A trademark search method based on cloud platform

Abstract

The present invention relates to the technical field of trademark information retrieval, and in particular to a trademark retrieval method based on a cloud platform, comprising: processing a target trademark and a number of retrieval trademarks that meet similarity standards into a standard target trademark and a number of standard retrieval trademarks; segmenting the standard target trademark and each of the standard retrieval trademarks according to a trademark segmentation rule to construct a target folder and a number of retrieval folders; respectively determining a similarity standard value of the standard target trademark and a similarity reference value corresponding to each of the standard retrieval trademarks; determining a similarity parameter and a similarity fluctuation parameter according to the similarity standard value and each of the similarity reference values; determining an output ranking of the corresponding standard retrieval trademark according to the similarity parameter and the similarity fluctuation parameter of each of the standard retrieval trademarks; and outputting the retrieval trademark according to the output ranking of each of the standard retrieval trademarks; the present invention improves the accuracy and reliability of trademark retrieval.

Description

Technical Field

[0001] The present invention relates to the technical field of trademark information retrieval, and in particular to a trademark retrieval method based on a cloud platform. Background Art

[0002] Trademark search is an important procedure before submitting a trademark registration application. Its purpose is to ensure that the trademark to be registered is not identical or similar to a registered trademark. As a mark for enterprises to identify goods or services, trademarks have important brand value and competitiveness. Through trademark search, applicants can avoid trademark conflicts and protect the intellectual property rights and market share of enterprises.

[0003] Trademark search faces the challenge of judging the similarity of text, graphics, colors, etc. Therefore, modern trademark search systems use advanced technologies such as image recognition technology, natural language processing and machine learning to improve the efficiency and accuracy of searches. Trademark search plays a key role in trademark registration and protection, helping companies establish and maintain their brand image.

[0004] Chinese patent publication / announcement number CN114706946A discloses a trademark category search platform and method, including: a trademark category world language unit, which is used to establish trademark category basic data according to trademark type groups and connect to the world's multi-language library; a search input language conversion unit, which is used to identify the language of trademark search text input and convert the trademark category search country language into the search target selection country language; a language country search engine unit, which is used to receive, identify and convert trademark search voice input, and start importing search target selection country language data and / or trademark search voice input conversion data to perform trademark category search by guiding the search data engine; a search result classification display unit, which is used to summarize the trademark category search results, automatically distinguish the text and voice output of the trademark category search results, and intelligently classify and display them. It can be seen that the invention has the following problems:

[0005] The invention does not consider optimizing the output ranking of searched trademarks according to their similarity with the target trademark, which leads to inaccurate search results. Summary of the invention

[0006] To this end, the present invention provides a trademark search method based on a cloud platform to overcome the problem in the prior art of inaccurate search results caused by not optimizing the output sorting of searched trademarks according to their similarities with target trademarks.

[0007] To achieve the above purpose, the present invention provides a trademark search method based on a cloud platform, comprising:

[0008] Determine several search trademarks that meet similarity standards obtained through the search of the target trademark;

[0009] Extracting the non-text parts of the target trademark and each of the search trademarks and pre-processing them to form a target image trademark and a search image trademark, and performing format assimilation processing on each of the search image trademarks according to the target image trademark to form a standard target trademark and a plurality of standard search trademarks;

[0010] Determine the identification code of the corresponding standard search trademark according to the search order of each search trademark;

[0011] The standard target trademark and each of the standard search trademarks are segmented according to the trademark segmentation rule to construct a target folder and a plurality of search folders, wherein:

[0012] The target sub-images obtained by segmentation are numbered according to a preset numbering rule and then stored in a target folder;

[0013] The plurality of search sub-images obtained by segmenting a single standard search trademark are numbered according to a preset numbering rule and stored in a search folder corresponding to the identification code of the standard search trademark;

[0014] Determine the trademark data of each target sub-image in the target folder, and determine the corresponding similarity standard value according to the trademark data of a single target sub-image;

