Music infringement detection method and system

Abstract

Embodiments of the present invention disclose a method and system for detecting music copyright infringement. The method includes identifying music information of a target piece of music to form a feature set of the target piece of music; wherein the music information of the target piece of music includes one or more of the following: song title, audio fingerprint, and lyrics; comparing the identified feature set of the target piece of music with feature sets of music in a copyright database; and determining whether the target piece of music is infringing based on the comparison results. The method and system for detecting music copyright infringement in embodiments of the present invention have wide coverage, low cost, and can promptly detect copyright infringement.

Description

Technical Field

[0001] The present invention relates to the technical field related to music infringement detection, and in particular to a music infringement detection method and system. Background Art

[0002] Since the advent of the "Internet Plus" era in 2013, the digital music industry, already thriving in the internet age, has experienced another surge. Advances in internet technology and its deepening integration with the digital music sector have ushered in a vibrant summer for digital music in this "Internet Plus" era. However, the global popularity of digital music has also been accompanied by a surge in digital music copyright infringement cases. Due to the inherent characteristics of this emerging music genre—the ease of copying and dissemination of digital music data and the increasingly user-friendly audio editing software—copyright protection for digital music in this "Internet Plus" era is particularly challenging.

[0003] At present, there are mainly the following types of music infringement:

[0004] 1. Illegal broadcasting. This includes illegal broadcasting on traditional music streaming platforms; illegal broadcasting by audio and video hosts; illegal broadcasting on public accounts; illegal broadcasting during shows and live broadcasts.

[0005] 2. Adaptation. This includes performing without the copyright holder's authorization; incorporating the work as background music into other audio or video works, etc.

[0006] 3. Use beyond the scope of authorization. For example, a platform may only have the right to play a song but allow users to download or forward it; or it may only have a one-year license for a song but still not remove the music after more than a year.

[0007] Current copyright monitoring systems are mostly manual, with limited coverage and high costs. They also fail to detect infringements promptly and have poor timeliness. Some monitoring systems rely on manual labor, combined with keyword screening, comparing information on music introduction pages with metadata from the target music program, such as the music title, artist name, and release year. These methods only address certain scenarios and are inadequate for background music, such as when used as background music in video programs or when streamers illegally broadcast copyrighted music as background music. Summary of the Invention

[0008] In view of this, one aspect of the present invention discloses a method for detecting music infringement.

[0009] The music infringement detection method comprises:

[0010] Identify music information of target music and form a feature set of the target music; wherein the music information of the target music includes one or more of song title, audio fingerprint and lyrics;

[0011] Comparing the identified feature set of the target music with the feature set of music in the copyright database;

[0012] Determine whether the target music is infringing based on the comparison results.

[0013] According to a preferred embodiment of the present invention, the music infringement detection method further comprises: generating a report if infringing music is detected.

[0014] According to a preferred embodiment of the present invention, identifying music information of target music and forming a feature set of target music includes:

[0015] Crawl music information of target music at target address;

[0016] Extract the text information of the target music and normalize the crawled text;

[0017] and / or

[0018] Extract the audio of the target music, obtain the audio fingerprint and save it.

[0019] According to a preferred embodiment of the present invention, comparing the identified feature set of the target music with the feature set of music in the copyright database includes:

[0020] Performing a text comparison between the identified target music text information and the corresponding text information in the copyright database; wherein the text information includes one or more of the target music song title, author, and lyrics;

[0021] and\or

[0022] Compare the audio fingerprint of the target music with the reference audio in the copyright database.

[0023] According to a preferred embodiment of the present invention, determining whether the target music is infringing based on the comparison result includes:

[0024] determining whether the text information and / or audio fingerprint of the target music is similar to the music in the copyright database based on the comparison result;

[0025] If similar, compare the usage of the target music with the authorization records of the music in the copyright database;

[0026] If the use of the target music does not match one or more of the playback, download and authorization time in the authorized records of the music in the copyright database, the platform's use of the target music will be judged as infringement.

