CN111666436B

CN111666436B - Data processing method and device and electronic equipment

Info

Publication number: CN111666436B
Application number: CN201910172732.1A
Authority: CN
Inventors: 谢泽华; 周泽南; 苏雪峰; 许静芳
Original assignee: Beijing Sogou Technology Development Co Ltd
Current assignee: Beijing Sogou Technology Development Co Ltd
Priority date: 2019-03-07
Filing date: 2019-03-07
Publication date: 2024-05-07
Anticipated expiration: 2039-03-07
Also published as: CN111666436A

Abstract

The embodiment of the invention provides a data processing method, a data processing device and electronic equipment, wherein the method comprises the following steps: receiving a search term; searching an online picture according to the search word to determine an online search result, and searching an offline search result from an offline picture library according to the search word, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures; ranking the offline search results back before the online search results; and then can arrange the high-quality picture in the front and demonstrate to improved the search result quality, and can also improve user experience.

Description

Data processing method and device and electronic equipment

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to a data processing method, a data processing device, and an electronic device.

Background

With the continuous development of internet technology and the development of search engine technology, users can perform information search through search applications, such as video search, music search, paper search, picture search, and so on.

In the process of searching pictures, a search engine usually uses the relevance as a ranking standard, and the pictures with high relevance are ranked before the pictures with low relevance, and then returned according to the ranking result. But the correlation may not be good, for example, the user inputs the search term "lake image" and searches for the image, and the lake image photographed by a general user is arranged before the lake image photographed by a photographing master in the returned search result page, i.e. the image photographed by the general user is higher in correlation than the image photographed by the photographing master, but the image photographed by the actual photographing master is more attractive to eyes than the image photographed by the general user; it can be seen that the existing picture search results are of low quality.

Disclosure of Invention

The embodiment of the invention provides a data processing method for improving the quality of search results.

Correspondingly, the embodiment of the invention also provides a data processing device and electronic equipment, which are used for ensuring the realization and application of the method.

In order to solve the above problems, an embodiment of the present invention discloses a data processing method, which specifically includes: receiving a search term; searching an online picture according to the search word to determine an online search result, and searching an offline search result from an offline picture library according to the search word, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures; and returning the offline search results before the online search results.

Optionally, searching the offline search result from the offline picture library according to the search word includes: searching a plurality of pictures from an offline picture library based on the search word searching mapping relation; and ordering the pictures in a descending order according to the quality scores corresponding to the pictures, and generating an offline search result.

Optionally, the determining the online search result by performing online picture search according to the search word includes: calculating the relevance scores of each picture in the online picture library and the search word; and ordering the pictures in a descending order according to the relevance scores of the pictures, and generating an online search result.

Optionally, the method further comprises the step of establishing an offline picture library: acquiring a first data set, the first data set comprising: the method comprises the steps of a plurality of first historical search words, a plurality of first pictures corresponding to each first historical search word and picture feature information corresponding to each first picture, wherein the picture feature information comprises: user behavior feature information, relevance feature information, and aesthetic feature information; determining the quality scores of the first pictures corresponding to the first historical search words according to the first data set and a preset model; and selecting a target picture with the quality score higher than a quality threshold value, and establishing an offline picture library according to the target picture and the corresponding first historical search word.

Optionally, the establishing an offline picture library according to the target picture and the corresponding first historical search word includes: respectively establishing mapping relations between each target picture and the corresponding first historical search word; and establishing an offline picture library by adopting each target picture, the mapping relation corresponding to each target picture and the quality score.

Optionally, the obtaining a plurality of first historical search terms includes: acquiring a plurality of historical search words from the picture search log, and counting the frequency corresponding to each historical search word; the first N historical search words with highest frequency are selected as first historical search words, and N is a positive integer.

Optionally, the obtaining the corresponding picture feature information of each first picture includes: acquiring user behavior characteristic information corresponding to each first picture from a picture search log, wherein the user behavior characteristic information comprises a plurality of dimensions; calculating the relevance scores of the first pictures corresponding to the first historical search words, and determining relevance characteristic information according to the relevance scores; and determining aesthetic characteristic information corresponding to each first picture by adopting an aesthetic quality evaluation model.

Optionally, the method further comprises the step of training the preset model: obtaining a second data set, the second data set comprising: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures; determining reference aesthetic feature information and reference correlation feature information of each second picture, and determining a reference quality score of each second picture according to the reference aesthetic feature information and the reference correlation feature information; and training the preset model according to the second data set and the reference quality score.

Optionally, the training the preset model according to the second data set and the reference quality score includes: inputting the second data set into the preset model to obtain a high-quality score; and comparing the high-quality score with a corresponding reference high-quality score, and adjusting the preset model.

The embodiment of the invention also discloses a data processing device, which specifically comprises: the receiving module is used for receiving the search word; the searching module is used for searching the online pictures according to the search words to determine online search results and searching the offline search results from an offline picture library according to the search words, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold value, and the quality scores are determined according to aesthetic values and relativity of the pictures; and the return module is used for returning the offline search results before the online search results.

