Song detection method and device, electronic equipment and computer program product

Abstract

The embodiment of the invention provides a song detection method and device, electronic equipment and a computer program product, and relates to the technical field of mode recognition. The method comprises the following steps: determining songs to be detected, wherein the songs to be detected comprise one or more of song library change songs, platform import songs and appointed circled songs; obtaining a pre-constructed original song library, wherein the original song library is constructed based on original songs of the music platform; performing similarity detection on the to-be-detected song and the original song to obtain a similarity detection result; and determining a song detection result of the to-be-detected song according to a pre-configured song detection reference condition and the similarity detection result. The invention provides an implementation scheme for automatically selecting the to-be-detected song and the original reference song and automatically pushing the song to complete imitation writing detection, so that the efficiency of an imitation writing song washing detection process is improved, a large number of song detection tasks can be quickly responded, and meanwhile, the accuracy of a song detection result is improved.

Description

Background Art

[0002] This section aims to provide background or context for the embodiments of the present disclosure stated in the claims. The description herein is not admitted to be prior art merely by virtue of being included in this section.

[0003] In the traditional song detection process, a batch of songs suspected of songwashing are usually exported from various channels. These songs may be due to user reports, or abnormal situations shown by the data analysis results of music platforms, such as abnormal fluctuations in the playback volume within a short period of time, or a large number of doubts about suspected songwashing in the comment area; or they may be included in the suspect list by music professionals based on their own experience and keen hearing.

[0004] After exporting the songs suspected of songwashing, the audio features and lyric features of the songs are first extracted through a simple algorithm, and the similarity is matched with the songs in the song protection library (referring to the song benchmark library that needs to be matched). If the similarity exceeds the preset threshold, there may be a suspicion of songwashing, and then it enters the manual review stage. Summary of the Invention

[0005] However, using manual review for songwashing detection in songs is time-consuming and laborious, the review results depend on the subjectivity of the reviewers, and it is difficult to handle large-scale data.

[0006] For this reason, the present disclosure proposes an improved song detection method, providing a solution for automatically selecting the songs to be detected and the original benchmark songs, and automatically pushing the songs to complete the imitation detection, improving the efficiency of the song imitation and songwashing detection process, being able to quickly respond to a large number of song detection tasks, and at the same time improving the accuracy of the song detection results based on the song detection reference conditions.

[0007] In this context, the embodiments of the present disclosure are expected to provide a song detection method, a song detection device, a computer-readable storage medium, an electronic device, and a computer program product.

[0008] In the first aspect of the embodiments of the present disclosure, a song detection method is provided, including: determining the song to be detected, where the song to be detected includes one or more of the songs with changed song libraries, the songs imported from the platform, and the specified selected songs; obtaining the pre-constructed original song library, where the original song library is constructed based on the original songs of the music platform; performing a similarity detection on the song to be detected and the original song to obtain a similarity detection result; and determining the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result.

[0009] In one embodiment of the present disclosure, the method for determining the song to be detected includes one or a combination of more than one of the following methods: obtaining a song library to be detected, where the song library to be detected includes initial detection songs; determining the song to be detected from the initial detection songs based on a song selection rule; using the song that has changed in the song library to be detected as the song to be detected; and using the song imported by the business personnel based on the song detection platform as the song to be detected.

[0010] In one embodiment of the present disclosure, obtaining a pre-constructed original song library includes: obtaining specified songs from a first music platform based on a preset time period to obtain first original songs, where the first original songs include the songs included in the song ranking list in the first music platform and / or the songs whose song evaluation indicators reach the specified indicator threshold; using the songs obtained from a second music platform as second original songs; constructing an original song library based on the first original songs and the second original songs, where the original song library includes an audio reference library and a lyrics reference library.

[0011] In one embodiment of the present disclosure, using the songs obtained from a second music platform as second original songs includes: using the songs currently included in the second music platform as initial songs; obtaining the songs imported through the second music platform as newly imported songs; obtaining a pre-configured song filtering condition; and performing song filtering processing on the initial songs and the newly imported songs based on the song filtering condition to obtain second original songs.

[0012] In one embodiment of the present disclosure, constructing an original song library based on the first original songs and the second original songs includes: performing a summarization process on the first original songs and the second original songs to determine initial incremental songs; determining target incremental songs and songs to be deleted based on the current songs in the original song library and the initial incremental songs; updating the audio reference library according to the target incremental songs and the songs to be deleted; obtaining original lyrics data based on a preset time period and updating the lyrics reference library according to the original lyrics data.

[0013] In one embodiment of the present disclosure, updating the audio reference library according to the target incremental songs and the songs to be deleted includes: determining the incremental song audio vectors corresponding to the target incremental songs, and adding the incremental song audio vectors to the audio reference library; deleting the to-be-deleted song audio vectors corresponding to the songs to be deleted from the audio reference library to update the audio reference library, where the audio reference library includes the original song audio vectors corresponding to the updated original songs.

[0014] In one embodiment of the present disclosure, similarity detection is performed on the song to be detected and the original song to obtain a similarity detection result, including: determining the basic song information corresponding to the song to be detected, where the basic song information includes the song version and the language type of the lyrics; determining the song detection type according to the basic song information, where the song detection type includes an audio detection type and a lyrics detection type; when the song detection type is the audio detection type, determining the audio vector of the song to be detected corresponding to the song to be detected; performing audio similarity detection processing based on the audio vector of the song to be detected and the audio vector of the original song corresponding to the original song to obtain a similarity detection result; when the song detection type is the lyrics detection type, performing audio similarity detection and lyrics similarity detection processing on the song to be detected to obtain a similarity detection result.

[0015] In one embodiment of the present disclosure, according to the pre-configured song detection reference conditions and the similarity detection result, the song detection result of the song to be detected is determined, including: obtaining the similarity detection result, where the similarity detection result includes the similarity feature value corresponding to the song similarity detection feature of the song to be detected; obtaining the pre-configured song detection executor, where the song detection executor includes an initial detection condition list composed of song detection reference conditions; sorting the initial detection condition list based on the detection priority to obtain a sorted detection condition list; and matching the similarity detection result with the song detection reference conditions included in the sorted detection condition list one by one to obtain the song detection result.

[0016] In one embodiment of the present disclosure, obtaining the pre-configured song detection executor includes: obtaining an initial rule executor, where the initial rule executor is a stateless executor; determining the similarity feature reference threshold corresponding to the song similarity detection feature, and generating song detection reference conditions based on one or more similarity feature reference thresholds; constructing an initial detection condition list according to at least one song detection reference condition; and configuring the initial rule executor based on the initial detection condition list to generate a song detection executor, where the song detection executor is used to perform song detection processing according to the initial detection condition list.

[0017] In one embodiment of the present disclosure, matching the similarity detection result with the song detection reference conditions included in the sorted detection condition list one by one to obtain the song detection result includes: obtaining the current detection condition from the sorted detection condition list, and determining the detection condition information corresponding to the current detection condition, where the detection condition information includes one or more of the detection index type, the detection index field, and the index field value; performing a matching process on the similarity detection result and the current detection condition based on the detection condition information to obtain a condition matching result; and determining the song detection result according to the obtained condition matching result.

[0018] In a second aspect of the embodiments of the present disclosure, a song detection device is provided, including: a song determination module for determining a song to be detected, where the song to be detected includes one or more of a song with a changed song library, a song imported from a platform, and a specified selected song; an original library construction module for obtaining a pre-constructed original song library, where the original song library is constructed based on the original songs of a music platform; a similarity detection module for performing a similarity detection on the song to be detected and the original song to obtain a similarity detection result; and a detection result determination module for determining the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result.

[0019] In an embodiment of the present disclosure, the song determination module includes a song determination unit for determining the song to be detected by one or more combinations of the following methods: obtaining a song library to be detected, where the song library to be detected includes initial detection songs; determining the song to be detected from the initial detection songs based on a song selection rule; using the song with changes in the song library to be detected as the song to be detected; and using the song imported by a businessperson based on a song detection platform as the song to be detected.