[0027] Another aspect of the present invention discloses a music infringement detection system.

[0028] The music infringement detection system includes:

[0029] an identification module for identifying music information of a target music and forming a feature set of the target music; wherein the music information of the target music includes one or more of a song title, an audio fingerprint, and lyrics;

[0030] A comparison module, configured to compare the feature set of the target music identified by the identification module with the feature set of music in the copyright database;

[0031] as well as

[0032] The judgment module is used to judge whether the target music is infringing according to the comparison result of the comparison module.

[0033] According to a preferred embodiment of the present invention, the music infringement detection system further includes a notification module; the notification module is used to generate a report.

[0034] According to a preferred embodiment of the present invention, the recognition module includes a crawling module, a text recognition module and an audio fingerprint extraction module;

[0035] The crawling module is used to crawl music information of target music in the target address;

[0036] The text recognition module is used to extract text information of the target music and normalize the crawled text; wherein the text information includes one or more of the song title, author and lyrics of the target music;

[0037] The audio fingerprint extraction module is used to extract the audio of the target music.

[0038] According to a preferred embodiment of the present invention, the audio fingerprint extraction module includes an audio data preprocessing end and an audio fingerprint extraction end;

[0039] The audio data preprocessing end is used to obtain the audio spectrum of the target music;

[0040] The audio fingerprint extraction end is used to extract frequency band points in the audio spectrum whose amplitude exceeds the amplitude threshold as feature points, record the amplitude, time sequence, and relationship between adjacent feature points of each feature point, generate an audio fingerprint, and save it.

[0041] According to a preferred embodiment of the present invention, the comparison module includes a text comparison module and an audio fingerprint comparison module;

[0042] The text comparison module is used to compare the text information identified by the text recognition module with the corresponding text information in the copyright database;

[0043] The audio fingerprint comparison module is used to perform audio fingerprint comparison between the audio of the target music and the control audio in the copyright database.

[0044] The above one or more technical solutions in the music infringement detection method and system provided by the embodiments of the present invention have at least one of the following technical effects:

[0045] The music infringement detection method and system according to the embodiment of the present invention have wide coverage, low cost, and can detect infringement behaviors in a timely manner.

[0046] Some additional features of the present invention may be set forth in the following description. Some additional features of the present invention will be apparent to those skilled in the art from an examination of the following description and accompanying drawings, or from the manufacture or operation of the embodiments. The disclosed features of the present invention may be realized and attained by practice or use of the various methods, means, and combinations of the specific embodiments described below. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute a limitation of the present invention. In each figure, the same reference numerals represent the same components.

[0048] Figure 1 1 is a schematic diagram of a judgment process of a music infringement detection method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0050] It should be noted that if the description and claims of the present invention and the above-mentioned drawings refer to the terms "first", "second", etc., they are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate to facilitate the embodiments of the present invention described herein. In addition, if the terms "including" and "having" and any variations thereof are involved, it is intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0051] In the present invention, when terms such as "upper," "lower," "left," "right," "front," "back," "top," "bottom," "inside," "outside," "center," "vertical," "horizontal," "transverse," and "longitudinal" are used, the orientations or positions they indicate are based on those shown in the accompanying drawings. These terms are primarily intended to better describe the present invention and its embodiments and are not intended to limit the devices, elements, or components indicated to having a specific orientation, or to being constructed or operated in a specific orientation.

[0052] Furthermore, some of the above terms may be used to express other meanings besides indicating a position or location. For example, the term "on" may also be used to indicate a dependency or connection in certain circumstances. Those skilled in the art will understand the specific meanings of these terms in the present invention based on the specific circumstances.

[0053] Furthermore, in this disclosure, the terms "installed," "disposed," "provided with," "connected," "connected," and "socketed" should be interpreted broadly. For example, these terms may refer to fixed connections, removable connections, or integral structures; mechanical connections or electrical connections; direct connections, indirect connections through an intermediary, or internal communication between two devices, elements, or components. Those skilled in the art will understand the specific meanings of these terms in this disclosure based on the specific circumstances.

