A trademark infringement analysis method based on big data

Through big data technology, the trademark content is analyzed, the trademark profile characteristics and RGB reference values ​​are extracted, and the overall matching parameters are calculated, which solves the problem of low efficiency in trademark infringement analysis and achieves efficient and accurate infringement identification and trademark rights protection.

CN118628769BActive Publication Date: 2024-12-03BEIJING HANTANG XINTONG CONSULTING CO LTD

Patent Information

Application Number
CN202410741698.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-11
Publication Date
2024-12-03
Estimated Expiration
2044-06-11

AI Technical Summary

Technical Problem

In the prior art, trademark infringement analysis is inefficient, and the comparison trademark of the target trademark cannot be quickly determined, resulting in low infringement analysis.

Method used

Using a trademark infringement analysis method based on big data, the trademark content of the target trademark is obtained, the positional relationship between its images and text is determined, the contour characteristic degree and RGB reference value are extracted, the overall matching parameters are calculated, and the trademark is accurately matched and compared, so as to improve the infringement analysis. Efficiency and accuracy.

Benefits of technology

It significantly improves the efficiency and accuracy of trademark infringement analysis, provides strong technical support, and can quickly and accurately screen out comparison trademarks similar to the target trademark, improving work efficiency and reliability of infringement identification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118628769B_ABST
    Figure CN118628769B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of network information processing technology, and in particular to a trademark infringement analysis method based on big data, including obtaining the trademark content of a target trademark; determining the positional relationship between the trademark image and the trademark text corresponding to the target trademark according to the trademark content, as well as the font corresponding to the trademark text of the target trademark; determining a number of corresponding comparison trademarks according to the trademark image; extracting the RGB channels of the target trademark and each comparison trademark to determine the RGB reference values ​​of the target trademark and each comparison trademark; comparing the target trademark and each comparison trademark to determine the graphic similarity, combining the RGB reference value and graphic similarity of a single comparison trademark to determine the corresponding overall matching parameter, and determining the infringement analysis result according to the overall matching parameter. The method performs comparative analysis in a large amount of trademark data, provides an efficient and reliable solution for infringement identification, protects trademark rights and interests, and improves work efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of network information processing, and in particular to a trademark infringement analysis method based on big data. Background Art

[0002] Trademark infringement analysis in the patent field involves in-depth understanding and application of trademark laws, trademark registration examination standards, trademark rights, trademark similarity, priority principles, market research, case law and professional opinions, and requires a comprehensive understanding of trademark laws and regulations, including trademark registration procedures and protection scope, as well as specific provisions of trademark rights. Among them, the similarity of trademark patterns is an important part of judging whether a trademark is infringing. Currently, there are many types of trademarks and a large number of trademarks. It is impossible to accurately, timely and effectively screen out potential similar trademarks by manually comparing and searching for infringing trademarks, thereby reducing the sensitivity of judging infringement.

[0003] Chinese Patent Publication No. CN114529784B discloses a trademark infringement analysis method and system for e-commerce data, which constructs a trademark detection model and a product detection model; uses the product detection model to extract the brand's standard product feature information, and determines which trademark classifications in the trademark detection model to be used for detection; obtains the product feature information to be judged for infringement and sets an infringement threshold, calculates the similarity between the standard product feature information of each product and the product feature information to be judged for infringement, and if there are products with a similarity greater than the infringement threshold, uses the trademark detection model to identify the product information to be judged for infringement, and if there is a corresponding trademark classification, judges that the trademark of this product is improperly used. The present invention also provides a trademark infringement analysis system, which analyzes trademark infringement by training a trademark detection model and a product detection model, and comprehensively analyzes trademark infringement with product information data related to the e-commerce platform, effectively improving the feasibility and accuracy of the trademark infringement judgment method for the e-commerce platform. It can be seen that the invention has the following problems:

[0004] The problem of low efficiency in trademark infringement analysis due to the inability to quickly identify comparative trademarks of the target trademark among a large number of trademarks is not considered. Summary of the invention

[0005] To this end, the present invention provides a trademark infringement analysis method based on big data to overcome the problem in the prior art of low efficiency of trademark infringement analysis due to failure to consider rapid determination of comparative trademarks of a target trademark.

