Multi-round garment design optimization method, system, device, medium

Abstract

This invention relates to a method, system, device, and medium for multi-round garment design optimization. The method includes: when the current optimization design information is explicit, obtaining a current optimization semantic vector based on the explicit optimization design information, and optimizing the current garment design drawing to be optimized based on the current optimization semantic vector and an AIGC model to obtain the current optimized garment design drawing; when the current optimization design information is indefinite, obtaining a current optimization semantic vector based on the current optimization design information, multiple target design information determined from all remaining historical design information after the latest filtering operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information, and optimizing the current garment design drawing to be optimized based on the current optimization semantic vector and an AIGC model to obtain the current optimized garment design drawing.

Description

Technical Field

[0001] Several embodiments of this specification relate to the field of clothing design, specifically to methods, systems, equipment, and media for multi-round clothing design optimization. Background Technology

[0002] With the continuous development of computer-aided design technology, Artificial Intelligence Generated Content (AIGC) technology is increasingly widely used in the field of apparel design, providing designers with efficient tools for realizing and modifying creative ideas. In the actual production process of apparel design, design intentions are often not achieved overnight, but require multiple rounds of iteration and optimization before finalization. Therefore, how to utilize AIGC models to efficiently and accurately respond to design needs has become a current research hotspot.

[0003] However, existing apparel design optimization technologies often have significant limitations. In current design processes, when optimizing apparel in each round, the system often considers only the optimization design requirements input in the latest round in isolation, ignoring the accumulated optimization requirements from previous rounds. This fragmented approach prevents the latest round of apparel optimization from comprehensively referencing and inheriting historical optimization information, easily leading to a lack of consistency in optimization direction or a break in design logic. Ultimately, this neglect of historical information makes it difficult for the apparel optimization results to fully and accurately match the user's true design intentions, affecting design efficiency and finished product quality.

[0004] In summary, existing technologies lack an effective and reasonable solution for optimizing clothing designs by incorporating historical optimization experiences. There is an urgent need to propose a new technical approach to address these issues. Summary of the Invention

[0005] This specification provides embodiments of a multi-round apparel design optimization method, system, equipment, and medium, offering an effective and reasonable multi-round apparel design optimization scheme that integrates historical optimization scenarios.

[0006] The technical solution is as follows:

[0007] Firstly, the embodiments of this specification provide a multi-round clothing design optimization method, including:

[0008] S1. Based on the AIGC model and the clothing design information input by the user, output multiple clothing design drawings;

[0009] S2. Responding to the user's selection information among multiple garment design drawings, determine the garment design drawing to be optimized;

[0010] S3. Obtain the current optimization design information input by the user in the current optimization round, and based on the design conflict degree between the current optimization design information and all the remaining historical design information, perform a filtering operation on all the remaining historical design information, and determine whether the current optimization design information is explicit optimization design information or implicit optimization design information.

[0011] S4. When the current optimization design information is explicit optimization design information, obtain the current optimization semantic vector based on the explicit optimization design information, and optimize the current garment design drawing to be optimized based on the current optimization semantic vector and AIGC model to obtain the current optimized garment design drawing. Then, return to step S3 after using the current optimized garment design drawing as the new garment design drawing to be optimized, until the optimization ends.

[0012] S5. When the current optimization design information is non-ambiguous optimization design information, based on the current optimization design information, multiple target design information determined from all remaining historical design information after the latest filtering operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information, obtain the current optimization semantic vector, and optimize the current garment design drawing to be optimized based on the current optimization semantic vector and the AIGC model to obtain the current optimized garment design drawing, and return to step S3 after using the current optimized garment design drawing as the new garment design drawing to be optimized, until the optimization ends.

[0013] As a preferred embodiment, step S5 includes:

[0014] When the current optimization design information is non-specific optimization design information, determine whether the current optimization design information is overall non-specific optimization design information or local design dimension non-specific optimization design information;

[0015] When the current optimization design information is an overall non-ambiguous optimization design information, the current optimization semantic vector is obtained based on the current optimization design information, multiple target design information determined from all historical design information remaining after the latest screening operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information.

[0016] When the current optimization design information is a local design dimension non-defined optimization design information, the current optimization semantic vector is obtained based on the current optimization design information, multiple target design information that are consistent with the design dimension involved in the current optimization design information or are related to the overall clothing design from all the remaining historical design information after the latest filtering operation, the optimization rounds corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information.

[0017] As a preferred approach, based on the current optimized design information, multiple target design information, the optimization rounds corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimized design information, the current optimization semantic vector is obtained, including:

[0018] Based on the optimization rounds corresponding to each of the multiple target design information and the degree of design conflict between each of the multiple target design information and the current optimization design information, obtain the fusion weights corresponding to each of the multiple target design information.

