Multimedia information recommendation method, device, storage medium, and apparatus

Abstract

The application discloses a multimedia information recommendation method and device, a storage medium and an apparatus, and relates to the technical field of multimedia information recommendation. The method comprises the following steps: determining to-be-browsed multimedia information according to a multimedia information browsing instruction, searching for to-be-recommended multimedia information corresponding to the to-be-browsed multimedia information from a preset mapping relationship, generating the preset mapping relationship based on feature vectors between different multimedia information, and sending the to-be-browsed multimedia information and the to-be-recommended multimedia information to a user terminal. Compared with the prior art of generating recommended multimedia information by semantic recognition according to the name of a POI point, the application searches for to-be-recommended multimedia information corresponding to to-be-browsed multimedia information from a preset mapping relationship, the preset mapping relationship is generated based on feature vectors between different multimedia information, the feature vectors of the multimedia information are directly associated with the multimedia information, the relevance of multimedia information recommendation is improved, the defect of poor recommendation effect is overcome, and the user experience is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet, and particularly relates to a multimedia information recommendation method, device, storage medium and apparatus. BACKGROUND

[0002] At present, when a user clicks a POI (Point of Interest) to view a map image, the server generates a recommended image according to the name of the POI point through semantic recognition, and sends the map image and the recommended image to the user terminal for display, so as to improve the image browsing experience of the user.

[0003] However, the above-mentioned technology lacks direct association between images, resulting in poor recommendation effect and poor user experience.

[0004] The above content is only used to assist in understanding the technical solutions of the present application and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0005] The main purpose of the present application is to provide a multimedia information recommendation method, device, storage medium and apparatus, which aims to solve the technical problem that the direct association between multimedia information is lacking in the prior art, resulting in poor recommendation effect.

[0006] To achieve the above-mentioned purpose, the present application provides a multimedia information recommendation method, which comprises the following steps:

[0007] When receiving a multimedia information browsing instruction sent by a user terminal, determining a to-be-browsed multimedia information according to the multimedia information browsing instruction;

[0008] Finding a to-be-recommended multimedia information corresponding to the to-be-browsed multimedia information from a preset mapping relationship, the preset mapping relationship being generated based on feature vectors between different multimedia information;

[0009] Sending the to-be-browsed multimedia information and the to-be-recommended multimedia information to the user terminal.

[0010] Optionally, before the step of determining the to-be-browsed multimedia information according to the multimedia information browsing instruction when receiving the multimedia information browsing instruction sent by the user terminal, the multimedia information recommendation method further comprises:

[0011] When receiving an information association instruction, determining a to-be-associated multimedia information according to the information association instruction;

[0012] Analyzing the to-be-associated multimedia information to obtain a feature vector of the to-be-associated multimedia information;

[0013] establish a preset mapping relationship between the to-be-associated multimedia information according to the feature vectors.

[0014] Optionally, the step of establishing the preset mapping relationship between the to-be-associated multimedia information according to the feature vectors specifically comprises:

[0015] grouping the to-be-associated multimedia information to obtain a plurality of multimedia information groups;

[0016] traversing the to-be-associated multimedia information, and taking the to-be-associated multimedia information traversed as to-be-processed multimedia information;

[0017] finding target multimedia information associated with the to-be-processed multimedia information in each multimedia information group according to a feature vector of the to-be-processed multimedia information;

[0018] establishing a corresponding relationship between the to-be-processed multimedia information and the target multimedia information;

[0019] after the traversal of the to-be-associated multimedia information ends, establishing a preset mapping relationship between the to-be-associated multimedia information according to the corresponding relationship.

[0020] Optionally, the step of finding target multimedia information associated with the to-be-processed multimedia information in each multimedia information group according to a feature vector of the to-be-processed multimedia information specifically comprises:

[0021] obtaining attribute information of the to-be-processed multimedia information, and screening the multimedia information groups according to the attribute information to obtain a target multimedia information group;

[0022] finding target multimedia information associated with the to-be-processed multimedia information in the target multimedia information group according to a feature vector of the to-be-processed multimedia information.

[0023] Optionally, the step of obtaining attribute information of the to-be-processed multimedia information, and screening the multimedia information groups according to the attribute information to obtain a target multimedia information group specifically comprises:

[0024] obtaining attribute information of the to-be-processed multimedia information, and matching the attribute information of the to-be-processed multimedia information with attribute information of the multimedia information groups to obtain a matching result;

[0025] screening the multimedia information groups according to the matching result to obtain a target multimedia information group.

[0026] Optionally, the step of searching the target multimedia information associated with the to-be-processed multimedia information from the target multimedia information group according to the feature vector of the to-be-processed multimedia information specifically comprises:

[0027] determining the relevance between the to-be-processed multimedia information and the multimedia information according to the feature vector of the to-be-processed multimedia information and the feature vector of each multimedia information in the target multimedia information group;

[0028] sorting the multimedia information according to the relevance, and screening the multimedia information according to the sorting result to obtain the target multimedia information.

[0029] Optionally, the step of establishing the corresponding relationship between the to-be-processed multimedia information and the target multimedia information specifically comprises:

[0030] merging the target multimedia information in each multimedia information group to obtain to-be-processed multimedia information;

[0031] sorting the to-be-processed multimedia information, and screening the to-be-processed multimedia information according to the sorting result to obtain the associated multimedia information;

[0032] establishing the corresponding relationship between the to-be-processed multimedia information and the associated multimedia information.

[0033] Optionally, the step of determining the to-be-associated multimedia information according to the information association instruction when the information association instruction is received specifically comprises:

[0034] determining to-be-processed map information according to the information association instruction when the information association instruction is received, and determining a map information point according to the to-be-processed map information;

[0035] searching an information point image corresponding to the map information point, and taking the information point image as the to-be-associated multimedia information.

[0036] Optionally, before the step of determining the to-be-browsed multimedia information according to the multimedia information browsing instruction when the multimedia information browsing instruction sent by the user terminal is received, the multimedia information recommendation method further comprises:

[0037] determining to-be-added multimedia information according to the multimedia information adding instruction when the multimedia information adding instruction is received;

[0038] analyzing the to-be-added multimedia information to obtain a feature vector of the to-be-added multimedia information;

[0039] establishing a preset mapping relationship between the to-be-added multimedia information and preset multimedia information according to the feature vector.

