Intelligent interaction-based wearing and building recommendation method and system, and storage medium

Abstract

The invention discloses a wearing and building recommendation method and system based on intelligent interaction and a storage medium, and relates to the field of artificial intelligence. Generating a questionnaire in combination with the time sequence-weather condition, the psychological state of the user at the current geographic position and the dressing preference; the current user fills in a questionnaire on the basis of a dressing requirement, the user preference test model captures key information of a questionnaire result and completes sentiment analysis and psychological feature portraits, and body health element information and dressing preference of the current user are obtained; face features and stature features of the current user are obtained, an appearance feature portrait of the current user is completed, and a plurality of wearing and fitting results are generated in combination with body health element information, dressing preferences and weather conditions of the current user; and matching the wearing and matching result with the appearance feature portrait of the current user to generate a virtual wearing and matching image, and selecting the most favorite wearing and matching result by the current user according to the virtual wearing and matching image. According to the method, the artificial intelligence wearing and building interactivity is improved, and the wearing and building result is accurately recommended.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and more specifically to an outfit recommendation method, system, and storage medium based on intelligent interaction. Background Art

[0002] As people's living standards improve, their clothing needs are no longer limited to warmth and coverage, but are increasingly focused on personalization, fashion, and adaptability to different scenarios. In an era of information explosion, consumers are faced with a vast array of clothing styles and matching options. Finding a quick and accurate outfit plan that meets their needs has become a major challenge.

[0003] Traditional outfit recommendations rely primarily on manual experience or simple rule-based algorithms. Manual matching requires specialized fashion knowledge and significant time and effort, making it incapable of meeting the real-time needs of large-scale users. Rule-based algorithms, on the other hand, are often too rigid and fail to account for complex and ever-changing factors such as user preferences, body types, and occasional needs. For example, recommendations based solely on season and clothing type fail to provide precise recommendations tailored to a user's unique temperament and specific daytime activities.

[0004] Meanwhile, while the rapid development of e-commerce platforms has greatly enriched the clothing supply, it has also exacerbated the difficulty users face in making decisions about clothing. Browsing through a vast array of clothing products, it's difficult for users to intuitively imagine how different items will look when combined. Furthermore, existing outfit recommendation systems lack effective user interaction and are unable to adjust recommendation strategies based on real-time user feedback.

[0005] With the rapid development of artificial intelligence (AI), intelligent interaction technology has been widely applied in various fields. Introducing intelligent interaction technology into the field of outfit recommendations can achieve more natural and flexible communication between users and recommendation systems, enabling the system to deeply understand user needs and provide more personalized and accurate outfit recommendations. This technology has broad development prospects and application value. Summary of the Invention

[0006] In view of this, the present invention provides a dressing recommendation method, system and storage medium based on intelligent interaction, which improves the interactivity of artificial intelligence dressing and accurately recommends dressing results to solve the problems existing in the background technology.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] An outfit recommendation method based on intelligent interaction includes the following steps:

[0009] Step 1: Locate the current geographic location, obtain the time series and weather conditions of the current geographic location, and generate a questionnaire based on the time series and weather conditions, the psychological state of the user at the current geographic location, and clothing preferences;

[0010] Step 2: The current user fills out a questionnaire based on their clothing needs. The user preference testing model captures key information from the questionnaire results, performs sentiment analysis and psychological profiles, and derives the current user's physical health information and clothing preferences.

[0011] Step 3: Obtain the current user's facial features and body features to complete the current user's appearance profile, and generate multiple outfit results based on the current user's health information, clothing preferences, and weather conditions;

[0012] Step 4: Match the outfit results with the current user's appearance feature portrait to generate a virtual outfit image. The current user selects the outfit result that he or she likes most based on the virtual outfit image.

[0013] Optionally, in step 2, a large language model and a knowledge graph structure are used to capture key information from the questionnaire structure, perform sentiment analysis and psychological profile analysis, and obtain the current user's physical health information and clothing preferences. The specific steps include:

[0014] Collect and integrate historical questionnaire results, obtain multi-dimensional survey data based on historical user physical conditions, and structure the multi-dimensional survey data;

[0015] Using deep learning algorithms, we extract keywords and corresponding relationship attributes from the structured survey data. We then use keywords and corresponding attribute relationships to construct a triple data chain and generate a knowledge graph of emotional and psychological characteristics. We then update the triple data chain based on the subsequent user-selected outfits.

[0016] The questionnaire results of the current user are input into the updated emotional and psychological characteristics knowledge graph to complete the emotional analysis and psychological portrait, and the clothing preference output is completed in combination with the current user's physical health factor information.

