Search result recall method and apparatus, electronic device, and readable storage medium

Abstract

The present disclosure provides a search result recall method, device, electronic equipment and readable storage medium. The search result recall method comprises: obtaining to-be-processed multimedia data, splitting the to-be-processed multimedia data into multiple sub-data; obtaining multiple search sub-results corresponding to each sub-data according to the data characteristics of each sub-data, each search sub-result containing first identification information and second identification information; dividing the search sub-results corresponding to the multiple sub-data into at least one candidate sequence according to the first identification information of the search sub-results and the order of the multiple sub-data; obtaining at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-results; and taking the search results corresponding to the at least one target sequence as the recall results of the to-be-processed multimedia data. The present disclosure can improve the recall rate and accuracy of search results.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of Internet, and in particular, to a search result recall method and device, an electronic device, and a readable storage medium. BACKGROUND

[0002] With the increasing number of multimedia data resources such as audio, video, and image, the way of searching engine to obtain recall results on the Internet according to the query statement input by a user has gradually failed to meet the user's demand in obtaining multimedia data. SUMMARY

[0003] According to a first aspect of the present disclosure, a search result recall method is provided, including: obtaining to-be-processed multimedia data, and splitting the to-be-processed multimedia data into a plurality of sub-data; obtaining a plurality of search sub-results corresponding to each sub-data according to the data feature of each sub-data, each search sub-result containing first identification information and second identification information; dividing the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-results and the order of the plurality of sub-data; obtaining at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-results; and taking the search results corresponding to the at least one target sequence as the recall results of the to-be-processed multimedia data.

[0004] According to a second aspect of the present disclosure, a search result recall device is provided, including: a splitting unit configured to obtain to-be-processed multimedia data, and split the to-be-processed multimedia data into a plurality of sub-data; a searching unit configured to obtain a plurality of search sub-results corresponding to each sub-data according to the data feature of each sub-data, each search sub-result containing first identification information and second identification information; a first processing unit configured to divide the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-results and the order of the plurality of sub-data; a second processing unit configured to obtain at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-results; and a recall unit configured to take the search results corresponding to the at least one target sequence as the recall results of the to-be-processed multimedia data.

[0005] According to a third aspect of the present disclosure, an electronic device is provided, including: at least one processor; and a memory communicatively connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described above.

[0006] According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to perform the method as described above.

[0007] According to a fifth aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the method as described above.

[0008] As can be seen from the above technical solutions, the present disclosure increases the richness of the obtained search sub-results by obtaining a plurality of search sub-results corresponding to different sub-data, thereby improving the recall rate of the search results. Moreover, the determination of the at least one target sequence according to the first identification information and the second identification information of the search sub-results can greatly avoid the problem of false recall, thereby improving the accuracy of the final obtained recall results.

[0009] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present disclosure, nor to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0010] The accompanying drawings are used to better understand the present scheme and do not constitute a limitation on the present disclosure. Among them:

[0011] Figure 1 is a schematic diagram according to the first embodiment of the present disclosure;

[0012] Figure 2 is a schematic diagram according to the second embodiment of the present disclosure;

[0013] Figure 3 is a schematic diagram according to the third embodiment of the present disclosure;

[0014] Figure 4 is a schematic diagram according to the fourth embodiment of the present disclosure;

[0015] Figure 5 is a block diagram of an electronic device used to implement the search result recall method of the embodiments of the present disclosure. DETAILED DESCRIPTION

[0016] Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, which include various details of the embodiments of the present disclosure to assist in understanding, and should be considered as merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Also, in order to be clear and concise, descriptions of well-known functions and mechanisms are omitted in the following description.

[0017] Figure 1is a schematic diagram according to the first embodiment of the present disclosure. As shown in Figure 1 The search result recall method of the present embodiment specifically comprises the following steps:

[0018] S101, obtaining multimedia data to be processed, and splitting the multimedia data to be processed into a plurality of sub-data;

[0019] S102, obtaining a plurality of search sub-results corresponding to each sub-data according to the data characteristics of each sub-data, each search sub-result containing first identification information and second identification information;

[0020] S103, dividing the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-results and the order of the plurality of sub-data;

[0021] S104, obtaining at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-results;

[0022] S105, taking the search results corresponding to the at least one target sequence as the recall results of the multimedia data to be processed.

