JP2022073872A - System, method, and program for extracting infringement information - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system, a method, and a program for extracting infringement information that extract infringement information which possibly and automatically infringes on the right of a target intellectual property.

SOLUTION: An infringement information extracting system is provided with an infringement information extracting server 10 that comprises: obtaining right information 22 indicating the right of a target intellectual property on the basis of a registration number; generating a model 20 by machine leaning learning data 18 in which data indicating the right of another intellectual property is processed; and analyzing the right information 22 indicating the right of the target intellectual property by the model 20 and extracting infringement information 28 which possibly infringes on the right of the target intellectual property, from a search result 26 obtained by a character search or an image search, using determination result 24 obtained by determination of a character or an image which is necessary for the search.

SELECTED DRAWING: Figure 2

COPYRIGHT: (C)2022,JPO&INPIT

Description

The present invention relates to an infringing information extraction system, a method and a program, and more specifically, to a technique for automatically extracting infringing information that may infringe the rights of the target intellectual property.

In recent years, techniques dealing with patents, trademarks, designs, and utility models have attracted attention. For example, there is a technique of machine learning the examination results examined in the past to generate correction information that increases the possibility of acquiring a right (Patent Document 1).

In addition, there is a technique of machine learning intellectual property information to calculate the possibility of acquiring intellectual property rights (Patent Document 2).

Japanese Unexamined Patent Publication No. 2019-179493 Japanese Unexamined Patent Publication No. 2019-179578

Even if you have acquired intellectual property and built a competitive advantage, if you leave the competitors who infringe the intellectual property rights unchecked, you should do business while maintaining the competitive advantage. I can't. Counterfeiting investigations are necessary to identify competitors that infringe intellectual property, but counterfeiting investigations of intellectual property require specialized skills and experience. There are many companies that cannot protect their competitive advantage because there is a problem that it takes a lot of labor and cost to conduct a counterfeit investigation using experts, and there is a problem that there is a shortage of experts in the first place. Therefore, infringement information that may automatically infringe the rights of the target intellectual property in order to protect the competitive advantage of the company by reducing labor and cost without the need for specialists. The technique of extracting is effective.

However, in the techniques of Patent Document 1 and Patent Document 2, the examination results examined in the past are machine-learned to generate correction information that increases the possibility of acquiring rights, and the intellectual property information is machine-learned to know. Although it is possible to calculate the possibility of acquiring the right of the target intellectual property, it is not possible to provide the technology to automatically extract the infringement information that may infringe the right of the target intellectual property.

Therefore, it is an object of the present invention to provide an infringing information extraction system, method and program for automatically extracting infringing information that may infringe the rights of the subject intellectual property.

The present invention obtains right information indicating the right of any one of the target patent, utility model, design and trademark from the registration number, and searches from the analysis means for analyzing the right information and the result of the analysis. Judgment means for determining the characters or images necessary for the determination, search means for searching with a search engine using the determined characters or images, and the target patents, utility models, and designs based on the search results. And an infringement information extraction system comprising an extraction means for extracting infringing information that may infringe any one of the rights of a trademark.

Further, in the present invention, from the step of acquiring the right information indicating the right of any one of the target patent, utility model, design and trademark from the registration number and analyzing the right information, and from the result of the analysis. From the step of determining the character or image necessary for the search, the step of searching with the search engine using the determined character or image, and the result of the search, the target patent, utility model, design and Provided is an infringement information extraction method comprising a step of extracting infringing information that may infringe any one of the rights of a trademark.

Further, the present invention comprises a step of acquiring a right information indicating the right of any one of a target patent, a utility model, a design and a trademark from a registration number to a computer and analyzing the right information, and the above-mentioned analysis. From the results, the step of determining the character or image necessary for searching, the step of searching with the search engine using the determined character or image, and the target patent or utility model from the search result. It provides a step of extracting infringing information that may infringe any one of the rights of a design and a trademark, and an infringing information extraction program for executing the step.

According to the present invention, the right information indicating the right of any one of the target patent, utility novel, design and trademark is acquired from the registration number, the right information is analyzed, and the search is performed from the analysis result. Determine the characters or images necessary for the purpose, search with a search engine using the determined characters or images, and from the search results, the right of any one of the applicable patents, utility novels, designs and trademarks. It was decided to extract infringing information that may be infringing.

In this way, there is an effect that infringing information that may infringe the rights of the target intellectual property can be automatically extracted.

It is a conceptual diagram which shows the whole structure of the infringement information extraction system of one Embodiment of this invention. It is a block diagram which shows the hardware configuration of the infringement information extraction server of the said embodiment. It is a block diagram which shows the functional structure of the infringement information extraction server of the said embodiment. It is a block diagram which shows the structure of the user terminal of the said embodiment. It is a figure which shows an example of the learning data of the said embodiment. It is a figure which shows another example of the learning data of the said embodiment. It is a figure which shows another example of the learning data of the said embodiment. It is a figure which shows an example of the synonym dictionary of the said embodiment. It is a figure which shows an example of the dictionary of the said embodiment. It is a figure which shows an example of the screen for acquiring the right information which shows the right of the target intellectual property which is displayed on the user terminal of the said embodiment. It is a figure which shows an example of the screen displayed on the user terminal of the said embodiment. It is a flowchart which shows an example of the infringement information extraction processing by the said embodiment.

The present invention is to acquire the right information indicating the right of any one of the target patent, utility model, design and trademark from the registration number, analyze the right information, and search from the analysis result. Determine the required characters or images, search with a search engine using the determined characters or images, and extract infringing information that may infringe the rights of the target intellectual property from the search results. Is what you do. Hereinafter, the best mode for carrying out the present invention will be described in detail based on examples.

<Basic Concept> ... First, the basic concept of the infringing information extraction technique of the present invention will be described. The present invention relates to an artificial intelligence technique that enhances the accuracy of extracting infringing information that may infringe the rights of an intellectual property subject to artificial intelligence. Specifically, the rights information indicating the rights of the target intellectual property (for example, the publication at the time of registration) is analyzed, and the characters or images necessary for searching from the analysis results (for example, the independent item of the patent) are analyzed. Judgment (such as the character described in the above, the image of the solid line and the broken line of the design, etc.), search with the search engine (character search or image search) using the determined character or image, and from the result of the search, Extract infringing information that may infringe the rights of the target intellectual property. For example, increasing the accuracy of extraction means that the infringing information to be extracted becomes infringing information to a high degree of usefulness.

Here, a model was generated by machine learning the training data (rounded image data or text data) processed by processing the data indicating the rights of other intellectual property (publication at the time of registration), and was generated in the analysis. The model may be used to analyze rights information indicating the rights of the target intellectual property. For example, in order to machine-learn image data to generate a model, deep learning, which is a method of machine learning, may be performed. Further, the synonyms of the determined characters may be specified and the search may be performed including the specified synonyms, or the determined characters may be translated into a foreign language and the search may be performed using the translated characters.

Hereinafter, the best mode for carrying out the present invention will be described in detail based on examples.

<Overall configuration> ... FIG. 1 is a conceptual diagram showing the overall configuration of the infringement information extraction system according to the present embodiment. As shown in FIG. 1, the infringement information extraction system 100 includes an infringement information extraction server 10 and a user terminal 82 of a user 80. These are capable of mutual data communication via a network such as the Internet (not shown). The user 80 may include members constituting the group described later and other persons.