[0015] Determine the trademark data of each of the search sub-images in each of the search folders, and determine a corresponding similarity reference value according to the trademark data of a single search sub-image;

[0016] Extracting the similarity standard value of each of the target sub-images and the similarity reference values ​​of a number of the retrieval sub-images having the same number as the target sub-image to determine the similarity difference between each of the retrieval sub-images having the same number and the target sub-image;

[0017] Determine the similarity parameter and similarity fluctuation parameter of the corresponding search folder according to the similarity difference of all the numbers in the single search folder;

[0018] Determine the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark;

[0019] Output the searched trademarks according to the output ranking of the searched trademarks according to each standard;

[0020] Among them, the number of the search folders is equal to the number of the search trademarks, the file name of the search folder is the identification code of the corresponding standard search trademark, the trademark data includes trademark area, trademark color and trademark shape, and the trademark shape includes only straight lines, only curves and coexistence of straight lines and curves.

[0021] Furthermore, the step of forming the standard target trademark and the plurality of standard search trademarks includes:

[0022] Extracting non-text parts of the target trademark and each of the search trademarks and pre-processing them to form a target image trademark and a search image trademark;

[0023] Determining format data of the target image trademark, wherein the format data includes resolution and physical size;

[0024] Adjusting the format data of each of the search image trademarks to be the same as the format data of the target trademark to form a standard search trademark;

[0025] Wherein, the target image trademark is determined as the standard target trademark.

[0026] Furthermore, the trademark segmentation rules include:

[0027] Determining the number of horizontal pixels and vertical pixels respectively according to the resolution of the standard target trademark;

[0028] Determining the number of segmentation columns of the standard target trademark according to the number of horizontal pixel points;

[0029] Determining the number of segmentation rows of the standard target trademark according to the number of vertical pixel points;

[0030] Evenly dividing the standard target trademark into a plurality of target sub-images according to the number of segmentation columns and the number of segmentation rows;

[0031] The standard search trademark is segmented according to the segmentation method of segmenting the standard target trademark into a plurality of target sub-images.

[0032] Furthermore, the preset numbering rules include:

[0033] Determining the number of segmentation columns and the number of segmentation rows of a single target sub-image or a single search sub-image;

[0034] Determine the segmentation column number as the first three digits of the serial number of a single target sub-image or a single search sub-image;

[0035] The segmentation row number is determined as the last three digits of the serial number of a single target sub-image or a single search sub-image.

[0036] Furthermore, the identification code of the corresponding standard search trademark is determined according to the search order of each search trademark, wherein:

[0037] The logo information of the standard search trademark is positively correlated with the search order of its corresponding search trademark.

[0038] Furthermore, the file name of the search folder is the same as the identification code of the corresponding standard search trademark.

[0039] Furthermore, the corresponding similarity standard value is determined by multiplying the trademark area, trademark color and trademark shape of a single target sub-image by the product of the corresponding weight coefficients.

[0040] Furthermore, the similarity fluctuation parameter of a single standard search trademark is calculated according to the following formula:

[0041]

[0042] Among them, A is the similarity fluctuation parameter, smi j is the similarity difference between the retrieval sub-image numbered ij in a single standard retrieval trademark and the target sub-image with the corresponding number, n is the number of segmentation columns, and m is the number of segmentation rows.

[0043] Furthermore, the step of determining the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark includes:

[0044] When the similarity parameters are different, the output ranking is determined based on the absolute value of the similarity parameters of the trademarks retrieved according to each standard;

[0045] The output ranking is positively correlated with the similarity parameter.

[0046] Furthermore, the step of determining the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark also includes:

[0047] When the similarity parameters are the same, the output ranking is determined based on the magnitude of the similarity fluctuation parameters of the trademarks retrieved according to each standard;

[0048] The output ranking is positively correlated with the similarity fluctuation parameter.