[0040] Optionally, the step of determining multimedia information to be browsed according to the multimedia information browsing instruction when the multimedia information browsing instruction sent by the user terminal is received specifically comprises the following steps.

[0041] When the multimedia information browsing instruction sent by the user terminal is received, the multimedia information browsing instruction is parsed to obtain an instruction identifier.

[0042] The multimedia information to be browsed is determined according to the instruction identifier.

[0043] Optionally, the step of finding the multimedia information to be recommended corresponding to the multimedia information to be browsed from a preset mapping relationship, the preset mapping relationship being generated based on feature vectors between different multimedia information specifically comprises the following steps.

[0044] The candidate multimedia information corresponding to the multimedia information to be browsed is found from a preset mapping relationship, the preset mapping relationship being generated based on feature vectors between different multimedia information.

[0045] The user portrait of the user terminal is obtained, the candidate multimedia information is screened according to the user portrait, and the multimedia information to be recommended is obtained.

[0046] Optionally, the step of sending the multimedia information to be browsed and the multimedia information to be recommended to the user terminal specifically comprises the following steps.

[0047] The device information of the user terminal is obtained, and an information display template corresponding to the device information is found.

[0048] The information display template, the multimedia information to be browsed and the multimedia information to be recommended are sent to the user terminal.

[0049] In addition, to achieve the above-mentioned purpose, the application further provides a multimedia information recommendation device, which comprises a memory, a processor and a multimedia information recommendation program stored in the memory and executable on the processor, and the multimedia information recommendation program is configured to implement the steps of the multimedia information recommendation method.

[0050] In addition, to achieve the above-mentioned purpose, the application further provides a storage medium, which stores a multimedia information recommendation program, and the multimedia information recommendation program implements the steps of the multimedia information recommendation method when executed by a processor.

[0051] In addition, to achieve the above-mentioned purpose, the application further provides a multimedia information recommendation device, which comprises a determination module, a finding module and a sending module.

[0052] The determining module is configured to determine multimedia information to be browsed according to a multimedia information browsing instruction when the multimedia information browsing instruction is received from the user terminal;

[0053] The searching module is configured to search for multimedia information to be recommended corresponding to the multimedia information to be browsed from a preset mapping relationship, the preset mapping relationship being generated based on feature vectors between different multimedia information;

[0054] The sending module is configured to send the multimedia information to be browsed and the multimedia information to be recommended to the user terminal.

[0055] Optionally, the multimedia information recommendation apparatus further includes an associating module.

[0056] The associating module is configured to determine multimedia information to be associated according to an information association instruction when the information association instruction is received;

[0057] The associating module is further configured to analyze the multimedia information to be associated to obtain feature vectors of the multimedia information to be associated;

[0058] The associating module is further configured to establish a preset mapping relationship between the multimedia information to be associated according to the feature vectors.

[0059] Optionally, the associating module is further configured to group the multimedia information to be associated to obtain a plurality of multimedia information groups.

[0060] The associating module is further configured to traverse the multimedia information to be associated and take the traversed multimedia information to be associated as multimedia information to be processed.

[0061] The associating module is further configured to search for target multimedia information associated with the multimedia information to be processed from each multimedia information group according to a feature vector of the multimedia information to be processed.

[0062] The associating module is further configured to establish a corresponding relationship between the multimedia information to be processed and the target multimedia information.

[0063] The associating module is further configured to establish a preset mapping relationship between the multimedia information to be associated according to the corresponding relationship after the traversal of the multimedia information to be associated ends.

[0064] Optionally, the associating module is further configured to obtain attribute information of the multimedia information to be processed and filter the multimedia information groups according to the attribute information to obtain a target multimedia information group.

[0065] The association module is also used to find target multimedia information associated with the multimedia information to be processed in the target multimedia information group based on the feature vector of the multimedia information to be processed.

[0066] Optionally, the association module is further configured to obtain attribute information of the multimedia information to be processed, and match the attribute information of the multimedia information to be processed with the attribute information of the multimedia information group to obtain a matching result;

[0067] The association module is also used to filter the multimedia information group based on the matching result to obtain the target multimedia information group.

[0068] Optionally, the association module is further configured to determine the correlation between the multimedia information to be processed and the multimedia information based on the feature vector of the multimedia information to be processed and the feature vector of each multimedia information in the target multimedia information group;

[0069] The association module is further configured to sort the multimedia information according to the relevance, and filter the multimedia information according to the sorting result to obtain target multimedia information.

[0070] This invention discloses a method for determining the multimedia information to be browsed based on a multimedia information browsing instruction received from a user terminal, searching for recommended multimedia information corresponding to the multimedia information to be browsed from a preset mapping relationship (generated based on feature vectors between different multimedia information), and sending the multimedia information to be browsed and the recommended multimedia information to the user terminal. Compared to existing methods that generate recommended multimedia information based on the name of a Point of Interest (POI) through semantic recognition, this invention improves the relevance of multimedia information recommendations by searching for recommended multimedia information corresponding to the multimedia information to be browsed from a preset mapping relationship (generated based on feature vectors between different multimedia information), and the feature vectors of multimedia information are directly associated with the multimedia information. This overcomes the shortcomings of poor recommendation performance and improves the user experience. Attached Figure Description

[0071] Figure 1 This is a schematic diagram of the structure of a multimedia information recommendation device in the hardware operating environment involved in the embodiments of the present invention;

[0072] Figure 2 This is a flowchart illustrating the first embodiment of the multimedia information recommendation method of the present invention;

[0073] Figure 3 This is a flowchart illustrating the second embodiment of the multimedia information recommendation method of the present invention;

[0074] Figure 4A flowchart of a third embodiment of the multimedia information recommendation method of the present application;

[0075] Figure 5 A structure block diagram of a first embodiment of the multimedia information recommendation device of the present application.

[0076] The implementation, functional features and advantages of the present application will be further explained with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0077] It should be understood that the specific embodiments described herein merely exemplify the present application and do not limit the present application.

[0078] Referring to Figure 1 , Figure 1 A structure diagram of a multimedia information recommendation device related to the hardware running environment of the embodiment of the present application.