As shown in FIG. 1, first, the infringement information extraction server 10 inputs the right information (registration number, publication at the time of registration, etc.) indicating the right of the target intellectual property by the user terminal 82. Get from. For example, the right information may be acquired by the registration number input by the user 80, or the content input by the user 80 may be acquired as the right information as it is (for example, the patent claim input by the user 80). Range etc.).

Further, the infringement information extraction server 10 machine-learns the learning data obtained by processing the data indicating the rights of other intellectual property (publication at the time of registration, etc.) as necessary (step S2), and generates a model. It may be (step S3). Specifically, the feature parts (keywords, feature points of figures, etc.) of the image showing the rights information of various intellectual property rights (patent rights, utility model rights, design rights, trademark rights) are rounded and processed. Dozens to tens of thousands of the trained data 18 are prepared, and the trained data 18 is machine-learned to generate a model. For example, in order to recognize what kind of field and what kind of content, the circle is performed on both the field information indicating the field and the content information indicating the content. For example, in order to machine-learn image data to generate a model, deep learning, which is a method of machine learning, may be performed.

Field information indicating a field is IPC classification, FI, F-term, etc. in the case of a patent, product / service classification in the case of a trademark, design classification, an article related to a design, etc. in the case of a design. Is. Utility models are the same as patents.

In addition, the content information indicating the content is a keyword of an independent term, a keyword of a dependent term, etc. in the case of a patent, a character or a figure of a registered trademark in the case of a trademark, and a character or a figure of a registered trademark in the case of a design. Solid lines, broken lines, etc. in the drawing. Utility models are the same as patents.

As the learning data, image learning data is preferable because it can be recognized corresponding to both characters and figures, but it does not prevent the use of text learning data.

When the above training data is machine-learned, a model customized for the training data is generated. That is, a machine-learned model is generated in which part of the gazette has field information, what kind of characters are described, which part has content information, and what kind of character or figure is. Will be done.

Next, the infringement information extraction server 10 analyzes the right information indicating the right of the target intellectual property input by the user 80 by the user terminal 82 (step S4). This analysis is performed, for example, to understand what is described in the publication. For example, the publication at the time of registration may be image-captured and image-analyzed, or the publication at the time of registration may be used as characters for text mining. In addition, text data including rights information indicating the rights of intellectual property may be acquired and analyzed.

Since field information indicating the field and content information indicating the content are described in a fixed place in the publication, analysis is possible without requiring machine learning, but in order to improve the accuracy of the analysis, the above The analysis may be performed using the model generated in step S3. For example, when analyzing using a model, if you enter an image (rights information) that shows the rights of the target intellectual property, which part of the rights information (keywords or graphic features that show field information or content information) Analyze whether to extract points etc.). For example, in the case of text data including rights information indicating the rights of intellectual property, it is analyzed where the rights information indicating the rights of intellectual property is located in the text data.

Then, the infringement information extraction server 10 determines a character or an image necessary for searching from the result of the analysis (step S5). Specifically, a search engine determines which character or image is required for a character search or image search. The characters or images required for the search are, for example, the field information determines the characters and the content information determines the characters or figures. Specifically, if it is field information, the characters of the IPC classification of the patent and the characters of the goods / services of the trademark are determined. If it is content information, the characters described in the independent section of the patent, the solid line and the broken line image of the design, etc. are judged. The determined characters or images may be changed (selected, added, deleted, etc.). For example, the determined character or image may be changed (selected, added, deleted, etc.) by the user.

Next, the infringement information extraction server 10 searches with a search engine using the determined characters or images (step S6). Using the characters or images determined in step S5, a character search or an image search is performed by a known search engine. For example, a search engine may search using characters or images modified by the user.

Next, the infringement information extraction server 10 extracts infringing information that may infringe the rights of the target intellectual property from the search results (step S7). Specifically, infringing information is extracted from the contents such as Web pages / videos / images displayed in the search results. It should be noted that these contents may be analyzed to extract infringing information, may be randomly extracted without analysis, or all may be extracted without analysis. For example, when the accuracy of analysis is high, it can be expected that the accuracy of extraction after the final reinforcement learning will be higher if the infringing information is extracted by analyzing the contents such as Web pages / videos / images.

In other words, the higher the accuracy of extraction means that the infringing information to be extracted becomes the infringing information to a high degree of usefulness.

In this way, it is possible to automatically extract infringing information that may infringe the rights of the target intellectual property.

<Hardware Configuration of Infringement Information Extraction Server> ... FIG. 2 is a block diagram showing a hardware configuration of the infringement information extraction server 10 of the present embodiment. The infringement information extraction server 10 includes, for example, a processor 12, a memory 14, a storage 16, and a communication unit 34, which are connected by a bus (not shown). The processor 12 is configured by, for example, a CPU (Central Processing Unit), and performs various processes by reading and executing various programs stored in the memory 14.

The memory 14 stores a program executed by the processor 12, and is composed of, for example, a ROM (Read Only Memory) or a RAM (Random Access Memory). For example, various means shown in FIG. 3 are stored. The storage 16 stores, for example, learning data 18, model 20, right information 22, determination result 24, search result 26, infringement information 28, dictionary 30, group information 32, a control program (not shown), and the like. The communication unit 34 performs data communication with the user terminal 82 and other external devices via the network.

The learning data 18 stored in the storage 16 is, for example, rounded with data (for example, a gazette at the time of registration) indicating various intellectual property rights (patent rights, utility model rights, design rights, trademark rights). It is processed by For example, tens to tens of thousands of training data 18 are prepared, and the training data 18 is machine-learned to generate a model 20. For example, in order to recognize what kind of field and what kind of content, the circle is applied to both the field information indicating the field and the content information indicating the content. For example, in order to machine-learn image data to generate a model, deep learning, which is a method of machine learning, may be performed.

FIG. 5 is a diagram showing an example of learning data related to field information indicating a field, (A) is learning data 18A related to patent rights, (B) is learning data 18B related to design rights, and (C) is learning related to trademark rights. An example of data 18C is shown. If it is a patent, the field information indicating the field is IPC classification, FI, F-term, or the like. In the example of FIG. 5A, the IPC classification and the FI term are rounded with the surrounding line C. The same applies to utility models. Further, the field information indicating the field is, if it is a design, a design classification, an article related to the design, etc., and in the example of FIG. 5B, the article related to the design, the design classification, the international design classification, and the D term are enclosed. It is rounded with line C. In addition, the field information indicating the field is, in the case of a trademark, the classification of goods / services, etc., and in the example of FIG. It is rounded.

FIG. 6 is a diagram showing an example of learning data 18A relating to content information indicating the content of the patent. In the case of a patent, the content information indicating the content of the right is a keyword of an independent term, a keyword of a dependent term, or the like in the claims. In the example of FIG. 6, an image showing the scope of claims is used as the learning data. Then, the image of the right range is rounded and processed to be the learning data 18A. As shown in FIG. 6, the circle is performed by enclosing the keyword with a surrounding line C. In the illustrated example, the box C indicates "person", "certify", "image", "validity", "simultaneous", "certificate with photo", "face", "user", and "match". Is surrounded by a box line C as a keyword. The same applies to utility models.

