KR100870265B1 - Combining Hash Technology and Contents Recognition Technology to identify Digital Contents, to manage Digital Rights and to operate Clearing House in Digital Contents Service such as P2P and Web Folder - Google Patents

Abstract

The present invention relates to a content identification method, a copyright DB (Database) management method, a bill payment method, and a settlement method necessary for efficiently performing P2P and web folder monetization. The present invention is used to add various functions to an existing P2P or web folder service to easily perform a monetization. In the past, P2P services have generally made it possible to exchange points between the giver and the receiver when downloading content. However, at present, this method is carried out in the absence of a content identification method that can clearly connect the content with the right holder to pay the fee for each download. Therefore, the biggest issue in the current P2P music fee is the clear music identification method. In general, content provided by an unspecified number is inaccurate and inconsistent, and thus, copyright management should be managed by using a reliable content identification method such as hash or acoustic fingerprint technology. The present invention utilizes only the advantages of both technologies by fusing a hash technology and a content recognition technology, thereby simultaneously providing the processing speed of the hash and the convenience of the content recognition technology. In the existing content recognition technology, the service operator inputs the initial content recognition DB information. However, in the present invention, when the content searched by the unspecified user is not included in the content recognition DB item, the search information itself is newly generated using the inquiry information. By providing items, an efficient content recognition DB construction method is provided, and when content recognition fails (when the same content is not recognized as the same content but is recognized as different content), a new content group is automatically registered to link two or more content recognition DB items. This also provides the effect of achieving 100% recognition accuracy. Existing P2P monetization solution applies copyright inquiry to the content exchange part. This is because all the contents of the shared folder are expensive to use copyright retrieval using the content recognition technology. However, the present invention allows the content recognition technology combined with the hash to reduce the cost per copyright inquiry to the lowest level so that the copyright inquiry can be performed not only for the content exchange but for the entire shared content. Therefore, instead of determining the copyright type at the time of content exchange, it provides a structure that dramatically improves usability by providing copyright type and price information at the time of content search. The present invention also provides a function in which the basic value of the music price is automatically set according to the sound quality.

P2P (Peer to Peer), Web Folders, Hash, Music Recognition Technology, Copyright Management, Content Identification, Billing, Settlement, Monetization

Description

Combining Hash Technology and Contents Recognition Technology to identify Digital Contents, to manage Digital Rights and to operate Clearing House in Digital Contents Service such as P2P and Web Folder}

1 is a system configuration diagram proposed by the present invention to perform a P2P / web folder monetization that the rights organization, P2P / web folder services and monetization solution companies participate. It shows the structure of providing contents identification means to existing P2P / Web folder functions, obtaining copyright information based on this information, processing it accordingly, and paying sales results to rights organizations. 100 and 500 are existing P2P / web folder services, which are partially modified by monetization, and the remaining parts are newly added.

2 is a flowchart illustrating a process of efficiently handling copyright retrieval by combining hash and content recognition technology.

3 is a view for showing a position for copyright inquiry. According to the present invention, the copyright inquiry can be performed on all the shared contents, and thus the existence of copyright and price information can be obtained in the search step.

4 is a conceptual configuration diagram of a copyright information DB. It shows a structure in which hash information and information generated by contents recognition technology are converged and managed.

5 is a flowchart for recording settlement information at the time of selling content as part of a contents settlement system.

6 is a flowchart of a content settlement system.

7 is a flowchart when a copyright information input system is used.

8 is an actual function screen of the copyright information input system.

9 is a diagram conceptually illustrating components required to perform copyright information inquiry and information registration.

10 shows a difference in processing when new content registration is required in a content recognition process.

11 shows a difference between content recognition using partial features and content recognition using all features.

Recently, all music rights organizations, P2P service, and monetization solution companies have been in charge of P2P monetization for about a year, and when P2P monetization is completed, the next step is to monetize web folders. For copyright processing, a clear music identification technology for linking specific music and copyright information has become a major issue of P2P payment. The monetization solution is responsible for providing a clear music identification means to the P2P service to provide a structure that can be settled for rights and all the sales music, the technical field of the present invention is related to the role of the monetization solution.

Means for identifying music have been around for a long time, but not all means can be used for copyright management. Music identification means most conveniently used in P2P retrieval are music file names and ID3 information. Such music identification means can be conveniently used in retrieval. However, they are often misleading, unstandardized and not easy to modify, so they are not suitable for copyright processing and are used only as an auxiliary means in copyright processing.

Currently, music identification techniques that can be used for copyright processing include hashing and music recognition techniques. Hash technologies such as MD5, SHA1, etc. have a very fast processing speed of about 0.1 seconds, but even small modifications are recognized as other files, and it is inconvenient to input copyright information into the modified files again.

Because of the above problems, music recognition technology is attracting attention as the most important factor in P2P pay. Music recognition technology makes it easy to manage copyrights by discriminating between small sound quality variations and the same music. Even if the music provided by the unspecified number is different from the file size, sound quality, etc., the music is discriminated as the same music so that the copyright DB can be easily operated.