[0023] The search result recall method of the present embodiment, after splitting the multimedia data to be processed into a plurality of sub-data, first obtains a plurality of search sub-results corresponding to each sub-data, and then obtains at least one target sequence according to the first identification information and the second identification information of the search sub-results, and further takes the search results corresponding to the at least one target sequence as the recall results of the multimedia data to be processed. The present embodiment increases the richness of the obtained search sub-results by obtaining a plurality of search sub-results corresponding to different sub-data, thereby improving the recall rate of the search results. Moreover, determining the at least one target sequence according to the first identification information and the second identification information of the search sub-results can greatly avoid the problem of false recall, thereby improving the accuracy of the final recall results.

[0024] When performing S101 to obtain multimedia data to be processed, the present embodiment can take the multimedia data input by the input end as the multimedia data to be processed, or take the multimedia data selected by the input end from the network as the multimedia data to be processed. The multimedia data to be processed obtained by the present embodiment performing S101 can be audio data, image data, video data, etc.

[0025] Since the multimedia data to be processed can be different types of multimedia data, the search result recall method provided in this embodiment can realize audio search (e.g., listening to music to identify the song), image search (e.g., searching for an image), or video search, that is, the final recalled search result can be multimedia data containing the multimedia data to be processed, or multimedia data similar or identical to the multimedia data to be processed.

[0026] To improve the rationality of splitting the multimedia data to be processed, when performing S101 to split the multimedia data to be processed into multiple sub-data, an optional implementation manner that can be adopted by this embodiment is to determine the data type of the multimedia data to be processed, such as audio type, image type, or video type, etc.; and split the multimedia data to be processed into multiple sub-data according to the splitting manner corresponding to the determined data type.

[0027] In this embodiment, when performing S101, if it is determined that the multimedia data to be processed is audio data, the multimedia data to be processed can be split into multiple sub-data according to a preset time length (e.g., 2s); if it is determined that the multimedia data to be processed is image data, the multimedia data to be processed can be split into multiple sub-data according to a preset size (e.g., 20 pixels x 20 pixels).

[0028] To improve the recall rate of the search result, when performing S101 to split the multimedia data to be processed into multiple sub-data, this embodiment can also obtain a preset overlap rate, and then split the multimedia data to be processed into multiple sub-data in combination with the obtained preset overlap rate.

[0029] For example, if the preset overlap rate is 50%, this embodiment can split 6s of audio data into 5 segments of 2s of sub-audio, which are 0-2s, 1s-3s, 2s-4s, 3s-5s, and 4s-6s; if directly according to the splitting manner corresponding to the audio type, this embodiment can split 6s of audio data into 3 segments of 2s of sub-audio, which are 0-2s, 2s-4s, and 4s-6s.

[0030] After performing S101 to split the multimedia data to be processed into multiple sub-data, this embodiment performs S102 to obtain multiple search sub-results corresponding to each sub-data according to the data features of each sub-data.

[0031] When performing S102 to obtain multiple search sub-results corresponding to each sub-data according to the data features of each sub-data, this embodiment can match the data features (e.g., image features, audio features, etc.) of each sub-data with the data features of the search sub-results contained in the database, and then arrange the matching scores of the search sub-results in the top N positions (N is a positive integer greater than or equal to 2) as the multiple search sub-results corresponding to each sub-data.

[0032] That is, the embodiment obtains the Top N search sub-results corresponding to each sub-data for each sub-data, and expands the number of search sub-results corresponding to the sub-data, and in the case of increasing the number of obtained search sub-results, the recall rate of the search results in the database can be improved.

[0033] It can be understood that the search sub-results pre-stored in the database in the embodiment are also obtained by splitting the corresponding search results according to the preset size or the preset time length, and the search sub-results have the same time length or the same size as the sub-data, thereby improving the accuracy of the obtained search sub-results corresponding to the sub-data.

[0034] In the search sub-results obtained by the embodiment performing S102, in addition to the image data of the preset size, the audio data of the preset time length or the video data, the first identification information and the second identification information can also be included.

[0035] The first identification information is the identification information of the search result corresponding to the search sub-result, and the search sub- results with the same first identification information correspond to the same search result; and the second identification information is the identification information of the search sub-result in the search result corresponding thereto, for example, the position information of the image search sub-result in the image search result corresponding thereto, or the time sequence information (for example, frame ID) of the audio search sub-result in the audio search result corresponding thereto.