[0020] In an embodiment of the present disclosure, the original library construction module includes an original library construction unit for: obtaining specified songs from a first music platform based on a preset time period to obtain first original songs, where the first original songs include the songs included in the song ranking list in the first music platform and / or the songs whose song evaluation indicators reach a specified index threshold; using the songs obtained from a second music platform as second original songs; and constructing an original song library based on the first original songs and the second original songs, where the original song library includes an audio reference library and a lyrics reference library.

[0021] In an embodiment of the present disclosure, the original library construction unit includes an original song acquisition subunit for: using the songs currently included in the second music platform as initial songs; obtaining the songs imported through the second music platform as newly imported songs; obtaining a pre-configured song filtering condition; and performing a song filtering process on the initial songs and the newly imported songs based on the song filtering condition to obtain second original songs.

[0022] In an embodiment of the present disclosure, the original library construction unit includes an original library construction subunit for: performing a summary process on the first original songs and the second original songs to determine initial incremental songs; determining target incremental songs and songs to be deleted based on the current songs in the original song library and the initial incremental songs; updating the audio reference library according to the target incremental songs and the songs to be deleted; obtaining original lyrics data based on a preset time period and updating the lyrics reference library according to the original lyrics data.

[0023] In one embodiment of the present disclosure, the original library construction subunit includes a song vector update subunit, which is used to: determine an incremental song audio vector corresponding to a target incremental song, and add the incremental song audio vector to the audio reference library; delete a to-be-deleted song audio vector corresponding to a to-be-deleted song from the audio reference library to update the audio reference library, where the audio reference library includes original song audio vectors corresponding to updated original songs.

[0024] In one embodiment of the present disclosure, the similarity detection module includes a similarity detection unit, which is used to: determine song basic information corresponding to a song to be detected, where the song basic information includes a song version and a song lyric language; determine a song detection type according to the song basic information, where the song detection type includes an audio detection type and a lyric detection type; when the song detection type is the audio detection type, determine a to-be-detected song audio vector corresponding to the song to be detected; perform an audio similarity detection process based on the to-be-detected song audio vector and an original song audio vector corresponding to the original song to obtain a similarity detection result; when the song detection type is the lyric detection type, perform an audio similarity detection and a lyric similarity detection process on the song to be detected to obtain a similarity detection result.

[0025] In one embodiment of the present disclosure, the detection result determination module includes a detection result determination unit, which is used to: obtain a similarity detection result, where the similarity detection result includes a similarity feature value corresponding to a song similarity detection feature of the song to be detected; obtain a pre-configured song detection executor, where the song detection executor includes an initial detection condition list composed of song detection reference conditions; sort the initial detection condition list based on a detection priority to obtain a sorted detection condition list; match the similarity detection result with the song detection reference conditions included in the sorted detection condition list one by one to obtain a song detection result.

[0026] In one embodiment of the present disclosure, the detection result determination unit includes a detector configuration subunit, which is used to, including: obtain an initial rule executor, where the initial rule executor is a stateless executor; determine a similarity feature reference threshold corresponding to a song similarity detection feature, and generate a song detection reference condition based on one or more similarity feature reference thresholds; construct an initial detection condition list according to at least one song detection reference condition; configure the initial rule executor based on the initial detection condition list to generate a song detection executor, where the song detection executor is used to perform a song detection process according to the initial detection condition list.

[0027] In one embodiment of the present disclosure, the detection result determination unit includes a detection result determination subunit for: obtaining the current detection condition from the sorted detection condition list, determining the detection condition information corresponding to the current detection condition, where the detection condition information includes one or more of the detection index type, detection index field, and index field value; based on the detection condition information, performing a matching process on the similarity detection result and the current detection condition to obtain a condition matching result; and determining the song detection result according to the obtained condition matching result.

[0028] In the third aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the song detection method as described above is implemented.

[0029] In the fourth aspect of the embodiments of the present disclosure, an electronic device is provided, including: a processor; and a memory, on which computer-readable instructions are stored, and when the computer-readable instructions are executed by the processor, the song detection method as described above is implemented.

[0030] According to the fifth aspect of the embodiments of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the song detection method as described in any one of the above is implemented.

[0031] According to the technical solution of the embodiments of the present disclosure, on the one hand, an implementation solution for detecting a song from the selection of a song to be detected and an original reference song to automatically push a song for imitation and song washing detection is provided, which improves the efficiency of song detection and can handle large-scale song detection tasks. On the other hand, during the song detection process, the song detection reference conditions can be quickly responded and adapted to obtain the song detection result, which improves the accuracy of the song detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present disclosure will become readily understood. In the drawings, several embodiments of the present disclosure are shown by way of illustration and not limitation, where:

[0033] Figure 1 A schematic block diagram of the system architecture of an exemplary application scenario according to some embodiments of the present disclosure is schematically shown;

[0034] Figure 2 A schematic flowchart of the song detection method according to some embodiments of the present disclosure is schematically shown;

[0035] Figure 3 A schematic overall flowchart of the song imitation and song washing detection according to some embodiments of the present disclosure is schematically shown;

[0036] Figure 4 Schematically shows a flowchart of selecting songs to be detected based on a submission library and performing song detection according to some embodiments of the present disclosure;

[0037] Figure 5 Schematically shows a flowchart of establishing an original song library in a song detection solution according to some embodiments of the present disclosure;

[0038] Figure 6 Schematically shows a flowchart of performing song detection based on a song detection actuator according to some embodiments of the present disclosure;

[0039] Figure 7 Schematically shows a schematic block diagram of a song detection device according to some embodiments of the present disclosure;

[0040] Figure 8 Schematically shows a schematic diagram of a storage medium according to an exemplary embodiment of the present disclosure;

[0041] Figure 9 Schematically shows a block diagram of an electronic device according to an exemplary embodiment of the invention.

[0042] In the drawings, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed embodiments

[0043] The principles and spirit of the present disclosure will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and then implement the present disclosure, and do not limit the scope of the present disclosure in any way. On the contrary, these embodiments are provided to make the present disclosure more thorough and complete, and to be able to convey the scope of the present disclosure fully to those skilled in the art.

[0044] Those skilled in the art know that the embodiments of the present disclosure can be implemented as a system, a device, an apparatus, a method, or a computer program product. Therefore, the present disclosure can be specifically implemented in the following forms, namely: completely hardware, completely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0045] According to the embodiments of the present disclosure, a song detection method, a song detection device, a computer-readable storage medium, an electronic device, and a computer program product are provided.

[0046] In this article, it is important to understand that the terms involved, such as song imitation, are a way of creating something based on an existing song, usually by retaining some elements of the original song, such as melody, rhythm or harmony, and then recreating other elements to form a new song. Song imitation may include but is not limited to melody imitation, rhythm imitation, lyrics imitation, etc. Song plagiarism refers to the presence of the same clips in the recording or lyrics as the original song. Plagiarism includes but is not limited to the same lyrics or the same melody in the main chorus section, the same accompaniment, etc. For example: cover, cover, partial sampling or the same iconic melody and lyrics of the chorus.

[0047] Low imitation / washed songs include the following six situations: 1) The song title is the same, including the situation where the lyrics are used as the song title. For example, add "official version, complete version, official version, DJ version, female version, male version, gentle version, chorus version, cute version, new version and other version-type labels" or use memorable lyrics (mostly lyrics fragments spread on short video platforms) directly as the song title to confuse users to misunderstand the song. 2) The first type of song title is the same + similar singer name. The original singer (or artist) name is deformed in various ways that are easy to confuse users; such as: replacing a certain character in the singer's name with a similar character, such as replacing a certain character in the singer's name with other characters with the same pronunciation but different characters; or replacing a certain character in the singer's name with other characters with similar meanings, etc. 3) The lyrics are similar in content, and the chorus or memorable lyrics are similar and appear multiple times. 4) The melody, accompaniment, and structure of the important memorable part of the song are the same or similar, and the lyrics of this part are also the same or similar in meaning and rhythm, which comprehensively presents a confusing effect. 5) The original melody is varied by transposition, imitation, embellishment, contraction, etc. 6) At least two of the rhythmic patterns, time values, and ornaments are the same or similar.