However, music recognition technology also has many problems, and existing technologies do not overcome these problems. Since music recognition technology requires much more resources than hash technology and requires time to analyze music, it has a problem of increasing user inconvenience and significantly increasing server operating costs due to a long time for searching copyright. In order to solve the problem of the music recognition technology, a large number of technologies simplify the music recognition technology and use only a part of the music file to recognize the music file. Improving the performance of the recognition technique has many problems.

Another problem of the existing music recognition technology is a content recognition accuracy problem. What people hear and recognize as the same music does not necessarily determine that content recognition technology is the same music. Also, different people have different criteria for listening to music and judging the same music. In fact, in the case of remake or live music, the music is almost the same, but in fact, the neighboring music holders are different and there are many cases in which the user has to pay a fee. Therefore, it is impossible to obtain 100% recognition accuracy using only content recognition technology, and some human intervention is required.

Another problem of the existing music recognition technology is a problem with the music registration method. Existing music recognition technology obtains music information by searching whether there is a part that matches the already registered music with a small feature point in the server operator inputs information about the music and then inquires the music. The problem is that in this way, it is not easy to cope with the failure of music recognition. In order to cope with the failure of music recognition, it is necessary to register the music whose recognition failed again as a new recognition object. The most efficient way to register a new recognition object is to register automatically. Manual registration is too expensive to manage because it is not easy to find the music that is coming out. However, as described above, the existing music recognition technology uses only a part of music when the client is actually recognized, and therefore, information related to the entire music is not uploaded to the server, so that automatic registration by the client is impossible. This lack of post-response capability is a problem as a result of the lack of perfection of the paid solution.

In addition, the existing technology has a problem that the music price or copyright type is not known at the search stage. This problem is caused by the performance of low music recognition technology. If the copyright inquiry is frequently performed, the number of processing of the music recognition is increased and the cost is high. In fact, the music recognition technology is used to decide whether to purchase the music at the time of downloading the music. This is because it tries to apply music recognition, which requires a lot of resources, only at the point of purchase, but it does not know whether it is priced, paid, or free before downloading, which causes a big problem in usability. It's like a clothing store can tell you the price of your clothes only after you try it on. Price and free of charge is information that must be provided in the search step before downloading music.

The present invention was developed to solve the above problems, and the technical problem of the present invention is to integrate content such as music or video by using a hash technology and a content recognition technology in a service environment where an unspecified number of contents such as P2P provides content. By efficiently identifying and linking the identified content and copyright information, it manages copyright, charges the service, and provides the owner with a method of efficiently calculating the content usage.

First, prior to explaining the configuration of the invention in detail the term to be used a lot in the description briefly described as follows.

Content identification technology-Content identification technology is used in this document as a broad concept. It is used to identify the same and different contents of contents such as how to process search terms using file names, how to use ID3 information in music, hash technology, and content recognition technology. It is used as a term that includes all methods that can be used directly or indirectly. In particular, content identification and content recognition are used in different terms. Content recognition is used here as one of the content identification methods. In the present invention, a new content identification technology that maximizes efficiency by fusing content recognition technology and hash technology is provided.

Content Recognition Technology-Content Recognition Technology refers to a technology that determines whether the content is the same based on the content of the content among the content identification technology.

Acoustic Fingerprint, which is a representative example of content recognition technology, is a technology that determines the same and different music by looking at the content of music, that is, the sound waveform of the music. This technique first generates a specific pattern value from the waveform of the music that is not changed by small noise or distortion. Next, find out the degree of distortion between the pattern values generated and the pattern values of the music you want to compare. If the degree of deformation does not make a big difference, and the level can be distinguished by the same music for human hearing, the same music is discriminated.

In general, music recognition technology extracts various feature values from music and stores them in the server, and when the client inquires whether they are the same as the music stored in the server, the server acquires the feature values from a small part of the music and stores the stored values of the server. Query whether there is a match with. This method has problems depending on the application. In the case of a person, if the one-fingerprint matches, the person can be identified, but in an environment where content is provided by an unspecified number of people such as P2P or web folders, it should not be judged the same even if a part matches the feature stored in the server. Because. For example, if there is a music file that exactly matches the front part but the back part is missing, it is not a music file that can be sold when actually paid. Therefore, if a part of a music file matches, it should not be treated as normal music. In general, it is extremely unusual for a person's arm or hand to be exchanged with another person, but fake files created by substituting or cutting out parts of music account for 10% of the music exchanged in P2P.

Since content recognition technology determines whether the content is identical based on the content of the content, there is a great difference in the applied field. For example, music recognition technology cannot be used for image recognition. As described above, the music recognition technology looks at the sound waveform of the music to determine the same and different contents because the image does not exist from the beginning. In this respect, it can be said that there is a big difference from hash technology, which is a technology that can be used regardless of what kind of content.