[0006] To achieve the above objectives, the present invention provides a trademark infringement analysis method based on big data, comprising:

[0007] Acquire the trademark content of the target trademark, wherein the trademark content includes a trademark image and trademark text;

[0008] Determining the positional relationship between the trademark image and the trademark text corresponding to the target trademark, and the font corresponding to the trademark text of the target trademark according to the trademark content;

[0009] Determining a number of corresponding comparative trademarks according to the trademark image, including:

[0010] Determine the number and positions of key points of the target trademark, determine the contour characteristic degree of the target trademark according to the number and positions of key points, determine the feature matching state of the target trademark according to the contour characteristic degree, and determine a number of comparison trademarks within a standard range of characteristic degree according to the feature matching state;

[0011] or, dividing the target trademark image into a plurality of sub-image shapes according to the chromaticity distribution of the target trademark image, respectively determining a plurality of candidate comparative trademarks containing a single sub-image shape, and determining whether a single candidate comparative trademark is a comparative trademark according to the number of different sub-images contained in the single candidate comparative trademark;

[0012] Extracting RGB channels of the target trademark and each of the comparison trademarks to determine RGB reference values ​​of the target trademark and each of the comparison trademarks;

[0013] The target trademark and each of the comparison trademarks are compared to determine the graphic similarity, the corresponding overall matching parameters are determined in combination with the RGB reference value and the graphic similarity of a single comparison trademark, and the infringement analysis result is determined based on the overall matching parameters.

[0014] Furthermore, the step of determining key points based on the closed image of the target trademark includes:

[0015] Obtaining the outer contour shape of the trademark image;

[0016] Determine the number of straight line segments and the number of curved line segments included in the outer contour shape;

[0017] Determine the sum of the number of the straight line segments and the number of the curve segments as the number of key points;

[0018] Determine the position of the key point of the straight line according to the two end points of each straight line segment;

[0019] Determine the position of the key point of the curve according to the two end points of each curve segment;

[0020] Determine the straight line key point position and the curve key point position as key point positions;

[0021] The different straight line segments satisfy that two adjacent straight line segments intersect, and the different curved line segments satisfy that two adjacent curved line segments have different roundness and / or different center points.

[0022] Furthermore, the contour characteristic degree of the target trademark is determined according to the number of key points and the positions of the key points, and the contour characteristic degree is calculated according to the following formula:

[0023]

[0024] Wherein, D is the contour feature, n is the number of key points, di is the distance between the ith key point and the center point of the trademark image, S is the area of ​​the target trademark, k1 is the quantity weight, k2 is the position weight, k3 is the area weight, and k1+k2+k3=1.

[0025] Furthermore, the quantity weight and the position weight are determined according to contour complexity, wherein:

[0026] When the contour complexity is greater than a preset contour complexity value, the quantity weight is positively correlated with the ratio of the contour complexity to the preset contour complexity value;

[0027] When the contour complexity is less than a preset contour complexity value, the position weight is positively correlated with the ratio of the contour complexity to the preset contour complexity value;

[0028] The contour complexity is determined according to the ratio of the perimeter to the area of ​​the contour.

[0029] Further, the feature matching state of the target trademark is determined according to the contour feature degree, wherein:

[0030] When the contour feature degree is greater than or equal to the contour feature reference value, the target trademark is determined to be in a strong feature matching state;

[0031] When the contour feature degree is less than the contour feature reference value, the target trademark is determined to be in a weak feature matching state.

[0032] Furthermore, when it is determined that the target trademark is in a strong feature matching state, several trademarks whose contour feature degrees are within a feature degree standard range are determined as comparison trademarks.

[0033] Furthermore, when it is determined that the target trademark is in a weak feature matching state, the target trademark image is divided into a plurality of sub-image shapes according to the chromaticity distribution of the target trademark image;

[0034] The difference in chromaticity values ​​of pixels in a single sub-image shape is less than a preset difference.

[0035] Further, several candidate trademarks containing a single sub-image shape are determined, and whether a single candidate comparative trademark is a comparative trademark is determined according to the number of different sub-images contained in the single candidate comparative trademark, wherein:

[0036] If the number of different sub-images contained in a single candidate comparative trademark is greater than or equal to a preset number, the candidate comparative trademark is determined to be a comparative trademark;

[0037] If the number of different sub-images included in a single candidate comparative trademark is less than a preset number, it is determined that the candidate comparative trademark is not a comparative trademark.