[0036] After the embodiment obtains the plurality of search sub-results corresponding to each sub-data by performing S102, the embodiment performs S103 to divide the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-result and the order of the plurality of sub-data.

[0037] That is, the embodiment obtains at least one candidate sequence by combining the order of the sub-data, the corresponding relationship between the sub-data and the search sub-result, and the first identification information of the search sub-result, so that different candidate sequences correspond to different first identification information, that is, the search sub- results included in one candidate sequence correspond to only one search result.

[0038] When the embodiment divides the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-result and the order of the plurality of sub-data by performing S103, an optional implementation manner that can be adopted by the embodiment is that for each first identification information, the search sub-result corresponding to the first identification information is selected from the plurality of search results corresponding to different sub-data respectively; and the selected search sub- results are arranged according to the order of the plurality of sub-data to obtain the candidate sequence corresponding to the first identification information.

[0039] That is, the embodiment selects the search sub-result corresponding to the first identification information from the search sub-results corresponding to different sub-data through the first identification information of the search result, so that each candidate sequence obtained finally corresponds to different first identification information, and the number of search sub-results contained in the obtained candidate sequence can be increased due to the expansion of the number of search sub-results, and the recall rate of the search result is further improved.

[0040] It can be understood that in the multiple search sub-results corresponding to the same sub-data, if the search sub- results with the same first identification information are multiple, the different search sub- results with the same first identification information usually have the same second identification information, for example, the sub-result G and the sub-result I in the following example, and arbitrarily selecting one of them will not affect the accuracy of the obtained candidate sequence.

[0041] Therefore, when performing S103 to select the search sub-result corresponding to each first identification information from the multiple search results corresponding to different sub-data respectively, the embodiment can further include the following content: for each sub-data and the multiple search sub- results corresponding thereto, in a case where it is determined that the number of search sub- results corresponding to the first identification information is multiple, one of the multiple search sub- results is selected as the search sub-result corresponding to the first identification information of the sub-data.

[0042] That is, the embodiment can ensure that there is no repeated search sub-result in the candidate sequence corresponding to different first identification information, improve the accuracy of the obtained candidate sequence, and in a case where there is no repeated candidate sub-result in one candidate sequence, the accuracy of the subsequently obtained target sequence can be improved.

[0043] For example, if the multiple sub-data are sub-data 1, sub-data 2 and sub-data 3 in order; if the search sub- results corresponding to the sub-data 1 are sub-result A, sub-result B and sub-result C, if the search sub- results corresponding to the sub-data 2 are sub-result D, sub-result E and sub-result F, if the search sub- results corresponding to the sub-data 3 are sub-result G, sub-result H and sub-result I; if the sub-result A, the sub-result F, the sub-result G and the sub-result I have the same first identification information 1, the sub-result B, the sub-result D and the sub-result H have the same first identification information 2, and the sub-result C and the sub-result E have the same first identification information 3; the embodiment can obtain the candidate sequence corresponding to the first identification information 1 as (sub-result A-sub-result F-sub-result G or sub-result I), the candidate sequence corresponding to the first identification information 2 as (sub-result B-sub-result D-sub-result H), and the candidate sequence corresponding to the first identification information 3 as (sub-result C-sub-result E) by performing S103.

[0044] After obtaining the at least one candidate sequence in S103, the embodiment can further include the following: for each candidate sequence, obtaining a matching score of each search sub-result in the candidate sequence, the matching score being calculated when obtaining the search sub-result according to the data feature; removing the search sub-result in the candidate sequence whose matching score is lower than a preset score threshold.

[0045] That is, the embodiment filters the search sub-results in the candidate sequence through the matching score, so that the search sub-results in the finally obtained candidate sequence all have a higher matching score, and thus the accuracy of the determined target sequence is improved.

[0046] After obtaining the at least one candidate sequence in S103, the embodiment performs S104 to obtain at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-result; wherein the target sequence obtained by the embodiment can be the candidate sequence itself, or a sub-sequence in the candidate sequence.