Content Pattern Value-Refers to a value obtained from the content to be able to recognize the degree to which the content recognition technology is tolerated by a small deformation that can be ignored and different from other content pattern values.

In general, the general content recognition technique is different from the pattern value stored in the server and the value to be retrieved. However, in the present invention, the same information can be used in both cases, so that if there is no matching information, the inquiry information itself can be used as registration information of the server.

Content Pattern ID Value-Refers to an identifier used to uniquely identify a particular group of content that has been identified as equal using a content recognition technique. In some cases, one typical pattern value among the content pattern values belonging to the content group is used, and in other cases, a serial number determined in the order of registration in the GUID (Global Unique IDentifier) or DB that is not significantly related to the content pattern value. There are various forms, such as the back.

The content pattern values often have the same content pattern value even if the sound quality or compression type is slightly different. This is because the content pattern value also has some ability to withstand minor variations. However, if there is more than a certain level of variation, the content pattern value changes. However, there is a range of deformation that a person feels with the same content even if the content pattern value is different depending on the strength of the deformation. If it is within this range, the content recognition technology determines that the contents are the same through a unique comparison method. In this case, contents found to be the same form one identical group and share the content pattern ID value. Therefore, it is normal for several content pattern values to share one content pattern ID value.

Hash Description-A very small summary value that is calculated from data to distinguish it from other data. Regardless of what kind of content it is, most cases can be applied to all digital files. When targeting a music file, processing can be performed 100 to 1000 times faster than normal music recognition technology. However, since the value is changed even with very few variations, it is recognized as a completely different content from the content before the transformation, and thus there is a problem in that it is difficult to manage copyright by using only the hash value.

In order to achieve the object of the present invention, a copyright inquiry step, loading new contents into a copyright inquiry server, returning an inquiry result, a copyright item registration preparation step when a query fails, a step of obtaining a content pattern ID value, a copyright item A copyright information inquiry / registration system is provided that includes a step of registering and completing a copyright inquiry, and a content settlement system including a step of storing sales information, a generation of a sales report, and a generation of royalty settlement information is provided, and the content to be entered / changed. A copyright information input system including an information retrieval step, copyright information input / change step, and hash-based information generation step is provided. Here, the copyright information inquiry / registration system and the copyright information input system are part of a copyright information management system, and are divided into a part that is automatically processed at the time of a copyright inquiry and a part that assists in inputting information that cannot be provided directly by a person. The present invention also provides a copyright information management system including an external service unit, a client unit, a data storage unit, a copyright information inquiry response unit, and a content identification technology convergence information generation unit.

The copyright retrieval step, which is the first step of the copyright information retrieval / registration system, generates a hash value of the content file and proceeds using this value. Since both the processing time on the client and the processing time on the server are extremely short compared to the content recognition technology, in the present invention in which the content recognition technology and the hash technology are combined, 'how much can the use of the content recognition technology' be reduced? Influences. Since all hash values are initially stored in the DB in association with the content pattern ID generated by the content recognition technology, the content having the hash value once used is prevented from using the content recognition technology again. The step of loading new contents to the copyright retrieval server is to cause the copyright retrieval server, which improves performance by using memory caching, to load the latest information when the copyright retrieval server does not have the latest copyright information. The copyright lookup server is a separate server for the purpose of improving performance only. The step of returning an inquiry result is to return corresponding copyright information if there is information on the hash value queried, and to return an inquiry failure if not. If the inquiry fails, the copyright item registration preparation step is a client-side operation, collecting information necessary to register a new item and requesting a registration to the copyright DB management server. The information prepared in this step is information such as song name, singer name, file name, and sound quality necessary for registering hash-based information and content pattern values used by the content recognition technology. Obtaining the content pattern ID value is a step in which the copyright DB management server inquires the content pattern ID value from the content recognition server using the content pattern value received from the client. The registration of copyright items and the completion of the copyright inquiry are performed by the copyright DB management server. The content pattern ID value is returned from the content recognition server to register a new hash value using the information received from the client and the server setting value. The step of completing the inquiry by returning the copyright information input for the pattern ID value or the copyright information set as the default value. As such, the copyright retrieval / registration system can be processed automatically without human intervention, and proceeds in a manner that fuses hash and content recognition technology. In view of the fact that the conventional technologies use only hash or only content recognition technology to identify contents, the efficiency is inferior. Therefore, the fusion of the hash technology and the content recognition technology proposed by the present invention can be called a groundbreaking technology by itself. . Unless otherwise specified, a call to a content-aware technique is not repeated for hash values once processed. Therefore, the number of times a content recognition technique is called does not exceed the total number of hash values.