[0054] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0055] One aspect of an embodiment of the present invention discloses a method for detecting music infringement.

[0056] like Figure 1 As shown, the music infringement detection method may mainly include:

[0057] Identify the music information of the target music and form a feature set of the target music, wherein the music information of the target music includes one or more of the song title, audio fingerprint and lyrics.

[0058] For example, identifying music information of a target piece of music and forming a feature set for the target piece of music may include: crawling the music information of the target piece of music at a target address; extracting text information of the target piece of music and normalizing the crawled text; and / or extracting audio of the target piece of music, obtaining an audio fingerprint, and saving the fingerprint.

[0059] Specifically, after crawling the music information of the target address, the name, author, lyrics and other text information of the target music are extracted and the crawled text is normalized, for example: the encoding is converted to UTF-8 encoding, spaces, placeholders, and encoded control characters are removed, full-width characters in the original text are converted to half-width characters, and all uppercase characters are converted to lowercase characters.

[0060] The audio data is sampled, filtered, short-time Fourier transformed, segmented, and converted to an audio spectrum, ultimately generating an audio spectrum. Frequency bands with amplitudes exceeding a threshold are extracted from the audio spectrum as feature points. The amplitude, timing, and relationships between adjacent feature points are recorded, and an audio fingerprint is generated. The audio fingerprint is stored in a (code, time) structure, where code and time are both continuous numbers. Code contains the frequency, distance, and amplitude of the feature point, and time contains the timing information of the feature point.

[0061] The feature set of the identified target music is compared with the feature set of music in the copyright database.

[0062] Exemplarily, comparing the feature set of the identified target music with the feature set of music in the copyright database may include:

[0063] Compare the identified text information of the target music with the corresponding text information in the copyright database. This text information may include one or more of the target music's song title, author, and lyrics. Alternatively, compare the audio fingerprint of the target music with the reference audio in the copyright database.

[0064] Specifically, the identified text information such as name, author, lyrics, etc. is compared with the corresponding name, author, lyrics in the company's copyright database.

[0065] For the name and author, you can directly match the name of the target music, the author's name in the company's copyright database, and the author field to determine whether there are any consistent items.

[0066] For lyrics, the following algorithm is used:

[0067] 1) Traverse the normalized lyrics text and extract the most frequently appearing characters as anchor points.

[0068] 2) Select the character string consisting of n characters before and after the anchor character, plus the anchor character itself as the feature window.

[0069] 3) Using the feature window as the basic unit, matching is performed in the lyrics database.

[0070] Set a matching value for all lyrics texts in the database. In the initial state, no matching has been done and all lyrics text matching values ​​are 0. If there is a string in the database whose matching consistency rate with a certain feature window exceeds the threshold, the matching value of the lyrics text corresponding to the string is increased by 1. The matching consistency rate is calculated as follows:

[0071] Match consistency rate = number of identical characters in the string / total number of characters in the feature window.

[0072] 4) After the matching is completed, if there is a matching value greater than the threshold, it is considered that there is music in the database with lyrics similar to the target music.

[0073] Compare the audio fingerprint of the audio to be identified with the reference audio in the company's copyright library. The algorithm is as follows:

[0074] 1) Compare the extracted audio fingerprint with the audio fingerprints of existing music in the database (that is, the comparison objects are the audio fingerprint of the target audio and the audio fingerprint of a certain music in the database), extract the feature points with the same code, and form multiple arrays [(code1, time1)(code1, time2)(code1, time3)…(code1, time n)], [(code2, time1)(code2, time2)(code2, time3)…(code2, time n)]…

[0075] 2) Calculate the difference between different times in the same array (same code, different time), and record them as ΔT1, ΔT2, ..., ΔTn.

[0076] When the number of identical ΔTs exceeds a threshold, it is considered that the audio fingerprint match is successful and there is music in the database that is similar to the target music audio fingerprint.

