Visual portrait data information matching method based on AI technology

Abstract

The invention relates to the technical field of data matching, in particular to an AI technology-based visual portrait data information matching method. According to the technical scheme, the method comprises the following steps: analyzing actual appearance information and a plurality of template databases to obtain subdivided personal information; the method comprises the steps of obtaining merchant demand information, analyzing the merchant demand information, determining matched subdivided personal information according to an analysis result, determining a matched design object according to the subdivided personal information, generating recommendation information, and recommending the design object closest to the merchant demand to a merchant according to the recommendation information. The technology provided by the invention comprises the steps of effectively matching the visual portrait data of the design object with the template database of the expected object, and recommending an anchor which most meets the requirements of merchants to the merchants, so that accurate propaganda and popularization are realized.

Description

Technical Field

[0001] The present application relates to the technical field of data matching, and in particular to an information matching method for visual portrait data based on AI technology. Background Art

[0002] At present, personal image design and transformation mainly focus on dressing and makeup, while ignoring important factors that constitute the overall personal image, such as consumption concepts, living habits, activity areas and footprints. These deep-seated factors are often only valued and carefully crafted by Internet celebrities, public figures or specific professional groups in order to shape their external image;

[0003] In existing technologies, some super-large e-commerce or service platforms try to collect and analyze users' browsing, interaction habits and personal information data, and use big data algorithms to convert them into labels, so as to give priority to recommending relevant consumer product content;

[0004] However, this recommendation mechanism based on big data does not actually analyze the target users (i.e. consumers) from a visual perspective. Therefore, even if the user's previous choice is not suitable for him / her, due to the inertial recommendation of the big data algorithm, the user may continue to receive wrong recommendation information, and thus be unable to effectively discover and try the consumption choices that are truly suitable for him / her.

[0005] In addition, existing technologies fail to fully consider the reality that "users do not necessarily understand the differences between themselves and social standards." This neglect leads to biased recommendation results produced by big data algorithms, affecting the accuracy, authenticity and suitability of the recommendations. Summary of the invention

[0006] The present application provides an information matching method for visual portrait data based on AI technology to solve the problem that the existing technology is unable to comprehensively consider the user's visual image and deep-level characteristics to achieve a precise matching and recommendation service model.

[0007] This application provides an information matching method for visual portrait data based on AI technology, including:

[0008] Get the basic information of several expected objects, and build a template database based on each basic information.

[0009] Acquire personal information of the design object, obtain a number of actual images, and determine actual appearance information of the design object based on the number of actual images;

[0010] Analyze the actual appearance information with several template databases to obtain segmented personal information;

[0011] Obtain merchant demand information, analyze the merchant demand information, determine matching segmented personal information based on the analysis results, determine matching design objects based on the segmented personal information, and generate recommendation information; based on the recommendation information, recommend the design object that is closest to the merchant's needs to the merchant.

[0012] The use of the above technology has the following advantages: it can effectively match the visual portrait data of the design object with the template database of the desired object, and recommend the anchor that best meets their needs to merchants, thereby achieving accurate publicity and promotion.

[0013] Optionally, the template database includes: expected appearance information and expected behavior information;

[0014] Analyze the actual appearance information with the desired appearance information respectively, determine the overall matching value between the actual appearance information and each desired appearance information, and obtain a number of overall matching values.

[0015] Obtaining the overall matching value with the highest value, generating similar appearance information according to the overall matching value with the highest value, generating an appearance tag according to the similar appearance information, and attaching the appearance tag to the personal information to obtain segmented personal information;

[0016] and determining a desired object matching the design object based on the similar appearance information, and then determining the similar object;

[0017] The expected behavior data of the similar object is obtained, and behavior guidance information is generated. The daily behavior of the design object is guided by the behavior guidance information, so that the design object is more consistent with the corresponding similar object in terms of personal style.

[0018] The use of the above technology has the following advantages: it can help the design objects find similarities with the expected objects in appearance, and can also make them more closely match the corresponding similar objects in personal style through behavioral guidance information.

[0019] Optionally, obtain local appearance information through actual appearance information.

[0020] Through some expected appearance information, some local expected information is obtained.