[0048] The above are the identification standards for song wash from the perspectives of song title, artist, lyrics, music, melody, arrangement, etc. The above situations 3) and 4) alone meet the standards, and situations 1), 2), 5) and 6) are additional standards. In the absence of situation 3) or 4), song wash cannot be established alone, but it may be suspected of riding on the popularity of the song.

[0049] Transfer includes the following two situations: 1) copying the audio of the original song; 2) making non-original adaptations at the mixing level based on the original audio, adjusting the audio amplitude, frequency, and phase, such as using equalizers (EQ) and distortion effects (Distortion).

[0050] In addition, any number of elements in the drawings is for illustration and not limitation, and any naming is only for distinction and does not have any limiting meaning.

[0051] The principle and spirit of the present disclosure are explained in detail below with reference to several representative embodiments of the present disclosure. Summary of the Invention

[0053] In the manual review for the song-copying detection solution, when professional music reviewers conduct manual reviews, by leveraging their professional knowledge and subjective judgment, they repeatedly listen to the songs suspected of song-copying and the original songs, comparing the melody trends, note combinations, repeated sections, etc. There may be a suspicion of song-copying if the melodies in the verse part are similar or there are consecutive identical measures; pay attention to the rhythm patterns, including the strength rules of drum beats and the beat changes. A high degree of similarity will raise alarm; analyze the harmony, including chord selection and conversion methods. Similarities require further review; review the lyrics content, themes, and expression methods. Even if the melodies are different but the lyrics are similar, it may also be plagiarism; finally, comprehensively consider the overall styles such as music types, singing styles, and arrangement features. If there are similarities in multiple aspects, the possibility of song-copying is high.

[0054] However, the above-mentioned manual review solution has the following problems: 1) Time-consuming and labor-intensive. Manual review requires a large amount of time and manpower input. Reviewers need to listen to songs one by one and conduct detailed analysis and comparison. Especially when the number of songs suspected of song-copying is large, the workload is very heavy. 2) Strong subjectivity. The results of manual review are affected by the personal music literacy, subjective preferences, and judgment criteria of the reviewers. Different reviewers may draw different conclusions for the same set of songs, lacking a certain degree of objectivity. 3) Difficult to handle large-scale data. In today's environment of a vast number of music works, the traditional manual review method is difficult to handle the detection tasks of a large number of songs suspected of song-copying. Low efficiency may lead to some song-copying behaviors not being discovered and processed in a timely manner.

[0055] Based on the above content, the basic idea of the present disclosure is to determine the song to be detected, where the song to be detected includes one or more of the songs with changed song libraries, the songs imported by the platform, and the designated circled songs; obtain the pre-constructed original song library, which is constructed based on the original songs of the music platform; perform similarity detection on the song to be detected and the original songs to obtain the similarity detection result; and determine the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result. The present disclosure provides an implementation solution for automatically circling the song to be detected and the original reference song and automatically pushing the songs to complete the imitation detection, improving the efficiency of the song imitation and song-copying detection process, being able to quickly respond to a large number of song detection tasks, and at the same time improving the accuracy of the song detection result based on the song detection reference conditions.

[0056] After introducing the basic principle of the present disclosure, the following specifically introduces various non-limiting implementation manners of the present disclosure.

[0057] Overview of Application Scenarios

[0058] First, refer to Figure 1 , Figure 1A schematic block diagram of a system architecture showing an exemplary application scenario of a song detection method and apparatus to which embodiments of the present disclosure can be applied.

[0059] As Figure 1 shown, the system architecture 100 may include a first music platform 111, a second music platform 112, a server 120, and a terminal device 130. Obtain the song to be detected from the music library of the first music platform 111; select the song to be detected from the inspection library of the second music platform 112 according to specified rules, and the operator can also manually import the song to be detected through the second music platform 112. Transmit the song to be detected to the server 120 through the network. After the server 120 performs song detection on the song to be detected, transmit the song detection result to the terminal device 130 through the network. The network is used as a medium to provide a communication link between the terminal device 130 and the server 120. The network may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc. The terminal device 130 may be various electronic devices with a display screen, including but not limited to desktop computers, portable computers, smartphones, and tablet computers, etc. It should be understood that Figure 1 the numbers of the terminal device, network, and server in

[0060] are merely illustrative. According to the implementation requirements, there may be any number of terminal devices, networks, and servers. For example, the server 120 may be a server cluster composed of multiple servers, etc.

[0061] It should be understood that Figure 1 the shown application scenario is only an example in which the embodiments of the present disclosure can be implemented. The scope of application of the embodiments of the present disclosure is not limited by any aspect of this application scenario.

[0062] Exemplary Method

[0063] The following will, in conjunction with Figure 1 the application scenarios, refer to Figure 2 to describe the song detection method according to the exemplary embodiments of the present disclosure. It should be noted that the above application scenarios are only shown for the convenience of understanding the spirit and principle of the present disclosure, and the embodiments of the present disclosure are not limited in this regard. On the contrary, the embodiments of the present disclosure can be applied to any applicable scenario.

[0064] The present disclosure first provides a song detection method. The execution subject of this method can be a terminal device or a server, and the present disclosure does not make any special limitations in this regard. In this exemplary embodiment, the case where the server executes this method is taken as an example for illustration.

[0065] Referring to Figure 2 as shown, the song detection method may include the following steps S210 to S240:

[0066] Step S210, determine the song to be detected, where the song to be detected includes one or more of the songs with changed song libraries, platform-imported songs, and specified selected songs;

[0067] Step S220, obtain the pre-constructed original song library, where the original song library is constructed based on the original songs of the music platform;

[0068] Step S230, perform a similarity detection on the song to be detected and the original songs to obtain a similarity detection result;

[0069] Step S240, determine the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result.

[0070] In the song detection method provided by this exemplary embodiment, on the one hand, an implementation solution for selecting from the songs to be detected and the original reference songs to automatically push songs for imitation and song washing detection is provided, which improves the efficiency of song detection and can handle large-scale song detection tasks. On the other hand, during the song detection process, the song detection reference conditions can be quickly responded to and adapted to obtain the song detection result, which improves the accuracy of the song detection result.

[0071] Next, the above steps of this exemplary embodiment will be described in more detail.

[0072] In step S210, determine the song to be detected, where the song to be detected includes one or more of the songs with changed song libraries, platform-imported songs, and specified selected songs.

[0073] In some example embodiments, the song to be detected may be a song for which song copying detection is to be performed, and the song to be detected may be obtained from one or more music platforms. The song with library change may be a song that has changed in the song library. The platform-imported song may be a song manually imported by relevant business personnel (such as operation personnel) into the music platform. The designated selected song may be a song determined from the music platform based on the designated selection rule.

[0074] In the song detection scenario, the music platform may pre-construct a song submission library, which can be used to store the songs to be detected for song copying detection. Refer to Figure 3 , Figure 3 which schematically shows the overall flowchart of song imitation and song copying detection according to some embodiments of the present disclosure. Figure 3 The song submission library in [[ ]] can be constructed from the songs selected from the music platform by adopting the designated rule.

[0075] Determine the song to be detected based on the pre-constructed song submission library. The song to be detected may be a song selected from the song submission library based on the pre-configured song selection rule; the song to be detected may also be a song that has changed in the song submission library. For example, when the audio file or lyric data of a certain song in the song submission library changes, that song can be used as the song to be detected; the song to be detected may also be a song manually imported by relevant business personnel (such as operation personnel) into the song submission library of the music platform.

[0076] In step S220, obtain the pre-constructed original song library, which is constructed based on the original songs of the music platform.

[0077] In some example embodiments, the original song library is also called the song protection library, abbreviated as the protection library, and may be a song library constructed from the original songs in one or more music platforms. The music platform may be an Internet platform for providing functions such as music playback, download, and social interaction. The original song may be a reference song for performing similarity detection on the song to be detected.