[0017] Optionally, in step 4, generating a virtual outfit image specifically includes the following steps:

[0018] A high-resolution camera is used to obtain the current user's planar features, and the user's body shape and facial features are modeled based on the current user's basic body parameters to obtain an initial virtual portrait model;

[0019] Use structured light to capture real human body surface data, construct a low-dimensional shape space based on the real human body surface data, and use principal component analysis to characterize body shape differences;

[0020] The low-dimensional shape space and body shape differences are used to perform precision correction on the initial virtual portrait model to obtain the final virtual portrait model.

[0021] Optionally, in step three, facial features and body features of the current user are obtained to complete the appearance feature portrait of the current user, which specifically includes the following steps:

[0022] Collect the facial features of the current user, use the key point detector to locate the key feature points of the face, and obtain the geometric structure features of the face;

[0023] Use deep learning models to detect the human body in the image and segment the human body parts. Based on the depth image or the 3D model reconstructed from multiple perspectives, calculate the key dimensions of the human body, such as height, shoulder width, waist circumference, and hip circumference.

[0024] Optionally, in step 1, the current geographic location is located using latitude and longitude information or the current geographic location is directly obtained from the Internet.

[0025] An outfit recommendation system based on intelligent interaction, including:

[0026] Basic information acquisition module: used to locate the current geographic location, obtain the time series and weather conditions of the current geographic location, and generate a questionnaire based on the time series and weather conditions, the psychological state of the user at the current geographic location, and clothing preferences;

[0027] User preference capture module: This module is used for users to fill out questionnaires based on their clothing needs. The user preference testing model captures key information from the questionnaire results, performs sentiment analysis and psychological profiles, and obtains the user's physical health information and clothing preferences.

[0028] Outfit recommendation module: used to obtain the current user's facial features and body features, complete the current user's appearance profile, and generate multiple outfit results based on the current user's physical health information, clothing preferences, and weather conditions;

[0029] Virtual generation module for outfit results: used to match the outfit results with the current user's appearance feature portrait to generate a virtual outfit image. The current user selects the outfit result that he or she likes most based on the virtual outfit image.

[0030] A computer storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any one of the methods for recommending outfits based on intelligent interaction.

[0031] It can be seen from the above technical solutions that, compared with the prior art, the present invention provides an outfit recommendation method, system, and storage medium based on intelligent interaction, which have the following beneficial effects:

[0032] 1. Utilizing a user preference testing model, we extract key information from the questionnaire results, analyze sentiment, and create a psychological profile to derive information about the user's health and clothing preferences. This model not only considers the user's external preferences but also internal factors such as their physical condition, further enhancing the accuracy and personalization of recommendations. For example, if we know a user is susceptible to cold, we can recommend clothing with warmer materials and styles.

[0033] 2. The user's facial and body features are captured to create a profile of their appearance, which is then matched with the outfit recommendations to generate a virtual avatar. This ensures that the recommended outfits not only meet the user's preferences and physical needs, but also complement their appearance, making the outfit more in line with their image and temperament. For example, for tall users, long garments that highlight their figure may be recommended, while for petite users, more delicate and petite styles may be recommended.

[0034] 3. Allowing users to make their own choices based on their virtual outfits, giving them full decision-making power, enhances their sense of participation and experience in the outfit recommendation process. Users can choose the outfit that best suits them from multiple recommendations based on their aesthetic preferences and actual needs, which strengthens their trust and favorability in the recommendation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0036] Figure 1 Schematic diagram of the method flow of the present invention;

[0037] Figure 2 Schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] The embodiment of the present invention discloses a method for recommending outfits based on intelligent interaction. Figure 1 As shown, the following steps are included:

[0040] Step 1: Locate the current geographic location, obtain the time series and weather conditions of the current geographic location, and generate a questionnaire based on the time series and weather conditions, the psychological state of the user at the current geographic location, and clothing preferences;

[0041] Step 2: The current user fills out a questionnaire based on their clothing needs. The user preference testing model captures key information from the questionnaire results, performs sentiment analysis and psychological profiles, and derives the current user's physical health information and clothing preferences.

[0042] Step 3: Obtain the current user's facial features and body features to complete the current user's appearance profile, and generate multiple outfit results based on the current user's health information, clothing preferences, and weather conditions;

[0043] Step 4: Match the outfit results with the current user's appearance feature portrait to generate a virtual outfit image. The current user selects the outfit result that he or she likes most based on the virtual outfit image.