Further, machine learning may be performed by associating the IPC classification with the learning data. Further, in addition to the keyword, an image in which the part corresponding to the keyword is rounded in the drawing that best reflects the scope of the patent right or the utility model right may be added as learning data.

FIG. 7 is a diagram showing an example of learning data 18B relating to content information indicating the content of the design. The content of the design right is determined by what is shown in the drawing, and the image of the drawing is used as the learning data. Then, the featured portion of the image in the right range is rounded and processed to be the learning data 18B. FIG. 7 is an example of a partial design, and the rounding is, for example, the entire "front and plane side perspective view", that is, the portion including the broken line portion surrounded by the surrounding line C1 and the portion to be registered as a design. That is, it is performed by surrounding the solid line portion with the surrounding line C2.

In addition, although only the "front and plane side perspective views" are shown in FIG. 7, the same rounding may be performed for other drawings. Further, in addition to the rounding of the drawing, the keyword may be rounded from "Explanation of design" to obtain the learning data 18B. Further, the classification of the design may be associated with the learning data for machine learning.

In the case of a character trademark, the learning data regarding the content of the trademark right is obtained by rounding the keywords in the same manner as the patent right and the utility model right. For example, the entire character of the trademark may be rounded as one keyword, or in the case of a combined trademark consisting of two or more words, in addition to rounding the entire trademark as a keyword, each word constituting the trademark may be rounded. It may be rounded as a keyword. Further, in addition to the above keywords, machine learning may be performed by associating the division with the learning data.

For trademarks other than character trademarks, images can be used in the same way, and the entire image can be circled and enclosed as a feature part, or the image part combined with characters can be circled and enclosed as a feature part to prepare learning data. good.

As the learning data, image learning data is preferable because it can be recognized corresponding to both characters and figures, but it does not prevent the use of text learning data.

The model 20 stored in the storage 16 is generated by machine learning the learning data 18, and is used for analysis of rights information indicating the rights of the target intellectual property acquired from the user 80.

The right information 22 is right information (registration number, claims, right holder information, etc.) indicating the right of intellectual property input from the user 80. For example, in the case of a patent right, the scope of claims, right holder information, IPC classification, etc. are stored in association with the registration number. The same applies to utility model rights.

If it is a design right, drawings, right holder information, classification, etc. are associated with the registration number and stored. If it is a trademark right, the trademark, right holder information, classification, etc. are associated with the registration number and stored.

The determination result 24 is determined as a character or an image necessary for searching for extracting infringement information of the target intellectual property input from the user 80. The determined characters or images may be changed (selected, added, deleted, etc.). For example, the determined character or image may be changed (selected, added, deleted, etc.) by the user.

The search result 26 is a result of a search (character search or image search) by a search engine using the character or image which is the determination result 24. Further, the search result 26 may be a result searched by a search engine using characters or images modified by the user.

The infringement information 28 is obtained by extracting infringement information that may infringe the rights of the target intellectual property from the search results 26 searched by using a search engine based on the judgment result 24. be. Since many contents such as Web pages / videos / images displayed in the search results are deleted, the extracted infringement information may be downloaded and stored in association with the downloaded contents and the date and time. .. Further, the infringement information is not limited to one, and when a plurality of infringement information is hit by a search, the plurality of infringement information is stored in association with the intellectual property right input from the user 80. In addition, since the storage cost becomes enormous if all the extracted infringement information is downloaded, for example, a save button is installed to save only the infringement information that the user presses the save button, and the infringement in which the save button is not pressed is not pressed. The information may be discarded after a certain period of time. On the contrary, only the URL of the extracted infringing information may be provided to the user, and only the infringing information for which the user presses the save button may be downloaded and saved.

The dictionary 30 is a synonym dictionary for specifying a synonym of a character which is a judgment result 24, or a foreign language for translation for translating the judged character into another language (for example, English, Chinese, Spanish, etc.). Includes language dictionaries, etc.

FIG. 8 is a diagram showing an example of a synonym dictionary for specifying a synonym of a character which is a determination result 24. In the synonym dictionary 30A shown in FIG. 8, "person" is a synonym for the judged character "principal", "judgment / rating / assessment" is a synonym for the judged character "certification", and "judgment / rating / assessment" is a synonym for the judged character "image". "Video / screen" is described.

」が記載されている。 FIG. 9 shows an example of a dictionary 30B for translating a character whose determination result 24 is, for example, from Japanese to a foreign language (English and Chinese in the illustrated example). In the dictionary 30B, "photo" is described as English for the judged character "photograph", "terrestrial" is described as Chinese, "yourself" is described as English for the judged character "person", and "this person" is described as Chinese. "Photograph" as English and "photograph" as Chinese for the judged character "certify"

"Is described.

Such a dictionary 30 may be on the infringing information extraction server 10, or may be linked with an external dictionary.

The group information 32 is information about a group composed of predetermined members. A group is composed of predetermined members such as a company's planning and management department, intellectual property department, XX team, and outside experts, and the combination of members can be adjusted. The extracted infringement information 26 is shared within the group.

<Functional Configuration of Infringement Information Extraction Server> ... FIG. 3 is a block diagram showing a functional configuration of the infringement information extraction server 10. The infringement information extraction server 10 includes analysis means 40, generation means 42, determination means 43, search means 44, extraction means 46, synonym identification means 48, translation means 49, task management means 50, notification means 51, download means 52, and storage. The means 54, the group creating means 56, the sharing means 58, the right holder information acquisition means 60, the grounds specifying means 62, the keyword receiving means 64, the management means 66, the output means 68, the calculation means 70, and the sorting means 72 are provided.

The analysis means 40 analyzes the right information indicating the right of the target intellectual property. This analysis is performed, for example, to understand what is described in the publication. For example, the publication at the time of registration may be image-captured and image-analyzed, or the publication at the time of registration may be used as characters for text mining. As for the right information indicating the right of the target intellectual property, for example, the right information of the right corresponding to the registration number may be acquired from the external database by the registration number input by the user 80, or by the user 80. The entered right scope as it is (for example, the wording of the scope of the patent claim, the drawing of the design, etc.) may be used as the right information. In addition, text data including rights information indicating the rights of intellectual property may be acquired and analyzed. In the case of text data including rights information indicating the rights of intellectual property, the location of the rights information indicating the rights of intellectual property is analyzed in the text data.

For the input by the user 80, the analysis means 40 displays a screen for causing the display unit 92 (see FIG. 1) of the user terminal 82 to input the right information indicating the right of the target intellectual property via the network. Then, the user 80 who sees the screen performs the operation by the input unit 90. By transmitting the input content to the infringement information extraction server 10 via the communication unit and the network, the infringement information extraction server 10 acquires the right information.

FIG. 10 shows an example of the screen 110 displayed on the display unit 92 of the user terminal 82. On the screen 110, a pull-down menu 112 for selecting patents, trademarks, designs, utility models, etc., a registration number input field 114, an IPC classification input field 116A for patents and utility models, and a classification input field 116B for design classification , The classification input field 116C for trademark classification is displayed.