[0079] As Figure 1 shown, the multimedia information recommendation device can include a processor 1001, such as a Central Processing Unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize the connection and communication among these components. The user interface 1003 can include a display screen (Display), and the optional user interface 1003 can further include a standard wired interface, a wireless interface, and the wired interface of the user interface 1003 can be a USB interface in the present application. The network interface 1004 can optionally include a standard wired interface, a wireless interface (such as a WIreless-FIdelity (WI-FI) interface). The memory 1005 can be a high-speed Random Access Memory (RAM) memory, or a stable Non-volatile Memory (NVM), such as a disk memory. The memory 1005 can also be a storage device independent of the aforementioned processor 1001.

[0080] Those skilled in the art can understand Figure 1 that the structure shown in the foregoing embodiments does not constitute a limitation on the multimedia information recommendation device, and can include more or fewer components than those shown in the drawings, or combine certain components, or different component arrangements.

[0081] As Figure 1 shown, the memory 1005, which is a kind of computer storage medium, can include an operating system, a network communication module, a user interface module, and a multimedia information recommendation program.

[0082] In Figure 1The network interface 1004 is mainly used for connecting a background server and communicating data with the background server; the user interface 1003 is mainly used for connecting a user terminal; the multimedia information recommendation device calls the multimedia information recommendation program stored in the memory 1005 through the processor 1001, and executes the multimedia information recommendation method provided by the embodiment of the application.

[0083] Based on the above hardware structure, the embodiment of the multimedia information recommendation method is provided.

[0084] Reference Figure 2 , Figure 2 The first embodiment of the multimedia information recommendation method is provided, and a flowchart of the first embodiment of the multimedia information recommendation method is shown.

[0085] In the first embodiment, the multimedia information recommendation method comprises the following steps.

[0086] Step S10: When receiving the multimedia information browsing instruction sent by the user terminal, determining the multimedia information to be browsed according to the multimedia information browsing instruction.

[0087] It should be understood that the execution subject of the embodiment is the multimedia information recommendation device, wherein the multimedia information recommendation device can be an electronic device such as a personal computer or a server, and can also be other devices that can realize the same or similar functions, and the embodiment does not limit this. In the embodiment, a server is taken as an example for illustration.

[0088] It should be noted that the multimedia information browsing instruction can be a control instruction input by the user through the interactive program of the user terminal. The multimedia information can be image, audio and text information, and the image is taken as an example for illustration in the embodiment and other embodiments.

[0089] In a specific implementation, for example, the user can click the image viewing button of the POI point on the map software to issue an image browsing instruction.

[0090] Further, in order to be able to pre-establish the preset mapping relationship between the multimedia information, so as to quickly generate the multimedia information to be recommended corresponding to the multimedia information to be browsed in subsequent user browsing, before the step of determining the multimedia information to be browsed according to the multimedia information browsing instruction, the method further comprises the following steps.

[0091] When receiving the information association instruction, determining the multimedia information to be associated according to the information association instruction, analyzing the multimedia information to be associated, obtaining the feature vector of the multimedia information to be associated, and establishing the preset mapping relationship between the multimedia information to be associated according to the feature vector.

[0092] In a specific implementation, the administrator of the server inputs an information association instruction through a management interface of the server. When the server receives the information association instruction, the server downloads all images of POIs through a picture bed to obtain to-be-associated images, performs image embedding on the to-be-associated images through a resnet model to obtain feature vectors of the to-be-associated images, and divides the to-be-associated images into m and n groups according to the province information and the category information of the POIs to which the to-be-associated images belong, respectively. For each group, an embedding vector of all images in the group is used to construct a faiss index. A to-be-processed image is selected from the to-be-associated images. For each to-be-processed image, the province and the category of the POI to which the to-be-processed image belongs are obtained, and the embedding vector of the to-be-processed image is used to retrieve top-k results with the highest relevance from the corresponding province faiss index and the category faiss index, respectively. Then, the results of the two indexes are merged, sorted, and target images are obtained, to establish a corresponding relationship between the to-be-processed image and the target image.

[0093] Further, in order to be able to pre-set the mapping relationship between the multimedia information when the multimedia information is added, when the user terminal sends a multimedia information browsing instruction, before determining the to-be-browsed multimedia information according to the multimedia information browsing instruction, the method further comprises:

[0094] When the multimedia information adding instruction is received, the to-be-added multimedia information is determined according to the multimedia information adding instruction, the to-be-added multimedia information is analyzed to obtain a feature vector of the to-be-added multimedia information, and a preset mapping relationship between the to-be-added multimedia information and a preset multimedia information is established according to the feature vector.

[0095] Step S20: searching for to-be-recommended multimedia information corresponding to the to-be-browsed multimedia information from a preset mapping relationship, the preset mapping relationship being generated based on feature vectors between different multimedia information.

[0096] It can be understood that after the to-be-recommended multimedia information corresponding to the to-be-browsed multimedia information is found, the to-be-browsed multimedia information can be sorted according to the relevance between the to-be-recommended multimedia information and the to-be-browsed multimedia information, and the display order is determined according to the sorting result. The relevance between the to-be-recommended multimedia information and the to-be-browsed multimedia information can be generated based on the feature vectors between different multimedia information, which is not limited in the embodiment.

[0097] Step S30: sending the to-be-browsed multimedia information and the to-be-recommended multimedia information to the user terminal.

[0098] It should be understood that the user terminal can directly display the multimedia information to be browsed and the multimedia information to be recommended after receiving the multimedia information to be browsed and the multimedia information to be recommended.

[0099] In the first embodiment, when receiving the multimedia information browsing instruction sent by the user terminal, the multimedia information to be browsed is determined according to the multimedia information browsing instruction, the multimedia information to be recommended corresponding to the multimedia information to be browsed is searched from a preset mapping relationship, the preset mapping relationship is generated based on the feature vectors between different multimedia information, the multimedia information to be browsed and the multimedia information to be recommended are sent to the user terminal; compared with the existing way of generating recommended multimedia information by semantic recognition based on the name of the POI point, since the multimedia information to be recommended corresponding to the multimedia information to be browsed is searched from the preset mapping relationship in the embodiment, the preset mapping relationship is generated based on the feature vectors between different multimedia information, and the feature vectors of the multimedia information are directly associated with the multimedia information, thereby improving the relevance of the multimedia information recommendation, overcoming the defect of poor recommendation effect, and further improving the user experience.

[0100] Reference Figure 3 , Figure 3 The flowchart of the second embodiment of the multimedia information recommendation method of the application is shown in the above Figure 2 based on the first embodiment shown in the above

[0101] In the second embodiment, before the step S10, the method further comprises:

[0102] Step S01: when receiving the information association instruction, determining the multimedia information to be associated according to the information association instruction.