[0044] Furthermore, in step 2, the large language model and knowledge graph structure are used to capture key information from the questionnaire structure, and sentiment analysis and psychological characteristic profiling are completed to obtain the current user's physical health factor information and clothing preferences. The specific steps include:

[0045] Collect and integrate historical questionnaire results, obtain multi-dimensional survey data based on historical user physical conditions, and structure the multi-dimensional survey data;

[0046] Using deep learning algorithms, we extract keywords and corresponding relationship attributes from the structured survey data. We then use keywords and corresponding attribute relationships to construct a triple data chain and generate a knowledge graph of emotional and psychological characteristics. We then update the triple data chain based on the subsequent user-selected outfits.

[0047] The questionnaire results of the current user are input into the updated emotional and psychological characteristics knowledge graph to complete the emotional analysis and psychological portrait, and the clothing preference output is completed in combination with the current user's physical health factor information.

[0048] Furthermore, in step one, the current geographic location is located using latitude and longitude information to complete the positioning or directly obtain the current geographic location from the Internet. By obtaining the time series of the current geographic location - weather conditions, user psychological state and clothing preferences to generate a questionnaire, the personalized needs of different users in different scenarios are comprehensively considered. For example, users in different regions have different clothing needs in the same season due to climate differences and personal preferences. For example, in the winter in the southern coastal areas, the temperature may be relatively high, and users may prefer light, warm and fashionable clothing; while in the northern regions, the winter is cold, and users pay more attention to thick and warm clothing. This method of generating questionnaires based on multiple factors can accurately capture user needs and lay the foundation for providing highly personalized clothing recommendations in the future.

[0049] Furthermore, in step three, the facial features and body features of the current user are obtained to complete the appearance feature portrait of the current user, which specifically includes the following steps:

[0050] Collect the facial features of the current user, use the key point detector to locate the key feature points of the face, and obtain the geometric structure features of the face;

[0051] A deep learning model is used to detect and segment human bodies in images. Key dimensions such as height, shoulder width, waist, and hip circumference are calculated based on depth images or multi-view reconstructed 3D models. Height is calculated by the distance between the highest and lowest points of a person in the depth image; shoulder width is calculated by the distance between two shoulder points in the 3D model.

[0052] Furthermore, in step 4, generating a virtual outfit image specifically includes the following steps:

[0053] A high-resolution camera is used to obtain the current user's planar features, and the user's body shape and facial features are modeled based on the current user's basic body parameters to obtain an initial virtual portrait model;

[0054] Use structured light to capture real human body surface data, construct a low-dimensional shape space based on the real human body surface data, and use principal component analysis to characterize body shape differences;

[0055] The low-dimensional shape space and body shape differences are used to perform precision correction on the initial virtual portrait model to obtain the final virtual portrait model.

[0056] Furthermore, in this embodiment, the vertex coordinates of the initial model are converted into a low-dimensional shape space, and then the shape of the initial model is adjusted based on the user's body shape difference parameters in the low-dimensional shape space. For example, if a user's parameter value for a principal component in the low-dimensional shape space is large, indicating that the user's corresponding body shape characteristics are significantly different from the average body shape, then the corresponding parts of the initial model are adjusted based on the body shape change pattern corresponding to this principal component. For example, if the principal component reflects differences in weight and the user's parameter value is large, the body width of the model is appropriately increased.

[0057] Regarding facial features, different body shapes may correspond to different facial fat distribution and muscle tone, thus affecting facial shape and contour. By analyzing the relationship between body shape parameters and facial features, the initial facial model is fine-tuned to better match the user's actual appearance. For other body parts, such as arms and legs, detailed optimizations are also made based on body shape differences, such as adjusting muscle ridges and joint thickness.

[0058] This final virtual portrait model not only incorporates the basic framework constructed from planar features and basic body parameters acquired from high-resolution cameras, but also incorporates body shape differences and detailed information captured from real human surface data captured using structured light, achieving high-precision virtual portrait modeling. It can be widely used in a variety of fields, including virtual reality, augmented reality, film and television production, game development, and fashion design, providing a foundation for realistic virtual character images.

[0059] and Figure 1 Corresponding to the method shown, the present invention also discloses a wear recommendation system based on intelligent interaction for Figure 1 The implementation of the method, the specific structure is as follows Figure 2 Shown, including:

[0060] Basic information acquisition module: used to locate the current geographic location, obtain the time series and weather conditions of the current geographic location, and generate a questionnaire based on the time series and weather conditions, the psychological state of the user at the current geographic location, and clothing preferences;

[0061] User preference capture module: This module is used for users to fill out questionnaires based on their clothing needs. The user preference testing model captures key information from the questionnaire results, performs sentiment analysis and psychological profiles, and obtains the user's physical health information and clothing preferences.