Optionally, the search module includes: the offline searching sub-module is used for searching a plurality of pictures from an offline picture library based on the search word searching mapping relation; and ordering the pictures in a descending order according to the quality scores corresponding to the pictures, and generating an offline search result.

Optionally, the search module includes: the online searching sub-module is used for calculating the relevance scores of each picture in the online picture library and the search word; and ordering the pictures in a descending order according to the relevance scores of the pictures, and generating an online search result.

Optionally, the apparatus further comprises: a first acquisition module for acquiring a first data set, the first data set comprising: the method comprises the steps of a plurality of first historical search words, a plurality of first pictures corresponding to each first historical search word and picture feature information corresponding to each first picture, wherein the picture feature information comprises: user behavior feature information, relevance feature information, and aesthetic feature information; the scoring module is used for determining the high-quality scores of the first pictures corresponding to the first historical search words according to the first data set and a preset model; and the picture library establishing module is used for selecting a target picture with the quality score higher than the quality threshold value and establishing an offline picture library according to the target picture and the corresponding first historical search word.

Optionally, the picture library building module is configured to build mapping relations between each target picture and a corresponding first historical search word; and establishing an offline picture library by adopting each target picture, the mapping relation corresponding to each target picture and the quality score.

Optionally, the acquiring module includes: the search word acquisition sub-module is used for acquiring a plurality of historical search words from the picture search log and counting the frequency corresponding to each historical search word; the first N historical search words with highest frequency are selected as first historical search words, and N is a positive integer.

Optionally, the acquiring module includes:

the characteristic information acquisition sub-module is used for acquiring user behavior characteristic information corresponding to each first picture from the picture search log, wherein the user behavior characteristic information comprises a plurality of dimensions; calculating the relevance scores of the first pictures corresponding to the first historical search words, and determining relevance information according to the relevance scores; and determining aesthetic characteristic information corresponding to each first picture by adopting an aesthetic quality evaluation model.

Optionally, the apparatus further comprises: a second acquisition module for acquiring a second data set, the second data set comprising: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures; the reference information determining module is used for determining reference aesthetic feature information and reference correlation information of each second picture, and determining a reference quality score of each second picture according to the reference aesthetic feature information and the reference correlation information; and the training module is used for training the preset model according to the second data set and the reference quality score.

Optionally, the training module is configured to input the second data set into the preset model to obtain a quality score; and comparing the high-quality score with a corresponding reference high-quality score, and adjusting the preset model.

The embodiment of the invention also discloses a readable storage medium, which enables the electronic device to execute the data processing method according to any one of the embodiments of the invention when the instructions in the storage medium are executed by the processor of the electronic device.

The embodiment of the invention also discloses an electronic device, which comprises a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, and the one or more programs comprise instructions for: receiving a search term; searching an online picture according to the search word to determine an online search result, and searching an offline search result from an offline picture library according to the search word, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures; and returning the offline search results before the online search results.

Optionally, the method further comprises the following steps of establishing an offline picture library: acquiring a first data set, the first data set comprising: the method comprises the steps of a plurality of first historical search words, a plurality of first pictures corresponding to each first historical search word and picture feature information corresponding to each first picture, wherein the picture feature information comprises: user behavior feature information, relevance feature information, and aesthetic feature information; determining the quality scores of the first pictures corresponding to the first historical search words according to the first data set and a preset model; and selecting a target picture with the quality score higher than a quality threshold value, and establishing an offline picture library according to the target picture and the corresponding first historical search word.

Optionally, instructions for training the preset model are further included: obtaining a second data set, the second data set comprising: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures; determining reference aesthetic feature information and reference correlation feature information of each second picture, and determining a reference quality score of each second picture according to the reference aesthetic feature information and the reference correlation feature information; and training the preset model according to the second data set and the reference quality score.

The embodiment of the invention has the following advantages:

In summary, in the embodiment of the present invention, after receiving a search word, an online picture may be searched according to the search word to determine an online search result, and an offline search result may be searched from an offline picture library according to the search word; the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures; and then the offline search results are returned before the online search results, and the high-quality pictures can be displayed in front, so that the quality of the search results is improved, and the user experience is also improved.

Drawings

FIG. 1 is a flow chart of steps of an embodiment of a data processing method of the present invention;

FIG. 2a is a diagram of a search results presentation interface according to an embodiment of the present invention;

FIG. 2b is a schematic diagram of another search results presentation interface according to an embodiment of the present invention;

FIG. 3 is a flowchart illustrating steps of an embodiment of a method for creating an offline picture library according to the present invention;

FIG. 4 is a flowchart illustrating steps of an embodiment of a method for training a preset model in accordance with the present invention;

FIG. 5 is a flowchart illustrating steps of an alternate embodiment of a data processing method of the present invention;

FIG. 6 is a block diagram of an embodiment of a data processing apparatus of the present invention;

FIG. 7 is a block diagram of an alternative embodiment of a data processing apparatus of the present invention;

FIG. 8 is a block diagram of an electronic device for data processing according to an exemplary embodiment;

Fig. 9 is a schematic diagram showing a structure of an electronic device for data processing according to another exemplary embodiment of the present invention.