[0038] Furthermore, the overall matching parameter between a single comparison trademark and a target trademark is calculated according to the following formula:

[0039]

[0040] Among them, W is the overall matching parameter, rgb1 is the RGB reference value of the comparison trademark, rgb0 is the RGB reference value of the target trademark, sm is the graphic similarity, r1 is the matching weight, r2 is the similarity weight, r3 is the trademark text weight, and r1+r2+r3=1.

[0041] Further, the RGB reference value is calculated according to the following formula,

[0042]

[0043] Among them, rgb is the RGB reference value of the image, R is the average value of each pixel in the red channel, G is the average value of each pixel in the green channel, B is the average value of each pixel in the blue channel, n is the number of pixels contained in a single trademark image, Ri is the value of the i-th pixel in the red channel, Gi is the value of the i-th pixel in the green channel, and Bi is the value of the i-th pixel in the blue channel.

[0044] Compared with the prior art, the beneficial effect of the present invention lies in that the big data trademark infringement analysis method provided by the present invention significantly improves the efficiency and accuracy of trademark infringement analysis by efficiently identifying trademark content and accurately matching and comparing trademarks, providing strong technical support for enterprises and users; the present invention can be used for infringement analysis when applying for new trademarks to avoid infringement of existing trademarks, and can also be applied to infringement analysis conducted by enterprises and users to protect existing trademarks.

[0045] Furthermore, the method uses big data technology to combine the image and text content of the target trademark to determine the positional relationship between the two, and determines how to quickly and accurately screen out comparison trademarks similar to the target trademark by considering the closedness of the trademark image; by combining the RGB channel reference values ​​of each comparison trademark and the target trademark and the graphic similarity of each comparison trademark and the target trademark determined by computer software and / or algorithms, the overall matching parameters of each comparison trademark and the target trademark are determined, so as to effectively perform infringement analysis based on the calculated values ​​of the overall matching parameters.

[0046] Furthermore, this method conducts comparative analysis on a large amount of trademark data, providing an efficient and reliable solution for infringement identification, thereby protecting trademark rights and interests while improving work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a step diagram of a trademark infringement analysis method based on big data according to an embodiment of the present invention;

[0048] Figure 2 A flow chart for determining a comparison trademark for an embodiment of the present invention;

[0049] Figure 3 A diagram showing the steps of determining key points of a target trademark according to an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0051] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0052] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0053] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0054] See also Figure 1 and Figure 2 As shown, they are respectively a step diagram of a trademark infringement analysis method based on big data according to an embodiment of the present invention and a flow chart of determining a comparative trademark according to an embodiment of the present invention. The present invention provides a trademark infringement analysis method based on big data, comprising:

[0055] Step S1, obtaining the trademark content of the target trademark, wherein the trademark content includes a trademark image and trademark text;

[0056] Step S2, determining the positional relationship between the trademark image and the trademark text corresponding to the target trademark, and the font corresponding to the trademark text of the target trademark according to the trademark content;

[0057] Step S3, determining a number of corresponding comparison trademarks according to the trademark image, including:

[0058] Step S31 determines the number and positions of the key points of the target trademark, and determines the contour characteristic of the target trademark according to the number and positions of the key points. It can be understood that the number and positions of the key points of the target trademark are accurately determined by acquiring the trademark image and text of the target trademark, thereby extracting the contour characteristic of the trademark, which improves the recognition accuracy and shortens the processing time of the target trademark; and determines the feature matching state of the target trademark according to the contour characteristic, and determines a number of comparison trademarks within the feature standard range according to the feature matching state;

[0059] Step S32 or, dividing the target trademark image into a plurality of sub-image shapes according to the chromaticity distribution of the target trademark image, respectively determining a plurality of candidate comparative trademarks containing a single sub-image shape, and determining whether a single candidate comparative trademark is a comparative trademark according to the number of different sub-images contained in the single candidate comparative trademark;

[0060] Step S4, extracting the RGB channels of the target trademark and each of the comparison trademarks to determine the RGB reference values ​​of the target trademark and each of the comparison trademarks;

[0061] Step S5, comparing the target trademark and each of the comparison trademarks by computer software and / or algorithm to determine the graphic similarity, combining the RGB reference value and the graphic similarity of a single comparison trademark to determine the corresponding overall matching parameter, and determining the infringement analysis result based on the overall matching parameter.