[0019] Based on the current optimization design information, multiple target design information, the preset fusion weights corresponding to the current optimization design information, and the fusion weights corresponding to each of the multiple target design information, the current optimization semantic vector is obtained.

[0020] As a preferred embodiment, the step of obtaining the fusion weights corresponding to each of the multiple target design information based on the optimization rounds corresponding to each of the multiple target design information and the design conflict degree between each of the multiple target design information and the current optimization design information includes:

[0021] Clustering operations are performed on multiple target design information to obtain multiple target design information clusters;

[0022] Based on the optimization rounds corresponding to each of the multiple target design information, the degree of design conflict between each of the multiple target design information and the current optimization design information, and the cluster size of the target design information cluster to which each of the multiple target design information belongs, the fusion weights corresponding to each of the multiple target design information are obtained.

[0023] As a preferred embodiment, the step of obtaining the fusion weights corresponding to each of the multiple target design information based on the optimization rounds corresponding to each of the multiple target design information, the design conflict degree between each of the multiple target design information and the current optimized design information, and the cluster size of the target design information cluster to which each of the multiple target design information belongs, includes:

[0024] When the design conflict between the target design information and the current optimized design information is greater than the preset conflict degree, the fusion weight of the target design information is obtained based on the optimization round of the target design information and the design conflict degree between the target design information and the current optimized design information.

[0025] When the degree of design conflict between the target design information and the current optimized design information is no greater than the preset degree of conflict, the fusion weight of the target design information is obtained based on the optimization round of the target design information, the degree of design conflict between the target design information and the current optimized design information, and the cluster size of the target design information cluster to which the target design information belongs.

[0026] As a preferred embodiment, the fusion weight of the target design information is obtained based on the optimization round of the target design information, the design conflict degree between the target design information and the current optimized design information, and the cluster size of the target design information cluster to which the target design information belongs, including:

[0027] The fusion weight of the target design information is obtained based on the optimization round of the target design information, the degree of design conflict between the target design information and the current optimization design information, the cluster size of the target design information cluster to which the target design information belongs, and the optimization round of each target design information in the target design information cluster to which the target design information belongs.

[0028] As a preferred approach, the degree of design conflict between design information includes:

[0029] Based on the semantic cosine similarity detection algorithm, the semantic cosine similarity between design information is obtained;

[0030] Based on the knowledge graph of the apparel industry, obtain the degree of graph conflict between design information;

[0031] The design conflict degree between design information is obtained based on the semantic cosine similarity and graph conflict degree between design information.

[0032] Secondly, embodiments of this specification provide a multi-round clothing design optimization system for implementing the multi-round clothing design optimization method described in the first aspect above, including:

[0033] The first processing module outputs multiple clothing design drawings based on the AIGC model and the clothing design information input by the user.

[0034] The response module responds to the user's selection of information from multiple garment design drawings to determine the garment design drawing to be optimized.

[0035] The judgment module obtains the current optimization design information input by the user in the current optimization round, and performs a filtering operation on all the remaining historical design information based on the design conflict degree between the current optimization design information and all the remaining historical design information, and determines whether the current optimization design information is explicit optimization design information or implicit optimization design information.

[0036] The optimization module, when the current optimization design information is explicit optimization design information, obtains the current optimization semantic vector based on the explicit optimization design information, and optimizes the current garment design drawing to be optimized based on the current optimization semantic vector and the AIGC model to obtain the current optimized garment design drawing;

[0037] The optimization module, when the current optimization design information is non-ambiguous, obtains the current optimization semantic vector based on the current optimization design information, multiple target design information determined from all remaining historical design information after the latest filtering operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information. Based on the current optimization semantic vector and the AIGC model, the current garment design drawing to be optimized is optimized to obtain the current optimized garment design drawing.

[0038] Thirdly, embodiments of this specification provide an electronic device, including a processor and a memory; the processor is connected to the memory; the memory is used to store executable program code; the processor reads the executable program code stored in the memory to run a program corresponding to the executable program code, so as to perform the steps described in the first aspect of the above embodiments.

[0039] Fourthly, embodiments of this specification provide a computer storage medium storing a plurality of instructions adapted for loading by a processor and executing the steps described in the first aspect of the above embodiments.