Detailed Description

In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.

One of the core concepts of the embodiment of the invention is that when a search word is received, an online picture search and an offline picture search are simultaneously carried out, wherein the offline picture search can be carried out from an offline picture library containing pictures which meet both relevance and aesthetic value; and then, the pictures searched by the offline pictures are returned before the pictures searched by the online pictures, so that the high-quality pictures can be displayed in front in the follow-up process, and the quality of the search results is improved.

Referring to fig. 1, a flowchart illustrating steps of an embodiment of a data processing method according to the present invention may specifically include the following steps:

Step 102, receiving search words.

In the embodiment of the invention, a user can search pictures on a search platform, wherein search words can be input in a picture search interface, and then search operation is executed; and then the search platform can receive the search instruction corresponding to the search operation, can acquire the search word from the search instruction, and send the search word to a search engine, and call the search engine to search pictures. After the search engine receives the search word, search results matched with the search word can be searched, and the search results can comprise pictures, attribute information of the pictures and the like.

And 104, searching online pictures according to the search words to determine online search results, and searching offline search results from an offline picture library according to the search words, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures.

In the embodiment of the invention, the search engine can perform online picture searching and offline picture searching simultaneously in the picture searching process, wherein online picture searching can be performed from an online picture library, online search results matched with the search word are searched, such as the correlation between each picture in the online picture library and the search word is calculated, and the matched online search results are determined according to the correlation; the online picture library may be a picture library generated from pictures crawled from the whole network in advance. And searching the offline picture from the offline picture library, searching the offline search result matched with the search word, for example, searching the matched offline search result according to the mapping relation between the search word and the picture. The off-line picture library is generated in advance, the generation process is described later, the off-line picture library comprises pictures with quality scores higher than a quality threshold, the quality scores are determined according to aesthetic values and relativity of the pictures, and the quality threshold can be set according to requirements; the relevance may refer to the relevance between a picture and a corresponding search term, and if both the aesthetic value and the relevance of the picture are higher, the quality score corresponding to the picture is higher, and the picture is better.

Step 106, the offline search results are returned before the online search results.

Wherein the online search results may be ranked after the online search results are determined, and the offline search results may be ranked after the offline search results are determined; the ranked offline search results may then be returned to the search platform before the ranked online search results. After the search platform receives the search results, the corresponding pictures can be displayed according to the arrangement sequence of the search results.

In one example of the invention, a search term such as "sunrise" is received, then an online picture search is performed according to the search term to determine an online search result, and an offline search result is searched from an offline picture library according to the search term; and then the offline search results are returned before the online search results. As shown in fig. 2a, a schematic diagram of a picture search result display interface according to the present invention is shown, in which 5 pictures in the first row are offline search results, 10 pictures in the second row-third row are online search results, the pictures in the first row are better than the composition of the pictures in the second and third rows, and the offline search results are better than the online search results.

In another example of the present invention, a search term such as "Hu Ge" is received, then an online picture search is performed according to the search term to determine an online search result, and an offline search result is searched from an offline picture library according to the search term; and then the offline search results are returned before the online search results. FIG. 2b is a schematic diagram of a picture search result presentation interface according to the present invention, wherein the first row of pictures are offline search results and the second row of pictures are online search results; the first line of pictures is better than the shooting angle of the second line of pictures, and the offline search results are better than the online search results in the visible.

In summary, in the embodiment of the present invention, after receiving a search word, an online picture may be searched according to the search word to determine an online search result, and an offline search result may be searched from an offline picture library according to the search word; the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures; and then, by arranging the offline search results before the online search results and returning, the high-quality pictures can be arranged in front for display, so that the quality of the search results is improved, and the user experience can be improved.

In another embodiment of the present invention, a process of creating an offline picture library is described.

Referring to fig. 3, a flowchart illustrating steps of an embodiment of a method for creating an offline picture library according to the present invention may specifically include the following steps:

step 302, acquiring a first data set, wherein the first data set comprises: the method comprises the steps of a plurality of first historical search words, a plurality of first pictures corresponding to each first historical search word and picture feature information corresponding to each first picture, wherein the picture feature information comprises: user behavior feature information, relevance feature information, and aesthetic feature information.

In the embodiment of the invention, a plurality of first historical search words and pictures corresponding to the first historical search words can be obtained, and then an offline picture library is built for the pictures corresponding to the first historical search words; wherein, the obtaining the plurality of first historical search terms may include the sub-steps of:

And step 22, acquiring a plurality of historical search words from the picture search log, and counting the frequency corresponding to each historical search word.

And 24, selecting the first N historical search words with highest frequency as first historical search words.

In the embodiment of the invention, a plurality of history search words can be obtained from the picture search log, wherein the history search words can be search words input when a whole network user searches pictures before the current moment; then counting the corresponding frequency of each historical search word, and selecting the first N historical search words with the highest frequency as first historical search words; and then the high-frequency search word of the whole network user can be selected, so that the offline search result searched from the offline picture library can meet the requirements of most users.