[0077] Determine whether the target music is infringing based on the comparison results.

[0078] Exemplarily, determining whether the target music is infringing based on the comparison results may include:

[0079] Based on the comparison results, the platform determines whether the text information and / or audio fingerprint of the target music are similar to those in the copyright database. If so, the platform compares the target music's usage with the authorization records for that music in the copyright database. If the target music's usage does not match one or more of the play, download, and authorization times in the copyright database, the platform determines that the target music's use is infringing.

[0080] Specifically, if music with similar lyrics is found, or the audio fingerprint module determines that music with similar audio fingerprints exists, it is considered that similar music exists.

[0081] After determining that the target music and the control audio in the copyright library are similar audio, further identify the use mode of the music by the music source address, such as only playing, playing and downloading, etc.

[0082] After the recognition is complete, the result is returned for judgment. Combined with the authorization records in the copyright library, it is determined whether the source address is using the music legally, including the judgment of various authorization conditions, such as the authorization period, whether playback, downloading, adaptation, and cover are allowed, etc.

[0083] For example:

[0084] It is identified that the target address uses the target music in two ways: playing and downloading. In the authorization records of the copyright library, the music platform corresponding to the target address is only authorized to play the target music. Therefore, it is determined that the use of the target music by the music platform is infringing.

[0085] If it is identified that the target address uses the target music in two ways: playing and downloading, and the music platform corresponding to the target address has authorization to play and download the target music in the authorization record of the copyright library, then it is determined that the music platform has not infringed on the target music in its use.

[0086] At this time, it is further determined whether the time when the target platform uses the target music is within the authorized time period in the authorization record of the copyright library.

[0087] If so, it is determined that the music platform's use of the target music is legal.

[0088] If not, the music platform’s use of the target music is deemed to be infringing.

[0089] Furthermore, the music infringement detection method may also include: generating a report if infringing music is detected.

[0090] Specifically, if infringing music is detected, a report will be generated. The report content may include the music title, author, the address or platform where the music is used, and the specific violation method.

[0091] One aspect of an embodiment of the present invention discloses a music infringement detection system.

[0092] The music infringement detection system can mainly include:

[0093] The recognition module is used to identify the music information of the target music and form a feature set of the target music. The music information of the target music includes one or more of the song title, audio fingerprint and lyrics.

[0094] Exemplarily, the recognition module includes a crawling module, a text recognition module, and an audio fingerprint extraction module. The crawling module is used to crawl the music information of the target music in the target address. The text recognition module is used to extract the text information of the target music and normalize the crawled text. The text information includes one or more of the song title, author, and lyrics of the target music. The audio fingerprint extraction module is used to extract the audio of the target music. The audio fingerprint extraction module may include an audio data preprocessing end and an audio fingerprint extraction end. The audio data preprocessing end is used to obtain the audio spectrum of the target music. The audio fingerprint extraction end is used to extract the frequency band points in the audio spectrum whose amplitude exceeds the amplitude threshold as feature points, record the amplitude, time sequence, and relationship between adjacent feature points of each feature point, generate an audio fingerprint, and save it.

[0095] Specifically, in this embodiment, the crawling module is used to crawl the music information of the target address. The text recognition module is used to extract the name, author, lyrics and other text information of the target music and normalize the crawled text, for example: converting the encoding to UTF-8 encoding, removing spaces, placeholders, and encoded control characters, converting full-width characters in the original text to half-width characters, and converting all uppercase characters to lowercase characters. The audio fingerprint extraction module extracts the audio of the target music and uses the model to extract the audio fingerprint. Specifically, it can include an audio data preprocessing end and an audio fingerprint extraction end. Among them, the audio data preprocessing end is used to sample, filter, short-time Fourier transform, audio segmentation, audio spectrum conversion, and finally obtain the audio spectrum. Among them, the audio fingerprint extraction end is used to extract frequency band points in the audio spectrum whose amplitude exceeds the amplitude threshold as feature points, record the amplitude, timing, and relationship between adjacent feature points of each feature point, and generate an audio fingerprint; the audio fingerprint is saved in a (code, time) structure, where code and time are both a set of continuous numbers. Code contains the feature point frequency, the distance between feature points, and the feature point amplitude, and time contains the timing information of the feature point.