[0049] Compared with the prior art, the beneficial effect of the present invention is that the trademark search method based on the cloud platform provided by the present invention realizes efficient and accurate trademark similarity analysis through the above-mentioned fine processing steps: extracting and preprocessing the non-text parts of the target trademark and the search trademark, effectively eliminating text interference and improving the accuracy of image recognition; through format assimilation processing and numbering rules, the trademark data is compared under a unified standard, ensuring the consistency and comparability of the search trademark and the target trademark, and by segmenting the standard target trademark and the standard search trademark and storing the corresponding sub-images after segmentation in their respective corresponding folders, so that the image information of each trademark can be systematically managed and facilitate subsequent similarity calculations; using the standard target trademark similarity standard value and the standard search trademark similarity reference value, the similarity difference and fluctuation parameters are determined to accurately evaluate the similarity between each search trademark and the target trademark, and provide search results according to the output ranking. This systematic and refined processing method not only improves the accuracy and reliability of the search results, but also greatly improves the efficiency of trademark search, providing users with more convenient and intelligent trademark search services.

[0050] Furthermore, the present invention ensures the purity and accuracy of the trademark image by extracting the non-text part and performing preprocessing; unifies the resolution and physical size of the target image trademark and the retrieval image trademark through the format assimilation processing step, eliminates the errors caused by the difference in image resolution and size to determine the consistency of format data in the subsequent process to increase the accuracy of similarity evaluation; the formed standard target trademark and standard retrieval trademark can ensure that the trademark images are compared under the same benchmark, thereby improving the accuracy and reliability of image matching and retrieval.

[0051] Furthermore, the trademark segmentation rules ensure effective image segmentation of the standard target trademark and the standard retrieval trademark. By determining the number of horizontal and vertical pixels of the resolution, the size accuracy of the trademark image is ensured; according to the number of horizontal and vertical pixels, the number of segmentation columns and rows of the standard target trademark is determined, thereby achieving uniform segmentation of the trademark image; by evenly segmenting the standard target trademark into multiple target sub-images, the different parts of the trademark are effectively extracted, providing a basis for subsequent comparison and search; using the same segmentation method to segment the standard retrieval trademark ensures that the number of retrieval sub-images is consistent with the number of target sub-images, ensuring the accuracy of the subsequent matching process, improving the accuracy and consistency of trademark segmentation, and providing a reliable basis for matching and retrieval of trademark images.

[0052] Furthermore, the present invention determines the output ranking by comprehensively considering the similarity parameters and similarity fluctuation parameters of the standard search trademarks, thereby ensuring the accuracy and stability of the trademark search process; the output ranking is determined by the absolute value of the similarity parameter, which can directly reflect the matching degree between each standard search trademark and the target trademark, ensuring that trademarks with high similarity are displayed first; when the similarity parameters of multiple standard search trademarks are the same, ranking is performed according to the size of the similarity fluctuation parameter, which can further distinguish trademarks with consistent similarity parameters. By considering the fluctuation of the similarity, the output ranking not only considers the matching degree, but also takes into account the stability of the matching results, thereby improving the accuracy and reliability of trademark search, and ensuring that users can obtain more objective and detailed ranking results when comparing and selecting trademarks. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a step diagram of a trademark search method based on a cloud platform according to an embodiment of the present invention;

[0054] Figure 2 A diagram of steps for forming the standard target trademark and a plurality of the standard search trademarks according to an embodiment of the present invention;

[0055] Figure 3 4 is a step diagram of a trademark segmentation rule according to an embodiment of the present invention. DETAILED DESCRIPTION

[0056] 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.

[0057] 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.

[0058] 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.

[0059] 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.