Then, for each first historical search word in the N first historical search words, a first picture corresponding to the first historical search word can be selected from an online picture library; one method for obtaining the first picture corresponding to the first historical search word may be to obtain picture information of the first picture corresponding to the first historical search word from a picture search log; and then counting the frequency of each first picture information, determining the first M pieces of picture information with the highest frequency, and acquiring M pieces of first pictures from an online picture library according to the picture information, wherein M is a positive integer. Another method for obtaining the first pictures corresponding to the first historical search terms may be to calculate the relevance scores of the pictures corresponding to the first historical search terms in the online picture library, and select the first M first pictures with the highest relevance scores. Of course, another method for obtaining the first picture corresponding to the first historical search term may be to directly select the first picture with the relevance score higher than the relevance threshold from the online picture library, where the relevance threshold may be set according to the requirement, and the embodiment of the present invention is not limited in this way.

In the embodiment of the invention, after each first historical search word and the corresponding first picture are obtained, a high-quality first picture can be selected to establish an offline picture library; and acquiring corresponding picture characteristic information corresponding to each first picture, and selecting a high-quality picture according to the picture characteristic information. In an embodiment of the present invention, the image feature information includes: user behavior feature information, relevance feature information, and aesthetic feature information. The user behavior feature information corresponding to each first picture may be obtained from a picture search log, where the user behavior feature information may include multiple dimensions, such as clicking, page turning, sharing, residence time, and the like, and may of course include other dimensions. And calculating the relevance scores of the first pictures and the corresponding first historical search words, and determining the relevance characteristic information according to the relevance scores, wherein the relevance scores can be in direct proportion to the relevance, for example, the bigger the relevance scores are, the higher the relevance of the first pictures and the corresponding first historical search words are. In the embodiment of the invention, an aesthetic quality evaluation model can be trained in advance: a large number of training pictures can be received, and corresponding reference aesthetic quality scores can be marked for each training picture; then, for each training image, the training image can be input into the aesthetic quality evaluation model to obtain an aesthetic quality score output by the aesthetic quality evaluation model; and comparing the output aesthetic quality score with a reference aesthetic quality score corresponding to the training picture, and adjusting the weight of the aesthetic quality evaluation model until the training image is input into the aesthetic quality evaluation model, wherein the aesthetic quality score output by the aesthetic quality evaluation model approaches to the corresponding reference aesthetic quality score. Then, a trained aesthetic quality evaluation model can be adopted to determine aesthetic characteristic information corresponding to each first picture, wherein for each first picture, the first picture can be input into the trained aesthetic quality evaluation model to obtain an aesthetic quality score output by the aesthetic quality evaluation model; the aesthetic quality score output by the aesthetic quality evaluation model is then used as aesthetic feature information, which can be used to characterize the aesthetic value of the picture, wherein the aesthetic feature information can be proportional to the aesthetic value, and the larger the aesthetic feature information, the higher the aesthetic value.

And step 304, determining the quality scores of the first pictures corresponding to the first historical search words according to the first data set and the preset model.

Then, scoring the first pictures corresponding to each first historical search word by adopting a preset model, wherein the training process of the preset model is described later; for a first picture of each first historical search word, the first historical search word, a first picture corresponding to the first historical search word, and picture feature information corresponding to the first picture may be input into the preset model, the preset model may score the first picture corresponding to the first historical search word, and a high-quality score corresponding to the first picture corresponding to the first historical search word may be output. By adopting the method, the quality scores of the first pictures corresponding to the first historical search words in the first data set can be determined.

And 306, selecting a target picture with a high-quality score higher than a high-quality threshold value, and establishing an offline picture library according to the target picture and the corresponding first historical search word.

In the embodiment of the invention, in order to establish a high-quality offline picture library, a picture with a high-quality score higher than a high-quality threshold value corresponding to each first historical search word can be selected as a target picture; then, an offline picture library can be established according to each target picture and the corresponding first historical search word; the method specifically comprises the following substeps:

step 62, respectively establishing mapping relations between each target picture and the corresponding first historical search word;

and a sub-step 64 of establishing an offline picture library by adopting each target picture, the mapping relation corresponding to each target picture and the high-quality score.

And then searching for offline search results corresponding to the search words according to the mapping relation in the process of searching the offline pictures.

In another embodiment of the present invention, a training process of a preset model is described.

Referring to fig. 4, a flowchart illustrating steps of an embodiment of a method for training a preset model according to the present invention may specifically include the following steps:

Step 402, obtaining a second data set, the second data set comprising: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures are used for searching the second historical search words.

In the embodiment of the present invention, a second data set may be obtained in advance as a training set, and training is performed on a preset model, where the second data set may include: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures; the number of the second historical search words in the second data set may be smaller than the number of the first historical search words in the first data set, and the second historical search words in the second data set may or may not overlap with the first historical search words in the first data set.