In the sales information storing step of the content settlement system, when a payment is required for the content from the client, a hash value is obtained and a copyright inquiry is performed to obtain information such as a copyright type and a price, and the user is requested to pay the fee and the fee is charged from the user. After the payment is completed, the payment information is stored in the settlement server based on the hash value. The generation of the sales report is a step of organizing the sales information for each copyright holder by using the content pattern ID value and the copyright holder information connected to the hash value and generating the sales report for settlement. The royalty payment information generation step is a step of generating information that can be used to calculate the royalties to the copyright holders using the stored sales information for each copyright holder. As such, the content settlement system is a system that can generate information that can be used to efficiently calculate the royalties to the copyright holders by using the content pattern ID value connected to the hash value and sales information can be easily collected on a hash basis.

The content information retrieval step of input / change of the copyright information input system is based on information such as a file name, a song name, and a singer name provided by a copyright information manager or an owner of the content, and the like. Or (b) searching for the content to be changed for which the copyright related information is to be changed in units of content pattern IDs and notifying the user of the search result. The search result notified to the user includes not only related information for each searched content but also a link for downloading the actual file for each searched content, thereby providing a function for downloading the actual file. The copyright information input / change step is a step of inputting / changing copyright related information requested by a user in content pattern ID units among search results notified through the above input / change target content information search step. The hash-based information generation step is to allow input / modified copyright related information to be used in copyright inquiry through hash using hash and content pattern ID value connection information. As described above, the copyright information input system easily inputs copyright information in units of content pattern IDs, but this information is an efficient system that can use all hash values associated with content pattern ID values.

In terms of composition, the external service department of the copyright information management system refers to the existing P2P or web folder. The client unit is operated by being inserted as an API (Application Program Interface) to an external service. After receiving a copyright information inquiry request from an external service, the client first obtains a hash value and requests a copyright information inquiry response from the copyright information inquiry response unit to obtain copyright related information. When the received information is returned to the external service and the copyright inquiry fails in response, the content pattern value is generated along with the hash value obtained previously, and the content identification technology convergence information generation unit is contacted again. It serves to return. The data storage unit stores information such as a hash value, a content pattern ID value associated with it, a copyright holder per content pattern ID value (ie, per content group), a copyright type, and the like. The copyright information inquiry response unit uses the data stored in the data storage unit, and if the client requests copyright information inquiry through the hash value, the hash value and the content pattern ID value including the hash value are already linked. It responds with the copyright related information linked to the content pattern ID value and returns a search failure if there is no content pattern ID value associated with the hash value yet. The content identification technology convergence information generator obtains the content pattern ID value by applying the content recognition technology from the content pattern value when the client requests through the hash value and the content pattern value using the data stored in the data storage unit. After adding information that associates the hashed content pattern ID value with the hash, the copyright information corresponding to the content pattern ID value is returned.

Next, embodiments of the present invention will be described in detail with reference to the accompanying drawings. 1 is a system configuration diagram proposed by the present invention to perform a P2P / web folder monetization that the rights organization, P2P / web folder services and monetization solution companies participate. It shows the structure of providing contents identification means to existing P2P / Web folder functions, obtaining copyright information based on this information, processing it accordingly, and paying sales results to rights organizations. 100 and 500 are functions that existing P2P and web folders have, and the API program included in the present invention can be inserted and modified into a form that can be paid. The 200, 210, 220, 300, and 400 are the present invention functions. This is a new part to perform. In fact, because of the different features of P2P and Web Folders, each server and client has a different role. This section is based on P2P and explains the differences of web folders if necessary. In general, in the case of Web folders, the server provides content storage and has little role for the client part. Additions to monetization of existing P2P and web folders are copyright lookup and sales reporting. Both of these features are typically added to the client in the case of P2P, and on the server side in the case of Web folders. The present invention enables copyright retrieval for all shared content regardless of the location of the copyright retrieval. While the existing technology has very low performance of the content identification technology, it allows copyright lookup only in the content exchange stage where a purchase occurs in order to reduce the number of copyright lookups. However, the present invention can perform copyright lookup for all the shared content. Make the information available. If an inquiry about copyright information occurs, the copyright inquiry server 210 processes it. The copyright inquiry server 210 is a hash-based copyright information memory caching server, and the number of processes per second of one server is more than 10,000, and since caching is a main role, it can be easily expanded to several. In fact, if you need high performance, you can easily improve performance by using multiple search servers. When designing the present invention for the first time, since the function is separated from the copyright DB server 200 for the purpose of enabling high speed processing through this server, if there is no need for high speed processing, integration with the copyright DB server 200 is not possible. It is possible. In terms of functionality, the copyright retrieval server 210 is a hash-based server, the copyright DB server 200 has a structure in which hash and content recognition technology are fused, and the content recognition server 220 is operated based on the content recognition function. The copyright DB server 200 is a server holding both D1 (hash based information) and D2 (content pattern ID based information) shown in FIG. As such, since D1 and D2 are fused together, fast processing in hash units is possible, and convenient copyright information management in content pattern ID units is possible. The settlement server 300 is a server that performs a function of receiving and processing a sales report when a sale occurs in a P2P / web folder. This server stores all sales data and organizes them by the copyright holders to perform sales reporting and settlement to the owner or the owner group. The owner server 400 receives the sales data from the settlement server as servers of the rights organization.