In addition, a keyword input field 118 for customizing the judgment accuracy by adding keywords, a right holder information input field 120 for collectively acquiring the registration number of the intellectual property right owned by the designated right holder, and a registration gazette. The registration gazette input field 122 for directly inputting is displayed. 118 to 122 in these input fields will be described later.

The generation means 42 generates a model by machine learning the learning data obtained by processing data indicating the rights of other intellectual property (publication at the time of registration, etc.). For example, in order to machine-learn image data to generate a model, deep learning, which is a method of machine learning, may be performed. Specifically, tens to tens of thousands of learning data 18 of images obtained by rounding the images of publications at the time of registration of various intellectual property rights are prepared, and the learning data 18 is machine-learned. Generate model 20. For example, in order to recognize what kind of field and what kind of content, the circle is applied to both the field information indicating the field and the content information indicating the content.

For the learning data 18, for example, for the patent right, as shown in the learning data 18A of FIG. 5A, the IPC classification, FI, F term, etc. are circled with a circle line C as field information indicating the field. And, as shown in the learning data 18A of FIG. 6, as the content information indicating the content, the keyword of the independent term, the keyword of the dependent term, etc. in the range of the claim, which are circled with a circled line C, are used. To. The same applies to character trademarks. The generation means 42 performs machine learning by associating the IPC classification with the learning data 18A for patents and practical novels, and machine learning by associating the classification with the learning data for character trademarks. In addition, keywords and their synonyms extracted by rounding of patents and character trademarks may also be machine-learned together.

As for the design right, as shown in the learning data 18B of FIG. 5B, the field information indicating the field is circled with a box line C such as a design classification or an article related to the design. As shown in the training data 18B of No. 7, as the content information indicating the content, the entire image and the characteristic graphic portion in the image are rounded with the surrounding lines C1 and C2. The figure (characteristic point) extracted by the surrounding lines C1 and C2 and the similar figure (similar feature point) may also be machine-learned together. For the graphic portion of the graphic trademark or the combined trademark in which the graphic and the character are combined, the rounded portion and the feature points similar to the portion may be collectively machine-learned. The generation means 42 may perform machine learning by associating the classification of the design with the learning data 18B, or may perform machine learning by associating the classification of the trademark with the learning data of the trademark right.

The circles shown in FIGS. 5 to 7 are examples, and the feature parts (characteristic points such as keywords and figures) of the image showing the scope of rights are highlighted or separated from other parts by color coding. May be good. Of course, these processing examples are also examples, and these may be combined, or may be appropriately changed within the range in which the same effect is obtained.

When the learning data 18 as described above is machine-learned, a model 20 customized to the learning data 18 is generated. That is, a model 20 is generated in which the field information is found in which part of the publication, what characters are described, the content information is found in which part, and what kind of character or figure is learned. Will be done. The generated model 20 is stored in the storage 16.

The analysis means 40 may use the generated model 20 to analyze rights information indicating the rights of the target intellectual property. For example, an image of rights information showing the rights of the target intellectual property (an image showing the rights information of the intellectual property rights held by the user 80) is input to the model 20, and which part is extracted from the rights information is selected. To analyze. The extracted part is, for example, a characteristic part such as a keyword or a figure.

The determination means 43 determines a character or an image necessary for searching from the analysis result of the analysis means 40. Specifically, a search engine determines which character or image is required for a character search or image search. The characters or images required for the search are, for example, the field information determines the characters and the content information determines the characters or figures. Specifically, the characters described in the independent section of the patent and the images of the solid line and the broken line of the design are judged. The determination result 24 is stored in the storage 16. The determined characters or images may be changed (selected, added, deleted, etc.). For example, the determined character or image may be changed (selected, added, deleted, etc.) by the user. The changed result may be stored in the storage 16.

The search means 44 uses a search engine to perform a search using characters or images determined by the determination means 43. The search result 26 is stored in the storage 16. For example, a search engine may be used to search using characters or images modified by the user, or the results may be stored as the search result 26 in the storage 16.

The extraction means 46 extracts infringing information that may infringe the rights of the target intellectual property from the results of the search by the search means 44. Specifically, infringing information is extracted from the contents such as Web pages / videos / images displayed in the search results. It should be noted that these contents may be analyzed to extract infringing information, may be randomly extracted without analysis, or all may be extracted without analysis. For example, when the accuracy of analysis is high, it can be expected that the accuracy of extraction after the final reinforcement learning will be higher if the infringing information is extracted by analyzing the contents such as Web pages / videos / images. The extracted infringement information 28 is stored in the storage 16.

When searching with a search engine, for example, noise that does not become infringing information such as academic papers or the right holder's own product sales site may be picked up. Therefore, the extraction means 46 has a filter function.

For example, when removing noise by keyword, if you search by a search engine, the page title will be displayed on the search list page, so if the page title contains academic or sales keywords, it will be excluded and noise will be removed. Perform removal.

Alternatively, a URL list to be excluded may be created, and if the URL in the list is applicable, the URL may be excluded from the extracted infringing information. Of course, it does not prevent noise removal by other methods.

FIG. 11 shows how the details of the infringement information 136 are shown on the screen 150 when the user 80 selects the infringement information 136 from the extracted plurality of infringement information 132 to 140. On the screen 150, the URL 152 of the infringement information page, the infringement information page 154, the sentence 156, the basis part 158 in which the part that is the basis of the infringement in the sentence 156 is shown by underline, bold, highlight, etc., and the like button. 160, the bad button 162 is displayed. It is just an example and may not be arranged as shown in FIG. For example, the like button 160 and the bad like button 162 may be arranged in the list of infringement information on the left side of FIG. For example, if the shown infringement information 136 is downloaded information, the URL 152 of the infringement information page may not be shown.

The synonym specifying means 48 identifies the synonyms of the characters determined by the determining means 43, and the search means 44 may search by a search engine using the specified synonyms. For the identification of the thesaurus, for example, the thesaurus 30A shown in FIG. 8 is referred to.

Next, the translation means 49 translates the characters determined by the determination means 43 into another language. In this case, the search means 44 searches with a search engine using characters translated into the other language. The translation by the translation means 49 is performed, for example, with reference to the dictionary 30B shown in FIG.

Further, the translation by the translation means 49 may correspond to the synonyms specified by the synonym specifying means 48. For example, all Japanese synonyms may be specified by referring to a Japanese dictionary and then translated into a foreign language (for example, Chinese), or conversely, by translation means 49, for example, Chinese. After translating into, you may refer to the Chinese dictionary to identify all Chinese synonyms.

Next, the task management means 50 manages the task for the infringement information 28 extracted by the extraction means 46. For example, it is a function of managing tasks (and progress status of tasks) such as exercise of rights / litigation / license negotiation for the extracted infringement information 28. If an imitation of another company is found, tasks such as exercise of rights / litigation / license negotiation will be performed in the subsequent process. Therefore, task management by the task management means 50 enhances convenience. The task management by the task management means 50 may be shared with, for example, a company acting as a litigation business agent to manage the tasks.