[0062] It can be understood that RGB channels refer to the channels representing the three basic colors of red, green, and blue in color images. In the RGB color mode, the color of each pixel is composed of the numerical values ​​of these three color channels. The value range of each channel is usually 0 to 255, where 0 represents the absence of this color component and 255 represents the maximum intensity of this color component. By adjusting the values ​​of these three channels, a variety of different colors can be produced, thus presenting rich colors.

[0063] In practice, the big data trademark infringement analysis method provided by the present invention can significantly improve the efficiency and accuracy of trademark infringement analysis by accurately matching the comparison trademark after efficiently identifying the trademark content, and provide strong technical support for enterprises and users; the present invention can be used for infringement analysis when applying for a new trademark to avoid infringement of an existing trademark, and can also be applied to infringement analysis conducted by enterprises and users to protect existing trademarks. This method uses big data technology to combine the image and text content of the target trademark to determine the positional relationship between the two, and by considering the closedness of the trademark image, it determines how to quickly and accurately screen out comparison trademarks similar to the target trademark; by combining the RGB channel reference values ​​of each comparison trademark and the target trademark and the graphic similarity of each comparison trademark and the target trademark determined by computer software and / or algorithms, the overall matching parameters of each comparison trademark and the target trademark are determined, so that infringement analysis can be effectively performed based on the calculated value of the overall matching parameters. This method performs comparative analysis in a large amount of trademark data, provides an efficient and reliable solution for infringement identification, and improves work efficiency while protecting trademark rights.

[0064] See also Figure 3 As shown, it is a step diagram of determining the key points of the target trademark according to an embodiment of the present invention. Specifically, in step S31, the step of determining the key points of the target trademark includes:

[0065] Step S311, obtaining the outer contour shape of the trademark image. In implementation, the Canny edge detection algorithm may be used to effectively identify the boundary in the image, thereby obtaining the outer contour shape of the trademark;

[0066] Step S312, determining the number of straight line segments and the number of curved line segments included in the outer contour shape;

[0067] Step S313, determining the sum of the number of the straight line segments and the number of the curve segments as the number of key points;

[0068] Step S314, determining the position of the key point of the straight line according to the two end points of each straight line segment;

[0069] Step S315, determining the position of the key point of the curve according to the two end points of each curve segment;

[0070] Step S316, determining the straight line key point position and the curve key point position as key point positions;

[0071] Among them, the different straight line segments satisfy that any straight line segment is not parallel to its adjacent straight line segment (if both ends of the straight line segment are adjacent to curved segments, then the straight line segment and the remaining straight line segments are different straight line segments; if both ends of a straight line segment are adjacent to a straight line segment and a curved segment, when these two straight line segments are not parallel, the straight line segment and all the remaining straight line segments are different straight line segments; if both ends of the straight line segment are adjacent to straight lines, when the straight line segment is not parallel to the two adjacent straight line segments, the straight line segment and all the remaining straight line segments are different straight line segments), and the different curved segments satisfy that the roundness and / or center of the two adjacent curved segments are different.

[0072] In implementation, this method determines the key points and key point positions of the target trademark through steps S311 to S316, and can comprehensively and accurately capture the characteristic points (key points) of the trademark image, providing a reliable basis for subsequent infringement analysis, thereby improving the accuracy and credibility of the analysis.

[0073] Specifically, in step S31, the contour characteristic degree of the target trademark is determined according to the number of key points and the positions of the key points. The contour characteristic degree is calculated according to the following formula:

[0074]

[0075] Wherein, D is the contour feature, n is the number of key points, di is the distance between the ith key point and the center point of the trademark image, S is the area of ​​the target trademark, k1 is the quantity weight, k2 is the position weight, k3 is the area weight, and k1+k2+k3=1.

[0076] It can be understood that the area weight ranges from 0.1 to 0.25, and preferably k3=0.2.