[0103] It should be noted that the information association instruction can be a control instruction input by the administrator of the multimedia information recommendation device through the management interface of the multimedia information recommendation device, and the embodiment does not limit this.

[0104] In a specific implementation, for example, the administrator of the server inputs the information association instruction through the management interface of the server, and the server downloads all the images of the POI through the image bed when receiving the information association instruction, and obtains the image to be associated.

[0105] Further, considering that in actual application, if the server downloads all the images of the POI every time, it will inevitably lead to too many images to process and too large computation. To overcome this defect, the step of determining the multimedia information to be associated according to the information association instruction when receiving the information association instruction comprises:

[0106] Upon receiving the information association instruction, determine the to-be-processed map information according to the information association instruction, determine a map information point according to the to-be-processed map information, find an information point image corresponding to the map information point, and take the information point image as the to-be-associated multimedia information.

[0107] Step S02: analyze the to-be-associated multimedia information to obtain a feature vector of the to-be-associated multimedia information.

[0108] It should be understood that the analysis of the to-be-associated multimedia information to obtain the feature vector of the to-be-associated multimedia information can be performed by a preset analysis model, and the preset analysis model can be pre-set by a manager of the multimedia information recommendation device. In this embodiment and other embodiments, a resnet network model is taken as an example for illustration.

[0109] In a specific implementation, for example, the resnet model is used to perform image embedding on the to-be-associated image to obtain the feature vector of the to-be-associated image.

[0110] It can be understood that after the to-be-associated multimedia information is determined in step S01, the to-be-associated multimedia information needs to be downloaded, and in step S02, the feature vector of the to-be-associated multimedia information needs to be analyzed.

[0111] It should be understood that since the multimedia information recommendation task is running in a cluster, the size of the cluster disk is uncontrollable, and the problem of insufficient disk space may occur when the to-be-associated multimedia information is downloaded and the feature vector of the to-be-associated multimedia information is analyzed. Therefore, in the present application, the multimedia information recommendation device performs multi-threading to download the to-be-associated multimedia information and analyze the feature vector of the to-be-associated multimedia information, that is, in the multimedia information recommendation device, the downloading of the to-be-associated multimedia information and the analysis of the feature vector of the to-be-associated multimedia information are performed simultaneously, so that the image after the analysis of the feature vector of the to-be-associated multimedia information can be deleted in time, and the problem of insufficient disk space is solved.

[0112] In a specific implementation, for example, in order to accelerate and increase robustness, the server performs multi-path concurrency through hadoop yarn, and the image download and embedding in each subtask are performed in parallel.

[0113] Step S03: establish a preset mapping relationship between the to-be-associated multimedia information according to the feature vector.

[0114] It should be understood that the preset mapping relationship between the to-be-associated multimedia information can be established according to the feature vectors, that is, the correlation between the to-be-associated multimedia information is determined according to the feature vectors, and the preset mapping relationship between the to-be-associated multimedia information with a correlation higher than a preset threshold is established. The preset threshold can be set by a manager of the multimedia information recommendation device, and the present embodiment is not limited in this regard.

[0115] Further, in order to improve the accuracy of the preset mapping relationship, the step of establishing the preset mapping relationship between the to-be-associated multimedia information according to the feature vectors comprises:

[0116] grouping the to-be-associated multimedia information to obtain a plurality of multimedia information groups, traversing the to-be-associated multimedia information, taking the traversed to-be-associated multimedia information as a to-be-processed multimedia information, searching for target multimedia information associated with the to-be-processed multimedia information in each multimedia information group according to the feature vector of the to-be-processed multimedia information, establishing a corresponding relationship between the to-be-processed multimedia information and the target multimedia information, and establishing the preset mapping relationship between the to-be-associated multimedia information according to the corresponding relationship after the traversal of the to-be-associated multimedia information ends.

[0117] In the second embodiment, by determining the to-be-associated multimedia information according to the information association instruction when the information association instruction is received, analyzing the to-be-associated multimedia information, obtaining the feature vector of the to-be-associated multimedia information, and establishing the preset mapping relationship between the to-be-associated multimedia information according to the feature vector, the preset mapping relationship between the multimedia information can be established in advance, so that the to-be-recommended multimedia information corresponding to the to-be-browsed multimedia information can be quickly generated when the user browses subsequently.

[0118] In the second embodiment, the step S10 comprises:

[0119] Step S101: When a multimedia information browsing instruction sent by a user terminal is received, the multimedia information browsing instruction is parsed to obtain an instruction identifier.

[0120] It should be noted that the multimedia information browsing instruction can be a control instruction input by the user through the interactive program of the user terminal. The multimedia information can be an image, an audio, and a text information, and the image is taken as an example for description in the present embodiment and other embodiments.

[0121] In a specific implementation, for example, the user can click the image viewing button of the POI point on the map software to issue an image browsing instruction.

[0122] Step S102: Determine the to-be-browsed multimedia information according to the instruction identifier.

[0123] It should be understood that determining the multimedia information to be browsed according to the instruction identifier can be finding the map information point corresponding to the instruction identifier, and finding the multimedia information to be browsed corresponding to the map information point.

[0124] In the second embodiment, by analyzing the multimedia information browsing instruction when receiving the multimedia information browsing instruction sent by the user terminal, obtaining the instruction identifier, and determining the multimedia information to be browsed according to the instruction identifier, the accuracy of the multimedia information to be browsed can be improved.

[0125] In the second embodiment, the step S20 comprises:

[0126] Step S201: finding the candidate multimedia information corresponding to the multimedia information to be browsed from the preset mapping relationship, the preset mapping relationship being generated based on the feature vectors between different multimedia information.

[0127] It should be understood that the multimedia information to be associated can be analyzed by determining the multimedia information to be associated according to the information association instruction when receiving the information association instruction, obtaining the feature vector of the multimedia information to be associated, and establishing the preset mapping relationship between the multimedia information to be associated based on the feature vector.

[0128] Step S202: obtaining the user portrait of the user terminal, and screening the candidate multimedia information according to the user portrait to obtain the multimedia information to be recommended.

[0129] It should be noted that the user portrait is pre-stored in the multimedia information recommendation device, and can be automatically generated by the multimedia information recommendation device according to user data, or can be pre-set by the administrator of the multimedia information recommendation device, and the present embodiment does not limit this.