[0060] See also Figure 1 As shown, it is a step diagram of a trademark search method based on a cloud platform according to an embodiment of the present invention. The present invention provides a trademark search method based on a cloud platform, comprising:

[0061] Step S1, determining a number of search trademarks that meet the similarity standard obtained by searching the target trademark; it can be understood that in the prior art, when determining whether a search trademark is a search trademark, each has its own conditions for meeting the similarity standard, and the present invention will not repeat them here;

[0062] Step S2, extracting the non-text parts of the target trademark and each of the search trademarks and pre-processing them to form a target image trademark and a search image trademark, and performing format assimilation processing on each of the search image trademarks according to the target image trademark to form a standard target trademark and a plurality of standard search trademarks;

[0063] Step S3, determining the identification code of the corresponding standard search trademark according to the search order of each search trademark; it can be understood that the identification codes of any two standard search trademarks are different; in practice, if there are x search marks, the identification code can be mark 1, mark 2, ..., mark x, or 1, 2, ..., x, or any name with numbers;

[0064] Step S4, segmenting the standard target trademark and each of the standard search trademarks according to the trademark segmentation rule to construct a target folder and a plurality of search folders, wherein:

[0065] Step S41, numbering the target sub-images obtained by segmentation according to a preset numbering rule and storing them in a target folder;

[0066] Step S42, a plurality of search sub-images obtained by segmenting a single standard search trademark are numbered according to a preset numbering rule and stored in a search folder corresponding to the identification code of the standard search trademark;

[0067] Step S5, determining the trademark data of each target sub-image in the target folder, and determining a corresponding similarity standard value according to the trademark data of a single target sub-image;

[0068] Step S6, determining the trademark data of each of the search sub-images in each of the search folders, and determining a corresponding similarity reference value according to the trademark data of a single search sub-image;

[0069] Step S7, extracting the similarity standard value of each target sub-image and the similarity reference values ​​of several retrieval sub-images with the same number as the target sub-image to determine the similarity difference between each retrieval sub-image with the same number and the target sub-image; it can be understood that, assuming that the similarity standard value of the target sub-image numbered 002003 and the similarity reference values ​​of all retrieval sub-images numbered 002003 are extracted, the similarity reference value of any retrieval sub-image numbered 002003 is subtracted from the similarity standard value to determine the similarity difference of the corresponding retrieval sub-image numbered 002003; the number of retrieval sub-images with the same number as a single target sub-image is equal to the number of retrieval trademarks;

[0070] Step S8, determining the similarity parameter and similarity fluctuation parameter of the corresponding search folder (corresponding standard search trademark, i.e. corresponding search trademark) according to the similarity difference of all serial numbers in the single search folder;

[0071] Step S9, determining the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark;

[0072] Step S10, outputting the searched trademarks according to the output ranking of the searched trademarks according to each standard;

[0073] Among them, the number of the search folders is equal to the number of the search trademarks, the file name of the search folder is the identification code of the corresponding standard search trademark, the trademark data includes trademark area, trademark color and trademark shape, and the trademark shape includes only straight lines, only curves and coexistence of straight lines and curves.

[0074] In practice, the trademark search method provided by the prior art will obtain a series of search trademarks that meet the similarity standards. After uploading the target trademark, the method provided by the present invention can optimize the output sorting of the search trademarks, and the trademarks with higher rankings are more closely aligned with the target trademark.

[0075] It can be understood that the trademark search method based on the cloud platform provided by the present invention realizes efficient and accurate trademark similarity analysis through the above-mentioned fine processing steps: extracting and preprocessing the non-text parts of the target trademark and the search trademark, effectively eliminating text interference and improving the accuracy of image recognition; through format assimilation processing and numbering rules, the trademark data is compared under a unified standard, ensuring the consistency and comparability of the search trademark and the target trademark, and by segmenting the standard target trademark and the standard search trademark and storing the corresponding sub-images after segmentation in their respective corresponding folders, so that the image information of each trademark can be systematically managed and facilitate subsequent similarity calculations; using the standard target trademark similarity standard value and the standard search trademark similarity reference value, the similarity difference and fluctuation parameters are determined to accurately evaluate the similarity between each search trademark and the target trademark, and provide search results according to the output ranking. This systematic and refined processing method not only improves the accuracy and reliability of the search results, but also greatly improves the efficiency of trademark search, providing users with a more convenient and intelligent trademark search service.