[0021] The local appearance information is compared with several local expected information respectively to obtain several local matching values, and the local matching value with the highest value is obtained to obtain the local similarity information. According to the local similarity information, a local tag is generated, and the local tag is attached to the basic information of the design object to obtain the local similarity information. The local similarity information is used to inform the staff which expected object the design object is more similar to in appearance, and then special design is performed on the part of the design object.

[0022] The use of the above technology has the following advantages: it successfully realizes personalized packaging recommendations for the design object, making its appearance closer to the star image expected by the merchants or fans.

[0023] Optionally, a plurality of appearance images of the similar object are acquired, salient local information of the similar object is acquired based on the plurality of appearance images, and a local acquisition instruction is generated based on the salient local information;

[0024] Through the local acquisition instruction, the corresponding local information of the design object is obtained, the corresponding local information is compared with the prominent local information to obtain the local similarity value, and the local similarity value is compared with the preset comparison value. If the local similarity value is less than the preset comparison value, avoidance information is generated and sent to the terminal device. The avoidance prompt information is generated through the terminal device to inform the merchant that the part of the design object is not very similar to the corresponding star.

[0025] The use of the above technology has the following advantages: packaging and promotion strategies can be formulated in a more targeted manner, which can fully utilize the similarities between the anchor and the celebrity to attract the audience and effectively avoid discordant factors in the promotion.

[0026] Optionally, obtain the actual information of the merchant and determine the merchant keyword data set based on the actual information of the merchant.

[0027] Analyze several expected behavior information respectively, and obtain expected keyword data according to the characteristics of each expected object;

[0028] Analyze the expected keyword data of the merchant keyword data set to obtain the keyword matching value.

[0029] The keyword matching value is compared with the preset matching threshold. If the keyword matching value is less than the preset matching threshold, a first elimination information is generated and sent to the terminal device.

[0030] If the keyword matching value is greater than the preset matching threshold, a reservation message is generated and sent to the terminal device.

[0031] According to the reserved information, the corresponding expected object is determined, according to the corresponding expected object, the corresponding appearance mark is retrieved, and according to the appearance mark, the corresponding design object is determined.

[0032] The use of the above technology has the following advantages: it achieves accurate matching of the desired object and the design object. This method not only improves the efficiency and accuracy of matching, but also helps merchants find the design object that best meets their needs, thereby achieving better publicity effects and commercial benefits.

[0033] Optionally, when the keyword matching value is greater than a preset matching threshold, a pre-selected tag is generated, and the pre-selected tag is attached to the corresponding expected keyword data to obtain the pre-selected keyword data.

[0034] Based on the pre-selected keyword data, an in-depth analysis instruction is generated. According to the in-depth analysis instruction, the pre-selected keyword data is analyzed through the merchant keyword data set. If the pre-selected keyword data contains at least one risk word, a second elimination information is generated, and the corresponding expected object is eliminated through the second elimination information.

[0035] The use of the above technology has the following advantages: it ensures that the image and style of the partner are consistent with the brand image, and avoids potential risks to the brand due to personal problems of the star.

[0036] Optionally, later, when the design object is evaluated again, some weight influence information is obtained according to some overall matching values, and evaluation correction information is obtained according to some weight influence information.

[0037] Before the actual appearance information is compared with the expected appearance information again, the evaluation correction information is added to the template database to obtain an improved template database, and the actual appearance information is evaluated through the improved template database.

[0038] The use of the above technology has the following advantages: using the historical data of the previous evaluation in the secondary evaluation can significantly improve the accuracy and reliability of the evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0040] Figure 1 A schematic diagram of an application scenario provided for an embodiment of the present application;

[0041] Figure 2 A flowchart of an information matching method for visual portrait data based on AI technology provided in one embodiment of the present application; DETAILED DESCRIPTION

[0042] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0043] In addition, the term "and / or" in this article is only a description of the association relationship of the associated objects by the information matching method of visual portrait data based on AI technology, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally indicates that the previous and next associated objects are in an "or" relationship of the information matching method of visual portrait data based on AI technology.