[0040] The beneficial effects of the technical solutions provided in some embodiments of this specification include at least the following:

[0041] By comprehensively considering historical optimization information to improve design fit, this invention does not process the current input in isolation during each optimization round. Instead, it innovatively introduces a management and filtering mechanism for historical design information. By calculating the design conflict degree and filtering out historical design information, and combining this with multiple target design information identified from all remaining historical design information after the latest filtering operation, the optimization rounds corresponding to each target design information, and the design conflict degree, a current optimization semantic vector is generated. This approach allows each optimization decision to "remember" and reasonably reference historical optimization trajectories, avoiding the design logic gaps caused by focusing only on the design optimization needs of a single round in existing technologies. This ensures that the optimized clothing design drawings more comprehensively and accurately match the user's true design intentions.

[0042] This solution addresses design optimization needs with a differentiated approach, balancing accuracy and timeliness. It creatively categorizes optimization design information into "explicit" and "implicit" types, employing different processing strategies for each. For explicit optimization design information, targeted optimization is performed directly, ensuring both efficiency and accuracy in design modifications. For implicit optimization design information, multi-dimensional data such as optimization rounds, target design information, and conflict levels are fully utilized to guide the construction of an optimization semantic vector. This differentiated processing mechanism satisfies users' decisive and clear modification commands while intelligently integrating historical optimization trajectories when user design intentions are ambiguous, effectively solving optimization challenges in complex design scenarios. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A flowchart illustrating a multi-round clothing design optimization method according to some embodiments of this disclosure is shown.

[0045] Figure 2 A schematic diagram of the structure of a multi-round clothing design optimization system according to some embodiments of the present disclosure is shown.

[0046] Figure 3 A schematic block diagram of an electronic device according to some embodiments of the present disclosure is shown. Detailed Implementation

[0047] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings.

[0048] The terms "first," "second," "third," etc., in the description, claims, and accompanying drawings are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such processes, methods, products, or apparatus.

[0049] The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made to the function and arrangement of the described elements without departing from the scope of this specification. Various processes or components may be appropriately omitted, substituted, or added to the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.

[0050] Figure 1 A flowchart illustrating a multi-round clothing design optimization method according to some embodiments of this disclosure is shown. It should be understood that the numbers in the flowchart do not indicate the order in which these steps are executed; some or all of these steps can be executed in parallel, or their execution order can be interchanged, and this disclosure does not limit this. Furthermore, Figure 1 The methods described may also include additional steps not shown and / or the steps shown may be omitted, and the scope of this disclosure is not limited in this respect.

[0051] like Figure 1 As shown, multi-round clothing design optimization methods can include at least:

[0052] S1. Based on the AIGC model and the clothing design information input by the user, output multiple clothing design drawings;

[0053] S2. Responding to the user's selection information among multiple garment design drawings, determine the garment design drawing to be optimized;

[0054] S3. Obtain the current optimization design information input by the user in the current optimization round, and based on the design conflict degree between the current optimization design information and all the remaining historical design information, perform a filtering operation on all the remaining historical design information, and determine whether the current optimization design information is explicit optimization design information or implicit optimization design information.

[0055] S4. When the current optimization design information is explicit optimization design information, obtain the current optimization semantic vector based on the explicit optimization design information, and optimize the current garment design drawing to be optimized based on the current optimization semantic vector and AIGC model to obtain the current optimized garment design drawing. Then, return to step S3 after using the current optimized garment design drawing as the new garment design drawing to be optimized, until the optimization ends.

[0056] S5. When the current optimization design information is non-ambiguous optimization design information, based on the current optimization design information, multiple target design information determined from all remaining historical design information after the latest filtering operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information, obtain the current optimization semantic vector, and optimize the current garment design drawing to be optimized based on the current optimization semantic vector and the AIGC model to obtain the current optimized garment design drawing, and return to step S3 after using the current optimized garment design drawing as the new garment design drawing to be optimized, until the optimization ends.

[0057] This specification provides a multi-round clothing design optimization method through several embodiments:

[0058] By comprehensively considering historical optimization information to improve design fit, this invention does not process the current input in isolation during each optimization round. Instead, it innovatively introduces a management and filtering mechanism for historical design information. By calculating the design conflict degree and filtering out historical design information, and combining this with multiple target design information identified from all remaining historical design information after the latest filtering operation, the optimization rounds corresponding to each target design information, and the design conflict degree, a current optimization semantic vector is generated. This approach allows each optimization decision to "remember" and reasonably reference historical optimization trajectories, avoiding the design logic gaps caused by focusing only on the design optimization needs of a single round in existing technologies. This ensures that the optimized clothing design drawings more comprehensively and accurately match the user's true design intentions.

[0059] This solution addresses design optimization needs with a differentiated approach, balancing accuracy and timeliness. It creatively categorizes optimization design information into "explicit" and "implicit" types, employing different processing strategies for each. For explicit optimization design information, targeted optimization is performed directly, ensuring both efficiency and accuracy in design modifications. For implicit optimization design information, multi-dimensional data such as optimization rounds, target design information, and conflict levels are fully utilized to guide the construction of an optimization semantic vector. This differentiated processing mechanism satisfies users' decisive and clear modification commands while intelligently integrating historical optimization trajectories when user design intentions are ambiguous, effectively solving optimization challenges in complex design scenarios.