In the embodiment of the invention, a plurality of history search words can be obtained from the picture search log, and then P history search words can be randomly selected as second history search words, wherein P is a positive integer smaller than N. And then, obtaining second pictures corresponding to each second search word, wherein Q pictures can be selected as the second pictures by calculating the relevant scores of the pictures in the online picture library and the second historical search words for each second historical search word in the P second historical search words, wherein Q is a positive integer smaller than M. One way to select the Q second pictures may be to sort all the pictures corresponding to the second historical search word according to the relevant scores, and divide consecutive L pictures into a group in turn, so as to divide all the pictures into K groups; extracting H pictures from a group of pictures, wherein L, K, H is a positive integer, L is the total number of pictures corresponding to the historical search word, and K is H=M; and then, the Q second pictures, namely the pictures with high correlation scores and the pictures with low correlation scores, are improved to determine the high-quality scores by the preset model. The first picture in the first data set may or may not overlap with the second picture in the second data set, which is not limited in this embodiment of the present invention.

Step 404, determining reference aesthetic feature information and reference correlation feature information of each second picture, and determining a reference quality score of each second picture according to the reference aesthetic feature information and the reference correlation feature information.

In the embodiment of the invention, for each second picture, the reference correlation characteristic information of the second picture and the corresponding second historical search word and the reference aesthetic characteristic information of the second picture can be determined; and then, carrying out weighted calculation on the reference correlation characteristic information and the reference aesthetic characteristic information, and determining the reference quality score of the second picture. Wherein the reference merit score, reference aesthetic feature information may be manually determined, wherein the reference aesthetic feature information may be considered in terms of composition, theme, contrast, color saturation, and the like.

Step 406, training the preset model according to the second data set and the reference quality score.

Then the second data set can be input into the preset model to obtain a quality score; comparing the high-quality score with a corresponding reference high-quality score, and adjusting the preset model; and aiming at a second picture of a second historical search word, inputting the second historical search word, the second picture corresponding to the historical search word and picture characteristic information corresponding to the second picture into the preset model to obtain a high-quality score. And then comparing the high-quality score with a reference high-quality score corresponding to the second picture, and adjusting the preset model to enable the high-quality score output by the preset model to approach the reference high-quality score after the second historical search word, the second picture corresponding to the historical search word and the picture characteristic information corresponding to the second picture are input into the preset model.

Referring to fig. 5, a flowchart illustrating steps of an alternative embodiment of a data processing method of the present invention may specifically include the steps of:

Step 502, receiving a search term.

In the embodiment of the invention, after a user executes a search operation, a search platform can receive a search instruction corresponding to the search operation, then can acquire a search word from the search instruction, and send the search word to a search engine, and call the search engine to search pictures. After the search engine receives the search term, on the one hand, an online picture search may be performed, which may include steps 504-506, and on the other hand, an offline picture search may be performed, which may include steps 508-510.

And 504, calculating the correlation scores of the pictures in the online picture library and the search words.

And 506, sorting the pictures in a descending order according to the relevant scores of the pictures, and generating an online search result.

In the embodiment of the invention, the related scores of the pictures and the search words in the online picture library can be calculated, for example, the related scores of the pictures and the search words are calculated by adopting a BM25 formula; and then, ordering the pictures in a descending order according to the related scores of the pictures to generate an online search result.

And step 508, searching a plurality of pictures from an offline picture library based on the search word searching mapping relation.

And 510, sorting the pictures in a descending order according to the quality scores corresponding to the pictures, and generating an offline search result.

In the embodiment of the invention, a plurality of pictures corresponding to the search word can be searched according to the mapping relation in the established offline picture library, and then the pictures are sorted in a descending order according to the quality scores corresponding to the pictures to generate an offline search result.

Step 512, ranking the offline search results before the online search results is returned.

In the embodiment of the invention, the offline search result comprises the pictures which meet the relevance and the aesthetic value, so that after the offline search result and the online search result are obtained, the offline search result can be returned to the search platform before the online search result. After the search platform receives the search results, the search results are displayed according to the arrangement sequence of the search results, and then the offline search results in the search result page are displayed before the online search results, so that a user can conveniently and quickly find out high-quality pictures, and the quality of the search results, the search efficiency and the user experience are improved.

In summary, after receiving a search word, the embodiment of the invention can perform online picture search according to the search word to determine an online search result, and search an offline search result from an offline picture library according to the search word; the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity; and then the offline search results are returned before the online search results, and the high-quality pictures can be displayed in front, so that the quality of the search results is improved, and the user experience is also improved.

Secondly, in the process of performing offline searching, the embodiment of the invention can search the mapping relation based on the search word and search a plurality of pictures from an offline picture library; and ordering the pictures in a descending order according to the high-quality scores corresponding to the pictures to generate an offline search result, and further arranging the pictures with high-quality scores before the pictures with low-quality scores, thereby further improving user experience.

Thirdly, in the process of establishing an offline picture library, a preset model scores each first picture of each first historical search word according to user behavior characteristic information, correlation characteristic information and aesthetic characteristic information, and then establishes the offline picture library according to pictures with high quality scores; in the searching process, the offline searching result searched from the offline picture library not only meets the correlation and aesthetic value, but also can meet better user requirements, and the value of the searching result is further improved.

It should be noted that, for simplicity of description, the method embodiments are shown as a series of acts, but it should be understood by those skilled in the art that the embodiments are not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the embodiments. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred embodiments, and that the acts are not necessarily required by the embodiments of the invention.