[0047] When obtaining the at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-result in S104, the embodiment can use the following optional implementation manner: for each candidate sequence, selecting at least one longest increasing sequence from the candidate sequence according to the order of the second identification information of the search sub-result in the candidate sequence; determining the at least one target sequence according to the at least one longest increasing sequence.

[0048] When obtaining the at least one target sequence according to the at least one longest increasing sequence in S104, the embodiment can further include the following: for each longest increasing sequence, obtaining a difference value between the second identification information of adjacent search sub-sequences in the longest increasing sub-sequence; in a case where it is determined that the obtained difference value does not exceed a preset interval threshold, taking the longest increasing sub-sequence as a target sequence.

[0049] That is, the embodiment obtains the target sequence through the second identification information of the search sub-result in the candidate sequence, ensures that the search sub-sequences in the target sequence are arranged in order according to the second identification information, and improves the accuracy of the obtained target sequence.

[0050] For example, if the candidate sequence is (sub-result 64-sub-result 65-sub-result 67-sub-result 76-sub-result 69-sub-result 70-sub-result 120-sub-result 75), the number after the sub-result represents the second identification information of the sub-result; in combination with the preset interval threshold (for example, 6), the embodiment performs S104 to determine the target sequence as (sub-result 64-sub-result 65-sub-result 67-sub-result 69-sub-result 70-sub-result 75); since the difference between the sub-result 70 and the sub-result 120 in the longest increasing sequence (sub-result 64-sub-result 65-sub-result 67-sub-result 69-sub-result 70-sub-result 120) exceeds the preset interval threshold, the longest increasing sequence is not used as the target sequence.

[0051] After the embodiment performs S104 to determine at least one target sequence from the at least one candidate sequence, the embodiment performs S105 to use the search result corresponding to the at least one target sequence as the recall result of the to-be-processed multimedia data.

[0052] When the embodiment performs S105 to use the search result corresponding to the at least one target sequence as the recall result of the to-be-processed multimedia data, the optional implementation manner that can be used by the embodiment is: for each target sequence, the number of search sub-results in the target sequence is obtained; in a case where it is determined that the obtained number exceeds the preset number threshold, the search result corresponding to the target sequence is used as the recall result of the to-be-processed multimedia data.

[0053] That is, the embodiment can determine whether the target sequence is available by using the number of search sub-results contained in the target sequence, and then in a case where the target sequence contains more search sub-results, the recall result of the to-be-processed multimedia data is obtained according to the target sequence, which can improve the accuracy of the obtained recall result.

[0054] When the embodiment performs S105, the search result corresponding to the first identification information of the search sub-result in the target sequence can be used as the recall result of the to-be-processed multimedia data.

[0055] After the embodiment performs S105 to obtain the recall result of the to-be-processed multimedia data, the obtained recall result can be returned to the input end to display the recall result on the input end.

[0056] Figure 2 is a schematic diagram according to the second embodiment of the present disclosure. Figure 2The flowchart of the recall method of the search result of the embodiment is shown in FIG. 1. After inputting the audio data, the audio data is split into multiple 2s of sub-data; each 2s of sub-data is searched (for example, searched using the Faiss system), to obtain Top N (N can be 5) search sub-results corresponding to each sub-data, and the obtained search sub-results contain first identification information, second identification information and a matching score; according to the first identification information of the search sub-results, the multiple search sub-results are divided into at least one candidate sequence, and the first identification information of the search sub-results contained in each candidate sequence is the same; for each candidate sequence, the search results contained in the candidate sequence are filtered according to the matching score; after the filtering is completed, according to the second identification information of the search sub-results, the longest increasing sub-sequence is selected from the candidate sequence as a target sequence; and finally, according to the number of search sub-results contained in the target sequence, the recall result of the audio data is obtained.