Next, in the notification means 51, the extraction means 46 periodically (for example, once a day / once a week / once every other week / once a month) extracts infringing information, and the infringing information extracted this time is the previous time. When there is a change from the extracted infringement information, the user 80 is notified that there is a change in the extraction result. The change of the infringement information is, for example, the case where 5 infringement information is extracted in the previous extraction and 7 infringement information is extracted in the current extraction. In addition, as cases where the extraction result changes even if it is extracted regularly, the accuracy of extraction becomes higher and it becomes possible to extract infringing information that could not be extracted so far, or new infringing information can be extracted. In some cases, it was extracted because someone posted it on the Internet. The information notified to the user 80 side may include not only the change in the infringing information but also the means for accessing the changed infringing information. By periodically judging in this way and notifying if there is a change, it is possible to have a monitoring notification function. It is useful for updating the infringement information that the user 80 knows.

The download means 52 downloads the infringing information extracted by the extraction means 46. The storage means 54 associates the content downloaded by the download means 52 with the date and time, and automatically stores the content as infringement information 28 in the storage 16. In addition, since the storage cost becomes enormous if all the extracted infringement information is downloaded, for example, a save button is installed to save only the infringement information for which the user 80 presses the save button, and the save button is not pressed. Infringing information may be destroyed after a certain period of time. On the contrary, only the URL of the extracted infringing information may be provided to the user 80, and only the infringing information for which the user 80 presses the save button may be downloaded and saved.

The group creating means 56 creates a group composed of predetermined members. For example, the display unit 92 of the user terminal 82 is displayed with a screen for designating a member of the group, and each predetermined member input to the screen is stored in the storage 16 as group information 32. The group is composed of predetermined members such as, for example, the planning and management department of a company, the intellectual property department, the XX team, and outside experts, but is not limited to this. You can also create a chat room where you can chat about infringing information within the created group.

The sharing means 58 shares the extracted infringement information 26 within the group created by the group creating means 56. As a result, internal sharing, team sharing, sharing with external experts, and the like become possible, and infringement information 28 can be efficiently supplied.

The right holder information acquisition means 60 acquires information on the right holder of the target intellectual property from the user 80. For example, the right holder information input in the right holder information input field 120 displayed on the screen 110 for acquiring the right information from the user 80 is acquired via the network. Further, the information of the right holder of the target intellectual property may be acquired from the registration number input by the user 80.

Then, the analysis means 40 collectively analyzes the right information indicating the intellectual property held by the acquired right holder. Since the right information indicating the intellectual property held by the right holder is analyzed collectively, it is possible to save the trouble of inputting the registration number one by one. Specifically, the right information is analyzed from the registration gazette of the right specified by the type of intellectual property right and the registration number. In the analysis, the model 20 may be used if necessary. When an image of the analyzed right information or the like is input to the generated model 20, the analysis means 40 collectively analyzes which keyword is extracted or which figure or feature point is extracted, and the search means 44 determines. Based on the analysis result, a search is performed using a known search engine, and the extraction means 46 collectively collects infringing information that may infringe the intellectual property rights owned by the acquired right holder from the search result. Extract with.

Next, the evidence specifying means 62 identifies in which part of the infringement information extracted by the extraction means 46 has the evidence of infringement. For example, on the screen 130 shown in FIG. 11, when the user 80 selects the infringement information 136, the URL 152 of the infringement information 136, the infringement information page 154, and the infringement information page 154 are shown as shown in the screen 150 on the right side of FIG. The sentence 156 described in is displayed. The rationale specifying means 62, for example, represents the portion of the sentence 156 that is the basis of infringement by underlining, bold, highlight, or the like, and clearly indicates the rationale portion 158.

The keyword receiving means 64 accepts the designation of a keyword from the user 80. For example, as shown in FIG. 10, in the keyword input field 118 displayed on the screen 110 for acquiring the right information from the user 80, the user 80 accepts the designation of the keyword input by the user terminal 82. The designated keyword is used so that the extraction means 46 reflects the keyword when extracting the infringing information. Judgment accuracy can be customized by adding keywords. Further, the characters determined by the determination means 43 can be changed or added. For example, it is possible to extract the content in which the specified keyword is described from the search results.

The management means 66 manages the URL of the infringing information extracted by the extraction means 46. The output means 68 outputs a report listing a plurality of URLs managed by the management means 66. For example, if the URLs of all the infringing information extracted are listed, it will be enormous. Therefore, a predetermined period may be set and a report listing the URLs of the infringing information extracted during that period may be output. The file format of the output report is preferably CSV format or PDF format, but it does not interfere with other file formats.

The calculation means 70 calculates the possibility of infringement of the infringement information 28 extracted by the extraction means 46. For example, in the case of a patent right or a utility model right, the possibility of infringement is calculated that the larger the number including the same keyword, the higher the possibility of infringement, and the smaller the number containing the same keyword, the lower the possibility of infringement. For example, in the case of a graphic design right or trademark right, it is calculated that the higher the degree of similarity of the characteristic points of the figure, the higher the possibility of infringement, and the lower the degree of similarity of the characteristic points of the figure, the lower the possibility of infringement. do. The calculation result of such possibility of infringement may be shown as a percentage, for example.

The sorting means 72 sorts the extracted infringement information 26 in the order of high possibility of infringement calculated by the determination means 70. For example, in the screen 130 on the left side of FIG. 11, the infringement information 132 has the highest possibility of infringement, is displayed at the top, and the possibility of infringement decreases as the infringement information 134 or later becomes. By such sorting, the user 80 can confirm the details in order from the infringement information having a high possibility of infringement.

The user may be made to register as a member and analyze the right information indicating the right of any one of the patent, utility model, design and trademark specified by the user. For example, a user who has been registered as a member specifies the registration number of the target intellectual property right, and analyzes the right information indicating the right of any one of the patent, utility model, design, and trademark of the registration number. do.

Of the extracted infringement information, the specified infringement information may be registered in the bookmark, and the infringement information registered in the bookmark may be displayed in a list. For example, the user can browse the extracted infringement information and register the infringement information that he / she is interested in in the bookmark, and the infringement information registered in the bookmark is displayed in a list. Further, the list of infringement information registered in the bookmark may be output as a file. In addition, the list of bookmarks may be organized and displayed for each target intellectual property right. It is also possible to record a memo from the user for the infringing information displayed in the bookmark list. In addition, the infringing information displayed in the bookmark list can be shared by the group.

You may set a schedule to be extracted and search with a search engine using the determined characters or images based on the set schedule. For example, depending on the characters and images to be searched, it may be desired to set a search schedule such as daily search, weekly search, or monthly search. In order to search for the same target intellectual property right with multiple character or image patterns, it is possible to set a schedule for extracting each of multiple character or image patterns. For example, it is possible to prepare a plurality of search patterns for the intellectual property rights of the same registration number and set a schedule for extracting each search pattern.

You may accept the change of the determined character or image and search by the search engine using the changed character or image. You can change (select, add, delete, etc.) the judged characters or images. For example, the determined character or image may be changed (selected, added, deleted, etc.) by the user. It can be searched by a search engine using characters or images modified by the user.

<Structure of User Terminal> ... FIG. 4 is a block diagram showing the configuration of the user terminal 82 of the present embodiment. The user terminal 82 includes, for example, a processor 84, a memory 86, a storage 88, an input unit 90, a display unit 92, and a communication unit 94, which are connected by a bus (not shown). The processor 84 is configured by, for example, a CPU (Central Processing Unit), and performs various processes by reading and executing various programs stored in the memory 86.