[0077] In implementation, S is the area of ​​the target trademark after scaling the trademark to a size of 10*10 cm; if the target trademark is not a square, its short side is scaled to 10 cm, and its long side is scaled in proportion to the short side.

[0078] Specifically, the quantity weight and the position weight are determined according to the contour complexity, wherein:

[0079] When the contour complexity is greater than a preset contour complexity value, the quantity weight is positively correlated with the ratio of the contour complexity to the preset contour complexity value;

[0080] When the contour complexity is equal to the contour complexity preset value, the quantity weight is equal to the position weight, that is, k1=k2=0.4;

[0081] When the contour complexity is less than a preset contour complexity value, the position weight is positively correlated with the ratio of the contour complexity to the preset contour complexity value;

[0082] In the implementation, k1∈[0.2,0.65], k2∈[0.15,0.6];

[0083] When the contour complexity is greater than the preset contour complexity value, the influence range of the number of key points on the contour feature degree is greater than the influence range of their position on the contour feature degree. The more key points there are, the clearer the contour is. Therefore, the quantity weight is calculated first and then the position weight is determined according to the quantity weight, k1 = contour complexity ÷ contour complexity preset value × 0.4, k2 = 1-k1;

[0084] When the contour complexity is less than the preset contour complexity value, the influence range of the key point position on the contour feature is greater than the influence range of its quantity on the contour feature. The key point position affects the contour clarity. Therefore, the position weight is calculated first and then the quantity weight is determined according to the position weight. k2 = contour complexity ÷ contour complexity preset value × 0.6, k1 = 1-k2;

[0085] The contour complexity is determined according to the ratio of the perimeter to the area of ​​the contour, that is, contour complexity=contour perimeter÷area enclosed by the contour.

[0086] It is understandable that when calculating contour complexity, no units are used, only ratios are calculated.

[0087] In implementation, the contour complexity preset value is set to In the range of

[0088] It can be understood that the present method dynamically adjusts the quantity weight and position weight according to the contour complexity, so that it can better adapt to trademark images with different contour complexities: when the contour complexity is high, the quantity weight is positively correlated with the ratio of the contour complexity to the preset value, so as to pay more attention to the complexity of the trademark shape; when the contour complexity is low, the position weight is positively correlated with the ratio of the contour complexity to the preset value, so as to pay more attention to the positional relationship of the trademark shape; when the contour complexity is equal to the preset value, the quantity weight and the position weight are equal, so as to balance the influence of the number and position of key points in the matching process. This flexible weight adjustment mechanism enables the present method to adapt to different types of trademarks more effectively and improves the accuracy and robustness of matching.

[0089] Specifically, in step S31, the feature matching state of the target trademark is determined according to the contour feature degree, wherein:

[0090] When the contour feature degree is greater than or equal to the contour feature reference value, the target trademark is determined to be in a strong feature matching state;

[0091] When the contour feature degree is less than the contour feature reference value, the target trademark is determined to be in a weak feature matching state.

[0092] It can be understood that the contour feature value (D value) calculated according to the formula is when the contour feature reference value is set to k1=k2=0.4 and k3=0.2.

[0093] In implementation, by comparing the contour feature degree of the target trademark with the contour feature reference value, the target trademark is classified into a strong feature matching state or a weak feature matching state, which can distinguish the degree of similarity between trademarks, thereby playing an important role in trademark infringement detection and brand protection: a strong feature matching state indicates that the trademarks have highly similar contour features, and there may be potential infringements, which require further in-depth investigation and verification; while a weak feature matching state indicates that the similarity between the trademarks is low, which may be due to other factors. Similarity requires more detailed analysis and comparison to determine the comparison trademark; this method helps to improve the efficiency and accuracy of trademark protection to promote the effective maintenance of trademark rights.

[0094] Specifically, when determining that the target trademark is in a strong feature matching state, several trademarks whose contour feature degrees are within the feature degree standard range are determined as comparison trademarks.