[0076] See also Figure 2 As shown, it is a step diagram of forming the standard target trademark and several standard search trademarks in an embodiment of the present invention. Specifically, in step S2, the step of forming the standard target trademark and several standard search trademarks includes:

[0077] Step S21, extracting the non-text parts of the target trademark and each of the search trademarks and pre-processing them to form a target image trademark and a search image trademark;

[0078] Step S22, determining the format data of the target image trademark, wherein the format data includes resolution and physical size;

[0079] Step S23, adjusting the format data of each of the search image trademarks to be the same as the format data of the target trademark to form a standard search trademark;

[0080] Wherein, the target image trademark is determined as the standard target trademark.

[0081] It can be understood that the steps of forming a standard target trademark and several standard search trademarks determined in steps S21 to S23 ensure the purity and accuracy of the trademark image by extracting the non-text part and performing preprocessing; the resolution and physical size of the target image trademark and the search image trademark are unified through the format assimilation processing step, and the errors caused by the difference in image resolution and size are eliminated to determine the consistency of format data in the subsequent process to increase the accuracy of similarity evaluation; the formed standard target trademark and standard search trademark can ensure that the trademark images are compared under the same benchmark, thereby improving the accuracy and reliability of image matching and retrieval.

[0082] In implementation, in step S21, the preprocessing step includes:

[0083] Step S211, denoising the trademark image to reduce noise in the image and improve image quality

[0084] Step S212, smoothing the trademark image using a filter, such as mean filtering, Gaussian filtering, etc.

[0085] Step S213: sharpen the smoothed image to enhance the edges and details of the image, such as by using a Laplacian operator.

[0086] See also Figure 3 As shown, it is a step diagram of the trademark segmentation rule of an embodiment of the present invention. Specifically, in step S4, the steps of the trademark segmentation rule include:

[0087] Step S401, determining the number of horizontal pixels and vertical pixels respectively according to the resolution of the standard target trademark; it can be understood that the resolution is composed of horizontal pixels and vertical pixels; in one implementation, the resolution of the standard target trademark and the standard search trademark is 1920x1080, the horizontal pixels are 1920, and the vertical pixels are 1080; in one implementation, the resolution of the standard target trademark and the standard search trademark is 720x720 pixels, the horizontal pixels are 720, and the vertical pixels are 720;

[0088] Step S402, determining the number of segmentation columns of the standard target trademark according to the number of horizontal pixel points;

[0089] Step S403, determining the number of segmentation lines of the standard target trademark according to the number of vertical pixel points;

[0090] Step S404, dividing the standard target trademark into a plurality of target sub-images evenly according to the number of segmentation columns and the number of segmentation rows; it can be understood that the number of target sub-images = the number of segmentation columns × the number of segmentation rows;

[0091] Step S405, the standard search trademark is segmented into a number of target sub-images according to the segmentation method of the standard target trademark; it can be understood that the number of search sub-images obtained after segmentation of a single standard search trademark = the number of target sub-images = the number of segmentation columns × the number of segmentation rows.

[0092] It can be understood that the resolution of a single (target / retrieval) sub-image should be between 10×10 and 30×30; preferably, the resolution of a single (target / retrieval) sub-image is 10×10; therefore, the number of segmentation columns = horizontal pixels ÷ 30 ~ horizontal pixels ÷ 10, the number of segmentation rows = vertical pixels ÷ 30 ~ vertical pixels ÷ 10, that is, the number of segmentation columns is selected between 0.03 times the horizontal pixels and 0.1 times the horizontal pixels, and the number of segmentation rows is selected between 0.03 times the vertical pixels and 0.1 times the vertical pixels; preferably, the number of segmentation columns is 0.1 times the horizontal pixels, and the number of segmentation rows is 0.1 times the vertical pixels (10).