The memory 86 stores a program executed by the processor 84, and is composed of, for example, a ROM (Read Only Memory) or a RAM (Random Access Memory). The storage 88 stores, for example, information acquired from the infringement information extraction server 10, a control program (not shown), and the like. The input unit 90 includes, but is not limited to, a keyboard and a touch panel. The display unit 92 is a display or a touch panel. The communication unit 94 performs various data communications with the infringement information extraction server 10 via the network. Of course, data communication may be performed with other external devices as needed.

As such a user terminal 82, a smartphone, a personal computer, a tablet, or the like is used, but the user terminal 82 is not limited thereto. Further, different types of user terminals 82 may be used for each user 80.

In the present embodiment, the various means described above are provided in the infringement information extraction server 10, but all or part of them may be provided in the user terminal 82 or another terminal or server on the cloud. ..

<Infringement Information Extraction Process> ... Next, an example of the infringement information extraction process by the infringement information extraction system 100 will be described with reference to FIG. FIG. 12 is a flowchart showing an example of the infringement information extraction process of the present embodiment.

First, the analysis means 40 of the infringement information extraction server 10 analyzes the right information indicating the right of the target intellectual property (step S20). The right information (registration number, publication at the time of registration, etc.) indicating the right of the target intellectual property is acquired from the content input by the user 80 by the user terminal 82. For example, the right information may be acquired by the registration number input by the user 80, or the content input by the user 80 may be acquired as the right information as it is (for example, the patent claim input by the user 80). Range etc.). This analysis is performed to understand what is described in the publication. For example, the publication at the time of registration may be image-captured and image-analyzed, or the publication at the time of registration may be used as characters for text mining. In addition, text data including rights information indicating the rights of intellectual property may be acquired and analyzed. In the case of text data including rights information indicating the rights of intellectual property, the location of the rights information indicating the rights of intellectual property is analyzed in the text data.

FIG. 10 shows an example of the screen 110 displayed on the display unit 92 of the user terminal 82. On the screen 110, a pull-down menu 112 for selecting patents, trademarks, designs, utility models, etc., a registration number input field 114, an IPC classification input field 116A for patents and utility models, and a classification input field 116B for design classification , The classification input field 116C for trademark classification is displayed.

In addition, a keyword input field 118 for customizing the judgment accuracy by adding keywords, a right holder information input field 120 for collectively acquiring the registration number of the intellectual property right owned by the designated right holder, and a registration gazette. The registration gazette input field 122 for directly inputting is displayed. When the user 80 inputs or selects desired information on the screen 110, the infringement information extraction server 10 can acquire the right information indicating the right of the target intellectual property input from the user 80. can.

In the analysis by the analysis means 40, the infringement information extraction server 10 uses the generation means 42 to process the learning data indicating the rights of other intellectual property (publication at the time of registration, etc.) as necessary. You may train and generate the model 20. Specifically, the feature parts (keywords, feature points of figures, etc.) of the image showing the rights information of various intellectual property rights (patent rights, utility model rights, design rights, trademark rights) are rounded and processed. Dozens to tens of thousands of the trained data 18 are prepared, and the trained data 18 is machine-learned to generate a model 20. For example, in order to recognize what kind of field and what kind of content, the circle is applied to both the information indicating the field and the information indicating the content. For example, in order to machine-learn image data to generate a model, deep learning, which is a method of machine learning, may be performed.

Field information indicating a field is IPC classification, FI, F-term, etc. in the case of a patent, product / service classification in the case of a trademark, design classification, an article related to a design, etc. in the case of a design. Is. Utility models are the same as patents.

In addition, the content information indicating the content is a keyword of an independent term, a keyword of a dependent term, etc. in the case of a patent, a character or a figure of a registered trademark in the case of a trademark, and a character or a figure of a registered trademark in the case of a design. Solid lines, broken lines, etc. in the drawing. Utility models are the same as patents.

As the learning data, image learning data is preferable because it can be recognized corresponding to both characters and figures, but it does not prevent the use of text learning data.

When the above training data is machine-learned, a model customized for the training data is generated. That is, a model 20 is generated in which the field information is found in which part of the publication, what characters are described, the content information is found in which part, and what kind of character or figure is learned. Will be done. The analysis means 40 inputs an image (right information) showing the right of the target intellectual property into the generated model 20 and analyzes which part (keyword, feature point of the figure, etc.) is extracted from the right information. do.

Next, the determination means 43 of the infringement information extraction server 10 determines a character or an image necessary for searching from the analysis result of the analysis means 40 (step S22). Specifically, a search engine determines which character or image is required for a character search or image search. The characters or images required for the search are, for example, the field information determines the characters and the content information determines the characters or figures. Specifically, if it is field information, the characters of the IPC classification of the patent and the characters of the goods / services of the trademark are determined. If it is content information, the characters described in the independent section of the patent, the solid line and the broken line image of the design, etc. are judged. The determined characters or images may be changed (selected, added, deleted, etc.). For example, the determined character or image may be changed (selected, added, deleted, etc.) by the user.

Next, the search means 44 of the infringement information extraction server 10 searches with a search engine using the determined characters or images (step S24). Using the characters or images determined in step S22, a character search or an image search is performed by a known search engine. For example, a search engine may search using characters or images modified by the user.

At this time, the synonym specifying means 48 may specify the synonyms of the characters determined by the determining means 43, and the search means 44 may search by the search engine using the specified synonyms. For the identification of the thesaurus, for example, the thesaurus 30A shown in FIG. 8 is referred to.

Further, the characters determined by the determination means 43 may be translated into another language by the translation means 49, and the search means 44 may search by a search engine using the characters translated into the other language. .. The translation by the translation means 49 is performed, for example, with reference to the dictionary 30B shown in FIG.

The translation by the translation means 49 may correspond to the synonyms specified by the synonym specifying means 48. For example, all Japanese synonyms may be specified by referring to a Japanese dictionary and then translated into a foreign language (for example, Chinese), or conversely, by translation means 49, for example, Chinese. After translating into, you may refer to the Chinese dictionary to identify all Chinese synonyms.

Then, the extraction means 46 of the infringement information extraction server 10 extracts infringing information that may infringe the rights of the target intellectual property from the search result (step S26). Specifically, infringing information is extracted from the contents such as Web pages / videos / images displayed in the search results. It should be noted that these contents may be analyzed to extract infringing information, may be randomly extracted without analysis, or all may be extracted without analysis. .. For example, when the accuracy of analysis is high, it can be expected that the accuracy of extraction after the final reinforcement learning will be higher if the infringing information is extracted by analyzing the contents such as Web pages / videos / images.

When searching with a search engine, for example, noise that does not become infringing information such as academic papers or the right holder's own product sales site may be picked up. Therefore, noise is removed by the filter function provided in the extraction means 46.

For example, when removing noise by keyword, if you search by a search engine, the page title will be displayed on the search list page, so if the page title contains academic or sales keywords, it will be excluded and noise will be removed. Perform removal. Alternatively, a URL list to be excluded may be created, and if the URL in the list is applicable, the URL may be excluded from the extracted infringing information. Of course, it does not prevent noise removal by other methods.