[0095] In implementation, the characteristic degree standard range is determined according to the characteristic degree D and comparison accuracy of the target trademark. The higher the comparison accuracy, the more comparison trademarks are screened out to increase the accuracy of infringement analysis. Therefore, the value range corresponding to the comparison accuracy is generally 0.1D~0.5D; preferably, the value range corresponding to the comparison accuracy is set to 0.3D, that is, the value range of the characteristic degree standard range is 0.7D~1.3D. Therefore, all existing trademarks with contour characteristic degrees between 0.7D and 1.3D are set as comparison trademarks.

[0096] It can be understood that when the target trademark is determined to be in a strong feature matching state, the corresponding feature standard range is determined according to its contour feature, and a series of trademarks within the range are selected as comparison trademarks based on the feature standard range to ensure that the selected comparison trademarks have similar contour features to the target trademark, which provides strong support for further comparison and analysis, helps to accurately judge the degree of similarity between trademarks, and thus effectively identify potential infringements.

[0097] Specifically, in step S32, when it is determined that the target trademark is in a weak feature matching state, the target trademark image is divided into a plurality of sub-image shapes according to the chromaticity distribution of the target trademark image;

[0098] The difference in chromaticity values ​​of pixels in a single sub-image shape is less than a preset difference.

[0099] It can be understood that the chromaticity value of a single pixel = the value of the pixel in the red channel + the value of the pixel in the green channel + the value of the pixel in the blue channel.

[0100] In implementation, the preset difference is generally set to 30-60, preferably, set to 45.

[0101] It is understandable that when the target trademark is determined to be in a weak feature matching state, the characteristics of the target trademark can be captured more finely by dividing the target trademark image into multiple sub-image shapes according to its chromaticity distribution; this segmentation method can better reflect the color and shape characteristics of the trademark image, provide more detailed information for subsequent comparison and analysis, and help to accurately judge the similarity between trademarks, thereby providing more comprehensive support and guidance for brand protection and infringement detection.

[0102] Specifically, in step S32, several candidate trademarks containing a single sub-image shape are determined, and whether a single candidate comparative trademark is a comparative trademark is determined according to the number of different sub-images contained in the single candidate comparative trademark, wherein:

[0103] If the number of different sub-images contained in a single candidate comparative trademark is greater than or equal to a preset number (half of the number of sub-images), the candidate comparative trademark is determined to be a comparative trademark;

[0104] If the number of different sub-images included in a single candidate comparative trademark is less than a preset number (half of the number of sub-images), it is determined that the candidate comparative trademark is not a comparative trademark.

[0105] In implementation, the preset number is set to one quarter to three quarters of the number of sub-images, preferably one half; it is understandable that a larger preset number means more sub-images overlap between the comparison trademark and the target trademark, and fewer comparison trademarks are determined.

[0106] In implementation, the alternative comparative trademark is determined based on whether the existing trademarks searched through big data contain a single sub-image shape. If an existing trademark contains at least one sub-image shape, the trademark is an alternative trademark.

[0107] It is understandable that by analyzing the number of different sub-images contained in the alternative comparative trademarks to determine whether to use them as comparative trademarks, comparative trademarks with sufficient characteristic information can be effectively screened out, thereby improving accuracy and efficiency in the trademark matching and infringement detection process; by reasonably setting the preset number, it can ensure that the comparative trademarks have sufficient characteristic information, which helps to more accurately identify potential trademark infringements and protect trademark rights.

[0108] Specifically, the overall matching parameter between a single comparison trademark and a target trademark is calculated according to the following formula:

[0109]

[0110] Among them, W is the overall matching parameter, rgb1 is the RGB reference value of the comparison trademark, rgb0 is the RGB reference value of the target trademark, sm is the graphic similarity, r1 is the matching weight, r2 is the similarity weight, r3 is the trademark text weight, and r1+r2+r3=1.

[0111] In practice, the color and shape of the image have the same influence on the overall matching parameter, ie, r1=r2.

[0112] Specifically, the RGB reference value is calculated according to the following formula,

[0113]

[0114] Among them, rgb is the RGB reference value of the image, R is the average value of each pixel in the red channel, G is the average value of each pixel in the green channel, B is the average value of each pixel in the blue channel, n is the number of pixels contained in a single trademark image, Ri is the value of the i-th pixel in the red channel, Gi is the value of the i-th pixel in the green channel, and Bi is the value of the i-th pixel in the blue channel.