[0057] Figure 3 is a schematic diagram according to the third embodiment of the present disclosure. Figure 3 The schematic diagram of the multiple search sub-results corresponding to different sub-data obtained by the embodiment is shown in FIG. 1: Figure 3 A circle in FIG. 1 represents a search sub-result, and the search sub-results in different rows correspond to different sub-data (sub-data 1 to sub-data 10), and the search sub-results in different columns correspond to different Top levels (the first column is Top1, the second column is Top2, and the third column is Top3); different gray values of the circles represent different first identification information, for example, the Top1 circle with the largest gray value in the first row can represent the first identification information 1, the Top2 circle with the second largest gray value represents the first identification information 2, and the Top3 circle with the smallest gray value represents the first identification information 3; the numbers in each circle represent the second identification information, for example, the circle in the second column of the first row represents the search sub-result (audio clip) with the frame ID of 64 in the search result (audio) corresponding to the first identification information 1. In the prior art, only the first column in FIG. 1 is usually obtained, and then the recall result is obtained through one column of search sub-results, resulting in a low recall rate of the search result, while the embodiment can effectively improve the recall rate of the search result by additionally obtaining the second column and the third column. Figure 3

[0058] Figure 4 is a schematic diagram according to the fourth embodiment of the present disclosure. As shown in Figure 4 The recall device 400 of the search result of the embodiment includes:

[0059] The splitting unit 401 is configured to obtain to-be-processed multimedia data and split the to-be-processed multimedia data into multiple sub-data;

[0060] ​The retrieval unit 402 obtains a plurality of search sub-results corresponding to each sub-data according to the data characteristics of each sub-data, and each search sub-result contains the first identification information and the second identification information;

[0061] The first processing unit 403 divides the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-result and the order of the plurality of sub-data;

[0062] The second processing unit 404 obtains at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-result;

[0063] The recall unit 405 takes the search result corresponding to the at least one target sequence as the recall result of the multimedia data to be processed.

[0064] When the split unit 401 obtains the multimedia data to be processed, the multimedia data input by the input end can be taken as the multimedia data to be processed, or the multimedia data selected from the network by the input end can be taken as the multimedia data to be processed. The multimedia data to be processed obtained by the split unit 401 can be audio data, image data, video data, etc.

[0065] In order to improve the rationality of splitting the multimedia data to be processed, when the split unit 401 splits the multimedia data to be processed into a plurality of sub-data, an optional implementation manner that can be adopted is as follows: determining the data type of the multimedia data to be processed; and splitting the multimedia data to be processed into a plurality of sub-data according to the splitting mode corresponding to the determined data type.

[0066] In order to improve the recall rate of the search result, when the split unit 401 splits the multimedia data to be processed into a plurality of sub-data, a preset overlap rate can be further obtained, and then the multimedia data to be processed is split into a plurality of sub-data in combination with the obtained preset overlap rate.

[0067] After the split unit 401 splits the multimedia data to be processed into a plurality of sub-data, the retrieval unit 402 obtains a plurality of search sub-results corresponding to each sub-data according to the data characteristics of each sub-data.

[0068] When the retrieval unit 402 obtains a plurality of search sub-results corresponding to each sub-data according to the data characteristics of each sub-data, for each sub-data, the data characteristics (such as image characteristics, audio characteristics, etc.) of the sub-data are matched with the data characteristics of the search sub-results contained in the database, and then the search sub- results with the highest matching scores are taken as the plurality of search sub-results corresponding to each sub-data.

[0069] That is, the searching unit 402 obtains the Top N search sub-results corresponding to each sub-data, which expands the number of search sub-results corresponding to the sub-data, and in the case of increasing the number of obtained search sub-results, the recall rate of search results in the database can be improved.

[0070] It can be understood that the search sub-results pre-stored in the database in the embodiment are also obtained by splitting the corresponding search results according to the preset size or the preset time length, and the search sub-results have the same time length or the same size as the sub-data, thereby improving the accuracy of the obtained search sub-results corresponding to the sub-data.

[0071] In the search sub-results obtained by the searching unit 402, in addition to the image data of the preset size, the audio data of the preset time length or the video data, the first identification information and the second identification information can also be included.

[0072] The first identification information is the identification information of the search result corresponding to the search sub-result, and the search sub-results having the same first identification information correspond to the same search result; and the second identification information is the identification information of the search sub-result in the search result corresponding thereto, for example, the position information of the image search sub-result in the image search result corresponding thereto, or the time sequence information (for example, frame ID) of the audio search sub-result in the audio search result corresponding thereto.

[0073] After the searching unit 402 obtains the plurality of search sub-results corresponding to each sub-data, the first processing unit 403 divides the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-result and the order of the plurality of sub-data.