[0130] In the second embodiment, by finding the candidate multimedia information corresponding to the multimedia information to be browsed from the preset mapping relationship, the preset mapping relationship being generated based on the feature vectors between different multimedia information, obtaining the user portrait of the user terminal, and screening the candidate multimedia information according to the user portrait to obtain the multimedia information to be recommended, the user relevance of the multimedia information to be recommended can be further improved.

[0131] In the second embodiment, the step S30 comprises:

[0132] Step S301: obtaining the device information of the user terminal, and finding the information display template corresponding to the device information.

[0133] It should be noted that the device information can be device model, device identifier and the like, and the present embodiment does not limit this.

[0134] It should be understood that the information display template corresponding to the device information can be searched in a preset template library. The preset template library includes the correspondence between the device information and the information display template. The correspondence between the device information and the information display template can be preset by a manager of the multimedia information recommendation device or by a user. The present example does not limit this.

[0135] Step S302: sending the information display template, the multimedia information to be browsed, and the multimedia information to be recommended to the user terminal.

[0136] It can be understood that the user terminal can aggregate the information display template and the multimedia information to be recommended according to the information display template after receiving the information display template, the multimedia information to be browsed, and the multimedia information to be recommended, obtain the multimedia information to be displayed, and display the multimedia information to be displayed.

[0137] In the second embodiment, the device information of the user terminal is obtained, the information display template corresponding to the device information is searched, the information display template, the multimedia information to be browsed, and the multimedia information to be recommended are sent to the user terminal, so that the information display template can be generated in advance and sent to the user terminal, so that the user terminal can successfully display the multimedia information.

[0138] Reference Figure 4 , Figure 4 The figure is a flowchart of a third embodiment of the multimedia information recommendation method of the present application, based on the second embodiment shown in the above Figure 3 The third embodiment of the multimedia information recommendation method of the present application is proposed based on the second embodiment shown in the above

[0139] In the third embodiment, the step S01 includes:

[0140] Step S011: upon receiving the information association instruction, determining the to-be-processed map information according to the information association instruction, and determining the map information point according to the to-be-processed map information.

[0141] It should be noted that the information association instruction can be a control instruction input by a manager of the multimedia information recommendation device through a management interface of the multimedia information recommendation device. The present example does not limit this.

[0142] It should be understood that determining the to-be-processed map information according to the information association instruction can include parsing the information association instruction to obtain a map identifier, searching for a to-be-processed map corresponding to the map identifier, and obtaining the to-be-processed map information of the to-be-processed map. The map information can include map POI points and other information. The present example does not limit this.

[0143] It should be noted that the map information point can be a map POI point. In the map, a POI point can be a house, a shop, a mailbox or a bus stop, etc. The present embodiment does not limit this.

[0144] Step S012: searching for an information point image corresponding to the map information point, and taking the information point image as the multimedia information to be associated.

[0145] It should be understood that searching for an information point image corresponding to the map information point can be searching for an information point image corresponding to the map information point in a preset map relationship table. The preset map relationship table includes the correspondence between the map information point and the information point image. The correspondence between the map information point and the information point image can be entered by the administrator of the multimedia information recommendation device when setting the map information point. The present embodiment does not limit this.

[0146] In the third embodiment, when receiving an information association instruction, the map information to be processed is determined according to the information association instruction, the map information point is determined according to the map information to be processed, the information point image corresponding to the map information point is searched, and the information point image is taken as the multimedia information to be associated. In the present embodiment, the user selects the map information to be processed through the instruction, so that all map information does not need to be processed, thereby reducing the computational amount and improving the processing efficiency.

[0147] In the third embodiment, the step S03 includes:

[0148] Step S031: grouping the multimedia information to be associated to obtain a plurality of multimedia information groups.

[0149] It should be understood that grouping the multimedia information to be associated to obtain a plurality of multimedia information groups can be searching for a map information point corresponding to the multimedia information to be associated, obtaining attribute information of the map information point, and grouping the multimedia information to be associated according to the attribute information to obtain a plurality of multimedia information groups. The attribute information can be province information, category information, name information and coordinate information, etc. The present embodiment does not limit this.

[0150] In specific implementation, for example, the images to be associated are divided into m and n groups according to the province information and the category information of the POI to which the images belong, respectively. For each group, the embedding vectors of all images in the group are used to construct the faiss index.

[0151] Step S032: traversing the multimedia information to be associated, and taking the traversed multimedia information to be associated as the multimedia information to be processed.

[0152] It can be understood that the traversing of the to-be-associated multimedia information can be random or according to a preset order. The preset order can be set by a manager of the multimedia information recommendation device, which is not limited in the embodiment.

[0153] Step S033: searching for target multimedia information associated with the to-be-processed multimedia information in each multimedia information group according to the feature vector of the to-be-processed multimedia information.

[0154] It should be understood that searching for target multimedia information associated with the to-be-processed multimedia information in each multimedia information group according to the feature vector of the to-be-processed multimedia information can be determining the relevance of the to-be-processed multimedia information and multimedia information in each multimedia information group according to the feature vector of the to-be-processed multimedia information and the feature vector of the multimedia information in each multimedia information group, sorting the multimedia information according to the relevance, and determining the target multimedia information according to the sorting result.

[0155] It can be understood that determining the target multimedia information according to the sorting result can be taking a preset number of multimedia information at the front of the sorting result as the target multimedia information. The preset number can be set by a manager of the multimedia information recommendation device, which is not limited in the embodiment.

[0156] Further, in actual applications, if the to-be-processed multimedia information is matched with all multimedia information groups, it will inevitably lead to too many processed images and too large computation. To overcome this defect, the step S033 comprises:

[0157] obtaining attribute information of the to-be-processed multimedia information, and screening the multimedia information groups according to the attribute information to obtain a target multimedia information group;

[0158] searching for target multimedia information associated with the to-be-processed multimedia information in the target multimedia information group according to the feature vector of the to-be-processed multimedia information.

[0159] It should be understood that obtaining attribute information of the to-be-processed multimedia information can be searching for map information points corresponding to the to-be-processed multimedia information, and obtaining attribute information of the map information points, that is, the attribute information of the to-be-processed multimedia information can be attribute information of the map information points corresponding to the to-be-processed multimedia information, which is not limited in the embodiment.

