] Acquire multimedia content to be published;
[0325] Extract features of the multimedia content to be published, and obtain content features corresponding to the multimedia content to be published;
[0326] Determine candidate titles of the multimedia content to be published according to content features;
[0327] According to target attribute information and candidate titles of the multimedia content to be published, a target number of target titles are generated, and a target number of target titles are recommended for the multimedia content to be published, wherein the target attribute information includes at least one of scene information, location information, time information and face information of the multimedia content to be published.
[0328] The specific implementation methods and corresponding beneficial effects of the above operations can be found in the detailed description of the title recommendation method above, which will not be elaborated here.
[0329] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by a computer program, or by controlling related hardware through a computer program. The computer program may be stored in a computer-readable storage medium and loaded and executed by a processor.
[0330] To this end, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored, and the computer program can be loaded by a processor to execute the steps in any of the title recommendation methods provided in the embodiments of the present application. For example, the computer program can execute the following steps:
[0331] Acquire multimedia content to be published;
[0332] Extract features of the multimedia content to be published, and obtain content features corresponding to the multimedia content to be published;
[0333] Determine candidate titles of the multimedia content to be published according to content features;
[0334] According to target attribute information and candidate titles of the multimedia content to be published, a target number of target titles are generated, and a target number of target titles are recommended for the multimedia content to be published, wherein the target attribute information includes at least one of scene information, location information, time information and face information of the multimedia content to be published.
[0335] The specific implementation methods and corresponding beneficial effects of the above operations can be found in the previous embodiments, which will not be described in detail here.
[0336] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.
[0337] Since the computer program stored in the computer-readable storage medium can execute the steps in any one of the title recommendation methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any one of the title recommendation methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.
[0338] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the above-mentioned title recommendation method.
[0339] The above is a detailed introduction to a title recommendation method, device, electronic device and computer storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
Claims
1. A title recommendation method, characterized in that: include: Acquire multimedia content to be published; Extracting features of the multimedia content to be published to obtain content features corresponding to the multimedia content to be published; Determining candidate titles of the multimedia content to be published according to the content features; A target number of target titles is generated based on the target attribute information of the multimedia content to be published and the candidate titles, and the target number of target titles is recommended for the multimedia content to be published, wherein the target attribute information includes at least one of scene information, location information, time information and face information of the multimedia content to be published.
2. The title recommendation method according to claim 1, characterized in that: The target attribute information includes location information, and generating a target number of target titles according to the target attribute information of the to-be-published multimedia content and the candidate titles includes: Acquire location information corresponding to the image and / or video in the media content to be published; Adding the location information to the target position of the first at least part of the candidate title in the candidate title to obtain a first candidate title; A target number of target titles is determined based on the first candidate title and a second at least part of the candidate titles among the candidate titles.
3. The title recommendation method according to claim 2, characterized in that: The target attribute information also includes scene information, and the step of obtaining location information corresponding to the image and / or video in the to-be-published media content includes: Matching the scene information with preset scene information in a scene set; If there is preset scene information matching the scene information, then location information corresponding to the image and / or video in the to-be-published media content is obtained.
4. The title recommendation method according to claim 2, characterized in that: The obtaining of location information corresponding to the image and / or video in the to-be-published media content includes: Determining whether the text information in the to-be-published media content includes location information; If the text information in the to-be-published media content does not include location information, then the location information corresponding to the image and / or video in the to-be-published media content is obtained.
5. The title recommendation method according to claim 4, characterized in that: After determining whether the text information in the to-be-published media content includes location information, the method further includes: If the text information in the to-be-published media content includes location information, determining whether the scene information corresponding to the image and / or video in the to-be-published media content is target scene information; If the scene information corresponding to the image and / or video in the to-be-published media content is target scene information, then the location information corresponding to the image and / or video in the to-be-published media content is acquired.
6. The title recommendation method according to claim 2, characterized in that: The step of adding the location information to the target position of the first at least part of the candidate titles in the candidate titles to obtain the first candidate title comprises: Determining whether the length of the location information meets a preset length condition; If the length of the location information meets the preset length condition, the location information is added to the target position of the first at least part of the candidate title in the candidate title to obtain the first candidate title.
7. The title recommendation method according to claim 2, characterized in that: The step of adding the location information to the target position of the first at least part of the candidate titles in the candidate titles to obtain the first candidate title comprises: Adding the location information at the target position of the first at least part of the candidate titles in the candidate titles to obtain an initial title; The first candidate title is determined based on the candidate titles corresponding to the first initial title and the second initial title, wherein the first initial title is the initial title corresponding to a length that meets a first preset length condition, and the second initial title is the initial title corresponding to a length that does not meet the first preset length condition.
8. The title recommendation method according to claim 2, characterized in that: Acquiring location information corresponding to the video in the media content to be published, including: Acquire location information corresponding to a portion of frame images of the video in the to-be-published media content, where the portion of frame images includes a cover image and / or a preset number of frame images of the video in the to-be-published media content; The location information corresponding to the partial frame images is determined as the location information corresponding to the video in the media content to be published.
9. The title recommendation method according to claim 1, characterized in that: The target attribute information includes time information, and generating a target number of target titles according to the target attribute information of the to-be-published multimedia content and the candidate titles includes: Matching the time information with candidate time information in the candidate title, the time information comprising at least one of shooting time information, editing time information, and historical publishing time information of the image and / or video in the to-be-published multimedia content; A target number of target titles is determined based on candidate titles corresponding to the candidate time information matching the time information.