[0044] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0045] Existing technologies cannot truly analyze target users (i.e., consumers) from a visual perspective. Therefore, even if the user's previous choice is not suitable for them, due to the inertial recommendation of the big data algorithm, the user may continue to receive wrong recommendation information, and thus be unable to effectively discover and try consumption choices that are truly suitable for them.

[0046] Based on this, the present application provides an information matching method for visual portrait data based on AI technology.

[0047] It can effectively match the visual portrait data of the design object with the template database of the desired object, and recommend the anchor that best meets their needs to merchants, thereby achieving accurate publicity and promotion.

[0048] Figure 1 A schematic diagram of an application scenario of an information matching method for visual portrait data based on AI technology provided in this application.

[0049] When packaging and modifying a design object in need, the method of the present application is applied to analyze the object involved, and the information matching method of visual portrait data is applicable to an interactive platform, which includes a terminal device and a server.

[0050] Figure 2 This is a flow chart of an information matching method for visual portrait data based on AI technology provided in one embodiment of the present application. Figure 2 As shown, the method includes:

[0051] S201, obtaining basic information of several desired objects, and establishing a template database according to each of the basic information,

[0052] The personal information of the design object is acquired to obtain a plurality of actual images, and the actual appearance information of the design object is determined according to the plurality of actual images.

[0053] Specifically, we obtained basic information of several desired objects (i.e. existing stars) from public channels or partners, including but not limited to their names, ages, genders, occupations, appearance characteristics, behavioral habits, etc. Based on this information, we established independent template databases for each desired object. These databases recorded their appearance information and behavioral characteristics in detail, providing data support for subsequent information matching.

[0054] Next, the personal information of the design objects (i.e. the anchors who needed packaging) was obtained, their actual images were taken using high-definition cameras and other equipment, and the actual appearance information of the design objects, such as facial features, body proportions, skin color, etc., was determined based on these images.

[0055] S202: Analyze the actual appearance information with the template databases to obtain segmented personal information.

[0056] Specifically, the actual appearance information of the design object is compared and analyzed in detail with the appearance information of the expected object in several template databases, and the similarity between them is calculated through an AI algorithm to obtain detailed personal information about the design object. This information reveals which expected objects the design object is closest to in appearance features, as well as the differences between them.

[0057] S203, obtaining merchant demand information, analyzing the merchant demand information, determining the matching segmented personal information based on the analysis result, determining the matching design object based on the segmented personal information, and generating recommendation information, and recommending the design object that is closest to the merchant demand to the merchant based on the recommendation information.

[0058] Specifically, we further obtained the merchants' demand information, including their desired anchor type, target audience, promotional style, etc. By analyzing this information, we determined the segmented personal information that best matched the merchants' needs, and determined the corresponding design objects accordingly.

[0059] Finally, recommendation information is generated and the design objects that are closest to the merchant’s needs are recommended to the merchant. This information is presented in the form of an intuitive report, including the actual image of the design object, similarity analysis with the expected object, possible publicity strategies, etc., providing powerful decision-making support for the merchant.

[0060] Through the method provided in this embodiment, the visual portrait data of the design object can be effectively matched with the template database of the desired object, and the anchor who best meets their needs can be recommended to the merchants, thereby achieving accurate publicity and promotion.

[0061] In some embodiments, the template database includes: expected appearance information and expected behavior information; analyzing the actual appearance information separately with several expected appearance information, determining the overall matching value of the actual appearance information and each expected appearance information, obtaining several overall matching values, obtaining the overall matching value with the highest value, generating similar appearance information according to the overall matching value with the highest value, generating an appearance tag according to the similar appearance information, attaching the appearance tag to the personal information, and obtaining segmented personal information; and determining the expected object matching the design object according to the similar appearance information, and then determining the similar object; obtaining the expected behavior data of the similar object, generating the behavior guidance information, and guiding the daily behavior of the design object through the behavior guidance information, so that the design object is more in line with the corresponding similar object in personal style.

[0062] Specifically, basic information of several desired objects (stars) is obtained, including their appearance characteristics (such as face shape, facial features, hairstyle, etc.) and behavior characteristics (such as speech and behavior, clothing style, social habits, etc.). Based on this information, a template database containing expected appearance information and expected behavior information is established for each desired object.