[0160] Further, to improve the reliability of the target multimedia information group, the obtaining attribute information of the to-be-processed multimedia information, and screening the multimedia information groups according to the attribute information to obtain a target multimedia information group comprises:

[0161] obtaining attribute information of the to-be-processed multimedia information, and matching the attribute information of the to-be-processed multimedia information with attribute information of the multimedia information groups to obtain a matching result;

[0162] screening the multimedia information groups according to the matching result to obtain a target multimedia information group.

[0163] It can be understood that the screening of the multimedia information groups according to the matching result to obtain a target multimedia information group can be that the multimedia information group with a matching result of matching success is taken as the target multimedia information group, and the present example does not limit this.

[0164] Further, in order to improve the reliability of the target multimedia information, the searching of the target multimedia information associated with the to-be-processed multimedia information in the target multimedia information group according to the feature vector of the to-be-processed multimedia information comprises:

[0165] determining a relevance degree of the to-be-processed multimedia information and the multimedia information according to the feature vector of the to-be-processed multimedia information and the feature vector of each multimedia information in the target multimedia information group;

[0166] sorting the multimedia information according to the relevance degree, and screening the multimedia information according to a sorting result to obtain a target multimedia information.

[0167] It should be understood that the sorting of the multimedia information according to the relevance degree and the screening of the multimedia information according to the sorting result to obtain a target multimedia information can be that the multimedia information is sorted in descending order of the relevance degree, and a preset number of multimedia information with a high ranking is taken as the target multimedia information.

[0168] Step S034: establishing a corresponding relationship between the to-be-processed multimedia information and the target multimedia information.

[0169] It can be understood that the establishment of the corresponding relationship between the to-be-processed multimedia information and the target multimedia information can be that the to-be-processed multimedia information is associated with the target multimedia information, so as to be used subsequently.

[0170] Further, considering that there is target multimedia information in each group, if the corresponding relationship between the to-be-processed multimedia information and the target multimedia information is directly established, the number of corresponding relationship groups is too large, in order to overcome the above defects, the step S034 comprises:

[0171] merging the target multimedia information in each multimedia information group to obtain to-be-processed multimedia information;

[0172] The multimedia information to be sorted is sorted, and the sorted multimedia information is filtered according to a sorting result to obtain associated multimedia information.

[0173] A corresponding relationship between the multimedia information to be processed and the associated multimedia information is established.

[0174] In a specific implementation, for example, the images to be processed are selected from the images to be associated, for each image to be processed, a POI province and a category to which the image to be processed belongs are obtained, embedding vectors of the image to be processed are used to retrieve top-k results with the highest relevance from corresponding province faiss indexes and category faiss indexes respectively, and then results of the two indexes are combined, sorted, and target images are obtained, and a corresponding relationship between the image to be processed and the target image is established.

[0175] Step S035: After the iteration of the multimedia information to be associated ends, a preset mapping relationship between the multimedia information to be associated is established according to the corresponding relationship.

[0176] It should be understood that after the iteration of the multimedia information to be associated ends, it is indicated that all the multimedia information to be associated has been processed, at this time, the preset mapping relationship between the multimedia information to be associated can be directly established according to the corresponding relationship.

[0177] In a specific implementation, the preset mapping relationship between the multimedia information to be associated can be stored in redis.

[0178] In the third embodiment, the multimedia information to be associated is grouped to obtain a plurality of multimedia information groups, the multimedia information to be associated is iterated, the multimedia information to be associated iterated is taken as multimedia information to be processed, target multimedia information associated with the multimedia information to be processed in each multimedia information group is searched according to a feature vector of the multimedia information to be processed, a corresponding relationship between the multimedia information to be processed and the target multimedia information is established, and after the iteration of the multimedia information to be associated ends, a preset mapping relationship between the multimedia information to be associated is established according to the corresponding relationship, so that the preset mapping relationship between the multimedia information to be associated can be quickly and accurately established.

[0179] In addition, an embodiment of the present application further provides a storage medium, the storage medium stores a multimedia information recommendation program, and the multimedia information recommendation program is executed by a processor to implement steps of the multimedia information recommendation method.

[0180] In addition, with reference to Figure 5 , an embodiment of the present application further provides a multimedia information recommendation device, the multimedia information recommendation device comprises a determination module 10, a search module 20 and a sending module 30.

[0181] The determining module 10 is configured to determine multimedia information to be browsed according to the multimedia information browsing instruction when receiving the multimedia information browsing instruction sent by the user terminal.

[0182] It should be noted that the multimedia information browsing instruction can be a control instruction input by the user through the interactive program of the user terminal. The multimedia information can be image, audio and text information, and in this embodiment and other embodiments, the image is taken as an example for description.

[0183] In a specific implementation, for example, the user can click the image viewing button of the POI point on the map software to issue an image browsing instruction.

[0184] Further, in order to be able to pre-establish a preset mapping relationship between the multimedia information, so as to quickly generate the to-be-recommended multimedia information corresponding to the to-be-browsed multimedia information in subsequent user browsing, the multimedia information recommendation apparatus further comprises an association module.

[0185] The association module is configured to determine to-be-associated multimedia information according to an information association instruction when receiving the information association instruction, analyze the to-be-associated multimedia information, obtain a feature vector of the to-be-associated multimedia information, and establish a preset mapping relationship between the to-be-associated multimedia information according to the feature vector.

[0186] In a specific implementation, the administrator of the server inputs an information association instruction through the management interface of the server, and the server downloads all images of POIs through a picture bed to obtain to-be-associated images when receiving the information association instruction, performs image embedding on the to-be-associated images through a resnet model to obtain a feature vector of the to-be-associated images, divides the to-be-associated images into m and n groups according to the province information and the category information of the POIs to which the to-be-associated images belong respectively, constructs a faiss index using the embedding vectors of all images in each group, selects to-be-processed images from the to-be-associated images, obtains the province and the category of the POI to which each to-be-processed image belongs, uses the embedding vector of the to-be-processed image to retrieve the top-k results with the highest relevance from the corresponding province faiss index and category faiss index respectively, then merges the results of the two indexes, sorts the results, obtains target images, and establishes a corresponding relationship between the to-be-processed images and the target images.