10. The title recommendation method according to claim 1, characterized in that: The step of generating a target number of target titles according to the target attribute information of the to-be-published multimedia content and the candidate titles includes: generating target titles according to the target attribute information of the multimedia content to be published and the candidate titles, wherein the number of the target titles is greater than the target number; Determine the target title corresponding to the maximum feature similarity among the feature similarities corresponding to the target titles greater than the target number as the first target title; According to a random rule, a first target number of second target titles are selected from the remaining target titles, wherein the remaining target titles are target titles other than the first target title among the target titles greater than the target number; The target number of target titles is determined based on the first target title and the second target title.
11. The title recommendation method according to claim 10, characterized in that: The step of selecting a first target number of second target titles from the remaining target titles according to a random rule comprises: Determine the preset type corresponding to each remaining target title; According to a random rule, a first target number of second target titles are screened out from the remaining target titles corresponding to the preset type matching the target preset type.
12. The title recommendation method according to claim 1, characterized in that: The step of determining the candidate titles of the multimedia content to be published according to the content features includes: Determining feature similarity between the content feature and features corresponding to titles in a preset title library; The titles corresponding to the features that meet the feature similarity of the preset similarity condition are determined as candidate titles.
13. The title recommendation method according to claim 12, characterized in that: Before determining the feature similarity between the content feature and the feature corresponding to the title in the preset title library, the method further includes: Get the title information published historically in the client; Performing expansion processing according to the title information to obtain an expanded title, wherein the expansion processing includes keyword replacement processing and / or sentence replacement processing; Generate a preset title library based on the expanded titles.
14. The title recommendation method according to claim 12, characterized in that: Before determining the feature similarity between the content feature and the feature corresponding to the title in the preset title library, the method further includes: Get the title information published historically in the client; Performing filtering and / or deduplication processing on the title information to obtain a processed title, wherein the filtering processing includes rule filtering processing and / or model filtering processing, and the deduplication processing includes character deduplication processing and / or semantic deduplication processing; A preset title library is generated according to the processed titles.
15. The title recommendation method according to any one of claims 1 to 14, characterized in that: The step of recommending the target number of target titles for the multimedia content to be published includes: Displaying the multimedia content to be published and the target number of target titles via a publishing interface; In response to a selection operation on the target title, displaying the target title corresponding to the selection operation at a title position of the to-be-published multimedia content; In response to a publishing operation on the to-be-published media content, a target title corresponding to the selection operation is used as a title of the to-be-published multimedia content and is published together with the to-be-published multimedia content.
16. The title recommendation method according to claim 15, characterized in that: The multimedia content to be published includes a plurality of objects, the objects include images and / or videos, the target number of target titles is a title determined according to the first object of the plurality of objects, and after the multimedia content to be published and the target number of target titles are displayed via the publishing interface, the method further includes: In response to a replacement operation on the first object, obtaining a first object after replacement; According to the first object after the replacement, updating the target title of the target number; An updated target title of the target quantity is displayed via the publishing interface.
17. The title recommendation method according to claim 15, characterized in that: After the target title corresponding to the selection operation is displayed at the title position of the to-be-published multimedia content in response to the selection operation of the target title, the method further includes: Hide other target titles, where the other target titles are target titles among the target number of target titles except the target title corresponding to the selection operation; In response to a triggering operation on a target title at the title position, displaying the target title in an editing state and displaying the other target titles; In response to an editing operation on the target title in the editing state, editing the target title in the editing state to obtain an edited target title; The edited target title is displayed at the title position.
18. A method for recommending a title, characterized in that: include: In response to a to-be-published instruction for multimedia content to be published, recommending a target number of target titles for the multimedia content to be published, the target titles being titles determined according to target attribute information of the multimedia content to be published, the target attribute information including at least one of scene information, location information, time information, and face information of the multimedia content to be published; Displaying the multimedia content to be published and the target number of target titles via a publishing interface; In response to a selection operation on the target title, displaying the target title corresponding to the selection operation at a title position of the to-be-published multimedia content; In response to a publishing operation on the to-be-published media content, a target title corresponding to the selection operation is used as a title of the to-be-published multimedia content and is published together with the to-be-published multimedia content.
19. The title recommendation method according to claim 18, characterized in that: After the target title corresponding to the selection operation is displayed at the title position of the to-be-published multimedia content in response to the selection operation of the target title, the method further includes: Hide other target titles, where the other target titles are target titles among the target number of target titles except the target title corresponding to the selection operation; In response to a triggering operation on a target title at the title position, displaying the target title in an editing state and displaying the other target titles; In response to an editing operation on the target title in the editing state, editing the target title in the editing state to obtain an edited target title; The edited target title is displayed at the title position.
20. A title recommendation device, characterized in that: include: An acquisition module, used to acquire multimedia content to be published; An extraction module, used for extracting features of the multimedia content to be released, and obtaining content features corresponding to the multimedia content to be released; A determination module, used to determine candidate titles of the multimedia content to be published according to the content features; A generation module is used to generate a target number of target titles based on the target attribute information of the multimedia content to be released and the candidate titles, and recommend the target number of target titles for the multimedia content to be released, wherein the target attribute information includes at least one of scene information, location information, time information and face information of the multimedia content to be released.
21. An electronic device, characterized in that: The invention comprises a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the title recommendation method according to any one of claims 1 to 19.
22. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the title recommendation method according to any one of claims 1 to 19.
Citation Information
Patent Citations
Medium title generation method and device, electronic equipment and readable medium
CN110134931A
Video title generation method and device, electronic equipment and storage medium
CN118075573A