[0063] Next, we obtain the personal information of the design object (new anchor) and obtain their actual images by shooting or uploading them. We use AI technology to extract the actual appearance information of the design object, such as face shape, facial features, etc., from these actual images.

[0064] Then, the actual appearance information of the design object is analyzed and compared with the expected appearance information in the template database, and the overall matching value is calculated. This value reflects the similarity between the design object and each expected object in appearance. The overall matching value with the highest value is selected, and it is considered that the expected object corresponding to this value is most similar to the design object in appearance. Thus, similar appearance information is generated, and an appearance tag is generated based on this information. This tag is attached to the personal information of the design object to form segmented personal information.

[0065] In addition, the expected object that best matches the design object, i.e., the similar object, is determined based on the similar appearance information. Then, the expected behavior data of the similar object, such as speech, behavior, and clothing style, is obtained to generate behavior guidance information. This information is sent to the design object through the terminal device to guide their daily behavior so that they are more in line with the corresponding similar object in terms of personal style.

[0066] When promoting suitable design objects, they can search and screen out the design objects that best meet their requirements on the platform based on their own demand information. Since the design objects have been attached with appearance tags and behavior guidelines similar to the expected objects, merchants can more easily find design objects that are similar to the expected objects in appearance and behavior, thereby meeting their own promotion needs.

[0067] The method provided in this embodiment can help the design objects find similarities with the expected objects in appearance, and can also help them better fit the corresponding similar objects in personal style through behavioral guidance information, thereby improving the matching degree between the design objects and the expected objects and meeting the promotion needs of merchants.

[0068] In some embodiments, local appearance information is obtained through actual appearance information, and some local expected information is obtained through some expected appearance information. The local appearance information is compared with some local expected information respectively to obtain some local matching values. The local matching value with the highest value is obtained to obtain local similarity information. According to the local similarity information, a local tag is generated. The local tag is attached to the basic information of the design object to obtain local similarity information. The local similarity information is used to inform the staff which expected object the local appearance of the design object is more similar to, and then the local design of the design object is specially designed.

[0069] Specifically, the actual appearance information of the design object is obtained, including its overall appearance and detailed images of each part, such as eyes, nose, mouth, etc. Then, these local images are recognized and analyzed using AI technology to extract key local appearance features.

[0070] Next, the expected appearance information of multiple expected objects is retrieved from the template database. Similarly, the local appearance information of these expected objects is extracted and analyzed. In this way, a database containing the local appearance features of multiple expected objects is obtained.

[0071] Then, the local appearance information of the design object is compared with the local appearance information of the expected object one by one, and by calculating the similarity between the two, a number of local matching values ​​are obtained. These values ​​reflect the similarity between the design object and each expected object in local appearance.

[0072] After obtaining all local matching values, the one with the highest value is selected, which represents the highest similarity between the design object and an expected object in local appearance. Based on this highest similarity, local similarity information is generated, and a local mark is added to the design object accordingly. This mark clearly indicates in which local features the design object is most similar to which expected object.

[0073] Through the method provided in this embodiment, personalized packaging recommendations for the design object are successfully implemented, making the appearance of the design object closer to the star image expected by merchants or fans.

[0074] In some embodiments, several appearance images of similar objects are obtained, and based on the several appearance images, salient local information of the similar objects is obtained, and based on the salient local information, a local acquisition instruction is generated; through the local acquisition instruction, corresponding local information of the design object is obtained, and the corresponding local information is compared with the salient local information to obtain a local similarity value, and the local similarity value is compared with a preset comparison value. If the local similarity value is less than the preset comparison value, avoidance information is generated and sent to a terminal device, and avoidance prompt information is generated through the terminal device to inform the merchant that the part of the design object does not look very similar to the corresponding star.

[0075] Specifically, the platform collects and organizes basic information of several desired objects (for example, well-known stars) to build a template database. This database includes not only the overall appearance information of each star, but also their respective prominent local information, such as a star's particularly slender legs or delicate facial contours.

[0076] Then, the platform obtains the personal information of the anchor who needs to be packaged, and collects the anchor's actual image through a camera or image library. These images will be used to determine the anchor's actual appearance information, including the overall image and various local features.