[0078] During the song detection process, the original songs included in the original song library can be used as reference songs. Continue to refer to Figure 3 , Figure 3 the original song library in [[ ]] can be created based on the original songs in one or more music platforms. Perform similarity detection on the song to be detected based on the original songs in the original song library to determine whether the song to be detected has the suspicion of song copying.

[0079] In step S230, perform similarity detection on the song to be detected and the original song to obtain the similarity detection result.

[0080] In some example embodiments, the similarity detection result may be a detection result used to characterize the similarity between the song to be detected and the original song matched in the original song library.

[0081] Continuing to refer to Figure 3 , the determined song to be detected is automatically submitted for inspection, and the song to be detected is processed for similarity detection through an algorithm cluster. For example, the original song corresponding to the song to be detected is determined from the original song library, and then the similarity between the song to be detected and the original song is calculated. For example, it is calculated whether the performing singers of the song to be detected and the original song are the same, whether the song names are the same, and the audio similarity and lyric similarity between the two can also be judged. By calculating the similarity of multiple song similarity detection features, the similarity detection result is obtained.

[0082] In step S240, according to the pre-configured song detection reference conditions and the similarity detection result, the song detection result of the song to be detected is determined.

[0083] In some example embodiments, the song detection reference conditions may be detection reference conditions used to judge whether there are song-washing operations such as imitation, plagiarism, and transfer for the song to be detected. The song detection result may be a detection result for determining whether there is a suspicion of song imitation.

[0084] When obtaining the similarity detection result corresponding to the song to be detected, the similarity detection result usually consists of specific values corresponding to multiple song similarity detection features. The above-mentioned feature values returned by the machine are compared with the pre-configured song detection reference conditions, and finally the corresponding type label is obtained as the final song detection result.

[0085] In an embodiment of the present disclosure, for step S210, the determination method of the song to be detected includes one or more combinations of the following methods: obtaining a song library to be detected, where the song library to be detected includes initial detection songs; determining the song to be detected from the initial detection songs based on a song selection rule; using the song that has changed in the song library to be detected as the song to be detected; and using the song imported by the business personnel based on the song detection platform as the song to be detected.

[0086] Among them, the song library to be detected, also called the inspection library, may be a database containing songs to be detected, and the song library to be detected may contain one or more songs to be detected. The initial detection song may be a song initially included in the song library to be detected. The song selection rule may be a specific rule used to select the song to be detected from the initial detection songs.

[0087] Refer to Figure 4 , Figure 4A flowchart of selecting songs to be detected based on a submission library and performing song detection according to some embodiments of the present disclosure is schematically shown. Obtain a pre-constructed library of songs to be detected (i.e., the submission library), and use the song selection rules configured on the Rule-based Selection Platform (CIO Platform) to determine a batch of songs as the songs to be detected for submission. The song selection rules can be determined based on multiple dimensions such as song popularity, song release time, song style genre, performing singer, and song version.

[0088] Specifically, for the dimension of song popularity, songs with relatively high play counts on music platforms can be selected as the songs to be detected, such as songs with a certain number of play counts in the recent week or month. Such songs have a wide spread and a high degree of attention, and are more likely to have disputes such as songwashing that need to be detected. In addition, songs on various music charts on music platforms, which belong to current popular songs, can be used as the songs to be detected.

[0089] For the dimension of release time, recently newly released songs can be mainly selected, such as songs released within the recent month, as the songs to be detected. Songs released during specific music events, seasons, or events can also be selected, such as songs released during a certain large-scale music event or during specific festival seasons, as the songs to be detected.

[0090] For the dimension of style genre, currently popular music styles are selected, such as currently popular Electronic Dance Music (EDM), Rhythm and Blues (R&B or RnB, meaning rhythm and blues), Chinese national style, Jazz music, etc. Since there is a large demand from listeners and many creators, such songs need to be paid attention to. In addition, for songs with niche styles, songwashing detection can also be carried out on such songs.

[0091] For the dimension of performing singer, the performing songs of newly emerging singers or controversial singers are selected. Newly emerging singers may have a certain motivation to plagiarize or wash songs in order to quickly gain attention and popularity. Therefore, the works of newly debuted singers can be selected as the songs to be detected. For singers who have had bad records such as plagiarism or songwashing or other disputes, their newly released songs should be mainly selected for detection.

[0092] For the dimension of song copyright, songs with unclear copyright ownership, rumors of copyright disputes, or those already involved in copyright lawsuits should be key detected. For songs without clear copyright information or unknown sources, especially some songs that are widely spread on the Internet but the copyright ownership cannot be determined, they also need to be used as the songs to be detected for songwashing detection.

[0093] For the song library to be detected, if it is detected that a certain song has changed, such as changes in musical elements such as melody, rhythm, harmony, lyrics, etc., which can also include changes in singing styles, production techniques, etc., the songs with such changes are regarded as the songs to be detected; in addition, the songs that are taken off the shelves or have changes in the temporary copyright locking status of the songs from the specified channels in the song library can also be regarded as the songs to be detected and automatically sent for inspection.

[0094] In addition, the songs imported by relevant business personnel (such as operation personnel) through the imitation and songwashing detection background of the music platform are regarded as the songs to be detected. Through the above various methods, the range of songs to be sent for inspection can be automatically selected, and then it can be automatically pushed to the song detection task.

[0095] In an embodiment of the present disclosure, for step S220, obtain a pre-constructed original song library, including: obtaining specified songs from a first music platform based on a preset time period to obtain first original songs, where the first original songs include the songs included in the song ranking list in the first music platform and / or the songs whose song evaluation indicators reach the specified index threshold; regarding the songs obtained from the second music platform as second original songs; constructing an original song library based on the first original songs and the second original songs, and the original song library includes an audio reference library and a lyrics reference library.

[0096] Among them, the preset time period can be a pre-configured time period. For example, the preset time period can be 24 hours (hour, h), 12h, etc. The first music platform can be other music platforms except the second music platform, that is, an off-site music platform. The first original songs can be the original songs included in the first music platform. The song ranking list, also known as the song chart, is a music list used to display the currently most popular or most influential songs. The song evaluation indicators can be the indicators used to evaluate the comprehensive level of a song work, and the song evaluation indicators can include but are not limited to tune and melody, lyrics and theme, arrangement and performance skills, sound effects and production quality, innovation and originality, etc. The second music platform can be the current music platform that supports the songwashing detection function, that is, an in-site music platform. The second original songs can be the original songs in the second music platform. The audio reference library, also known as the audio protection library, can be a database composed of the song audios in the original song library. The lyrics reference library, also known as the lyrics protection library, can be a database composed of the original song lyrics in the original song library.

[0097] Reference Figure 5 , Figure 5A flowchart for establishing an original song library in a song detection solution according to some embodiments of the present disclosure is schematically shown. When establishing the original song library (i.e., the protection library), specified songs can be obtained from an off-site music platform at a preset time period (such as daily) as the first original songs. For example, the first original songs can include the songs included in various song charts in the first music platform. Song charts are usually compiled by institutions such as music radio stations, online music platforms, or professional music magazines, and are ranked according to multiple indicators such as the play volume, download volume, radio play times, and album sales of the songs.

[0098] The first original songs can also include the songs whose song evaluation indicators in the off-site music platform reach the specified indicator thresholds, such as the popular songs and soaring songs of each day. Through data communication technology, popular songs, soaring songs, and other songs that meet the song evaluation indicators are downloaded from the off-site music platform every day as the first original songs.

[0099] In addition, the songs obtained from the second music platform (i.e., the in-site music platform) are used as the second original songs. After obtaining the first original songs and the second original songs, the original song library can be constructed based on the first original songs and the second original songs. Since a song usually includes two parts: audio data and lyric data, the audio data corresponding to each obtained original song is stored in the audio reference library, and the lyric data is stored in the lyric reference library. The constructed original song library can include the audio reference library and the lyric reference library. Subsequently, the original songs in the original song library are used as the benchmark for song detection.