Referring to FIG. 6, a block diagram illustrating an embodiment of a data processing apparatus of the present invention may include the following modules:

A receiving module 602, configured to receive a search term;

The searching module 604 is configured to perform online picture searching according to the search word to determine an online search result, and search for an offline search result from an offline picture library according to the search word, where the offline picture library includes pictures with a quality score higher than a quality threshold, and the quality score is determined according to an aesthetic value and a relevance;

a return module 606 for returning the offline search results before the online search results.

Referring to FIG. 7, a block diagram of an alternative embodiment of a data processing apparatus of the present invention is shown.

In an alternative embodiment of the present invention, the search module 604 includes:

An offline searching sub-module 6042, configured to search for a plurality of pictures from an offline picture library based on the search term searching mapping relationship; and ordering the pictures in a descending order according to the quality scores corresponding to the pictures, and generating an offline search result.

An online searching submodule 6044 for calculating the relevant scores of each picture and the search word in the online picture library; and ordering the pictures in a descending order according to the relevant scores of the pictures, and generating an online search result.

In an alternative embodiment of the present invention, the apparatus further comprises:

A first acquisition module 608 for acquiring a first data set, the first data set comprising: the method comprises the steps of a plurality of first historical search words, a plurality of first pictures corresponding to each first historical search word and picture feature information corresponding to each first picture, wherein the picture feature information comprises: user behavior feature information, relevance feature information, and aesthetic feature information;

a scoring module 610, configured to determine, according to the first data set and a preset model, a high-quality score of each first picture corresponding to each first historical search term;

The picture library establishment module 612 is configured to select a target picture with a quality score higher than a quality threshold, and establish an offline picture library according to the target picture and the corresponding first historical search word.

In an optional embodiment of the present invention, the picture library building module 612 is configured to build mapping relations between each target picture and a corresponding first historical search term; and establishing an offline picture library by adopting each target picture, the mapping relation corresponding to each target picture and the quality score.

In an alternative embodiment of the present invention, the obtaining module 608 includes:

The search word acquisition sub-module 6082 is configured to acquire a plurality of historical search words from the picture search log, and count the frequency corresponding to each historical search word; the first N historical search words with highest frequency are selected as first historical search words, and N is a positive integer.

the feature information obtaining sub-module 6084 is configured to obtain, from the picture search log, user behavior feature information corresponding to each first picture, where the user behavior feature information includes multiple dimensions; calculating the correlation scores corresponding to the first pictures and the corresponding first historical search words, and determining correlation characteristic information according to the correlation scores; and determining aesthetic characteristic information corresponding to each first picture by adopting an aesthetic quality evaluation model.

A second acquisition module 614 for acquiring a second data set, the second data set comprising: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures;

A reference information determining module 616, configured to determine reference aesthetic feature information and reference correlation feature information of each second picture, and determine a reference quality score of each second picture according to the reference aesthetic feature information and the reference correlation feature information;

and a training module 618, configured to train the preset model according to the second data set and the reference quality score.

In an alternative embodiment of the present invention, the training module 618 is configured to input the second data set into the preset model to obtain a quality score; and comparing the high-quality score with a corresponding reference high-quality score, and adjusting the preset model.

In summary, in the embodiment of the present invention, after receiving a search word, an online picture may be searched according to the search word to determine an online search result, and an offline search result may be searched from an offline picture library according to the search word; the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity; and then the offline search results are returned before the online search results, and the high-quality pictures can be displayed in front, so that the quality of the search results is improved, and the user experience is also improved.

For the device embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments for relevant points.

Fig. 8 is a block diagram illustrating a configuration of an electronic device 800 for data processing, according to an example embodiment. For example, electronic device 800 may be a mobile phone, computer, digital broadcast terminal, messaging device, game console, tablet device, medical device, exercise device, personal digital assistant, or the like.

Referring to fig. 8, an electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814, and a communication component 816.

The processing component 802 generally controls overall operation of the electronic device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. Processing element 802 may include one or more processors 820 to execute instructions to perform all or part of the steps of the methods described above. Further, the processing component 802 can include one or more modules that facilitate interactions between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate interaction between the multimedia component 808 and the processing component 802.

The memory 804 is configured to store various types of data to support operations at the device 800. Examples of such data include instructions for any application or method operating on the electronic device 800, contact data, phonebook data, messages, pictures, videos, and so forth. The memory 804 may be implemented by any type or combination of volatile or nonvolatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

The power component 806 provides power to the various components of the electronic device 800. Power components 806 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for electronic device 800.

The multimedia component 808 includes a screen between the electronic device 800 and the user that provides an output interface. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or slide action, but also the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 808 includes a front camera and/or a rear camera. When the electronic device 800 is in an operational mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

The audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a Microphone (MIC) configured to receive external audio signals when the electronic device 800 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may be further stored in the memory 804 or transmitted via the communication component 816. In some embodiments, audio component 810 further includes a speaker for outputting audio signals.