[0077] Next, the platform will use AI technology to match and analyze the host's actual appearance information with the celebrity appearance information in the template database. In this process, it not only considers the overall appearance match between the host and the celebrity, but also pays special attention to the similarity of local features.

[0078] For example, after analysis, the platform finds that a host and a star have a high degree of overall appearance match, but there are big differences in leg features. At this time, the platform will generate a local acquisition instruction based on the prominent local information of the star's legs to obtain local information of the corresponding part of the host. By comparing the leg features of the host and the star, the platform calculates the local similarity value and finds that the value is lower than the preset comparison value.

[0079] Therefore, the platform will generate an avoidance message based on the comparison result and send it to the terminal device. On the terminal device, this avoidance message will be converted into an intuitive avoidance prompt message, such as "Please try to avoid showing the host's leg features during promotion to avoid causing a sense of disobedience."

[0080] Ultimately, after receiving the platform’s recommendation information and avoidance tips, merchants can formulate packaging and promotion strategies more specifically.

[0081] Through the method provided in this embodiment, packaging and promotion strategies can be formulated in a more targeted manner, which can fully utilize the similarities between the anchor and the star to attract the audience and effectively avoid discordant factors in the promotion.

[0082] In some embodiments, the actual information of the merchant is obtained, and a merchant keyword data set is determined based on the actual information of the merchant, and a plurality of expected behavior information is analyzed respectively, and expected keyword data is obtained according to the characteristics of each expected object; the merchant keyword data set is used to analyze the expected keyword data respectively to obtain a keyword matching value, and the keyword matching value is compared with a preset matching threshold value, and if the keyword matching value is less than the preset matching threshold value, a first rejection information is generated, and the first rejection information is sent to a terminal device; if the keyword matching value is greater than the preset matching threshold value, a retention information is generated, and the retention information is sent to a terminal device, and a corresponding expected object is determined according to the retention information, and a corresponding appearance mark is retrieved according to the corresponding expected object, and a corresponding design object is determined according to the appearance mark.

[0083] Specifically, we obtain the actual information of the merchants, which may include but are not limited to the business scope, brand concept, target consumer groups, and promotional focus of the merchants. Based on this information, we use natural language processing technology to extract the keyword data set most relevant to the merchants, such as "fashion", "youth", "vitality", etc.

[0084] Next, we analyze the expected behavior information of the expected objects (i.e. existing stars). This information may come from the stars’ public activities, social media content, interview records, etc., and can reflect the stars’ personal style, life attitude, and values. By deeply mining this information, we extract the expected keyword data that matches their characteristics for each star. For example, the expected keyword data of a star may include “fashionable”, “love of life”, “positive energy”, etc.

[0085] Then, the merchant keyword dataset is compared and analyzed with the expected keyword data of each star, and the compatibility between the merchant and the star is measured by calculating the keyword matching value. This step may involve complex algorithms and models to ensure the accuracy and reliability of the matching results.

[0086] After obtaining the keyword matching value, it is compared with the preset matching threshold. If the keyword matching value of a star is lower than the threshold, it means that the star is not compatible with the merchant. Therefore, the first rejection information is generated and sent to the terminal device. In this way, in the subsequent recommendation process, the star can be avoided from being recommended to the merchant as a potential design object.

[0087] On the contrary, if the keyword matching value of a star is higher than the threshold, it means that the star has a high degree of compatibility with the merchant, so retention information is generated and sent to the terminal device. Based on the retention information, the corresponding expected object can be determined, and further design objects with similar appearance can be retrieved. This step may involve comparative analysis of the actual appearance information of the design object and the expected appearance information of the star to determine the design object that best meets the merchant's needs.

[0088] Through the method provided in this embodiment, accurate matching of the desired object (star) and the design object (anchor) is achieved. This method not only improves the efficiency and accuracy of matching, but also helps merchants find the design object that best meets their needs, thereby achieving better publicity effects and commercial benefits.