[0060] Here, the explicit optimization design information and the implicit optimization design information described in this invention are specifically defined as follows:

[0061] The explicit optimization design information refers specifically to user-issued instructions with a clear modification target and a specific execution path. This type of information typically includes explicit object attributes and modification status, such as "change the sleeves to puff sleeves" or "change the color to blue." Its characteristics are clear semantics and a single target, allowing the AIGC model to directly execute deterministic local modifications based on these instructions.

[0062] The aforementioned non-specific optimization design information refers to adjustment instructions issued by users that lean towards design style, subjective feelings, or abstract dimensions. This type of information only indicates the direction or trend of optimization, without specifying concrete modification methods, such as "adjust the overall design style to be more youthful" or "adjust the sleeve design style to be more business-like." Its characteristic is semantic ambiguity; the model cannot directly pinpoint a single modification solution based solely on the current instruction.

[0063] The classification of explicit and implicit optimization design information can be achieved through a pre-trained intelligent model. Specifically, a large number of historical design instruction samples with labeled types can be used to train the neural network model, enabling it to learn the semantic features and classification boundaries of different instructions. In practical applications, the current optimization design information input by the user is fed into the trained model, and the model can automatically output the corresponding classification result, thereby achieving rapid and accurate identification of optimization instruction types.

[0064] It should be noted that the core technical effect of the filtering operation in step S3 is that by dynamically monitoring and filtering out historical design information that conflicts too much with the current optimization intention, interference items and logical contradictions in the historical design information database are effectively eliminated. This filtering mechanism ensures the semantic compatibility of the subsequently generated optimization semantic vectors and avoids deviations in the optimization direction caused by forcibly merging conflicting semantics.

[0065] It should also be noted that step S3 filters out historical design information with completely opposite design intentions. The following example illustrates this:

[0066] Strong conflicts (e.g., long sleeves vs. short sleeves): These are logically mutually exclusive, with no intermediate state. If the weights are reduced during fusion, the model might generate "three-quarter sleeves," which violates the user's latest modification intent and must therefore be removed.

[0067] Weak conflict (e.g., business vs. casual): This is often reconcilable. For example, if a user's latest design intention is to "make the clothing more casual," completely deleting all historical design information related to "making the clothing more business-like" might cause the image to lose its original texture. However, reducing the weight of the blend might create a "business casual" style, which is usually an acceptable transition for users.

[0068] Therefore, it is understandable that the design information filtered out in step S3 based on the degree of design conflict is historical design information that is completely contrary to the design intent presented by the current optimized design information.

[0069] It is also understandable that in step S5, multiple target design information can be determined from all the remaining historical design information after the latest filtering operation, taking into account the optimization rounds of historical design information and the degree of design conflict between historical design information and current optimization design information. That is, historical design information with too early optimization rounds and too high design conflict with current optimization design information can be excluded, and the remaining historical optimization design information can be regarded as target design information.

[0070] The calculation of design conflict degree will be described in detail in the following sections.

[0071] In some embodiments of this specification, step S5 includes:

[0072] When the current optimization design information is non-specific optimization design information, determine whether the current optimization design information is overall non-specific optimization design information or local design dimension non-specific optimization design information;

[0073] When the current optimization design information is an overall non-ambiguous optimization design information, the current optimization semantic vector is obtained based on the current optimization design information, multiple target design information determined from all historical design information remaining after the latest screening operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information.

[0074] When the current optimization design information is a local design dimension non-defined optimization design information, the current optimization semantic vector is obtained based on the current optimization design information, multiple target design information that are consistent with the design dimension involved in the current optimization design information or are related to the overall clothing design from all the remaining historical design information after the latest filtering operation, the optimization rounds corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information.

[0075] First, we will explain the overall non-specific optimization design information and the non-specific optimization design information at the local design dimension:

[0076] Overall non-specific optimization design information refers to design optimization instructions initiated at the overall level of clothing. Its scope covers the entire garment, rather than being limited to a specific component. For example, "adjust the overall design style to be more youthful." Such instructions aim to reshape the overall visual appeal of clothing from a macro perspective.

[0077] Non-specific optimization design information refers to design optimization instructions issued for specific parts or details of clothing. Their scope is clearly defined, such as "adjust the sleeve design style to be more business-like." These instructions do not involve overall adjustments; they only provide abstract style adjustment requirements for specific components.