The I/O interface 812 provides an interface between the processing component 802 and peripheral interface modules, which may be a keyboard, click wheel, buttons, etc. These buttons may include, but are not limited to: homepage button, volume button, start button, and lock button.

The sensor assembly 814 includes one or more sensors for providing status assessment of various aspects of the electronic device 800. For example, the sensor assembly 814 may detect an on/off state of the device 800, a relative positioning of the components, such as a display and keypad of the electronic device 800, the sensor assembly 814 may also detect a change in position of the electronic device 800 or a component of the electronic device 800, the presence or absence of a user's contact with the electronic device 800, an orientation or acceleration/deceleration of the electronic device 800, and a change in temperature of the electronic device 800. The sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an acceleration sensor, a gyroscopic sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 816 is configured to facilitate communication between the electronic device 800 and other devices, either wired or wireless. The electronic device 800 may access a wireless network based on a communication standard, such as WiFi,2G, or 3G, or a combination thereof. In one exemplary embodiment, the communication part 814 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component 814 further includes a Near Field Communication (NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the electronic device 800 may be implemented by one or more Application Specific Integrated Circuits (ASICs), digital Signal Processors (DSPs), digital Signal Processing Devices (DSPDs), programmable Logic Devices (PLDs), field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic elements for executing the methods described above.

In an exemplary embodiment, a non-transitory computer readable storage medium is also provided, such as memory 804 including instructions executable by processor 820 of electronic device 800 to perform the above-described method. For example, the non-transitory computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

A non-transitory computer readable storage medium, which when executed by a processor of an electronic device, causes the electronic device to perform a data processing method, the method comprising: receiving a search term; searching an online picture according to the search word to determine an online search result, and searching an offline search result from an offline picture library according to the search word, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures; and returning the offline search results before the online search results.

Fig. 9 is a schematic structural view of an electronic device 900 for data processing according to another exemplary embodiment of the present invention. The electronic device 900 may be a server that may vary widely in configuration or performance and may include one or more central processing units (central processing units, CPU) 922 (e.g., one or more processors) and memory 932, one or more storage media 930 (e.g., one or more mass storage devices) that store applications 942 or data 944. Wherein the memory 932 and the storage medium 930 may be transitory or persistent. The program stored in the storage medium 930 may include one or more modules (not shown), each of which may include a series of instruction operations on a server. Still further, the central processor 922 may be arranged to communicate with a storage medium 930, and execute a series of instruction operations in the storage medium 930 on a server.

The server(s) may also include one or more power supplies 926, one or more wired or wireless network interfaces 950, one or more input/output interfaces 958, one or more keyboards 956, and/or one or more operating systems 941, such as Windows ServerTM, mac OS XTM, unixTM, linuxTM, freeBSDTM, and the like.

An electronic device comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs comprising instructions for: receiving a search term; searching an online picture according to the search word to determine an online search result, and searching an offline search result from an offline picture library according to the search word, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures; and returning the offline search results before the online search results.

In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described by differences from other embodiments, and identical and similar parts between the embodiments are all enough to be referred to each other.

Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiment and all such alterations and modifications as fall within the scope of the embodiments of the invention.

Finally, it is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or terminal device that comprises the element.

The foregoing has described in detail a data processing method, a data processing apparatus and an electronic device according to the present invention, and specific examples have been provided herein to illustrate the principles and embodiments of the present invention, the above examples being provided only to assist in understanding the method and core idea of the present invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.

Claims

1. A method of data processing, comprising:

Receiving a search term;

Searching an online picture according to the search word to determine an online search result, and searching an offline search result from an offline picture library according to the search word, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold, and the quality scores are determined according to aesthetic values and relativity of the pictures;

Ranking the offline search results back before the online search results;

the aesthetic value of the picture is determined by the following steps:

inputting a first picture into a trained aesthetic quality evaluation model aiming at the first picture to obtain an aesthetic quality score output by the aesthetic quality evaluation model;

and taking the aesthetic quality score output by the aesthetic quality evaluation model as aesthetic characteristic information, wherein the aesthetic characteristic information is used for representing the aesthetic value of the first picture.

2. The method of claim 1, wherein searching for offline search results from an offline picture library according to the search term comprises:

Searching a plurality of pictures from an offline picture library based on the search word searching mapping relation;

and ordering the pictures in a descending order according to the quality scores corresponding to the pictures, and generating an offline search result.

3. The method of claim 1, wherein the determining the online search result from the online picture search based on the search term comprises:

Calculating the relevance scores of each picture in the online picture library and the search word;

and ordering the pictures in a descending order according to the relevance scores of the pictures, and generating an online search result.

4. The method of claim 1, further comprising the step of creating an offline picture library:

Acquiring a first data set, the first data set comprising: the method comprises the steps of a plurality of first historical search words, a plurality of first pictures corresponding to each first historical search word and picture feature information corresponding to each first picture, wherein the picture feature information comprises: user behavior feature information, relevance feature information, and aesthetic feature information;

Determining the quality scores of the first pictures corresponding to the first historical search words according to the first data set and a preset model;

and selecting a target picture with the quality score higher than a quality threshold value, and establishing an offline picture library according to the target picture and the corresponding first historical search word.