[0089] In some embodiments, when the keyword matching value is greater than a preset matching degree threshold, a pre-selected tag is generated, and the pre-selected tag is attached to the corresponding expected keyword data to obtain the pre-selected keyword data. According to the pre-selected keyword data, an in-depth analysis instruction is generated. According to the in-depth analysis instruction, the pre-selected keyword data is analyzed through the merchant keyword data set. If the pre-selected keyword data contains at least one risk word, a second elimination information is generated, and the corresponding expected object is eliminated through the second elimination information.

[0090] Specifically, the basic information of the desired object (i.e., the star) is collected, and a template database is established based on this information. Then, the platform obtains the personal information of the design object (i.e., the potential co-host) and extracts the actual image to determine its actual appearance information.

[0091] Then, the platform uses the information analysis module to compare the actual appearance information of the design object with the expected appearance information in the template database, determines the overall matching value, and selects several stars that match the platform's promotion needs. At this time, the platform generates pre-selected tags and attaches these tags to the corresponding expected keyword data to form pre-selected keyword data.

[0092] Next, the platform generated in-depth analysis instructions based on these pre-selected keyword data, and conducted in-depth analysis of the pre-selected keyword data through the merchant keyword data set (that is, the data set of keywords such as outdoor sports and healthy lifestyle collected by the platform). During the analysis process, the platform found that the pre-selected keyword data of one of the stars contained risky words such as "negative news" and "controversial events". After further verification, the platform found that the star had participated in a controversial event in the past that was inconsistent with his healthy image.

[0093] Therefore, the second elimination information was generated, and the star was eliminated from potential cooperation partners. In the end, the platform selected several other stars who were more in line with the brand image and had no risky words as cooperation partners, and recommended designers who were more compatible with these stars in appearance and personal style for cooperation.

[0094] The method provided in this embodiment not only ensures that the image and style of the cooperation partner are consistent with the brand image, but also avoids potential risks to the brand due to personal problems of celebrities. At the same time, this also reflects the advantages of this method in merchant information matching and risk control.

[0095] In some embodiments, when the design object is evaluated again later, some weight influence information is obtained according to some overall matching values, and evaluation correction information is obtained according to some weight influence information. Before the actual appearance information is compared with some expected appearance information again, the evaluation correction information is added to the template database to obtain an improved template database. The actual appearance information is evaluated by the improved template database.

[0096] Specifically, the platform first compares the actual appearance information of the anchor with the expected appearance information in the template database, and obtains a number of overall matching values, which reflect the similarity between the anchor and each star.

[0097] After the initial evaluation, the platform records these overall matching values ​​and generates corresponding weighted impact information. The weighted impact information is determined based on the relative importance of the matching degree between each anchor and the star. For example, if the matching value between an anchor and a star is particularly high, then the weight of this matching value in subsequent evaluations may increase accordingly.

[0098] Next, when it is necessary to conduct a second evaluation of the anchor, the platform will use the historical data of the previous evaluation to generate evaluation correction information. The evaluation correction information is adjusted based on the weight influence information of the initial evaluation, which reflects the influence of the previous evaluation on the current evaluation.

[0099] Before comparing the actual appearance information with the expected appearance information again, the platform will add the evaluation correction information to the template database to obtain an improved template database. This improved template database not only contains the original expected appearance information, but also incorporates the weight influence information and evaluation correction information of the previous evaluation.

[0100] Finally, the platform uses the improved template database to conduct a secondary evaluation of the host’s actual appearance information. Since the historical data of the previous evaluation is added, the result of this evaluation will be more accurate and reliable. The platform can choose the host who best matches the product image to cooperate with based on the results of the secondary evaluation, thereby achieving better promotion effects.

[0101] Through the method provided in this embodiment, the accuracy and reliability of the evaluation can be significantly improved by using the historical data of the previous evaluation in the secondary evaluation. This method is not only suitable for the matching evaluation of anchors and celebrities, but can also be widely used in other scenarios that require multiple evaluations.

[0102] The device of this embodiment can be used to execute the method of any of the above embodiments. Its implementation principle and technical effects are similar and will not be described in detail here.