[0096] The comparison module is used to compare the feature set of the target music identified by the recognition module with the feature set of the music in the copyright database.

[0097] Exemplarily, the comparison module includes a text comparison module and an audio fingerprint comparison module. The text comparison module is used to compare the text information identified by the text recognition module with the corresponding text information in the copyright database. The audio fingerprint comparison module is used to compare the audio fingerprints of the target music with the reference audio in the copyright database.

[0098] Specifically, in this embodiment, the text comparison module compares the text information such as the name, author, lyrics, etc. identified by the text recognition module with the corresponding name, author, lyrics in the company's copyright database.

[0099] For names and authors, the text comparison module directly matches the name of the target music, the name of the author in the company's copyright database, and the author field to determine whether there are any consistent items.

[0100] For lyrics, use the following method:

[0101] 5) Traverse the normalized lyrics text and extract the most frequently appearing characters as anchor points.

[0102] 6) Select the character string consisting of n characters before and after the anchor character, plus the anchor character itself as the feature window.

[0103] 7) Using the feature window as the basic unit, matching is performed in the lyrics database.

[0104] Set a matching value for all lyrics texts in the database. In the initial state, no matching has been done and all lyrics text matching values ​​are 0. If there is a string in the database whose matching consistency rate with a certain feature window exceeds the threshold, the matching value of the lyrics text corresponding to the string is increased by 1. The matching consistency rate is calculated as follows:

[0105] Match consistency rate = number of identical characters in the string / total number of characters in the feature window.

[0106] 8) After the matching is completed, if there is a matching value greater than the threshold, it is considered that there is music in the database with lyrics similar to the target music.

[0107] The audio fingerprint comparison module compares the audio to be identified with the reference audio in the company's copyright library. The method is as follows:

[0108] 3) Compare the audio fingerprint extracted by the fingerprint extraction module with the audio fingerprints of the music already in the database (that is, the comparison objects are the audio fingerprint of the target audio and the audio fingerprint of a certain music in the database), extract the feature points with the same code, and form multiple arrays [(code1, time1)(code1, time2)(code1, time3)…(code1, timen)], [(code2, time1)(code2, time2)(code2, time3)…(code2, time n)]…

[0109] 4) Calculate the difference between different times in the same array (same code, different time), and record them as ΔT1, ΔT2, ..., ΔTn.

[0110] 5) When the number of identical ΔTs exceeds a threshold, it is considered that the audio fingerprint match is successful and there is music in the database that is similar to the target music audio fingerprint.

[0111] The judgment module is used to judge whether the target music is infringing according to the comparison result of the comparison module.

[0112] Exemplarily, the judgment module determines whether the text information and / or audio fingerprint of the target music is similar to the music in the copyright database based on the comparison result; if similar, the usage method of the target music is compared with the authorization record of the music in the copyright database; if the usage method of the target music is inconsistent with one or more of the playback, download and authorization time in the authorization record of the music in the copyright database, the platform is judged to have infringed the use of the target music.

[0113] Specifically, in this embodiment, if the text comparison module determines that there is music with similar lyrics after judging the lyrics, or if the audio fingerprint module determines that there is music with similar audio fingerprints, it is considered that there is similar music.

[0114] After determining that the target music and the control audio in the copyright library are similar audio, the judgment module notifies the recognition module to further identify the usage mode of the music source address for the music, such as only playing, playing and downloading, etc.

[0115] After the recognition is completed, the result is returned to the judgment module.

[0116] The judgment module combines the authorization records of the copyright library to determine whether the source address's use of the music is legal, including the judgment of various authorization situations, such as the authorization time, whether playback, downloading, adaptation, cover, etc. are allowed.