[0187] Still further, in order to be able to set the mapping relationship between the multimedia information in advance when the multimedia information is added, the association module is further configured to, when receiving a multimedia information adding instruction, determine multimedia information to be added according to the multimedia information adding instruction, analyze the multimedia information to be added, obtain a feature vector of the multimedia information to be added, and establish a preset mapping relationship between the multimedia information to be added and preset multimedia information according to the feature vector.

[0188] The search module 20 is configured to search for the multimedia information to be recommended corresponding to the multimedia information to be browsed from the preset mapping relationship, and the preset mapping relationship is generated based on the feature vectors between different multimedia information.

[0189] It can be understood that after the multimedia information to be recommended corresponding to the multimedia information to be browsed is searched, the multimedia information to be browsed can also be sorted according to the relevance of the multimedia information to be recommended and the multimedia information to be browsed, and the display order is determined according to the sorting result. The relevance of the multimedia information to be recommended and the multimedia information to be browsed can be generated based on the feature vectors between different multimedia information, which is not limited in the embodiment.

[0190] The sending module 30 is configured to send the multimedia information to be browsed and the multimedia information to be recommended to the user terminal.

[0191] It should be understood that the user terminal can directly display the multimedia information to be browsed and the multimedia information to be recommended after receiving the multimedia information to be browsed and the multimedia information to be recommended.

[0192] In the embodiment, when receiving the multimedia information browsing instruction sent by the user terminal, the multimedia information to be browsed is determined according to the multimedia information browsing instruction, the multimedia information to be recommended corresponding to the multimedia information to be browsed is searched from the preset mapping relationship, the preset mapping relationship is generated based on the feature vectors between different multimedia information, the multimedia information to be browsed and the multimedia information to be recommended are sent to the user terminal. Compared with the existing method of generating recommended multimedia information by semantic recognition according to the name of the POI point, since the multimedia information to be recommended corresponding to the multimedia information to be browsed is searched from the preset mapping relationship in the embodiment, the preset mapping relationship is generated based on the feature vectors between different multimedia information, and the feature vector of the multimedia information is directly associated with the multimedia information, thereby improving the relevance of the multimedia information recommendation, overcoming the defect of poor recommendation effect, and further improving the user experience.

[0193] Other embodiments or specific implementation manners of the multimedia information recommendation device can refer to the above-mentioned method embodiments, which will not be described here.

[0194] It should be noted that, in this document, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0195] The above-mentioned embodiment numbers of the present application are only for description, and do not represent the advantages or disadvantages of the embodiments. In the unit claims in which several devices are listed, several of the devices can be embodied by the same hardware item. The use of the words first, second, and third, etc. does not represent any order, and the words can be interpreted as names.

[0196] From the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be realized by means of software and necessary general hardware platforms, and of course, can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as a Read Only Memory image (ROM) / Random Access Memory (RAM), a magnetic disk, or an optical disk) and includes a number of instructions for making an end device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device) execute the methods described in the various embodiments of the present application.

[0197] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent process transformation, or direct or indirect application in other related technical fields, is also included in the patent protection scope of the present application.

Claims

1. A multimedia information recommendation method characterized by comprising: The multimedia information recommendation method comprises the following steps: Upon receiving a multimedia information browsing instruction sent by a user terminal, determining multimedia information to be browsed according to the multimedia information browsing instruction, wherein the multimedia information browsing instruction is an instruction issued by the user when clicking an image viewing button of a map interest point on a map software; Finding multimedia information to be recommended corresponding to the multimedia information to be browsed from a preset mapping relationship, wherein the preset mapping relationship is generated based on feature vectors between different multimedia information, the different multimedia information is to-be-associated images, and the to-be-associated images are images of all map interest points downloaded through a picture bed; Sending the multimedia information to be browsed and the multimedia information to be recommended to the user terminal.

2. The multimedia information recommendation method of claim 1, wherein, Before the step of determining multimedia information to be browsed according to a multimedia information browsing instruction sent by a user terminal upon receiving the multimedia information browsing instruction, the multimedia information recommendation method further comprises: Upon receiving an information association instruction, determining multimedia information to be associated according to the information association instruction; Analyzing the multimedia information to be associated to obtain feature vectors of the multimedia information to be associated; Establishing a preset mapping relationship between the multimedia information to be associated according to the feature vectors.

3. The multimedia information recommendation method of claim 2, wherein, The step of establishing a preset mapping relationship between the multimedia information to be associated according to the feature vectors comprises: Grouping the multimedia information to be associated to obtain a plurality of multimedia information groups; Iterating the multimedia information to be associated, and taking the iterated multimedia information to be associated as multimedia information to be processed; Finding target multimedia information associated with the multimedia information to be processed in each multimedia information group according to the feature vectors of the multimedia information to be processed; Establishing a corresponding relationship between the multimedia information to be processed and the target multimedia information; After the iteration of the multimedia information to be associated ends, establishing a preset mapping relationship between the multimedia information to be associated according to the corresponding relationship.

4. The multimedia information recommendation method of claim 3, wherein, The step of finding target multimedia information associated with the multimedia information to be processed in each multimedia information group according to the feature vectors of the multimedia information to be processed comprises: Obtaining attribute information of the multimedia information to be processed, and screening the multimedia information groups according to the attribute information to obtain target multimedia information groups; Finding target multimedia information associated with the multimedia information to be processed in the target multimedia information groups according to the feature vectors of the multimedia information to be processed.

5. The multimedia information recommendation method of claim 4, wherein, The step of obtaining attribute information of the multimedia information to be processed, and screening the multimedia information groups according to the attribute information to obtain target multimedia information groups comprises: Obtaining attribute information of the multimedia information to be processed, and matching the attribute information of the multimedia information to be processed with attribute information of the multimedia information groups to obtain a matching result; Screening the multimedia information groups according to the matching result to obtain target multimedia information groups. The step of obtaining attribute information of the multimedia information to be processed, and screening the multimedia information groups according to the attribute information to obtain target multimedia information groups comprises: Obtaining attribute information of the multimedia information to be processed, and matching the attribute information of the multimedia information to be processed with attribute information of the multimedia information groups to obtain a matching result; Screening the multimedia information groups according to the matching result to obtain target multimedia information groups.

6. The multimedia information recommendation method of claim 4, wherein, The step of searching the target multimedia information associated with the to-be-processed multimedia information from the target multimedia information group according to the feature vector of the to-be-processed multimedia information specifically comprises: Determining the relevance of the to-be-processed multimedia information and the multimedia information according to the feature vector of the to-be-processed multimedia information and the feature vector of each multimedia information in the target multimedia information group; According to the relevance, the multimedia information is sorted, and the target multimedia information is obtained by screening the multimedia information according to the sorting result.