[0078] Furthermore, for optimizing local design dimensions, it is not necessary to integrate all historical design information. It only needs to integrate historical design information consistent with the design dimension and historical design information specific to the overall garment design (note: historical design information specific to the overall garment design affects the design of local design dimensions; therefore, when optimizing local design dimensions, historical design information specific to the overall garment design must also be considered). Therefore, when the current optimized design information is of an unclear local design dimension, the current optimization semantic vector is obtained based on multiple target design information points that are consistent with the design dimension involved in the current optimized design information or are specific to the overall garment design, identified from all remaining historical design information after the latest filtering operation.

[0079] In some embodiments of this specification, a current optimization semantic vector is obtained based on the current optimized design information, multiple target design information, the optimization rounds corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimized design information, including:

[0080] Based on the optimization rounds corresponding to each of the multiple target design information and the degree of design conflict between each of the multiple target design information and the current optimization design information, obtain the fusion weights corresponding to each of the multiple target design information.

[0081] Based on the current optimization design information, multiple target design information, the preset fusion weights corresponding to the current optimization design information, and the fusion weights corresponding to each of the multiple target design information, the current optimization semantic vector is obtained.

[0082] Understandably, the earlier the optimization round of the target design information (note: the earlier the optimization round, the smaller the optimization round number), the smaller its fusion weight should be. Conversely, the higher the design conflict between the target design information and the current optimized design information, the smaller its fusion weight should also be. In other words, there is a positive correlation between the optimization round of the target design information and its corresponding fusion weight, and a negative correlation between the design conflict between the target design information and the current optimized design information and its corresponding fusion weight.

[0083] In some embodiments of this specification, obtaining the fusion weights corresponding to each of the multiple target design information based on the optimization rounds corresponding to each of the multiple target design information and the design conflict degree between each of the multiple target design information and the current optimization design information includes:

[0084] Clustering operations are performed on multiple target design information to obtain multiple target design information clusters;

[0085] Based on the optimization rounds corresponding to each of the multiple target design information, the degree of design conflict between each of the multiple target design information and the current optimization design information, and the cluster size of the target design information cluster to which each of the multiple target design information belongs, the fusion weights corresponding to each of the multiple target design information are obtained.

[0086] Understandably, this embodiment significantly improves the accuracy of understanding multi-round design intent by clustering the target design information and further introducing cluster size as a key factor in adjusting the fusion weight. The larger the cluster size of the target design information, the more times the user has emphasized or semantically repeated the design intent corresponding to that target design information during historical interactions, reflecting the intensity of that design intent. Based on this, adaptively increasing the fusion weight of this information based on the cluster size can effectively strengthen the expression of core design features in the final semantic vector, avoiding the dilution or loss of key design intent due to time decay or interference from new information, thereby ensuring that the generated clothing design drawing more accurately matches the user's core needs.

[0087] In some embodiments of this specification, obtaining the fusion weights corresponding to each of the multiple target design information based on the optimization rounds corresponding to each of the multiple target design information, the design conflict degree between each of the multiple target design information and the current optimized design information, and the cluster size of the target design information cluster to which each of the multiple target design information belongs, includes:

[0088] When the design conflict between the target design information and the current optimized design information is greater than the preset conflict degree, the fusion weight of the target design information is obtained based on the optimization round of the target design information and the design conflict degree between the target design information and the current optimized design information.

[0089] When the degree of design conflict between the target design information and the current optimized design information is no greater than the preset degree of conflict, the fusion weight of the target design information is obtained based on the optimization round of the target design information, the degree of design conflict between the target design information and the current optimized design information, and the cluster size of the target design information cluster to which the target design information belongs.

[0090] This embodiment differentiates the weight calculation strategy by setting a preset conflict level, effectively ensuring the rigor of the design logic. When the design conflict level between the target design information and the current optimization design information is greater than the preset conflict level, it indicates a significant contradiction between the target design information and the current design intent. In this case, the influence of cluster size on the weight is discarded, avoiding interference with the current design decision due to repeated emphasis on a conflicting design intent in the user's history (i.e., large cluster size). This mechanism ensures that the current optimization intent has absolute dominance when facing logically mutually exclusive scenarios, preventing the "reverse hijacking" of results by historically high-frequency design intents, thereby significantly improving the accuracy and flexibility of the system in handling design changes.

[0091] In some embodiments of this specification, obtaining the fusion weight of the target design information based on the optimization round of the target design information, the degree of design conflict between the target design information and the current optimized design information, and the cluster size of the target design information cluster to which the target design information belongs, includes:

[0092] The fusion weight of the target design information is obtained based on the optimization round of the target design information, the degree of design conflict between the target design information and the current optimization design information, the cluster size of the target design information cluster to which the target design information belongs, and the optimization round of each target design information in the target design information cluster to which the target design information belongs.