[0100] In an embodiment of the present disclosure, using the songs obtained from the second music platform as the second original songs includes: using the songs currently included in the second music platform as the initial songs; obtaining the songs imported through the second music platform as the newly imported songs; obtaining the pre-configured song filtering conditions; and performing song filtering processing on the initial songs and the newly imported songs based on the song filtering conditions to obtain the second original songs.

[0101] Among them, the initial songs can be the original songs that are already in the second music platform. The newly imported songs can be the newly added songs manually imported by relevant business personnel. The song filtering conditions can be the screening conditions for filtering out some non-compliant songs.

[0102] Continue to refer to Figure 5 , for the in-site music platform, first obtain the full amount of protection library data in the second music platform, and then select some songs using the song selection rule as the original songs determined from the protection library of the second music platform as the initial songs. In addition, the operation personnel can also manually import songs through the second music platform, and use the songs imported through the second music platform as the newly imported songs.

[0103] In order to improve the song quality of the original songs, song filtering conditions can be pre-configured to filter and screen the songs before storing them in the original song library. For example, the song filtering conditions can be to filter out songs with mismatched song status, or to filter out songs with mismatched song recording copyright status and other song status. By using the above song filtering conditions to perform song filtering processing on the initial songs and newly imported songs, the second original songs are obtained. By filtering the songs through the song filtering conditions, some non-compliant songs that do not meet the requirements of being original songs can be further eliminated, improving the accuracy of the song detection results.

[0104] In an embodiment of the present disclosure, constructing an original song library based on the first original song and the second original song includes: summarizing the first original song and the second original song to determine the initial incremental songs; determining the target incremental songs and the songs to be deleted based on the current songs in the original song library and the initial incremental songs; updating the audio reference library according to the target incremental songs and the songs to be deleted; obtaining the original lyric data based on a preset time period, and updating the lyric reference library according to the original lyric data.

[0105] Among them, the initial incremental songs can be the incremental songs directly determined after summarizing the first original song and the second original song. The target incremental songs can be the incremental songs determined after repeatedly comparing the initial incremental songs with the songs in the original song library. The songs to be deleted can be the songs that need to be deleted from the original song library. The original lyric data can be the lyric data corresponding to the original songs.

[0106] After determining the first original song and the second original song, summarize the incremental songs obtained from the in-station or out-station music platforms as the initial incremental songs. Then, according to a preset time period, such as obtaining the song data (i.e., the current songs) already in the original song library at a fixed time every day, and after determining the initial incremental songs, the full amount of song data that should be in the original song library. Compare the initial incremental songs with the current songs to determine the songs that actually need to be added to the original song library as the target incremental songs.

[0107] Since the song information of some songs in the original song library may change and they are no longer suitable to be used as the reference songs for song detection, in this case, such songs can be used as the songs to be deleted and deleted from the original song library. After determining the target incremental songs and the songs to be deleted, add the target incremental songs to the original song library, delete the songs to be deleted from the original song library, and update the original song library.

[0108] Continue to refer to Figure 5, the original song library can be composed of an audio reference library and a lyrics reference library. The audio reference library can be updated based on the audio data in the target incremental songs and the songs to be deleted. Additionally, the music platform can obtain the original lyrics data corresponding to the original songs at preset time intervals. For example, the music platform constructs the full lyrics data corresponding to all original songs daily, and updates the lyrics reference library based on the full lyrics data to obtain the original song library. For instance, constructing the full lyrics protection library data and pushing it to the remote file server, the algorithm node will automatically obtain the data in the lyrics protection library for lyrics detection. By constructing the audio reference library and the lyrics reference library respectively with the audio data and lyrics data of the original songs, a benchmark song library for matching can be constructed for subsequent song detection.

[0109] In an embodiment of the present disclosure, updating the audio reference library according to the target incremental songs and the songs to be deleted includes: determining the incremental song audio vector corresponding to the target incremental song, and adding the incremental song audio vector to the audio reference library; deleting the to-be-deleted song audio vector corresponding to the song to be deleted from the audio reference library to update the audio reference library, where the audio reference library includes the original song audio vectors corresponding to the updated original songs.

[0110] Among them, the incremental song audio vector can be the audio vector corresponding to the incremental song to be added to the original song library. The to-be-deleted song audio vector can be the audio vector corresponding to the song to be deleted from the original song library. The original song audio vector can be the audio vector corresponding to the original songs included in the original song library.

[0111] For the update operation of the audio protection library, the target incremental songs and the songs to be deleted can be determined by comparing the difference between the initial incremental songs and the current songs in the original song library; then determining the incremental song audio vector corresponding to the target incremental song, and adding the incremental song audio vector to the audio reference library; deleting the to-be-deleted song audio vector corresponding to the song to be deleted from the audio reference library to complete the update operation of the audio reference library. The updated audio reference library includes the original song audio vectors corresponding to the latest original songs. The updated audio reference library can be used as the data basis for subsequent wash song detection of the audio data of the song to be detected.

[0112] In one embodiment of the present disclosure, for step S230, a similarity detection is performed between the song to be detected and the original song to obtain a similarity detection result, including: determining basic song information corresponding to the song to be detected, the basic song information including the song version and the language of the lyrics; determining a song detection type based on the basic song information, the song detection type including an audio detection type and a lyrics detection type; when the song detection type is an audio detection type, determining an audio vector of the song to be detected corresponding to the song to be detected; performing audio similarity detection processing based on the audio vector of the song to be detected and the original song audio vector corresponding to the original song to obtain a similarity detection result; when the song detection type is a lyrics detection type, performing audio similarity detection and lyrics similarity detection processing on the song to be detected to obtain a similarity detection result.

[0113] Among them, the basic information of the song can be the basic information contained in the song to be detected. The song version can be a specific interpretation version of the song to be detected. The language of the lyrics can be the language corresponding to the lyrics of the song to be detected, such as the language of the lyrics can include various world languages ​​such as Chinese, English, Cantonese, etc. The song detection type can be the specific type of the song to be detected that needs to be detected. The audio detection type can be a type that only needs to perform audio similarity detection on the audio data of the song to be detected. The lyrics detection type can be a type that needs to perform audio similarity detection and song similarity detection on the song to be detected when the lyrics of the song to be detected change. The audio vector of the song to be detected can be a feature vector corresponding to the audio data of the song to be detected. The audio similarity detection process can be a process of performing similarity detection between the audio vector of the song to be detected and the audio vector of the matched original song.

[0114] After determining the song to be detected, the basic information of the song to be detected can be obtained. For example, the basic information of the song can include but is not limited to the song version and the language of the lyrics. The basic information of the song can be determined based on the characteristics of the song to be detected, and the audio feature vector and the lyrics feature vector corresponding to the song to be detected are obtained respectively. According to the characteristics of the song itself, it is automatically determined whether to use the audio detection type or the lyrics detection type. For example, the song can be pushed to the audio algorithm detection service, and the audio similarity detection and the lyrics similarity detection are performed respectively. At the same time, the system can archive the inspection records for subsequent machine review and manual review.

[0115] When the song version of the song to be detected changes, such as changes in melody, rhythm, or performance style, etc., but the lyrics content remains unchanged, the song detection type of the song to be detected can be determined as the audio detection type. In this case, the audio vector of the song to be detected is obtained, and from the audio reference library of the original song library, the original song audio vector used for similarity comparison with the audio vector of the song to be detected is matched. Then, audio similarity detection processing is performed on the audio vector of the song to be detected and the matched original song audio vector to obtain the similarity detection result of the song to be detected.

[0116] When the lyrics content of the song to be detected changes, such as the language of the lyrics changes from Chinese to Cantonese, or the song melody is the same but the lyrics content is completely different, the song detection type can be determined as the lyrics detection type. At this time, audio similarity detection and lyrics similarity detection processing need to be performed on the song to be detected respectively, that is, the audio data and lyrics data of the song to be detected need to be subjected to similarity detection with the corresponding original songs respectively, and the similarity detection result of the song is obtained. At this time, the similarity detection result can include the audio similarity detection result and the lyrics similarity detection result. By determining the song detection type, similarity detection can be performed on the song to be detected in a targeted manner. Subsequently, the similarity detection result returned by the audio algorithm detection will be subjected to a multi-dimensional machine review process based on the song detection reference conditions.