7. The multimedia information recommendation method of claim 3, wherein, The step of establishing the corresponding relationship between the to-be-processed multimedia information and the target multimedia information specifically comprises: Merging the target multimedia information in each multimedia information group to obtain to-be-sequenced multimedia information; The to-be-sequenced multimedia information is sorted, and the associated multimedia information is obtained by screening the to-be-sequenced multimedia information according to the sorting result; Establishing the corresponding relationship between the to-be-processed multimedia information and the associated multimedia information.

8. The multimedia information recommendation method of claim 2, wherein, The step of determining the to-be-associated multimedia information according to the information association instruction when the information association instruction is received specifically comprises: When the information association instruction is received, the to-be-processed map information is determined according to the information association instruction, and the map information point is determined according to the to-be-processed map information; Searching for the information point image corresponding to the map information point, and taking the information point image as the to-be-associated multimedia information.

9. The multimedia information recommendation method according to any one of claims 1-8, wherein, The multimedia information recommendation method further comprises: When the multimedia information adding instruction is received, the to-be-added multimedia information is determined according to the multimedia information adding instruction; The to-be-added multimedia information is analyzed to obtain the feature vector of the to-be-added multimedia information; The preset mapping relationship between the to-be-added multimedia information and the preset multimedia information is established according to the feature vector of the to-be-added multimedia information.

10. The multimedia information recommendation method according to any one of claims 1-8, wherein, The step of determining the to-be-viewed multimedia information according to the multimedia information browsing instruction when the multimedia information browsing instruction sent by the user terminal is received specifically comprises: When the multimedia information browsing instruction sent by the user terminal is received, the instruction identifier is obtained by analyzing the multimedia information browsing instruction; The to-be-viewed multimedia information is determined according to the instruction identifier.

11. The multimedia information recommendation method according to any one of claims 1 to 8, wherein, The step of searching the to-be-viewed multimedia information corresponding to the to-be-recommended multimedia information from the preset mapping relationship, which is generated based on the feature vectors between different multimedia information, specifically comprises: The candidate multimedia information corresponding to the to-be-viewed multimedia information is searched from the preset mapping relationship, which is generated based on the feature vectors between different multimedia information; The user portrait of the user terminal is obtained, and the candidate multimedia information is screened according to the user portrait to obtain the to-be-recommended multimedia information.

12. The multimedia information recommendation method according to any one of claims 1 to 8, wherein, The step of sending the to-be-viewed multimedia information and the to-be-recommended multimedia information to the user terminal specifically comprises: acquire device information of the user terminal, and find an information display template corresponding to the device information; send the information display template, the multimedia information to be browsed, and the multimedia information to be recommended to the user terminal.

13. A multimedia information recommendation device characterized by comprising: The multimedia information recommendation device includes a memory, a processor, and a multimedia information recommendation program stored on the memory and executable on the processor. When the multimedia information recommendation program is executed by the processor, the steps of the multimedia information recommendation method according to any one of claims 1 to 12 are implemented.

14. A storage medium, characterized by The storage medium stores a multimedia information recommendation program. When the multimedia information recommendation program is executed by the processor, the steps of the multimedia information recommendation method according to any one of claims 1 to 12 are implemented.

15. A multimedia information recommendation apparatus characterized by comprising: The multimedia information recommendation device includes a determination module, a finding module, and a sending module. The determination module is configured to determine multimedia information to be browsed according to a multimedia information browsing instruction sent by a user terminal when the multimedia information browsing instruction is received. The multimedia information browsing instruction is an instruction issued by a user when clicking an image viewing button of a map interest point on a map software. The finding module is configured to find multimedia information to be recommended corresponding to the multimedia information to be browsed from a preset mapping relationship. The preset mapping relationship is generated based on feature vectors between different multimedia information. The different multimedia information is image to be associated, and the image to be associated is an image of all map interest points downloaded through a picture bed. The sending module is configured to send the multimedia information to be browsed and the multimedia information to be recommended to the user terminal.

16. The multimedia information recommendation apparatus of Claim 15, characterized by comprising: The multimedia information recommendation device further includes an association module. The association module is configured to determine multimedia information to be associated according to an information association instruction when the information association instruction is received. The association module is further configured to analyze the multimedia information to be associated to obtain a feature vector of the multimedia information to be associated. The association module is further configured to establish a preset mapping relationship between the multimedia information to be associated based on the feature vector.

17. The multimedia information recommendation apparatus of Claim 16, characterized by comprising: The association module is further configured to group the multimedia information to be associated to obtain a plurality of multimedia information groups. The association module is further configured to traverse the multimedia information to be associated, and take the traversed multimedia information to be associated as multimedia information to be processed. The association module is further configured to find target multimedia information associated with the multimedia information to be processed in each multimedia information group based on a feature vector of the multimedia information to be processed. The association module is further configured to establish a corresponding relationship between the multimedia information to be processed and the target multimedia information. The association module is further configured to establish a preset mapping relationship between the multimedia information to be associated based on the corresponding relationship after the traversal of the multimedia information to be associated ends.

18. The multimedia information recommendation apparatus of Claim 17, characterized by comprising: The association module is further configured to acquire attribute information of the multimedia information to be processed, and filter the multimedia information groups based on the attribute information to obtain a target multimedia information group. The association module is further configured to search for target multimedia information associated with the to-be-processed multimedia information from the target multimedia information group according to the feature vector of the to-be-processed multimedia information.

19. The multimedia information recommendation apparatus of Claim 18, characterized by comprising: The association module is further configured to acquire attribute information of the to-be-processed multimedia information, and match the attribute information of the to-be-processed multimedia information with attribute information of the multimedia information group to obtain a matching result. The association module is further configured to screen the multimedia information group according to the matching result to obtain a target multimedia information group.

20. The multimedia information recommendation apparatus of Claim 18, characterized by comprising: The association module is further configured to determine a relevance degree between the to-be-processed multimedia information and the multimedia information according to the feature vector of the to-be-processed multimedia information and feature vectors of the multimedia information in the target multimedia information group. The association module is further configured to sort the multimedia information according to the relevance degree, and screen the multimedia information according to a sorting result to obtain a target multimedia information.

Images