[0062] Outfit recommendation module: used to obtain the current user's facial features and body features, complete the current user's appearance profile, and generate multiple outfit results based on the current user's physical health information, clothing preferences, and weather conditions;

[0063] Virtual generation module for outfit results: used to match the outfit results with the current user's appearance feature portrait to generate a virtual outfit image. The current user selects the outfit result that he or she likes most based on the virtual outfit image.

[0064] This embodiment also discloses a computer storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the methods for recommending outfits based on intelligent interaction are implemented.

[0065] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0066] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for recommending outfits based on intelligent interaction, characterized in that: The following steps are involved: Step 1: Locate the current geographic location, obtain the time series and weather conditions of the current geographic location, and generate a questionnaire based on the time series and weather conditions, the psychological state of the user at the current geographic location, and clothing preferences; Step 2: The current user fills out a questionnaire based on their clothing needs. The user preference testing model captures key information from the questionnaire results, performs sentiment analysis and psychological profiles, and derives the current user's physical health information and clothing preferences. Step 3: Obtain the current user's facial features and body features to complete the current user's appearance profile, and generate multiple outfit results based on the current user's health information, clothing preferences, and weather conditions; Step 4: Match the outfit results with the current user's appearance feature portrait to generate a virtual outfit image. The current user selects the outfit result that he or she likes most based on the virtual outfit image.

2. The method for recommending outfits based on intelligent interaction according to claim 1, characterized in that: In step 2, we use a large language model and knowledge graph structure to capture key information from the questionnaire structure, perform sentiment analysis and psychological profile analysis, and derive the current user's health information and clothing preferences. This includes the following steps: Collect and integrate historical questionnaire results, obtain multi-dimensional survey data based on historical user physical conditions, and structure the multi-dimensional survey data; Using deep learning algorithms, we extract keywords and corresponding relationship attributes from the structured survey data. We then use keywords and corresponding attribute relationships to construct a triple data chain and generate a knowledge graph of emotional and psychological characteristics. We then update the triple data chain based on the subsequent user-selected outfits. The questionnaire results of the current user are input into the updated emotional and psychological characteristics knowledge graph to complete the emotional analysis and psychological portrait, and the clothing preference output is completed in combination with the current user's physical health factor information.

3. The method for recommending outfits based on intelligent interaction according to claim 1, characterized in that: In step 4, generating a virtual outfit image specifically includes the following steps: A high-resolution camera is used to obtain the current user's planar features, and the user's body shape and facial features are modeled based on the current user's basic body parameters to obtain an initial virtual portrait model; Use structured light to capture real human body surface data, construct a low-dimensional shape space based on the real human body surface data, and use principal component analysis to characterize body shape differences; The low-dimensional shape space and body shape differences are used to perform precision correction on the initial virtual portrait model to obtain the final virtual portrait model.

4. The method for recommending outfits based on intelligent interaction according to claim 1, characterized in that: In step three, the facial features and body features of the current user are obtained to complete the appearance feature portrait of the current user, which specifically includes the following steps: Collect the facial features of the current user, use the key point detector to locate the key feature points of the face, and obtain the geometric structure features of the face; Use deep learning models to detect the human body in the image and segment the human body parts. Based on the depth image or the 3D model reconstructed from multiple perspectives, calculate the key dimensions of the human body, such as height, shoulder width, waist circumference, and hip circumference.

5. The method for recommending outfits based on intelligent interaction according to claim 1, characterized in that: In step 1, the current geographic location is located using latitude and longitude information or directly obtained from the Internet.

6. An outfit recommendation system based on intelligent interaction, characterized in that: include: Basic information acquisition module: used to locate the current geographic location, obtain the time series and weather conditions of the current geographic location, and generate a questionnaire based on the time series and weather conditions, the psychological state of the user at the current geographic location, and clothing preferences; User preference capture module: This module is used for users to fill out questionnaires based on their clothing needs. The user preference testing model captures key information from the questionnaire results, performs sentiment analysis and psychological profiles, and obtains the user's physical health information and clothing preferences. Outfit recommendation module: used to obtain the current user's facial features and body features, complete the current user's appearance profile, and generate multiple outfit results based on the current user's physical health information, clothing preferences, and weather conditions; Virtual generation module for outfit results: used to match the outfit results with the current user's appearance feature portrait to generate a virtual outfit image. The current user selects the outfit result that he or she likes most based on the virtual outfit image.

7. A computer storage medium, characterized in that The computer storage medium stores a computer program, which, when executed by a processor, implements the steps of a clothing recommendation method based on intelligent interaction as described in any one of claims 1 to 5.

Images