[0117] In an embodiment of the present disclosure, for step S240, according to the pre-configured song detection reference conditions and the similarity detection result, the song detection result of the song to be detected is determined, including: obtaining the similarity detection result, where the similarity detection result includes the similarity feature value corresponding to the song similarity detection feature of the song to be detected; obtaining the pre-configured song detection executor, where the song detection executor includes an initial detection condition list composed of song detection reference conditions; sorting the initial detection condition list based on the detection priority to obtain a sorted detection condition list; and matching the similarity detection result with the song detection reference conditions included in the sorted detection condition list one by one to obtain the song detection result.

[0118] Among them, the song similarity detection feature can be a detection feature applied to evaluate the similarity degree between two songs. The similarity feature value can be the specific value corresponding to the song similarity detection feature of the song to be detected. The song detection executor can be an executor used to execute the song similarity detection task. The initial detection condition list can be a condition list composed of multiple song detection conditions. The detection priority can be the priority corresponding to each song detection condition. The sorted detection condition list can be a list obtained by re-sorting the song detection conditions included in the initial detection condition list according to the detection priority.

[0119] Continue to refer to Figure 3After the song detection algorithm cluster performs similarity detection processing on the song to be detected, it outputs the corresponding similarity detection result. The similarity detection result may include the similarity feature value corresponding to the song similarity detection feature of the song to be detected. For example, the song similarity detection feature may include, but is not limited to, the audio similarity between songs, the number of audio hits, the audio hit rate, the lyric similarity, the number of lyric hits, the audio hit rate, etc. The similarity detection result will include the specific values corresponding to the above features.

[0120] To determine the song detection result based on the similarity detection result, a pre-configured song detection executor can be obtained. The song detection executor includes an initial detection condition list. Refer to Figure 6 , Figure 6 Schematically shows a flowchart of song detection based on a song detection executor according to some embodiments of the present disclosure. In step S610, an initial detection condition list is obtained. The initial detection condition list may be a list composed of one or more song detection reference conditions. In step S620, the initial detection condition list is sorted according to the detection priority. After obtaining the initial detection condition list, in order to further improve the song detection efficiency, the initial detection condition list can be sorted based on the detection priority of each song detection reference condition to obtain a sorted detection condition list.

[0121] In step S630, the sorted detection condition list is traversed. The specific similarity feature values included in the similarity detection result are matched one by one with the song detection reference conditions included in the sorted detection condition list to calculate whether the song to be detected hits the songs in the protection library, and the review label of the song to be detected is determined as the song detection result. According to the song detection result, the song label can be updated in real time for subsequent business expansion to use the song efficiently and reasonably.

[0122] In an embodiment of the present disclosure, obtaining a pre-configured song detection executor includes: obtaining an initial rule executor, where the initial rule executor is a stateless executor; determining a similarity feature reference threshold corresponding to the song similarity detection feature, and generating a song detection reference condition based on one or more similarity feature reference thresholds; constructing an initial detection condition list according to at least one song detection reference condition; and configuring the initial rule executor based on the initial detection condition list to generate a song detection executor, where the song detection executor is used to perform song detection processing according to the initial detection condition list.

[0123] Among them, the initial rule executor can be a rule executor in the initial state. The stateless executor can be an executor for providing stateless services. A stateless service means that when the container is running, no data is saved in the container, but the data is uniformly saved outside the container. All processes that the server side can handle must come entirely from the information carried by the request, as well as other public information saved by the server side itself and available for all requests. The similarity feature reference threshold can be a reference value used to compare with each similarity feature value.

[0124] For example, the song detection executor for song review can be generated through the following steps: Obtain an initial rule executor, and the initial rule executor can be a stateless executor; the stateless executor is used to provide stateless services. When performing data processing, one or more song detection reference conditions can be pre-configured to construct an initial detection condition list, so as to configure the initial rule executor through the initial detection condition list to generate a song detection executor. Subsequently, the song detection executor can perform song review processing on the song to be detected based on the initial detection condition list to obtain the final song review label.

[0125] The construction of the initial detection condition list can be carried out through the following steps: Determine the similarity feature reference threshold corresponding to the song similarity detection feature, and generate song detection reference conditions based on one or more similarity feature reference thresholds.

[0126] For example, the song detection reference conditions can include reference conditions corresponding to two song similarity detection features. For example, the detection condition for judging whether a certain song has the risk of being carried is that the condition number of this detection condition is "2" and the priority is "1". When judging whether a certain song has the risk of being carried, the reference threshold of audio similarity can be configured to 0.9; the detection condition for judging whether a certain song has the risk of plagiarism, the condition number of this detection condition is "3" and the priority is "4". When judging whether a certain song has the risk of plagiarism, the reference threshold of lyric hit similarity can be configured to 0.0.

[0127] In some other exemplary embodiments of the present disclosure, corresponding song detection reference conditions can also be configured according to specific song detection requirements, and an initial detection condition list can be constructed according to the configured one or more song detection reference conditions. Then, the initial rule executor is configured based on the initial detection condition list to generate a song detection executor, and the subsequent song detection executor will perform song detection processing on the song to be detected according to the initial detection condition list.

[0128] The song detection executor in this disclosure is a general stateless service. Through unified configuration and a rules engine, it outputs machine-reviewed song tags by combining the above song detection reference conditions. Through this executor, it is possible to change the song detection reference conditions without modifying the code and immediately take effect in the system, quickly responding to changes in online requirements.

[0129] In an embodiment of this disclosure, the similarity detection results are matched one by one with the song detection reference conditions included in the sorting detection condition list to obtain song detection results, including: obtaining the current detection condition from the sorting detection condition list, determining the detection condition information corresponding to the current detection condition, where the detection condition information includes one or more of the detection index type, detection index field, and index field value; based on the detection condition information, performing a matching process on the similarity detection results and the current detection condition to obtain a condition matching result; and determining the song detection result according to the obtained condition matching result.

[0130] Among them, the current detection condition can be the song detection condition currently obtained from the sorting detection condition list. The detection condition information can be the relevant information corresponding to the current detection condition. The detection index type can be the specific type corresponding to the detection index. The detection index field can be the field used to represent the specific name of different detection indexes. The index field value can be the specific value corresponding to each detection index. The condition matching result can be the matching result obtained by comparing each similarity feature value in the similarity detection results with the similarity feature reference threshold in the current detection condition.

[0131] Continue to refer to Figure 6 , during the song detection process, in step S640, traverse the detection conditions in the sorting detection condition list. Obtain the song detection reference conditions from the sorting detection condition list in order, and use the currently obtained song detection reference condition as the current detection condition. Determine the detection condition information corresponding to the current detection condition. For example, the detection condition information can include but is not limited to the detection index type, detection index field, and index field value.

[0132] Specifically, the detection index type can be to detect whether there are operations such as piracy, plagiarism, imitation, or other types of song washing in the song to be detected; the detection index field can be the specific field used to complete this detection index type. For example, to determine whether there is a piracy risk in the song to be detected, it can be done by looking at the specific value of the audio similarity feature of the song to be detected; to determine whether there is a plagiarism risk in the song to be detected, it can be done by looking at the specific value of the lyric similarity feature of the song to be detected. The index field value can be the reference threshold corresponding to the field. For example, the reference threshold for audio similarity can be 0.9, etc.

[0133] In step S650, it is determined whether the condition is matched. When the detection condition information corresponding to the current detection condition is obtained, the similarity detection result and the current detection condition can be matched based on the above detection condition information to determine whether the similarity detection result matches the current detection condition, and a condition matching result is obtained.