[0093] Understandably, in addition to considering the optimization rounds of the target design information itself, we can also add the optimization rounds corresponding to each of the target design information in the target design information cluster to which the target design information belongs.

[0094] Furthermore, the fusion weight of the target design information, obtained based on the optimization rounds of the target design information, the design conflict degree between the target design information and the current optimized design information, the cluster size of the target design information cluster to which the target design information belongs, and the optimization rounds corresponding to each of the target design information clusters to which the target design information belongs, may include:

[0095] Based on the optimization rounds corresponding to each of the target design information in the target design information cluster to which the target design information belongs, the average number of rounds for the target is obtained.

[0096] The fusion weight of the target design information is obtained based on the optimization rounds based on the target design information, the average rounds of the target, the degree of design conflict between the target design information and the current optimization design information, and the cluster size of the target design information cluster to which the target design information belongs.

[0097] Understandably, the smaller the average number of rounds in the target, the smaller the fusion weight of the target design information should be; that is, there is a positive correlation between the average number of rounds in the target and the fusion weight of the target design information.

[0098] In some embodiments of this specification, the determination of design conflict between design information includes:

[0099] Based on the semantic cosine similarity detection algorithm, the semantic cosine similarity between design information is obtained;

[0100] Based on the knowledge graph of the apparel industry, obtain the degree of graph conflict between design information;

[0101] The design conflict degree between design information is obtained based on the semantic cosine similarity and graph conflict degree between design information.

[0102] Understandably, this embodiment achieves multi-dimensional and accurate quantification of design conflict by integrating semantic cosine similarity detection and a knowledge graph of the apparel industry. Semantic cosine similarity can quickly capture the semantic correlation between design information at the text vector level, while the apparel industry knowledge graph can delve deeper into the logically mutually exclusive relationships of design intentions within the professional domain. By comprehensively considering both the shallow semantic connections and the deep logic at the industry level, this method effectively overcomes the limitations of single-dimensional judgment, thereby more accurately and reasonably obtaining the true degree of design conflict between the target design information and the current optimized design information, providing a reliable basis for subsequent weight allocation.

[0103] The following examples illustrate this:

[0104] Suppose historical design information involves "changing the sleeves to long sleeves," and current optimized design information involves "changing the sleeves to short sleeves." These two texts are very similar, differing by only one word, and are very close in semantic space, resulting in a high calculated semantic cosine similarity (potentially above 0.8). If only semantic cosine similarity is considered, the design conflict would be deemed low. However, the clothing industry knowledge graph defines a mutual exclusion relationship between "long sleeves" and "short sleeves," assigning specific degrees of mutual exclusion. Therefore, the graph conflict between the two is high. Without incorporating the clothing industry knowledge graph, relying solely on semantic similarity would lead the system to highly fuse "changing the sleeves to long sleeves" and "changing the sleeves to short sleeves," resulting in "mid-length sleeves" or "strange mixed sleeve styles" in the final clothing design image, contradicting the user's latest modification intention of "changing to short sleeves." By introducing the clothing industry knowledge graph, the system can accurately identify this strong conflict, thus directly filtering out such historical design information in step S3 or assigning it a lower fusion weight in subsequent steps.

[0105] Conversely, if only graph conflict degree is considered without considering semantic cosine similarity, the following situation may occur: Suppose historical design information involves "Please adjust my clothing to be more business formal," and current optimized design information involves "Please adjust my clothing to be more slim-fitting." In the clothing industry knowledge graph, the nodes "business formal" and "slim-fitting" may be independent of each other. Therefore, the design conflict degree between the two cannot be directly determined by the clothing industry knowledge graph alone. However, in a large fashion corpus, "business formal" and "slim-fitting" often co-occur frequently (e.g., "slim-fitting business suit"), meaning the design conflict degree between the two should be relatively small. Therefore, in the embodiments of this specification, the semantic cosine similarity is calculated using a semantic cosine similarity detection algorithm. This reveals that the vector angle between the two is small, indicating a high semantic cosine similarity. Finally, by comprehensively considering semantic cosine similarity and graph conflict degree, the true design conflict degree between the design information can be obtained more accurately.

[0106] Figure 2 The diagram illustrates the structure of a multi-round clothing design optimization system according to some embodiments of this disclosure. The various embodiments in this specification are described in a progressive manner; similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the multi-round clothing design optimization system embodiments are basically similar to the multi-round clothing design optimization method embodiments, so the description is relatively simple; relevant parts can be referred to in the description of the multi-round clothing design optimization method embodiments.