[0074] That is, the first processing unit 403 combines the order of the sub-data, the corresponding relationship between the sub-data and the search sub-result, and the first identification information of the search sub-result to obtain at least one candidate sequence, so that different candidate sequences correspond to different first identification information, that is, the search sub- results included in one candidate sequence correspond to only one search result.

[0075] When the first processing unit 403 divides the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence according to the first identification information of the search sub-result and the order of the plurality of sub-data, an optional implementation manner that can be adopted is: for each first identification information, the search sub-result corresponding to the first identification information is selected from the plurality of search results corresponding to different sub-data respectively; and the selected search sub- results are arranged according to the order of the plurality of sub-data to obtain the candidate sequence corresponding to the first identification information.

[0076] That is, the first processing unit 403 selects the search sub-result corresponding to the first identification information from the search sub-results corresponding to different sub-data through the first identification information of the search result, so that each candidate sequence finally obtained corresponds to different first identification information. Due to the expansion of the number of search sub-results, the number of search sub-results contained in the obtained candidate sequence can be improved, and the recall rate of the search result is improved.

[0077] It can be understood that among the multiple search sub-results corresponding to the same sub-data, if there are multiple search sub-results with the same first identification information, different search sub-results with the same first identification information usually have the same second identification information, and selecting one arbitrarily will not affect the accuracy of the obtained candidate sequence.

[0078] Therefore, when the first processing unit 403 selects the search sub-result corresponding to the first identification information from the multiple search results corresponding to different sub-data for each first identification information, it can also include the following content: for each sub-data and the multiple search sub-results corresponding thereto, if it is determined that the number of search sub-results corresponding to the first identification information is multiple, one is selected from the multiple search sub-results as the search sub-result corresponding to the first identification information of the sub-data.

[0079] That is, the first processing unit 403 can ensure that there is no repeated search sub-result in the candidate sequence corresponding to different first identification information, improve the accuracy of the obtained candidate sequence, and in the case that there is no repeated candidate sub-result in a candidate sequence, the accuracy of the subsequently obtained target sequence can be improved.

[0080] After obtaining at least one candidate sequence, the first processing unit 403 can also include the following content: for each candidate sequence, obtaining the matching score of each search sub-result in the candidate sequence, the matching score being calculated when obtaining the search sub-result according to the data characteristics; removing the search sub-result in the candidate sequence whose matching score is lower than a preset score threshold.

[0081] That is, the first processing unit 403 filters the search sub-result in the candidate sequence through the matching score, so that the search sub-result in the finally obtained candidate sequence has a higher matching score, and the accuracy of the determined target sequence is improved.

[0082] After obtaining at least one candidate sequence by the first processing unit 403, the second processing unit 404 obtains at least one target sequence from the at least one candidate sequence according to the second identification information of the search sub-result; wherein the target sequence obtained by the second processing unit 404 can be the candidate sequence itself, or a sub-sequence in the candidate sequence.

[0083] When the second processing unit 404 obtains at least one target sequence from at least one candidate sequence based on the second identifier information of the search sub-result, the optional implementation method may be as follows: for each candidate sequence, select at least one longest increasing sequence from the candidate sequence according to the order of the second identifier information of the search sub-result in the candidate sequence; determine at least one target sequence based on the at least one longest increasing sequence.

[0084] When the second processing unit 404 obtains at least one target sequence based on at least one longest increasing sequence, it may further include the following: for each longest increasing sequence, obtaining the difference between the second identification information of adjacent search sub-results in the longest increasing sub-sequence; if it is determined that the obtained difference does not exceed a preset interval threshold, the longest increasing sub-sequence is taken as the target sequence.

[0085] In other words, the second processing unit 404 obtains the target sequence through the second identifier information of the search sub-results in the candidate sequence, ensuring that the search sub-results in the target sequence are arranged sequentially according to the order of the second identifier information, thereby improving the accuracy of the obtained target sequence.

[0086] In this embodiment, after the second processing unit 404 determines at least one target sequence from at least one candidate sequence, the recall unit 405 uses the search results corresponding to the at least one target sequence as the recall result of the multimedia data to be processed.

[0087] When the recall unit 405 uses the search results corresponding to at least one target sequence as the recall result of the multimedia data to be processed, the optional implementation method may be as follows: for each target sequence, obtain the number of search sub-results in the target sequence; if it is determined that the obtained number exceeds a preset number threshold, use the search results corresponding to the target sequence as the recall result of the multimedia data to be processed.