[0115] It can be understood that: (1) if the relative position relationship between the trademark image and trademark text of a single comparison trademark is the same as the relative position relationship between the trademark image and trademark text of the target trademark, and the trademark text font of the single comparison trademark is the same as the trademark text font of the target trademark, then r3 is 0.2, and r1 = r2 = 0.4; (2) if the relative position relationship between the trademark image and trademark text of a single comparison trademark is the same as the relative position relationship between the trademark image and trademark text of the target trademark, but the trademark text font of the single comparison trademark is different from the trademark text font of the target trademark, then r3 is 0.12, and r1 = r2 = 0. 44; (3) If the relative position relationship between the trademark image and trademark text of a single comparison trademark is different from the relative position relationship between the trademark image and trademark text of the target trademark, but the trademark text font of the single comparison trademark is the same as the trademark text font of the target trademark, then r3 is 0.08, and at this time r1=r2=0.46; (4) If the relative position relationship between the trademark image and trademark text of a single comparison trademark is different from the relative position relationship between the trademark image and trademark text of the target trademark, and the trademark text font of the single comparison trademark is also different from the trademark text font of the target trademark, then r3=0, and at this time r1=r2=0.5.

[0116] In the implementation, the image similarity is determined using open source software or algorithms, such as: (1) the open source computer vision and machine learning software library OpenCV, which includes a variety of image similarity algorithms (value hash algorithm, difference hash algorithm, and perceptual hash algorithm, etc.); (2) the Java-based open source library J ImageHash, which provides a variety of hash algorithms (such as Average Hash, Difference Hash, and Perceptua l Hash) for calculating and comparing the similarity between two images; (3) the open source image comparison tool MegSpot, which provides efficient and professional image and video comparison functions.

[0117] In practice, the infringement analysis result is determined based on the value of the overall matching parameter: when the overall matching parameter between the comparison trademark and the target trademark is greater than or equal to the overall matching threshold, infringement is determined; when the overall matching parameter between the comparison trademark and the target trademark is less than the overall matching threshold, non-infringement is determined. When determining infringement, the degree of infringement is determined based on the difference between the overall matching parameter and the overall matching threshold, where the degree of infringement is positively correlated with the difference between the overall matching parameter and the overall matching threshold.

[0118] In implementation, the overall matching threshold is generally set to be greater than 0.85, preferably set to 0.9.

[0119] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A trademark infringement analysis method based on big data, It is characterized in that include: Acquire the trademark content of the target trademark, wherein the trademark content includes a trademark image and trademark text; Determining the positional relationship between the trademark image and the trademark text corresponding to the target trademark, and the font corresponding to the trademark text of the target trademark according to the trademark content; Determining a number of corresponding comparative trademarks according to the trademark image, including: Determine the number and positions of key points of the target trademark, determine the contour characteristic degree of the target trademark according to the number and positions of key points, determine the feature matching state of the target trademark according to the contour characteristic degree, and determine a number of comparison trademarks within a standard range of characteristic degree according to the feature matching state; Or, dividing the target trademark image into a plurality of sub-image shapes according to the chromaticity distribution of the target trademark image, determining a plurality of candidate comparison trademarks containing a single sub-image shape respectively, and determining whether a single candidate comparison trademark is a comparison trademark according to the number of different sub-images contained in the single candidate comparison trademark, wherein the difference in chromaticity values ​​of each pixel point in the single sub-image shape is less than a preset difference value, and the chromaticity value of a single pixel point is the sum of the value of the pixel point in the red channel, the value of the pixel point in the green channel, and the value of the pixel point in the blue channel; Extracting RGB channels of the target trademark and each of the comparison trademarks to determine RGB reference values ​​of the target trademark and each of the comparison trademarks; Comparing the target trademark with each of the comparison trademarks to determine graphic similarity, determining corresponding overall matching parameters in combination with the RGB reference value and the graphic similarity of a single comparison trademark, and determining infringement analysis results based on the overall matching parameters; The contour characteristic of the target trademark is determined according to the number of key points and the positions of the key points. The contour characteristic is calculated according to the following formula: Wherein, D is the contour feature, n is the number of key points, di is the distance between the ith key point and the center point of the trademark image, S is the area of ​​the target trademark, k1 is the quantity weight, k2 is the position weight, k3 is the area weight, and k1+k2+k3=1.