[0107] like Figure 2 As shown, a multi-round clothing design optimization system can include at least:

[0108] The first processing module outputs multiple clothing design drawings based on the AIGC model and the clothing design information input by the user.

[0109] The response module responds to the user's selection of information from multiple garment design drawings to determine the garment design drawing to be optimized.

[0110] The judgment module obtains the current optimization design information input by the user in the current optimization round, and performs a filtering operation on all the remaining historical design information based on the design conflict degree between the current optimization design information and all the remaining historical design information, and determines whether the current optimization design information is explicit optimization design information or implicit optimization design information.

[0111] The optimization module, when the current optimization design information is explicit optimization design information, obtains the current optimization semantic vector based on the explicit optimization design information, and optimizes the current garment design drawing to be optimized based on the current optimization semantic vector and the AIGC model to obtain the current optimized garment design drawing;

[0112] The optimization module, when the current optimization design information is non-ambiguous, obtains the current optimization semantic vector based on the current optimization design information, multiple target design information determined from all remaining historical design information after the latest filtering operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information. Based on the current optimization semantic vector and the AIGC model, the current garment design drawing to be optimized is optimized to obtain the current optimized garment design drawing.

[0113] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this specification are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., Digital Versatile Discs (DVDs)), or semiconductor media (e.g., Solid State Disks (SSDs)).

[0114] Figure 3 A block diagram of an electronic device 300 that can implement various embodiments of the present disclosure is shown. For example... Figure 3As shown, the electronic device 300 includes a processor 310, a disk drive 320, an input / output interface 330, a network interface 340, and a memory 350. The processor 310, disk drive 320, input / output interface 330, network interface 340, and memory 350 can communicate with each other via a communication bus 360.

[0115] The processor 310 can be implemented using a general-purpose CPU, microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits to execute relevant programs in order to implement the technical solution provided in this application.

[0116] The memory 350 can be implemented in the form of ROM (Read Only Memory), RAM (Read Access Memory), static memory, dynamic storage devices, etc. The memory 350 can store the operating system 351 used to control the operation of the electronic device 300, and the basic input / output system (BIOS) 352 used to control the low-level operations of the electronic device 300. Additionally, it can store a web browser 353, a data storage management system 354, etc. In summary, when the technical solution provided in this application is implemented through software or firmware, the relevant program code is stored in the memory 350 and is called and executed by the processor 310.

[0117] Input / output interface 330 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0118] Network interface 340 is used to connect a communication module (not shown in the figure) to enable communication and interaction between the device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0119] Bus 360 includes a pathway for transmitting information between various components of the device, such as processor 310, disk drive 320, input / input interface 330, network interface 340, and memory 350.

[0120] It should be noted that although the above-described device only shows the processor 310, disk drive 320, input / output interface 330, network interface 340, memory 350, bus 360, etc., in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the method of this application, and does not necessarily include all the components shown in the figures.

[0121] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0122] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. Furthermore, although operations are depicted in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the foregoing discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0123] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A multi-round clothing design optimization method, characterized in that, include: S1. Based on the AIGC model and the clothing design information input by the user, output multiple clothing design drawings; S2. Responding to the user's selection information among multiple garment design drawings, determine the garment design drawing to be optimized; S3. Obtain the current optimization design information input by the user in the current optimization round, and based on the design conflict degree between the current optimization design information and all the remaining historical design information, perform a filtering operation on all the remaining historical design information, and determine whether the current optimization design information is explicit optimization design information or implicit optimization design information. S4. When the current optimization design information is explicit optimization design information, obtain the current optimization semantic vector based on the explicit optimization design information, and optimize the current garment design drawing to be optimized based on the current optimization semantic vector and AIGC model to obtain the current optimized garment design drawing. Then, return to step S3 after using the current optimized garment design drawing as the new garment design drawing to be optimized, until the optimization ends. S5. When the current optimization design information is non-ambiguous optimization design information, based on the current optimization design information, multiple target design information determined from all remaining historical design information after the latest filtering operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information, obtain the current optimization semantic vector, and optimize the current garment design drawing to be optimized based on the current optimization semantic vector and the AIGC model to obtain the current optimized garment design drawing, and return to step S3 after using the current optimized garment design drawing as the new garment design drawing to be optimized, until the optimization ends.

2. The multi-round clothing design optimization method according to claim 1, characterized in that, Step S5 includes: When the current optimization design information is non-specific optimization design information, determine whether the current optimization design information is overall non-specific optimization design information or local design dimension non-specific optimization design information; When the current optimization design information is an overall non-ambiguous optimization design information, the current optimization semantic vector is obtained based on the current optimization design information, multiple target design information determined from all historical design information remaining after the latest screening operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information. When the current optimization design information is a local design dimension non-defined optimization design information, the current optimization semantic vector is obtained based on the current optimization design information, multiple target design information that are consistent with the design dimension involved in the current optimization design information or are related to the overall clothing design from all the remaining historical design information after the latest filtering operation, the optimization rounds corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information.