[0088] In other words, the recall unit 405 can determine whether the target sequence is available by the number of search sub-results contained in the target sequence. Then, when the target sequence contains a large number of search sub-results, the recall result of the multimedia data to be processed can be obtained based on the target sequence, which can improve the accuracy of the obtained recall result.

[0089] The recall unit 405 can use the search results corresponding to the first identifier information of the search sub-results in the target sequence as the recall results of the multimedia data to be processed.

[0090] After obtaining the recall result of the multimedia data to be processed, the recall unit 405 can return the obtained recall result to the input terminal for display on the input terminal.

[0091] In the technical solutions of the present disclosure, the acquisition, storage and application of user personal information comply with relevant laws and regulations and do not violate public order and good customs.

[0092] According to the embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0093] As Figure 5 shown, is a block diagram of an electronic device for a search result recall method according to an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit implementations of the present disclosure described and / or claimed in this document.

[0094] As Figure 5 shown, the device 500 includes a computing unit 501 that can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded into a random access memory (RAM) 503 from a storage unit 508. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0095] Various components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunications networks.

[0096] The computing unit 501 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 501 performs various methods and processes described above, such as the recall method of search results. For example, in some embodiments, the recall method of search results can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 508.

[0097] In some embodiments, portions or all of the computer program can be loaded onto the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded onto the RAM 503 and executed by the computing unit 501, one or more steps of the recall method of search results described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the recall method of search results by any other appropriate means, such as by means of firmware.

[0098] Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0099] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable search result recall apparatuses to produce a machine, such that the program code, when executed by the processor or controller, causes the machine to perform functions / operations specified in the flow diagrams and / or block diagrams. The program code can execute entirely on a machine, partly on a machine, as a stand-alone software package, partly on a machine and partly on a remote machine or entirely on a remote machine or server.

[0100] In the context of this disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. 2

[0101] 2To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input. 2

[0102] 2The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet. 2

[0103] 2The computer system can include clients and servers. This relationship can be. remote, where each server is stored on a remote computer from a client. The clients and the servers can be connected through a communication network. The relationship can be a client-server relationship over a network. A server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of large management difficulty and weak business scalability in traditional physical hosts and VPS services (Virtual Private Server, or VPS for short). The server can also be a server of a distributed system, or a server combined with a blockchain.

[0104] It should be understood that the steps shown above can be reordered, added to, or deleted from. For example, the steps described in the present disclosure can be executed in parallel, in sequence, or in a different order, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present disclosure is not limited herein.

[0105] The above detailed description does not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for recalling search results, comprising: Acquire multimedia data to be processed, and split the multimedia data to be processed into multiple sub-data; Based on the data characteristics of each sub-data, multiple search sub-results corresponding to each sub-data are obtained, and each search sub-result contains first identification information and second identification information; Based on the first identifier information of the search sub-result and the order of the plurality of sub-data, the search sub-results corresponding to the plurality of sub-data are divided into at least one candidate sequence; Based on the second identifier information of the search sub-result, at least one target sequence is obtained from the at least one candidate sequence; The search results corresponding to the at least one target sequence are used as the recall results of the multimedia data to be processed.

2. The method according to claim 1, wherein, The step of splitting the multimedia data to be processed into multiple sub-data includes: Determine the data type of the multimedia data to be processed; The multimedia data to be processed is divided into multiple sub-data according to the splitting method corresponding to the data type.

3. The method according to claim 1, wherein, The step of dividing the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence based on the first identifier information of the search sub-results and the order of the plurality of sub-data includes: For each first identifier, select the search sub-result corresponding to that first identifier from multiple search results corresponding to different sub-data; The selected search sub-results are arranged according to the order of the multiple sub-data to obtain the candidate sequence corresponding to the first identifier information.

4. The method according to claim 1, further comprising: After obtaining the at least one candidate sequence, for each candidate sequence, obtain the matching score of each search sub-result in the candidate sequence; Remove search sub-results from the candidate sequence whose matching score is lower than a preset score threshold.