2 is a flowchart illustrating a process of efficiently handling copyright retrieval by combining hash and content recognition technology. Existing technologies have incomplete structure where processing cannot be started without providing the operator's data and content recognition fails even if there is a problem such as content recognition failure. It can be done automatically. In addition, the existing technologies have a very slow processing speed and a large operating cost because the fast hash method and the content recognition technology are not fused, but this drawing shows the process of automatically combining the hash and the content recognition technology. Copyright inquiry A1 is performed based on the hash value. The hash value processing speed is 20-70 times faster on the client side and 100-1000 times faster on the server side than content recognition technology. Therefore, how much use of the content recognition function determines the overall operating cost. The present invention minimizes the use of the content recognition function because the content recognition function is used only for the purpose of associating the content pattern ID with all hash values. Currently, the total number of music hashes is estimated at 12 million, and the number of new hashes every day is estimated at less than 100,000. If we calculate the total number of music files to 1 billion and assume new copyright search every day because of copyright information change, 1 billion hash views are processed per day based on hash, and music recognition function is called to perform new hash registration process. Can be counted as 100,000 a day. In the existing technology, if the copyright inquiry is to be processed by using the content recognition function, it should cost 100 to 1000 times. For this reason, in order to minimize copyright inquiries, the existing technology only applies copyright inquiries only when content is purchased. Therefore, if the exchange is not carried out, it is a big inconvenience not to know whether it is paid or free or not. do. At A2, the copyright inquiry server checks whether it is a registered hash value and if it is a registered value, obtains copyright information from the corresponding table and returns it to the client (A3-1, A4). If it is not registered, hash information registration is performed (A3-2). The client extracts information necessary for registering the hash value from the content (B1) and generates a content pattern value for performing the content recognition function (B2) to request a registration to the copyright DB server 200. The copyright DB server 200 requests the content recognition server 220 a content pattern ID value for the corresponding content pattern value. If there is a content pattern ID value to which the corresponding content pattern value belongs (B4-1), the content recognition server generates a new content pattern ID value (B4-2) and returns it. When the content pattern ID value is obtained, the hash value, the hash related information, the content pattern value, the content pattern ID value, and the content pattern ID value related information are linked to each other to use the hash value, the content pattern value, or the content pattern ID value. The structure obtained is completed (B5). These values are stored in the copyright DB server. After the hash registration process, copyright information, which is information related to the content pattern ID value connected to the hash, is returned to the client (A4).

3 is a view for showing a position for copyright inquiry. C1 is a share folder on the giving side, and in the present invention, most of the copyright inquiries have occurred in the shared folder C1, so that the copyright information is held in advance. Therefore, the copyright type or price information may be included in the search result C3 in the search request C2. Existing technologies cannot include copyright type or price information in search results, which causes a problem of large usability. If the file is downloaded after the search (C4), it is stored in the temporary storage (C5), and existing technologies only search for copyright here. Usually, when payment is completed here, it is moved to a shared folder (C6) which is a regular storage. Existing technology is a copyright search in C5 to reduce the number of uses of the content recognition technology, but in the present invention can be a copyright search for all the shared content at any location. The copyright search for all contents in C1 results in a 100 times increase in the number of copyright searches compared to the copyright search in C5 alone. The present invention greatly improves the performance of copyright retrieval to enable copyright retrieval of all contents.

4 is a conceptual configuration diagram of a copyright information DB. It shows a structure in which hash information and information generated by content recognition technology are managed by fusion. The hash related information D1 has a much larger number of records than the content pattern ID value related information D2. The type of music hash value is estimated at 12 million and the content pattern ID value is estimated at about 2 million. Therefore, copyright lookup / registration based on the content pattern ID value has various advantages in terms of the total number of cases, but it is very slow in terms of processing speed. Therefore, when the two pieces of information are completely linked as shown in FIG. 4, two advantages can be used according to the situation.

5 is a flowchart for recording settlement information at the time of selling content as part of a contents settlement system. These treatments can be processed based on fast hash values to maximize efficiency. The hash value is linked to the content pattern ID value, which is necessary for organizing data for each copyright holder, so that the overall processing procedure is processed automatically without any inconvenience. When a purchase occurs, a hash value of the purchased file is obtained (F1), the copyright information is retrieved (F2), and the user is asked whether the purchase is made using the price of the retrieved content (F3, F4). If the purchase is refused, the downloaded content is deleted to prohibit use (F5-1). When the purchase is made, the payment is made (F5-2-1) and the sales information based on the hash value is stored (F5-2-2). The next step continues in FIG. 6.