The management means 66 of the infringement information extraction server 10 manages the URL of the infringement information extracted by the extraction means 46, and the output means 68 outputs a report listing a plurality of URLs managed by the management means 66. The file format of the output report is preferably CSV format or PDF format, but it does not interfere with other file formats.

At this time, if the number of cases is large when all the extracted infringement information is displayed, the extracted infringement information 26 is sorted by the rearrangement means 72 in the order of high possibility of infringement calculated by the calculation means 70, and infringement is possible. Only the best 10 with high characteristics may be displayed on the display unit 92 of the user terminal 82. Further, the output means 68 may output a list of URLs of all the extracted infringement information in CSV format, PDF format, or the like.

Then, as shown in the screen 130 of FIG. 11, the details of the infringement information 136 selected by the user 80 from the displayed plurality of infringement information may be displayed as shown in the screen 150 of FIG.

Here, the evidence specifying means 62 may specifically specify in which part of the infringement information extracted by the extraction means 46 has the evidence of infringement. For example, on the screen 130 shown in FIG. 11, when the user 80 selects the infringement information 136, the URL 152 of the infringement information 136, the infringement information page 154, and the infringement information page 154 are shown as shown in the screen 150 on the right side of FIG. The sentence 156 described in is displayed. The rationale specifying means 62, for example, represents the portion of the sentence 156 that is the basis of infringement by underlining, bold, highlight, or the like, and clearly indicates the rationale portion 158.

In the present embodiment, in the notification means 60, the extraction means 46 periodically (for example, once a day / once a week / once every other week / once a month) extracts infringing information, and the infringing information extracted this time is used. , If there is a change from the previously extracted infringement information, notify that there is a change in the extraction result (Yes in step S38, step S40). The change of the infringement information is, for example, the case where 5 infringement information is extracted in the previous extraction and 7 infringement information is extracted in the current extraction. In addition, as cases where the extraction result changes even if it is extracted regularly, the accuracy of extraction becomes higher and it becomes possible to extract infringing information that could not be extracted so far, or new infringing information can be extracted. In some cases, it was extracted because someone posted it on the Internet. The information notified to the user 80 side may include not only the change in the infringing information but also the means for accessing the changed infringing information. By periodically judging in this way and notifying if there is a change, it is possible to have a monitoring notification function. It is useful for updating the infringement information that the user 80 knows.

Further, in the present embodiment, the task management means 50 manages the task for the infringement information 28 extracted by the extraction means 46 (step S30). For example, it manages tasks such as exercising rights / litigation / license negotiation for the extracted infringement information 28. If an imitation of another company is found, tasks such as exercise of rights / litigation / license negotiation will be performed in the subsequent process. Therefore, task management by the task management means 50 enhances convenience. The task management by the task management means 50 may be shared with, for example, a company acting as a litigation business agent to manage the tasks.

Further, in the present embodiment, the download means 52 downloads the infringement information 28 extracted by the extraction means 46 (step S32), and the storage means 54 associates the downloaded contents with the date and time, and the infringement information is stored in the storage 16. It can also be automatically saved as 28 (step S34). Information on the network may be erased immediately if it is not saved, but according to this embodiment, it is automatically downloaded and saved, so even if the infringing information is erased from the network, evidence is provided. Can be left. In addition, since the storage cost becomes enormous if all the extracted infringement information is downloaded, for example, a save button is installed to save only the infringement information for which the user 80 presses the save button, and the save button is not pressed. Infringing information may be destroyed after a certain period of time. On the contrary, only the URL of the extracted infringing information may be provided to the user 80, and only the infringing information for which the user 80 presses the save button may be downloaded and saved.

In the present embodiment, the group creation means 56 of the infringement information extraction server 10 may be used to create a group consisting of predetermined members. For example, the display unit 92 of the user terminal 82 is displayed with a screen for designating a member of the group, and each predetermined member input to the screen is stored in the storage 16 as group information 32. The group is composed of predetermined members such as, for example, the planning and management department of a company, the intellectual property department, the XX team, and outside experts, but is not limited to this. You can also create a chat room where you can chat about infringing information within the created group.

The sharing means 58 shares the infringement information 28 extracted in step S26 within the group created by the group creating means 56. As a result, internal sharing, team sharing, sharing with external experts, and the like become possible, and infringement information 28 can be efficiently supplied.

Further, in the present embodiment, the keyword reception means 64 of the infringement information extraction server 10 may accept the designation of the keyword from the user 80. For example, as shown in FIG. 10, the designation of the keyword input in the keyword input field 118 displayed on the screen 110 for acquiring the right information from the user 80 is accepted. The designated keyword is used so that the extraction means 46 reflects the keyword when extracting the infringing information. Judgment accuracy can be customized by adding keywords. Further, the characters determined by the determination means 43 can be changed or added. For example, it is possible to extract the content in which the specified keyword is described from the search results.

Further, in the present embodiment, the right holder information may be acquired from the user 80 by the right holder information acquisition means 60. For example, the right holder information input in the right holder information input field 120 displayed on the screen 110 for acquiring the right information from the user 80 may be acquired via the network. Further, the information of the right holder of the target intellectual property may be acquired from the registration number input by the user 80.

Then, the analysis means 40 collectively analyzes the right information indicating the intellectual property held by the acquired right holder. Since the right information indicating the intellectual property held by the right holder is analyzed collectively, it is possible to save the trouble of inputting the registration number one by one. Specifically, the right information is analyzed from the registration gazette of the right specified by the type of intellectual property right and the registration number. Then, when an image of the analyzed right information or the like is input to the generated model 20, the analysis means 40 collectively analyzes which keyword is extracted or which figure or feature point is extracted, and the search means 44 is used. However, based on the analysis result, a search is performed using a known search engine, and the extraction means 46 may infringe the intellectual property right owned by the acquired right holder from the search result. Is extracted in a batch.

In addition, the user 80 may be made to register as a member in advance, and the right information indicating the right of any one of the patent, the utility model, the design and the trademark designated from the user terminal 82 may be analyzed. For example, a user who has been registered as a member specifies the registration number of the target intellectual property right, and analyzes the right information indicating the right of any one of the patent, utility model, design, and trademark of the registration number. do.

Further, among the extracted infringement information, the designated infringement information may be registered in the bookmark, and the infringement information registered in the bookmark may be displayed in a list on the user terminal 82 or the like. For example, among the extracted infringement information, the infringement information that the user browses and is interested in can be registered in the bookmark, and the infringement information registered in the bookmark is displayed in a list. Further, the list of infringement information registered in the bookmark may be output as a file on the user terminal 82 or the like. In addition, the list of bookmarks may be organized and displayed for each target intellectual property right. It is also possible to record a memo from the user for the infringing information displayed in the bookmark list. In addition, the infringing information displayed in the bookmark list can be shared by the group.

In addition, a "schedule" which is a date and time to be extracted or a period may be set, and a search engine may periodically search using the determined characters or images based on the set schedule. For example, depending on the characters and images to be searched, it may be desired to set a search schedule such as daily search, weekly search, or monthly search. In order to search for the same target intellectual property right with multiple character or image patterns, it is possible to set a schedule for extracting each of multiple character or image patterns. For example, it is possible to prepare a plurality of search patterns for the intellectual property rights of the same registration number and set a schedule for extracting each search pattern.