3. The multi-round clothing design optimization method according to claim 2, characterized in that, Based on the current optimized design information, multiple target design information, the optimization rounds corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimized design information, the current optimization semantic vector is obtained, including: Based on the optimization rounds corresponding to each of the multiple target design information and the degree of design conflict between each of the multiple target design information and the current optimization design information, obtain the fusion weights corresponding to each of the multiple target design information. Based on the current optimization design information, multiple target design information, the preset fusion weights corresponding to the current optimization design information, and the fusion weights corresponding to each of the multiple target design information, the current optimization semantic vector is obtained.

4. The multi-round clothing design optimization method according to claim 3, characterized in that, The process of obtaining the fusion weights corresponding to each of the multiple target design information based on the optimization rounds corresponding to each of the multiple target design information and the design conflict degree between each of the multiple target design information and the current optimization design information includes: Clustering operations are performed on multiple target design information to obtain multiple target design information clusters; Based on the optimization rounds corresponding to each of the multiple target design information, the degree of design conflict between each of the multiple target design information and the current optimization design information, and the cluster size of the target design information cluster to which each of the multiple target design information belongs, the fusion weights corresponding to each of the multiple target design information are obtained.

5. The multi-round clothing design optimization method according to claim 4, characterized in that, The process of obtaining the fusion weights for each of the multiple target design information based on the optimization rounds corresponding to each of the multiple target design information, the design conflict degree between each of the multiple target design information and the current optimization design information, and the cluster size of the target design information cluster to which each of the multiple target design information belongs, includes: When the design conflict between the target design information and the current optimized design information is greater than the preset conflict degree, the fusion weight of the target design information is obtained based on the optimization round of the target design information and the design conflict degree between the target design information and the current optimized design information. When the degree of design conflict between the target design information and the current optimized design information is no greater than the preset degree of conflict, the fusion weight of the target design information is obtained based on the optimization round of the target design information, the degree of design conflict between the target design information and the current optimized design information, and the cluster size of the target design information cluster to which the target design information belongs.

6. The multi-round clothing design optimization method according to claim 5, characterized in that, The fusion weight of the target design information is obtained based on the optimization round, the degree of design conflict between the target design information and the current optimization design information, and the cluster size of the target design information cluster to which the target design information belongs, including: The fusion weight of the target design information is obtained based on the optimization round of the target design information, the degree of design conflict between the target design information and the current optimization design information, the cluster size of the target design information cluster to which the target design information belongs, and the optimization round of each target design information in the target design information cluster to which the target design information belongs.

7. The multi-round clothing design optimization method according to claim 1, characterized in that, The degree of design conflict between design information is obtained, including: Based on the semantic cosine similarity detection algorithm, the semantic cosine similarity between design information is obtained; Based on the knowledge graph of the apparel industry, obtain the degree of graph conflict between design information; The design conflict degree between design information is obtained based on the semantic cosine similarity and graph conflict degree between design information.

8. A multi-round clothing design optimization system, based on the multi-round clothing design optimization method according to any one of claims 1 to 7, characterized in that, include: The first processing module outputs multiple clothing design drawings based on the AIGC model and the clothing design information input by the user. The response module responds to the user's selection of information from multiple garment design drawings to determine the garment design drawing to be optimized. The judgment module obtains the current optimization design information input by the user in the current optimization round, and performs a filtering operation on all the remaining historical design information based on the design conflict degree between the current optimization design information and all the remaining historical design information, and determines whether the current optimization design information is explicit optimization design information or implicit optimization design information. The optimization module, when the current optimization design information is explicit optimization design information, obtains the current optimization semantic vector based on the explicit optimization design information, and optimizes the current garment design drawing to be optimized based on the current optimization semantic vector and the AIGC model to obtain the current optimized garment design drawing; The optimization module, when the current optimization design information is non-ambiguous, obtains the current optimization semantic vector based on the current optimization design information, multiple target design information determined from all remaining historical design information after the latest filtering operation, the optimization round corresponding to each of the multiple target design information, and the design conflict degree between each of the multiple target design information and the current optimization design information. Based on the current optimization semantic vector and the AIGC model, the current garment design drawing to be optimized is optimized to obtain the current optimized garment design drawing.

9. An electronic device, characterized in that, include: One or more processors, and A memory associated with the one or more processors, the memory being used to store program information, which, when read and executed by the one or more processors, performs the steps of the method according to any one of claims 1-7.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1-7.

Images