6 is a flowchart of a content settlement system and illustrates a process in which sales information stored on a hash basis is organized into sales data for each copyright holder. 5 shows a procedure of storing sales information on a hash basis, which corresponds to E1 of FIG. 6. Sales information stored on a hash basis obtains copyright information using the content pattern ID value associated with the hash (E2), and the amount aggregated based on the content pattern ID value (E3) is soon aggregated by the copyright owner (E4). This copyright-specific information is used for settlement later (E5).

7 is a flowchart illustrating a process when a user uses a copyright information input system. Existing technology is a way that the operator can only collect and input the copyright information in advance, while the present invention proceeds in a manner that the operator supplements the specific information in the state that most of the information is automatically registered from the music shared in the P2P. Therefore, the first copyright information input does not proceed in the blank state but in the state where most of the information is collected. Initially, the operator requests a search for a target to enter or change copyright information using a file name, a song name, and a singer name (G1). The copyright information management system searches for input / change target content by content pattern ID based on information provided by the user, and lists the search results based on the content pattern ID value and notifies the user (G2). The operator usually downloads a file directly by using the link information included in the notified search result when it is common to input / change copyright information based on automatically input information (G4), or when the contents of the content need to be actually checked. G3). In practice, this function is performed by pressing the link information in the download column of FIG. The information entered as described above is input based on the content pattern ID value, but since this value is associated with hash values, the information can be automatically searched as a hash value (G5).

8 is an actual function screen of the copyright information input system. This function is primarily used to enter copyright information, but can also be used to change the copyright type (sales method) or change the price.

9 is a diagram conceptually illustrating components required to perform copyright information inquiry and information registration. The external service unit I1 has a meaning as an object for monetization in that the present invention is for monetizing a service such as P2P or web folder. In order for the external service unit to be paid, it is necessary to modify the external service unit. In the present invention, such an modification can be easily performed through the client unit I2 inserted into the external service unit and operated. This is because the external service unit simply needs to request the copyright-related information from the client unit and use the information returned. This is because the external service department can easily obtain the copyright information because the client unit handles the copyright inquiry role internally. The client unit first obtains only the hash value and then sends an inquiry request to the copyright information inquiry response unit I3. If the copyright information inquiry response unit responds with the copyright related information, the inquiry is successful. In this case, the client unit responds by returning the copyright related information obtained from the external service. If a lookup failure is returned, the content further generates a content pattern value used in the content recognition technique, and additionally other information, if necessary. Thereafter, the information is sent together with the hash value previously created and sent to the content identification technology convergence information generation unit I4. Then, when the content identification fusion information generation unit receives the copyright related information in response, the copyright inquiry step is completed by returning the information to the external service. The server portion includes a data storage unit I5, a copyright information inquiry response unit, and a content identification technology convergence information generation unit. First, the data storage unit is associated with copyright such as the hash value used by the copyright information inquiry response unit and the content identification technology convergence information generation unit, the content pattern ID value associated with it, the copyright holder per content pattern ID value (ie, per content group), and the copyright type. A data repository that stores information. Usually, data storage is accessed and managed by DBMS (Database Management System). The copyright information inquiry response unit and the content identification technology convergence information generation unit perform a substantial server operation based on the data storage unit, which is a data repository, and responds to the client's request. When the client inquires copyright-related information through a hash value, the copyright information inquiry response unit stores the hash value stored in the data storage unit, the content pattern ID value associated with it, and the copyright-related information per content pattern ID value (ie, per content group). This is the part that responds by using or caches in advance and responds quickly. The content identification technology convergence information generator is the main part that generates information that fuses the content recognition technology and the hash technology, and finds the content pattern ID value through the content pattern value sent from the client upon request and associates it with the hash value. Information stored in the data storage. Thereafter, it plays a role of sending a copyright related information response to the client through copyright related information per content pattern ID value (that is, per content group) written in the corresponding content pattern ID value.

10 shows a difference in processing when new content registration is required in a content recognition process. This picture is illustrated using music as an example. K1 is a database of existing technologies and usually the operator is responsible for registering new music. The information K4 used when the client inquires (K2) is not the same as the information stored in the server, but may be referred to as partial information. This information can't be used for registration because it doesn't have enough information to register with the server. If it doesn't know which group it belongs to, it simply returns a query failure. On the other hand, in the case of the new technology proposed by the present invention, the format of the information stored in the server DB K5 is the same as the format of the information K9 used when the client inquires (K7). Therefore, if the server item is not found in the query, it registers a new item (K6) and returns the same result as K6 (K8). As such, in the present invention, no lookup failure occurs, so that all hash values have corresponding content recognition values. This new procedure enables the automation of all registration processes, significantly reducing operating costs compared to traditional technologies with high human intervention.