2. According to the trademark infringement analysis method based on big data according to claim 1, It is characterized in that The steps to determine the key points of the target trademark include: Obtaining the outer contour shape of the trademark image; Determine the number of straight line segments and the number of curved line segments included in the outer contour shape; Determine the sum of the number of the straight line segments and the number of the curve segments as the number of key points; Determine the position of the key point of the straight line according to the two end points of each straight line segment; Determine the position of the key point of the curve according to the two end points of each curve segment; Determine the straight line key point position and the curve key point position as key point positions; Among them, different straight line segments satisfy that two adjacent straight line segments intersect, and different curved line segments satisfy that two adjacent curved line segments have different roundness and / or different center points.

3. According to the trademark infringement analysis method based on big data as claimed in claim 2, It is characterized in that The quantity weight and the position weight are determined according to the contour complexity, wherein: When the contour complexity is greater than a preset contour complexity value, the quantity weight is positively correlated with the ratio of the contour complexity to the preset contour complexity value; When the contour complexity is less than a preset contour complexity value, the position weight is positively correlated with the ratio of the contour complexity to the preset contour complexity value; The contour complexity is determined according to the ratio of the perimeter to the area of ​​the contour.

4. According to the trademark infringement analysis method based on big data as claimed in claim 2, It is characterized in that The feature matching state of the target trademark is determined according to the contour feature degree, wherein: When the contour feature degree is greater than or equal to the contour feature reference value, the target trademark is determined to be in a strong feature matching state; When the contour feature degree is less than the contour feature reference value, the target trademark is determined to be in a weak feature matching state.

5. According to the trademark infringement analysis method based on big data according to claim 4, It is characterized in that When determining that the target trademark is in a strong feature matching state, several trademarks whose contour feature degrees are within the feature degree standard range are determined as comparison trademarks.

6. The trademark infringement analysis method based on big data according to claim 4, It is characterized in that When it is determined that the target trademark is in a weak feature matching state, the target trademark image is divided into a plurality of sub-image shapes according to the chromaticity distribution of the target trademark image.

7. The trademark infringement analysis method based on big data according to claim 6, It is characterized in that Determine a number of candidate trademarks containing a single sub-image shape, and determine whether a single candidate comparative trademark is a comparative trademark according to the number of different sub-images contained in the single candidate comparative trademark, wherein: If the number of different sub-images contained in a single candidate comparative trademark is greater than or equal to a preset number, the candidate comparative trademark is determined to be a comparative trademark; If the number of different sub-images included in a single candidate comparative trademark is less than a preset number, it is determined that the candidate comparative trademark is not a comparative trademark.

8. The trademark infringement analysis method based on big data according to claim 1, It is characterized in that The overall matching parameter of a single comparison trademark and a target trademark is calculated according to the following formula: Among them, W is the overall matching parameter, rgb1 is the RGB reference value of the comparison trademark, rgb0 is the RGB reference value of the target trademark, sm is the graphic similarity, r1 is the matching weight, r2 is the similarity weight, r3 is the trademark text weight, and r1+r2+r3=1.

9. According to the trademark infringement analysis method based on big data as claimed in claim 8, It is characterized in that The RGB reference value is calculated according to the following formula, Among them, rgb is the RGB reference value of the image, R is the average value of each pixel in the red channel, G is the average value of each pixel in the green channel, B is the average value of each pixel in the blue channel, n is the number of pixels contained in a single trademark image, Ri is the value of the i-th pixel in the red channel, Gi is the value of the i-th pixel in the green channel, and Bi is the value of the i-th pixel in the blue channel.

Citation Information

Patent Citations

  • A Trademark Infringement Analysis Method and System for E-commerce Data

    CN114529784B

  • Automatic judgment method and device for trademark infringement, electronic equipment and storage medium

    CN112507160A

  • Product infringement judgment method and device, terminal equipment and storage medium

    CN113919852A

  • Human body contour point detection method and device, electronic equipment and storage medium

    CN114120352A

  • Image processing method, device and equipment

    CN115409881A

Cited By

  • Trademark infringement risk analysis method based on big data

    CN120894045A