[0134] If the condition is not matched, return to execute step S620; if the condition is matched, repeat the above steps until all the song detection reference conditions included in the sorting detection condition list are completed in the matching process. After the traversal operation of the sorting detection condition list is completed, if there is no detection condition in the list that does not match the similarity detection result, in step S660, return the first matched condition group label as the song detection label returned by the machine, that is, use this label as the song detection result of the song to be detected.

[0135] The machine review system will determine the machine review result of the song, such as plagiarism, etc., based on the song similarity detection features returned by the algorithm cluster and in combination with the standardized song detection executor. According to the pre-configured song detection reference conditions, the machine review result of the song to be detected can be obtained, and the song label can be updated in real time.

[0136] In practical applications, the present disclosure already supports functions such as automatically selecting songs for inspection, automatically constructing a song protection library, multi-dimensional machine review, and song tagging; the song detection method of the present disclosure shortens the human and time costs of the songs to be inspected from the original daily level to the minute level; for the currently connected inspection platforms, 22,198 songs have been inspected, compared with the in-station and out-of-station songs, 1,767 songs with the risk of song washing are found, accounting for about 7.96%; in addition, this solution can achieve more than 3,000 songs automatically inspected per day, with more than 100 songs being found, and the proportion of the inspected songs reaching more than 5%.

[0137] In summary, the song detection method of the present disclosure determines the song to be detected, and the song to be detected includes one or more of the songs with changed song libraries, the songs imported by the platform, and the specified selected songs; obtains the pre-constructed original song library, and the original song library is constructed based on the original songs of the music platform; performs similarity detection on the song to be detected and the original songs to obtain a similarity detection result; determines the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result. On the one hand, it closes the process of selecting songs to be inspected, automatically inspecting, machine reviewing, and song tagging of the original songs, improves the efficiency of the song imitation and washing detection process, can discover copyright problems in advance, and avoid copyright risks. On the other hand, by performing multi-dimensional review on the songs, the song review efficiency and the accuracy of the song review can be improved. On the other hand, through the design of the song detection executor, various rule configurations can be handled, and various detection conditions can be quickly responded and adapted.

[0138] Exemplary Device

[0139] After introducing the method of the exemplary embodiments of the present disclosure, next, reference is made to Figure 7 to describe the song detection device of the exemplary embodiments of the present disclosure.

[0140] In Figure 7 the song detection device 700 may include: a song determination module 710, an original library construction module 720, a similarity detection module 730, and a detection result determination module 740.

[0141] Among them, the song determination module 710 is used to determine the song to be detected, and the song to be detected includes one or more of the song library change songs, the platform imported songs, and the specified selected songs; the original library construction module 720 is used to obtain the pre-constructed original song library, and the original song library is constructed based on the original songs of the music platform; the similarity detection module 730 is used to perform similarity detection on the song to be detected and the original song to obtain a similarity detection result; the detection result determination module 740 is used to determine the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result.

[0142] In an embodiment of the present disclosure, the song determination module 710 includes a song determination unit, which is used to determine the song to be detected by one or more combinations of the following methods: obtain the song library to be detected, and the song library to be detected includes the initial detection songs; determine the song to be detected from the initial detection songs based on the song selection rule; use the song that has changed in the song library to be detected as the song to be detected; and use the song imported by the business personnel based on the song detection platform as the song to be detected.

[0143] In an embodiment of the present disclosure, the original library construction module 720 includes an original library construction unit, which is used to: obtain the specified songs from the first music platform based on the preset time period to obtain the first original songs, and the first original songs include the songs included in the song ranking list in the first music platform and / or the songs whose song evaluation indicators reach the specified index threshold; use the songs obtained from the second music platform as the second original songs; construct the original song library based on the first original songs and the second original songs, and the original song library includes an audio reference library and a lyrics reference library.

[0144] In an embodiment of the present disclosure, the original library construction unit includes an original song acquisition subunit, which is used to: use the songs currently included in the second music platform as the initial songs; obtain the songs imported through the second music platform as the newly imported songs; obtain the pre-configured song filtering conditions; perform song filtering processing on the initial songs and the newly imported songs based on the song filtering conditions to obtain the second original songs.

[0145] In one embodiment of the present disclosure, the original library construction unit includes an original library construction subunit, which is configured to: summarize the first original song and the second original song to determine the initial incremental song; determine the target incremental song and the song to be deleted based on the current song in the original song library and the initial incremental song; update the audio reference library according to the target incremental song and the song to be deleted; obtain the original lyric data based on a preset time period, and update the lyric reference library according to the original lyric data.

[0146] In one embodiment of the present disclosure, the original library construction subunit includes a song vector update subunit, which is configured to: determine the incremental song audio vector corresponding to the target incremental song, and add the incremental song audio vector to the audio reference library; delete the to-be-deleted song audio vector corresponding to the song to be deleted from the audio reference library to update the audio reference library, and the audio reference library includes the original song audio vectors corresponding to the updated original songs.

[0147] In one embodiment of the present disclosure, the similarity detection module 730 includes a similarity detection unit, which is configured to: determine the basic song information corresponding to the song to be detected, and the basic song information includes the song version and the language type of the lyrics; determine the song detection type according to the basic song information, and the song detection type includes an audio detection type and a lyric detection type; when the song detection type is the audio detection type, determine the to-be-detected song audio vector corresponding to the song to be detected; perform audio similarity detection processing based on the to-be-detected song audio vector and the original song audio vector corresponding to the original song to obtain a similarity detection result; when the song detection type is the lyric detection type, perform audio similarity detection and lyric similarity detection processing on the song to be detected to obtain a similarity detection result.

[0148] In one embodiment of the present disclosure, the detection result determination module 740 includes a detection result determination unit, which is configured to: obtain the similarity detection result, and the similarity detection result includes the similarity feature value corresponding to the song similarity detection feature of the song to be detected; obtain the pre-configured song detection executor, and the song detection executor includes an initial detection condition list composed of song detection reference conditions; sort the initial detection condition list based on the detection priority to obtain a sorted detection condition list; match the similarity detection result with the song detection reference conditions included in the sorted detection condition list one by one to obtain the song detection result.

[0149] In one embodiment of the present disclosure, the detection result determination unit includes a detector configuration subunit, which is used to: obtain an initial rule executor, where the initial rule executor is a stateless executor; determine a similarity feature reference threshold corresponding to the song similarity detection feature, and generate a song detection reference condition based on one or more similarity feature reference thresholds; construct an initial detection condition list according to at least one song detection reference condition; configure the initial rule executor based on the initial detection condition list to generate a song detection executor, and the song detection executor is used to perform song detection processing according to the initial detection condition list.

[0150] In one embodiment of the present disclosure, the detection result determination unit includes a detection result determination subunit, which is used to: obtain the current detection condition from the sorted detection condition list, and determine the detection condition information corresponding to the current detection condition, where the detection condition information includes one or more of a detection index type, a detection index field, and an index field value; based on the detection condition information, perform a matching process on the similarity detection result and the current detection condition to obtain a condition matching result; determine the song detection result according to the obtained condition matching result.

[0151] Since each functional module of the song detection device in the exemplary embodiments of the present disclosure corresponds to the steps in the exemplary embodiments of the above song detection method, for the details not disclosed in the device embodiments of the present disclosure, please refer to the embodiments of the above song detection method of the present disclosure, which will not be elaborated here.

[0152] It should be noted that although several modules or units of the song detection device are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of the two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0153] Exemplary Medium

[0154] After introducing the device of the exemplary embodiments of the present disclosure, next, reference is made to Figure 8 to describe the storage medium of the exemplary embodiments of the present disclosure.

[0155] In some embodiments, various aspects of the present disclosure can also be implemented as a medium having program code stored thereon, which is used to implement the steps in the song detection method according to various exemplary embodiments of the present disclosure described in the "Exemplary Method" section of this specification when the program code is executed by a processor of a device.