Further, the change of the character or the image determined by the determination means 43 may be accepted, and the search engine may search using the changed character or the image. That is, the characters or images once determined can be changed (selected, added, deleted, etc.) later. For example, the determined character or image may be changed (selected, added, deleted, etc.) by the user via the user terminal 82. It can be searched by a search engine using characters or images modified by the user.

<Effect> ... According to the embodiment described above, the infringement information extraction system 100 is necessary for analyzing the right information indicating the right of the target intellectual property and searching from the analysis result. To determine infringing information that may infringe the rights of the target intellectual property from the search results by judging the text or image and searching with the search engine using the judged text or image. It has the effect of automatically extracting infringing information that may infringe the rights of the target intellectual property.

The above-described embodiment is an example, and can be appropriately changed within a range in which the same effect is obtained. Further, the infringement information extraction server 10 may be a single computer, for example, a terminal. Further, the above-mentioned functional configuration may be a computer system (cloud) executed by different computers. Further, in the present embodiment, although many processes are executed by the infringing information extraction server 10, all or a part of the functions may be executed by the user terminal 82.

Further, the above-mentioned means and functions are realized by a computer (including a CPU, an information processing device, and various terminals) reading and executing a predetermined program. The program is provided, for example, in the form of being provided via a network from one or more computers (Cloud Service, SaaS: Software as a Service). The program is also provided, for example, in the form of being recorded on a computer-readable recording medium. In this case, the computer reads the program from the recording medium, transfers it to an internal recording device or an external recording device, records the program, and executes the program.

Furthermore, the invention may be provided as an invention of a method or program. Further, the program may be recorded in advance on a recording device (recording medium) such as a magnetic disk, an optical disk, or a magneto-optical disk, and the program may be provided from the recording device to a computer via a communication line.

According to the present invention, the rights information indicating the rights of the target intellectual property is analyzed, the characters or images necessary for searching are determined from the analysis results, and the determined characters or images are used. We decided to search with a search engine and extract infringing information that may infringe the rights of the target intellectual property from the search results.

In this way, since infringing information that may automatically infringe the rights of the target intellectual property can be automatically extracted, it can be applied to the use of the infringing information extraction system.

10: Infringement information extraction server 12: Processor 14: Memory 16: Storage 18, 18A, 18B: Learning data 20: Model 22: Right information 24: Judgment result 26: Search result 28: Infringement information 30, 30A, 30B: Dictionary 32 : Group information 34: Communication unit 40: Analysis means 42: Generation means 43: Judgment means 44: Search means 46: Extraction means 48: Synonyms specifying means 49: Translation means 50: Task management means 51: Notification means 52: Download means 54 : Storage means 56: Group creation means 58: Sharing means 60: Right holder information acquisition means 62: Grounds identification means 64: Keyword reception means 66: Management means 68: Output means 70: Calculation means 72: Sorting means 80: User 82 : User terminal 84: Processor 86: Memory 88: Storage 90: Input unit 92: Display unit 94: Communication unit 100: Infringement information extraction system 110: Screen 112: Pull-down menu 114: Registration number input field 116A: IPC classification input field 116B : Classification input field 116C: Classification input field 118: Keyword input field 120: Right holder information input field 122: Registration publication input field 130: Screen 132, 134, 136, 140: Infringement information 150: Screen 152: URL
154: Infringement information page 156: Rationale information 160: Like button 162: Bad button C, C1, C2: Surrounding line

Claims (18)

An analysis means for acquiring the right information indicating the right of any one of the target patent, utility model, design and trademark from the registration number and analyzing the right information.
Judgment means for determining characters or images necessary for searching from the results of the analysis, and
A search means for searching with a search engine using the determined characters or images, and
An extraction means for extracting infringing information that may infringe the rights of any one of the target patents, utility models, designs and trademarks from the search results.
Infringement information extraction system equipped with. Equipped with a synonym identification means for identifying the synonyms of the determined characters,
The search means searches with a search engine using the specified synonyms.
The infringement information extraction system according to claim 1. A translation means for translating the determined characters into other languages is provided.
The infringement information extraction system according to claim 1, wherein the search means searches with a search engine using characters translated into the other language. A task management means for managing tasks for the extracted infringement information, and
The infringement information extraction system according to claim 1. The extraction means periodically extracts and extracts.
Notification means for notifying that there is a change in the extraction result when the infringement information extracted this time is changed from the infringement information extracted last time.
The infringement information extraction system according to claim 1. Downloading means for downloading the extracted infringement information and
A storage method for saving the downloaded contents in association with the date and time,
The infringement information extraction system according to claim 1. Group creation means to create a group consisting of predetermined members,
A sharing means for sharing the extracted infringement information within the created group,
The infringement information extraction system according to claim 1. It is equipped with a right holder information acquisition means for acquiring information on any one of the above-mentioned target patents, utility models, designs and trademarks.
The infringement information extraction system according to claim 1, wherein the analysis means collectively analyzes the right information indicating any one of the patent, utility model, design and trademark held by the acquired right holder. Evidence identification means for analyzing the extracted infringement information and identifying which part has the basis for infringement,
The infringement information extraction system according to claim 1. Equipped with a keyword reception means that accepts keyword specifications
The extraction means reflects the designated keyword and extracts.
The infringement information extraction system according to claim 1. A management means for managing the URL of the extracted infringing information and
An output means for outputting a report listing a plurality of managed URLs, and
The infringement information extraction system according to claim 1. A calculation means for calculating the possibility of infringement of the extracted infringement information, and
A sorting means for sorting the extracted infringement information in descending order of the calculated infringement possibility.
The infringement information extraction system according to claim 1. Equipped with a membership method that allows users to register as a member
The infringement information extraction system according to claim 1, wherein the analysis means analyzes right information indicating the right of any one of a patent, a utility model, a design and a trademark designated by the user. Of the extracted infringement information, a bookmarking means for registering the specified infringement information in a bookmark, and
A display means for displaying the infringement information registered in the bookmark in a list, and
The infringement information extraction system according to claim 1. A schedule means for setting the extraction schedule is provided.
The infringement information extraction system according to claim 1, wherein the search means searches with a search engine using the determined characters or images based on the set schedule. A change receiving means for accepting changes in the determined characters or images is provided.
The infringement information extraction system according to claim 1, wherein the search means searches with a search engine using the changed characters or images. The step of acquiring the right information indicating the right of any one of the target patent, utility model, design and trademark from the registration number and analyzing the right information,
From the results of the analysis, the steps to determine the characters or images required for searching, and
The step of searching with a search engine using the determined characters or images, and
From the search results, a step of extracting infringing information that may infringe the rights of any one of the target patents, utility models, designs and trademarks, and
Infringement information extraction method. On the computer
The step of acquiring the right information indicating the right of any one of the target patent, utility model, design and trademark from the registration number and analyzing the right information,
From the results of the analysis, the steps to determine the characters or images required for searching, and
The step of searching with a search engine using the determined characters or images, and
From the search results, a step of extracting infringing information that may infringe the rights of any one of the target patents, utility models, designs and trademarks, and
Infringement information extraction program to execute.

Images