11 shows a difference between content recognition using partial features and content recognition using all features. This picture is illustrated using music as an example. The existing technology of recognizing content using the partial feature makes a content recognition request to the server using the partial feature M4 of the specific content M5 (M3). The result of the request should be different because M2 and M5 are actually different content. However, since the existing technology judges only partial information, M5 is the same as M2. This means that the content recognition technology returns an incorrect result (M6). This is a problem that partial file system cannot avoid. This may not be a big problem depending on the application of the content recognition technology, but it is a big problem in the reliability of the content recognition technology in the digital content of the P2P which partially replaces files. Usually fake files have this problem. It's also similar, like remake or live, but it's also a problem if the copyright holders are different. On the other hand, in the present invention, since the content recognition performance is greatly improved by hash fusion, this problem is eliminated by using the entire file in the situation where the content recognition is directly used. In the present invention, the recognition request (M10) is made using the entire file (M12), and only part of the same M8 is returned as the recognized content (M11).

As described above, the present invention fuses a hash means with excellent processing speed and a content recognition technology that facilitates copyright information management, thereby efficiently processing copyright inquiry, copyright information input and settlement, etc. without a problem of processing speed. . The time to generate the hash value on the client is about 0.1 second and the processing time on the server is 1/10000 second. On the other hand, music recognition technology, which is a content recognition technology used for music, takes about 6 seconds when calculated as a whole file on the client side, and server processing time also shows a big difference from 1/10 seconds to 1/100 seconds. Copyright information management efforts required about 2 million records management using music recognition technology, but more than 12 million records management using hash only. The present invention allows only 2 million records to be managed while having a hash processing speed effect. Since the present invention has the effect of making the speed of the content recognition technology to the hash level, it is possible to look up anywhere without restricting copyright lookup, thereby greatly improving usability. In other words, it is necessary to download the content from the existing technology to find out whether it is paid or free, thereby solving the problem of limiting usability. The present invention has brought qualitative improvements to the content recognition technology itself. In the case of the top-down method in which the operator registers the content in the server, the content retrieval fails if the operator has not registered in advance, but in the bottom-up method, which is automatically registered upon the appearance of new content of the present invention, the content is recognized. Failure does not occur, and even if new content appears, the default value is automatically registered, so that most copyright management can be handled automatically, so that copyright retrieval failure does not occur. The present invention minimizes human intervention in all procedures. In the case of music, the music price has a structure in which a default value can be automatically registered according to sound quality. In fact, all copyright management is automatically performed except for inputting individual copyright holder information required at the time of settlement. Information such as song name, singer name, etc. is extracted from a content file such as music and inputted, and the operator has a procedure to engage when there is a need to correct wrong information.

Claims (18)