[0156] For example, when the processor of the device executes the program code, it can implement as Figure 2In step S210 described above, determine the song to be detected, where the song to be detected includes one or more of the songs with changed song libraries, the songs imported from the platform, and the specified circled songs; step S220, obtain the pre-constructed original song library, where the original song library is constructed based on the original songs of the music platform; step S230, perform a similarity detection on the song to be detected and the original songs to obtain a similarity detection result; step S240, determine the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result.

[0157] Reference Figure 8 As shown, a program product 800 for implementing the above song detection method or implementing the above song detection method according to an embodiment of the present disclosure is described. It can be a portable compact disc read-only memory (CD-ROM) and includes program code, and can run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto.

[0158] The program product can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0159] The computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries the readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable signal medium can also be any readable medium other than the readable storage medium.

[0160] The program code for performing the operations of the present disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages - such as Java, C++, etc., and also including conventional procedural programming languages - such as the "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN).

[0161] Exemplary Computing Device

[0162] After introducing the song detection method, song detection device, and storage medium of the exemplary embodiments of the present disclosure, next, reference will be made to Figure 9 describe the electronic device of the exemplary embodiments of the present disclosure.

[0163] Those skilled in the art of the relevant technical field can understand that various aspects of the present disclosure can be implemented as a system, method, or program product. Therefore, various aspects of the present disclosure can be specifically implemented in the following forms, namely: a complete hardware embodiment, a complete software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects, which can be collectively referred to as "circuit", "module", or "system" here.

[0164] In some possible embodiments, the electronic device according to the present disclosure may at least include at least one processing unit and at least one storage unit. Among them, the storage unit stores program code, and when the program code is executed by the processing unit, the processing unit executes the steps in the song detection method according to various exemplary embodiments of the present disclosure described in the above "Exemplary Method" section of this specification. For example, the processing unit may execute steps such as Figure 2 shown in S210, determine the song to be detected, where the song to be detected includes one or more of the songs with changed song libraries, platform-imported songs, and designated circled songs; step S220, obtain the pre-constructed original song library, which is constructed based on the original songs of the music platform; step S230, perform a similarity detection on the song to be detected and the original songs to obtain a similarity detection result; step S240, determine the song detection result of the song to be detected according to the pre-configured song detection reference conditions and the similarity detection result.

[0165] Next, reference will be made to Figure 9 describe the electronic device 900 according to the exemplary embodiments of the present disclosure. Figure 9 The electronic device 900 shown is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.

[0166] As Figure 9 shown, the electronic device 900 is presented in the form of a general-purpose computing device. The components of the electronic device 900 may include, but are not limited to: the above at least one processing unit 901, the above at least one storage unit 902, a bus 903 connecting different system components (including the storage unit 902 and the processing unit 901), and a display unit 907.

[0167] The bus 903 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus structures.

[0168] The storage unit 902 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 921 and / or cache memory 922, and may further include read-only memory (ROM) 923.

[0169] The storage unit 902 may also include a program / utilities 925 having a set (at least one) of program modules 924. Such program modules 924 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment.

[0170] The electronic device 900 can also communicate with one or more external devices 904 (such as a keyboard, a pointing device, a Bluetooth device, etc.), can also communicate with one or more devices that enable a user to interact with the electronic device 900, and / or can communicate with any device that enables the electronic device 900 to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication can be carried out through an input / output (I / O) interface 905. Also, the electronic device 900 can communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 906. As shown in the figure, the network adapter 906 communicates with other modules of the electronic device 900 through the bus 903. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device 900, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0171] It should be noted that although several units / modules or sub-units / modules of the song detection device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more of the above-described units / modules can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0172] Moreover, although the operations of the methods of the present disclosure are depicted in the drawings in a particular order, this is not required or implied, i.e., these operations must be performed in this particular order, or that all illustrated operations must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step and performed, and / or one step may be decomposed into multiple steps and performed.

[0173] Although the spirit and principles of the present disclosure have been described with reference to several specific embodiments, it should be understood that the present disclosure is not limited to the specific embodiments disclosed, and the division of various aspects does not mean that the features in these aspects cannot be combined for benefits. This division is only for the convenience of expression. The present disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A song detection method, characterized in that: include: Determine songs to be tested, where the songs to be tested include one or more of songs in a library that have been changed, songs imported from a platform, and songs that have been designated and selected; Obtain a pre-built original song library, where the original song library is built based on original songs on a music platform; Performing a similarity test on the song to be tested and the original song to obtain a similarity test result; The song detection result of the song to be detected is determined according to the pre-configured song detection reference condition and the similarity detection result.

2. The method according to claim 1, characterized in that: The method for determining the song to be detected includes one or more combinations of the following methods: Acquire a library of songs to be detected, wherein the library of songs to be detected includes initial detection songs; Determining the song to be detected from the initial detected songs based on a song selection rule; The songs that have been changed in the to-be-detected song library are used as the to-be-detected songs; as well as The songs imported by the business personnel based on the song detection platform are used as the songs to be detected.

3. The method according to claim 1, characterized in that The obtaining of the pre-built original song library comprises: Acquire a specified song from a first music platform based on a preset time period to obtain a first original song, where the first original song includes songs included in a song ranking list in the first music platform and / or songs whose song evaluation index reaches a specified index threshold; The song obtained from the second music platform is used as the second original song; The original song library is constructed based on the first original song and the second original song, and the original song library includes an audio reference library and a lyrics reference library.

4. The method according to claim 3, characterized in that The constructing the original song library based on the first original song and the second original song comprises: Aggregating the first original song and the second original song to determine an initial incremental song; Based on the current songs in the original song library and the initial incremental songs, determining target incremental songs and songs to be deleted; Update the audio reference library according to the target incremental songs and the songs to be deleted; Original lyrics data is acquired based on a preset time period, and the lyrics reference library is updated according to the original lyrics data.

5. The method according to claim 1, characterized in that The determining the song detection result of the song to be detected according to the pre-configured song detection reference condition and the similarity detection result includes: Acquire the similarity detection result, wherein the similarity detection result includes a similarity feature value corresponding to the song similarity detection feature of the song to be detected; Obtaining a preconfigured song detection executor, the song detection executor including an initial detection condition list consisting of the song detection reference conditions; Sorting the initial detection condition list based on the detection priority to obtain a sorted detection condition list; The similarity detection result is matched one by one with the song detection reference conditions included in the sorted detection condition list to obtain the song detection result.

6. The method according to claim 5, characterized in that The step of obtaining a pre-configured song detection executor includes: Obtain an initial rule executor, where the initial rule executor is a stateless executor; Determine a similarity feature reference threshold corresponding to the song similarity detection feature, and generate the song detection reference condition based on one or more of the similarity feature reference thresholds; Constructing the initial detection condition list according to at least one of the song detection reference conditions; The initial rule executor is configured based on the initial detection condition list to generate the song detection executor, and the song detection executor is used to perform song detection processing according to the initial detection condition list.

7. The method according to claim 5, characterized in that The step of matching the similarity detection result with the song detection reference conditions included in the sorting detection condition list one by one to obtain the song detection result includes: Acquire the current detection condition from the sorted detection condition list, and determine the detection condition information corresponding to the current detection condition, wherein the detection condition information includes one or more of a detection indicator type, a detection indicator field, and an indicator field value; Based on the detection condition information, matching the similarity detection result with the current detection condition to obtain a condition matching result; The song detection result is determined according to the obtained condition matching result.

8. A song detection device, characterized in that: include: A song determination module, used to determine songs to be detected, wherein the songs to be detected include one or more of songs changed from a music library, songs imported from a platform, and songs selected by a designated circle; An original library construction module is used to obtain a pre-constructed original song library, where the original song library is constructed based on the original songs of the music platform; A similarity detection module is used to perform similarity detection on the song to be detected and the original song to obtain a similarity detection result; The detection result determination module is used to determine the song detection result of the song to be detected based on the pre-configured song detection reference conditions and the similarity detection result.

9. An electronic device, characterized in that: include: processor; as well as A memory having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by the processor, implement the song detection method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the song detection method according to any one of claims 1 to 7 is implemented.

Images