[0093] It can be understood that the trademark segmentation rule ensures effective image segmentation of the standard target trademark and the standard search trademark, and ensures the size accuracy of the trademark image by determining the number of horizontal and vertical pixels of the resolution; determines the number of segmentation columns and rows of the standard target trademark based on the number of horizontal and vertical pixels, thereby achieving uniform segmentation of the trademark image; effectively extracts different parts of the trademark by evenly segmenting the standard target trademark into multiple target sub-images, providing a basis for subsequent comparison and search; uses the same segmentation method to segment the standard search trademark, ensuring that the number of search sub-images is consistent with the number of target sub-images, ensuring the accuracy of the subsequent matching process, improving the accuracy and consistency of trademark segmentation, and providing a reliable basis for matching and retrieval of trademark images.

[0094] Specifically, in step S4, the preset numbering rules include:

[0095] Determining the number of segmentation columns and the number of segmentation rows of a single target sub-image or a single search sub-image;

[0096] Determine the segmentation column number as the first three digits of the serial number of a single target sub-image or a single search sub-image;

[0097] The segmentation row number is determined as the last three digits of the serial number of a single target sub-image or a single search sub-image.

[0098] Specifically, in step S3, the identification code of the corresponding standard search trademark is determined according to the search order of each search trademark, wherein:

[0099] The logo information of the standard search trademark is positively correlated with the search order of its corresponding search trademark.

[0100] Specifically, in step S42, the file name of the search folder is the same as the identification code of the corresponding standard search trademark.

[0101] Specifically, in step S5, the corresponding similarity standard value is determined according to the product of the trademark area, trademark color and trademark shape of a single target sub-image and the corresponding weight coefficients.

[0102] In implementation, the similarity standard value of a single target sub-image is calculated according to the following formula,

[0103] SM0=a1×S0+a2×C0+a3×L0,

[0104] Among them, SM0 is the similarity standard value, S0 is the trademark area in a single target sub-image, C0 is the trademark color of a single target sub-image, L0 is the trademark shape of a single target sub-image, a1 is the area weight, a2 is the color weight, and a3 is the shape weight.

[0105] It can be understood that the trademark color of a single target sub-image is the average chromaticity value of all pixels in the target sub-image, where 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.

[0106] It can be understood that the value selection rules of the trademark shape of a single target sub-image include: when the trademark shape is only a straight line or only a curve, L=1; when the trademark shape is a combination of straight lines and curves, L=2.

[0107] In implementation, a1+a2+a3=1, wherein a1=a2∈[0.4,0.45], a3∈[0.1,0.2]; preferably, a1=a2=0.4, a3=0.2.

[0108] In step S6, the corresponding similarity reference value is determined according to the trademark area, trademark color and trademark shape of a single search sub-image. The similarity reference value is calculated according to the following formula:

[0109] SMij=b1×Sij+b2×Cij+b3×Lij,

[0110] Among them, SMi j is the similarity reference value of the retrieval sub-image numbered ij, Si j is the trademark area in the retrieval sub-image, Ci j is the trademark color of the retrieval sub-image, Li j is the trademark shape of the retrieval sub-image, b1 is the area weight of the retrieval sub-image, b2 is the color weight of the retrieval sub-image, and b3 is the shape weight of the retrieval sub-image.

[0111] In implementation, b1∈[0.4,0.45], the value of b1 is determined based on a1, S0 and Si j: when Si j>S0, b1=Si j÷S0×a1; when Si j<S0, b1=a1; when the calculated value of b1 is greater than 0.45, b1 is taken as 0.45.

[0112] In implementation, b2∈[0.35,0.45], the value of b2 is determined based on a2, C0 and Ci j: b2=Ci j÷C0×a2; when the calculated value of b2 is greater than 0.45, b2 is 0.45, and when the calculated value of b2 is less than 0.35, b2 is 0.35.