Claims

1. An information matching method for visual portrait data based on AI technology, characterized in that: The information matching method is applied to an interactive platform, the interactive platform includes a terminal device and a server, and the method includes: Obtain basic information of several desired objects, and establish a template database according to each of the basic information. Acquire personal information of the design object, obtain a plurality of actual images, and determine actual appearance information of the design object based on the plurality of actual images; Analyze the actual appearance information with the template databases to obtain segmented personal information; The merchant demand information is obtained, the merchant demand information is analyzed, and according to the analysis result, the matching segmented personal information is determined, and according to the segmented personal information, the matching design object is determined, and recommendation information is generated, and according to the recommendation information, the design object that is closest to the merchant demand is recommended to the merchant.

2. The method according to claim 1, characterized in that The method further comprises: Wherein, the template database includes: expected appearance information and expected behavior information; Analyze the actual appearance information with the plurality of expected appearance information respectively, determine the overall matching value between the actual appearance information and each of the expected appearance information, and obtain the plurality of overall matching values, Obtaining the overall matching value with the highest value, generating similar appearance information according to the overall matching value with the highest value, generating an appearance tag according to the similar appearance information, and attaching the appearance tag to the personal information to obtain the segmented personal information; and determining the desired object matching the design object according to the similar appearance information, and then determining the similar object; The expected behavior data of the look-alike object is obtained, behavior guidance information is generated, and the daily behavior of the designed object is guided by the behavior guidance information, so that the designed object is more consistent with the corresponding look-alike object in terms of personal style.

3. The method according to claim 2, characterized in that The method further comprises: Obtaining local appearance information through the actual appearance information, By using the desired appearance information, obtaining some local desired information, The local appearance information is respectively compared with several local expected information to obtain several local matching values, and the local matching value with the highest value is obtained to obtain local similarity information. According to the local similarity information, a local tag is generated, and the local tag is attached to the basic information of the design object to obtain local similarity information. The local similarity information is used to inform the staff of which expected object the local appearance of the design object is more similar to, and then special design is performed on the part of the design object.

4. The method according to claim 3, characterized in that The method further comprises: Acquire a plurality of appearance images of the similar object, acquire salient local information of the similar object according to the plurality of appearance images, and generate a local acquisition instruction according to the salient local information; The local acquisition instruction is used to obtain the corresponding local information of the design object, and the corresponding local information is compared with the prominent local information to obtain a local similarity value. The local similarity value is compared with a preset comparison value. If the local similarity value is less than the preset comparison value, avoidance information is generated, and the avoidance information is sent to the terminal device. Avoidance prompt information is generated by the terminal device to inform the merchant that the part of the design object does not look very similar to the corresponding star.

5. The method according to claim 2, characterized in that: The method further comprises: Obtaining actual merchant information, and determining a merchant keyword data set based on the merchant actual information, Analyze the plurality of expected behavior information respectively, and obtain expected keyword data according to each expected object feature; Analyze the expected keyword data with the merchant keyword data set to obtain keyword matching values. The keyword matching value is compared with a preset matching threshold, and if the keyword matching value is less than the preset matching threshold, a first elimination information is generated, and the first elimination information is sent to the terminal device. If the keyword matching value is greater than the preset matching threshold, a reservation message is generated and sent to the terminal device. The corresponding expected object is determined according to the reserved information, the corresponding appearance mark is retrieved according to the corresponding expected object, and the corresponding design object is determined according to the appearance mark.

6. The method according to claim 5, characterized in that The method further comprises: When the keyword matching value is greater than the preset matching degree threshold, a pre-selected tag is generated, and the pre-selected tag is attached to the corresponding expected keyword data to obtain the pre-selected keyword data. Based on the pre-selected keyword data, an in-depth analysis instruction is generated. According to the in-depth analysis instruction, the pre-selected keyword data is analyzed through the merchant keyword data set. If the pre-selected keyword data contains at least one risk word, second elimination information is generated, and the corresponding expected object is eliminated through the second elimination information.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: Subsequently, when the design object is evaluated again, a plurality of weight influence information is obtained according to the plurality of overall matching values, and evaluation correction information is obtained according to the plurality of weight influence information. Before the actual appearance information is compared with the plurality of expected appearance information again, the evaluation correction information is added to the template database to obtain an improved template database, and the actual appearance information is evaluated through the improved template database.

Images