5. The method of claim 4, wherein the creating an offline picture library from the target picture and the corresponding first historical search term comprises:

respectively establishing mapping relations between each target picture and the corresponding first historical search word;

And establishing an offline picture library by adopting each target picture, the mapping relation corresponding to each target picture and the quality score.

6. The method of claim 4, wherein the obtaining a plurality of first historical search terms comprises:

acquiring a plurality of historical search words from the picture search log, and counting the frequency corresponding to each historical search word;

The first N historical search words with highest frequency are selected as first historical search words, and N is a positive integer.

7. The method according to claim 4, wherein the obtaining the corresponding picture feature information of each first picture includes:

acquiring user behavior characteristic information corresponding to each first picture from a picture search log, wherein the user behavior characteristic information comprises a plurality of dimensions;

calculating the relevance scores of the first pictures corresponding to the first historical search words, and determining relevance characteristic information according to the relevance scores;

And determining aesthetic characteristic information corresponding to each first picture by adopting an aesthetic quality evaluation model.

8. The method of claim 4, further comprising the step of training the pre-set model:

Obtaining a second data set, the second data set comprising: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures;

determining reference aesthetic feature information and reference correlation feature information of each second picture, and determining a reference quality score of each second picture according to the reference aesthetic feature information and the reference correlation feature information;

And training the preset model according to the second data set and the reference quality score.

9. The method of claim 8, wherein training the pre-set model based on the second data set and a reference merit score comprises:

inputting the second data set into the preset model to obtain a high-quality score;

and comparing the high-quality score with a corresponding reference high-quality score, and adjusting the preset model.

10. A data processing apparatus, comprising:

the receiving module is used for receiving the search word;

The searching module is used for searching the online pictures according to the search words to determine online search results and searching the offline search results from an offline picture library according to the search words, wherein the offline picture library comprises pictures with quality scores higher than a quality threshold value, and the quality scores are determined according to aesthetic values and relativity of the pictures; the aesthetic value of the picture is determined by the following steps: inputting a first picture into a trained aesthetic quality evaluation model aiming at the first picture to obtain an aesthetic quality score output by the aesthetic quality evaluation model; taking the aesthetic quality score output by the aesthetic quality evaluation model as aesthetic feature information, wherein the aesthetic feature information is used for representing the aesthetic value of the first picture;

and the return module is used for returning the offline search results before the online search results.

11. The apparatus of claim 10, wherein the search module comprises:

the offline searching sub-module is used for searching a plurality of pictures from an offline picture library based on the search word searching mapping relation; and ordering the pictures in a descending order according to the quality scores corresponding to the pictures, and generating an offline search result.

12. The apparatus of claim 10, wherein the search module comprises:

the online searching sub-module is used for calculating the relevance scores of each picture in the online picture library and the search word; and ordering the pictures in a descending order according to the relevance scores of the pictures, and generating an online search result.

13. The apparatus of claim 10, wherein said apparatus further comprises:

A first acquisition module for acquiring a first data set, the first data set comprising: the method comprises the steps of a plurality of first historical search words, a plurality of first pictures corresponding to each first historical search word and picture feature information corresponding to each first picture, wherein the picture feature information comprises: user behavior feature information, relevance feature information, and aesthetic feature information;

the scoring module is used for determining the high-quality scores of the first pictures corresponding to the first historical search words according to the first data set and a preset model;

and the picture library establishing module is used for selecting a target picture with the quality score higher than the quality threshold value and establishing an offline picture library according to the target picture and the corresponding first historical search word.

14. The apparatus of claim 13, wherein the device comprises a plurality of sensors,

The picture library establishing module is used for respectively establishing mapping relations between each target picture and the corresponding first historical search word; and establishing an offline picture library by adopting each target picture, the mapping relation corresponding to each target picture and the quality score.

15. The apparatus of claim 13, wherein the acquisition module comprises:

the search word acquisition sub-module is used for acquiring a plurality of historical search words from the picture search log and counting the frequency corresponding to each historical search word; the first N historical search words with highest frequency are selected as first historical search words, and N is a positive integer.

16. The apparatus of claim 13, wherein the acquisition module comprises:

17. The apparatus of claim 13, wherein said apparatus further comprises:

A second acquisition module for acquiring a second data set, the second data set comprising: the second historical search words, the second pictures corresponding to the second historical search words and the picture characteristic information corresponding to the second pictures;

The reference information determining module is used for determining reference aesthetic feature information and reference correlation information of each second picture, and determining a reference quality score of each second picture according to the reference aesthetic feature information and the reference correlation information;

And the training module is used for training the preset model according to the second data set and the reference quality score.

18. The apparatus of claim 17, wherein the device comprises a plurality of sensors,

The training module is used for inputting the second data set into the preset model to obtain a high-quality score; and comparing the high-quality score with a corresponding reference high-quality score, and adjusting the preset model.

19. A readable storage medium, characterized in that instructions in said storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the data processing method according to any one of the method claims 1-9.

20. An electronic device comprising a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs comprising instructions for performing the data processing method of any of claims 1-9.