[0113] In implementation, if the trademark shape of a single retrieval sub-image is exactly the same as the trademark shape of the corresponding target sub-image, then Li j=L0, and b3=a3; if the trademark shape of a single retrieval sub-image is completely different from the trademark shape of the corresponding target sub-image, then Li j=0, and b3=0; if the trademark shape of a single retrieval sub-image has the same part as the trademark shape of the corresponding target sub-image, then Li j=0.5×L0, and b3=0.5×a3.

[0114] Specifically, the similarity fluctuation parameter of a single standard search trademark is calculated according to the following formula:

[0115]

[0116] Among them, A is the similarity fluctuation parameter, smi j is the similarity difference between the retrieval sub-image numbered ij in a single standard retrieval trademark and the target sub-image with the corresponding number, n is the number of segmentation columns, and m is the number of segmentation rows.

[0117] In implementation, smi j = SMi j - SM0.

[0118] In implementation, the similarity parameter of a single search folder (corresponding to a standard search trademark, ie, corresponding to a search trademark) is the average value of the similarity differences in the folder.

[0119] Specifically, the step of determining the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark includes:

[0120] When the similarity parameters are different, the output ranking is determined based on the absolute value of the similarity parameters of the trademarks retrieved according to each standard;

[0121] The output ranking is positively correlated with the absolute value of the similarity parameter, that is, the smaller the absolute value of the similarity parameter (corresponding to the searched trademark), the smaller the output ranking (ie, the higher the ranking).

[0122] Specifically, the step of determining the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark also includes:

[0123] When the similarity parameters are the same, the output ranking is determined based on the magnitude of the similarity fluctuation parameters of the trademarks retrieved according to each standard;

[0124] The output ranking is positively correlated with the similarity fluctuation parameter, that is, for standard search trademarks with equal similarity parameters, the smaller the similarity fluctuation parameter (corresponding search trademark), the smaller the output ranking (ie, the higher the ranking).

[0125] It can be understood that the present invention determines the output ranking by comprehensively considering the similarity parameters and similarity fluctuation parameters of the standard search trademarks, thereby ensuring the accuracy and stability of the trademark search process; determining the output ranking by the absolute value of the similarity parameter can directly reflect the degree of matching between each standard search trademark and the target trademark, ensuring that trademarks with high similarity are displayed first; when the similarity parameters of multiple standard search trademarks are the same, ranking them according to the size of the similarity fluctuation parameter can further distinguish trademarks with consistent similarity parameters, and by considering the fluctuation of the similarity, the output ranking not only considers the degree of matching, but also takes into account the stability of the matching results, thereby improving the accuracy and reliability of trademark search, and ensuring that users can obtain more objective and detailed ranking results when comparing and selecting trademarks.

[0126] It can be understood that the similarity standard value, similarity reference value, similarity difference value, similarity parameter and similarity fluctuation parameter mentioned in the present invention are all dimensionless values, and the formulas corresponding to each value are calculated without dimension during calculation.

[0127] 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.