[0117] For example, the identification module recognizes that the target address uses the target music in two ways: playing and downloading. In the authorization record of the copyright library, the music platform corresponding to the target address only has authorization to play the target music. Therefore, it is judged that the use of the target music by the music platform is infringing.

[0118] If the identification module identifies that the target address uses the target music in two ways: playing and downloading, and the music platform corresponding to the target address has authorization to play and download the target music in the authorization record of the copyright library, then it is judged that the music platform has not infringed on the use of the target music. At this time, it is further judged whether the time when the target platform uses the target music is within the authorized time period in the authorization record of the copyright library. If so, it is judged that the music platform's use of the target music is legal. If not, it is judged that the music platform's use of the target music is infringing.

[0119] Furthermore, the music infringement detection system may further include a notification module, which is used to generate a report.

[0120] Specifically, if infringing music is detected, a report will be generated. The report content may include the music name, author, address or platform where the music is used, specific violation method, etc.

[0121] The music infringement detection method and system according to the embodiment of the present invention have wide coverage, low cost, and can detect infringement behaviors in a timely manner.

[0122] It should be noted that all features disclosed in this specification, or steps in all methods or processes disclosed, except for mutually exclusive features and / or steps, can be combined in any manner.

[0123] Furthermore, the above-described specific embodiments are merely illustrative. Persons skilled in the art may devise various solutions based on the disclosure of the present invention, and such solutions fall within the scope of the present invention and are intended to be protected by the present invention. Persons skilled in the art should understand that the present description and drawings are intended to be illustrative and not to limit the scope of the claims. The scope of protection of the present invention is defined by the claims and their equivalents.

Claims

1. A method for detecting music infringement, characterized in that: The music infringement detection method comprises: Identify music information of target music and form a feature set of the target music; wherein the music information of the target music includes one or more of song title, audio fingerprint and lyrics; The identifying of music information of target music to form a feature set of target music includes: Crawl music information of target music at target address; Extract the text information of the target music and normalize the crawled text; Extract the audio of the target music, obtain the audio fingerprint and save it; Comparing the identified feature set of the target music with the feature set of music in the copyright database; The comparing the identified feature set of the target music with the feature set of music in the copyright database includes: Performing a text comparison between the identified target music text information and the corresponding text information in the copyright database; wherein the text information includes one or more of the target music song title, author, and lyrics; Compare the audio fingerprint of the target music with the reference audio in the copyright database; Compare the text of the identified name, author, and lyrics with the corresponding name, author, and lyrics in the company's copyright database, including: For the name and author, directly match the name of the target music, the author's name in the company's copyright database, and the author field to determine whether there are any consistent items; For lyrics, the following algorithm is used: Traverse the normalized lyrics text and extract the most frequently appearing characters as anchor points; Select the string consisting of n characters before and after the anchor character, plus the anchor character itself as the feature window; Using the feature window as the basic unit, matching is performed in the lyrics database; Set a matching value for all lyrics texts in the database. In the initial state, no matching has been performed and all lyrics text matching values ​​are 0. If there is a string in the database whose matching consistency rate with a certain feature window exceeds the threshold, the lyrics text matching value corresponding to the string is increased by 1. The matching consistency rate is calculated as follows: Matching consistency rate = number of identical characters in the string / total number of characters in the feature window; After the matching is completed, if there is a matching value greater than the threshold, it is considered that there is music in the database with lyrics similar to the target music; Compare the target music's audio with the reference audio in the copyright database using the following algorithm: Compare the extracted audio fingerprints with the audio fingerprints of the music in the database, extract the feature points with the same code, and form multiple arrays; Calculate the difference between different times in the same array, recorded as ΔT1, ΔT2, ..., ΔTn; When the number of identical ΔTs exceeds the threshold, it is considered that the audio fingerprint is matched successfully and there is music in the database that is similar to the target music audio fingerprint; Determine whether the target music is infringing based on the comparison results.