5. The method according to claim 1, wherein, Obtaining at least one target sequence from the at least one candidate sequence based on the second identifier information of the search sub-result includes: For each candidate sequence, at least one longest increasing sequence is selected from the candidate sequence according to the order of the second identifier information of the search sub-results in the candidate sequence; The at least one target sequence is obtained based on the at least one longest increasing sequence.

6. The method according to claim 5, wherein, The step of obtaining the at least one target sequence based on the at least one longest increasing sequence includes: For each longest increasing sequence, obtain the difference between the second identifier information of adjacent search sub-results in the longest increasing sequence; If the difference does not exceed a preset interval threshold, the longest increasing sequence is taken as the target sequence.

7. The method according to claim 1, wherein, The step of using the search results corresponding to the at least one target sequence as the recall result of the multimedia data to be processed includes: For each target sequence, obtain the number of search sub-results in that target sequence; If the number exceeds a preset threshold, the search results corresponding to the target sequence will be used as the recall results of the multimedia data to be processed.

8. The method according to claim 7, wherein, The step of using the search results corresponding to the target sequence as the recall result of the multimedia data to be processed includes: The search results corresponding to the first identifier information of the search sub-results in the target sequence are used as the recall results of the multimedia data to be processed.

9. A search result recall device, comprising: A splitting unit is used to acquire multimedia data to be processed and split the multimedia data to be processed into multiple sub-data; The retrieval unit is used to obtain multiple search sub-results corresponding to each sub-data based on the data characteristics of each sub-data. Each search sub-result contains first identification information and second identification information. The first processing unit is configured to divide the search sub-results corresponding to the plurality of sub-data into at least one candidate sequence based on the first identification information of the search sub-results and the order of the plurality of sub-data; The second processing unit is used to obtain at least one target sequence from the at least one candidate sequence based on the second identification information of the search sub-result; The recall unit is used to take the search results corresponding to the at least one target sequence as the recall result of the multimedia data to be processed.

10. The apparatus according to claim 9, wherein, When the splitting unit splits the multimedia data to be processed into multiple sub-data, it specifically performs the following: Determine the data type of the multimedia data to be processed; The multimedia data to be processed is divided into multiple sub-data according to the splitting method corresponding to the data type.

11. The apparatus according to claim 9, wherein, When the first processing unit divides the search sub-results corresponding to the multiple sub-data into at least one candidate sequence based on the first identifier information of the search sub-results and the order of the multiple sub-data, it specifically performs the following: For each first identifier, select the search sub-result corresponding to that first identifier from multiple search results corresponding to different sub-data; The selected search sub-results are arranged according to the order of the multiple sub-data to obtain the candidate sequence corresponding to the first identifier information.

12. The apparatus of claim 9, wherein the first processing unit is further configured to perform: After obtaining the at least one candidate sequence, for each candidate sequence, obtain the matching score of each search sub-result in the candidate sequence; Remove search sub-results from the candidate sequence whose matching score is lower than a preset score threshold.

13. The apparatus according to claim 9, wherein, When the second processing unit obtains at least one target sequence from the at least one candidate sequence based on the second identifier information of the search sub-result, it specifically performs the following: For each candidate sequence, at least one longest increasing sequence is selected from the candidate sequence according to the order of the second identifier information of the search sub-results in the candidate sequence; The at least one target sequence is obtained based on the at least one longest increasing sequence.

14. The apparatus according to claim 13, wherein, When the second processing unit obtains the at least one target sequence based on the at least one longest increasing sequence, it specifically performs the following: For each longest increasing sequence, obtain the difference between the second identifier information of adjacent search sub-results in the longest increasing sequence; If the difference does not exceed a preset interval threshold, the longest increasing sequence is taken as the target sequence.

15. The apparatus according to claim 9, wherein, When the recall unit uses the search results corresponding to the at least one target sequence as the recall result of the multimedia data to be processed, it specifically performs the following: For each target sequence, obtain the number of search sub-results in that target sequence; If the number exceeds a preset threshold, the search results corresponding to the target sequence will be used as the recall results of the multimedia data to be processed.

16. The apparatus according to claim 15, wherein, When the recall unit uses the search results corresponding to the target sequence as the recall result of the multimedia data to be processed, it specifically performs the following: The search results corresponding to the first identifier information of the search sub-results in the target sequence are used as the recall results of the multimedia data to be processed.

17. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-8.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-8.

19. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-8.

Images