Generating a hash value at the client to perform copyright inquiry on the copyright inquiry server; Loading the changed contents from the copyright DB management server to the copyright inquiry server if the contents of the copyright inquiry server and the copyright DB management server which perform memory caching are different; If the copyright inquiry server has an item corresponding to the inquired hash value, returns copyright-related information such as copyright type information and content price of the corresponding content and completes the inquiry, and returns a query failure when there is no DB item; Requesting registration by the client generating a content pattern value and sending the information describing the content such as a song name, artist name, genre, and sound quality to the copyright DB management server when the inquiry failure is returned from the server; Inquiring, by the copyright DB management server, of the content pattern ID value from the client using the content pattern value received from the client; The copyright DB management server, which has returned the content pattern ID value from the content recognition server, registers a new hash value using the information received from the client and the server setting value, and sets the copyright information or default value previously inputted for the content pattern ID value. Returning copyright information to complete the inquiry; To Including copyright information inquiry / registration system. Generating a hash value at the client to perform copyright inquiry on the copyright inquiry server; Contents of a copyright inquiry server and a copyright DB management server that perform memory caching are different, and loading the changed contents from the copyright DB management server to the copyright inquiry server; If the copyright inquiry server has an item corresponding to the inquired hash value, returns copyright-related information such as copyright type information and content price of the corresponding content and completes the inquiry, and returns a query failure when there is no DB item; If the inquiry failure is returned from the server, the client ignores it and performs processing based on the default value; To Copyright information inquiry / registration system that includes and uses only the information accumulated by other systems without the role of copyright information registration. The copyright retrieval server according to claim 1, wherein the copyright retrieval server acts as a copyright retrieval server instead of a copyright retrieval server, which serves to cache and quickly respond to copyright-related information to be responsive. Copyright information inquiry / registration system, characterized in that. The method according to claim 1, when the content recognition server receives a request for querying the content pattern ID value from the copyright DB management server, if the content pattern ID value to which the content pattern value belongs belongs to the DB, the corresponding content pattern ID value is returned. And generating a pattern ID value and returning the value. The method according to claim 3, when the content recognition server receives a request for inquiry of the content pattern ID value from the copyright DB management server, if the content pattern ID value to which the content pattern value belongs belongs to the DB, the corresponding content pattern ID value is returned. And generating a pattern ID value and returning the value. The method of claim 1, 3, 4, or 5, wherein the content recognition server is requested to query the content pattern ID value, but the server fails to successfully search the content pattern ID value due to an error of the server or a limitation of the content recognition technology. Notifying the requesting party of the failure to retrieve the content pattern ID value; Responding to the client for copyright processing using the basic copyright information which is intended to be used at the time of failure of the copyright DB management server which has received the inquiry failure response of the content pattern ID value, instead of responding to the client as a simple inquiry failure; To Copyright information inquiry / registration system further comprises. The copyright DB management server does not perform the inquiry of the content pattern ID value by the content pattern value, but the client uses the content pattern value in advance to request the content pattern ID value to the content recognition server according to claim 1 or 3 or 4 or 5. Copyright information inquiry / registration system, characterized in that by sending the information that describes the content, such as song name, artist name, genre, sound quality, along with the content pattern ID value retrieved and returned to the copyright DB management server. The method according to claim 1 or 2 or 3 or 4 or 5, wherein the target content is a music file, the content pattern value is generated from the content, and the content pattern ID value corresponding to the specific content pattern value is determined. A copyright information inquiry / registration system, characterized by using music recognition technology that looks at the contents of music, that is, the sound waveform of the music, and determines the same and different contents. The sound of the video file in claim 1 or 2 or 3 or 4 or 5, wherein the target content is a video file, the content pattern value is generated from the content, and the content pattern ID value corresponding to the specific content pattern value is determined. Copyright information inquiry / registration system, characterized in that the use of music recognition technology to determine the same or different content of the contents, that is, the sound waveform of the video. The music video file according to claim 1 or 2 or 3 or 4 or 5, wherein the target content is a music video file, the content pattern value is generated from the content, and the music video file is determined in determining a content pattern ID value corresponding to the specific content pattern value. Copyright information inquiry / registration system, characterized in that using the music recognition technology to determine the same or different content of the content contained in the sound portion of the music video, that is, the music video. The copyright information inquiry / registration system according to claim 1, 2, 3, 4, or 5, wherein copyright inquiry / registration is performed on all shared contents in advance. The copyright information inquiry / registration system according to claim 1 or 3 or 4 or 5, wherein the music price is automatically determined and registered based on the sound quality value in the process of automatically registering the hash information. Claim information according to claim 1 or 3 or 4 or 5, not only the hash information registration is automatically performed, but also includes a function of registering new information manually or replacing existing information. Inquiry / Registration System. The method according to claim 1 or claim 3 or claim 4 or claim 5, if the same content has a different content pattern ID value, there is a function that can be assigned to the same content pattern ID value manually so that the same content is recognized as different content Copyright information inquiry / registration system characterized in that it is possible to correct errors. Claim 1 or claim 3 or claim 4 or claim 5 when the content pattern value to be searched does not belong to the content pattern ID to which the new content pattern ID value can be registered based on the content pattern value itself. A copyright information inquiry / registration system characterized by preventing a situation where there is no content pattern ID. If the client needs to pay for the content, the hash value is obtained and then the copyright inquiry is used to obtain information such as the copyright type and price, and the user is requested to pay the fee, and when the payment is completed from the user, the settlement is based on the hash value. Storing the sales information at the server; Using the content pattern ID value and copyright holder information linked to the hash value, arranging the sales information for each copyright holder and generating a sales report for settlement; Generating data for paying a royalty fee to the copyright holders using the stored copyright holder's sales information; To And a content settlement system that can provide real-time sales reports online. Searching for input / change target content in units of content pattern IDs based on information provided by the user and notifying the user of the search result; Inputting / changing copyright-related information requested by a user for input / change in units of content pattern IDs; Allowing input / modified copyright related information to be used in copyright inquiry through hash using hash and content pattern ID value linking information; To And a copyright information input system according to the content pattern ID value unit, characterized in that the actual file download function is provided for each content found in the notified search result of the first step of the above three steps. External service units such as P2P or web folders; It is operated by being inserted as an API (Application Program Interface) to an external service. After receiving a copyright information inquiry request from an external service, it first obtains a hash value and requests a copyright information inquiry response from a copyright information inquiry responder. In case of returning to the external service and receiving copyright failure as a response, it is necessary to generate the content pattern value along with the hash value obtained previously and to contact the content identification technology convergence information generation unit again and return the resulting copyright related information to the external service. A client unit characterized in that; A data storage unit for storing information such as a hash value, a content pattern ID value associated with the content, a copyright holder per content pattern ID value (ie, per content group), a copyright type, and the like; If the client requests to search for copyright information through the hash value using the data stored in the data storage, if the requested hash value is already associated with the content pattern ID value, the copyright related information is used using the content pattern ID value. A copyright information inquiry response unit which responds and returns an inquiry failure if there is no content pattern ID value associated with the hash value yet; When the client requests through the hash value and the content pattern value using the data stored in the data storage unit, the client obtains the content pattern ID value from the content pattern value and adds the information to associate the hash with the found content pattern ID value. A content identification technology fusion information generation unit for responding to copyright information corresponding to the content pattern ID value; To Including copyright information management system.

Images