2. The music infringement detection method according to claim 1, characterized in that: The music infringement detection method further includes: generating a report if infringing music is detected.

3. The music infringement detection method according to claim 1, characterized in that: The determination of whether the target music is infringing based on the comparison results includes: Determine whether the text information and audio fingerprint of the target music are similar to the music in the copyright database based on the comparison results; If similar, compare the usage of the target music with the authorization records of the music in the copyright database; If the use of the target music does not match one or more of the playback, download and authorization time in the authorized records of the music in the copyright database, the platform's use of the target music will be judged as infringement.

4. A music infringement detection system, characterized in that: It includes: an identification module for identifying music information of a target music and forming a feature set of the target music; wherein the music information of the target music includes one or more of a song title, an audio fingerprint, and lyrics; The recognition module includes a crawling module, a text recognition module and an audio fingerprint extraction module; The crawling module is used to crawl music information of target music in the target address; The text recognition module is used to extract text information of the target music and normalize the crawled text; wherein the text information includes one or more of the song title, author and lyrics of the target music; The audio fingerprint extraction module is used to extract the audio of the target music; A comparison module, configured to compare the feature set of the target music identified by the identification module with the feature set of music in the copyright database; The comparison module includes a text comparison module and an audio fingerprint comparison module; The text comparison module is used to compare the text information identified by the text recognition module with the corresponding text information in the copyright database; The comparing the identified feature set of the target music with the feature set of music in the copyright database includes: Performing a text comparison between the identified target music text information and the corresponding text information in the copyright database; wherein the text information includes one or more of the target music song title, author, and lyrics; Compare the audio fingerprint of the target music with the reference audio in the copyright database; Compare the text of the identified name, author, and lyrics with the corresponding name, author, and lyrics in the company's copyright database, including: For the name and author, directly match the name of the target music, the author's name in the company's copyright database, and the author field to determine whether there are any consistent items; For lyrics, the following algorithm is used: Traverse the normalized lyrics text and extract the most frequently appearing characters as anchor points; Select the string consisting of n characters before and after the anchor character, plus the anchor character itself as the feature window; Using the feature window as the basic unit, matching is performed in the lyrics database; Set a matching value for all lyrics texts in the database. In the initial state, no matching has been performed and all lyrics text matching values ​​are 0. If there is a string in the database whose matching consistency rate with a certain feature window exceeds the threshold, the lyrics text matching value corresponding to the string is increased by 1. The matching consistency rate is calculated as follows: Matching consistency rate = number of identical characters in the string / total number of characters in the feature window; After the matching is completed, if there is a matching value greater than the threshold, it is considered that there is music in the database with lyrics similar to the target music; The audio fingerprint comparison module is used to compare the audio of the target music with the control audio in the copyright database; Compare the target music's audio with the reference audio in the copyright database using the following algorithm: Compare the extracted audio fingerprints with the audio fingerprints of the music in the database, extract the feature points with the same code, and form multiple arrays; Calculate the difference between different times in the same array, recorded as ΔT1, ΔT2, ..., ΔTn; When the number of identical ΔTs exceeds the threshold, it is considered that the audio fingerprint is matched successfully and there is music in the database that is similar to the target music audio fingerprint; as well as The judgment module is used to judge whether the target music is infringing according to the comparison result of the comparison module.

5. The music infringement detection system according to claim 4, characterized in that: It also includes a notification module; the notification module is used to generate a report.

6. The music infringement detection system according to claim 4, characterized in that: The audio fingerprint extraction module includes an audio data preprocessing end and an audio fingerprint extraction end; The audio data preprocessing end is used to obtain the audio spectrum of the target music; The audio fingerprint extraction end is used to extract frequency band points in the audio spectrum whose amplitude exceeds the amplitude threshold as feature points, record the amplitude, time sequence, and relationship between adjacent feature points of each feature point, generate an audio fingerprint, and save it.

Images