[0128] 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 search method based on a cloud platform, characterized in that: include: Determine several search trademarks that meet similarity standards obtained through the search of the target trademark; Extracting the non-text parts of the target trademark and each of the search trademarks and pre-processing them to form a target image trademark and a search image trademark, and performing format assimilation processing on each of the search image trademarks according to the target image trademark to form a standard target trademark and a plurality of standard search trademarks; Determine the identification code of the corresponding standard search trademark according to the search order of each search trademark; The standard target trademark and each of the standard search trademarks are segmented according to the trademark segmentation rule to construct a target folder and a plurality of search folders, wherein: The target sub-images obtained by segmentation are numbered according to a preset numbering rule and then stored in a target folder; The plurality of search sub-images obtained by segmenting a single standard search trademark are numbered according to a preset numbering rule and stored in a search folder corresponding to the identification code of the standard search trademark; Determine the trademark data of each target sub-image in the target folder, and determine the corresponding similarity standard value according to the trademark data of a single target sub-image; The specific process of the similarity standard value is to determine the corresponding similarity standard value according to the product of the trademark area, trademark color and trademark shape of a single target sub-image and the corresponding weight coefficient respectively; Determine the trademark data of each of the search sub-images in each of the search folders, and determine a corresponding similarity reference value according to the trademark data of a single search sub-image; Extracting the similarity standard value of each of the target sub-images and the similarity reference values ​​of a number of the retrieval sub-images having the same number as the target sub-image to determine the similarity difference between each of the retrieval sub-images having the same number and the target sub-image; Determine the similarity parameter and similarity fluctuation parameter of the corresponding search folder according to the similarity difference of all the numbers in the single search folder; Among them, the similarity fluctuation parameter of a single standard search trademark is calculated according to the following formula: Among them, A is the similarity fluctuation parameter, smij is the similarity difference between the search sub-image numbered ij in a single standard search trademark and the target sub-image with the corresponding number, n is the number of segmentation columns, and m is the number of segmentation rows; Determine the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark; Output the searched trademarks according to the output ranking of the searched trademarks according to each standard; Among them, the number of the search folders is equal to the number of the search trademarks, the file name of the search folder is the identification code of the corresponding standard search trademark, the trademark data includes trademark area, trademark color and trademark shape, and the trademark shape includes only straight lines, only curves and coexistence of straight lines and curves.

2. The trademark search method based on the cloud platform according to claim 1, characterized in that: The steps of forming the standard target trademark and the plurality of standard search trademarks include: Extracting non-text parts of the target trademark and each of the search trademarks and pre-processing them to form a target image trademark and a search image trademark; Determining format data of the target image trademark, wherein the format data includes resolution and physical size; Adjusting the format data of each of the search image trademarks to be the same as the format data of the target trademark to form a standard search trademark; The target image trademark is determined as the standard target trademark.

3. The trademark search method based on the cloud platform according to claim 1, characterized in that: The trademark segmentation rules include: Determining the number of horizontal pixels and vertical pixels respectively according to the resolution of the standard target trademark; Determining the number of segmentation columns of the standard target trademark according to the number of horizontal pixel points; Determining the number of segmentation rows of the standard target trademark according to the number of vertical pixel points; Evenly dividing the standard target trademark into a plurality of target sub-images according to the number of segmentation columns and the number of segmentation rows; The standard search trademark is segmented according to the segmentation method of segmenting the standard target trademark into a plurality of target sub-images.

4. The trademark search method based on the cloud platform according to claim 3 is characterized in that: The preset numbering rules include: Determining the number of segmentation columns and the number of segmentation rows of a single target sub-image or a single search sub-image; Determine the segmentation column number as the first three digits of the serial number of a single target sub-image or a single search sub-image; The segmentation row number is determined as the last three digits of the serial number of a single target sub-image or a single search sub-image.

5. The trademark search method based on cloud platform according to claim 1, characterized in that: The identification code of the corresponding standard search trademark is determined according to the search order of each search trademark, wherein: The logo information of the standard search trademark is positively correlated with the search order of its corresponding search trademark.

6. The trademark search method based on cloud platform according to claim 5, characterized in that: The file name of the search folder is the same as the identification code of the corresponding standard search trademark.

7. The trademark search method based on the cloud platform according to claim 1, characterized in that: The step of determining the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark, include, When the similarity parameters are different, the output ranking is determined based on the absolute value of the similarity parameters of the trademarks retrieved according to each standard; The output ranking is positively correlated with the similarity parameter.

8. The trademark search method based on the cloud platform according to claim 7, characterized in that: The step of determining the output ranking of the corresponding standard search trademark according to the similarity parameter and similarity fluctuation parameter of each standard search trademark also includes: When the similarity parameters are the same, the output ranking is determined based on the magnitude of the similarity fluctuation parameters of the trademarks retrieved according to each standard; The output ranking is positively correlated with the similarity fluctuation parameter.

Images