Intelligent lightweight collaborative design method and system for digital media content

Abstract

The invention discloses a digital media content-oriented intelligent lightweight collaborative design method and system, and relates to the technical field of image data processing. Comprising the following steps of S1, collecting collaborative operation data in real time, and performing data preprocessing; s2, grouping the cooperative operations to form operation pairs, quantifying operation conflicts of each operation pair, and screening out conflicted operation pairs; s3, for the conflict operation pairs, generating content feature vectors after each operation, evaluating the fusion similarity of the operation pairs, and carrying out merging and branch processing on the operation pairs; and S4, performing comprehensive evaluation on cooperative processing, adjusting a cooperative strategy, allocating resources and optimizing a fusion scheme by combining an operation conflict quantification result, the fusion similarity and a merging and branch processing result of the operation pair. The problems that in a digital media content collaborative design scene, an existing collaborative design technology can only solve operation layer conflicts, and semantic layer conflicts generated when multiple users edit the same media content at the same time are difficult to recognize and process are solved.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology, specifically to an intelligent, lightweight collaborative design method and system for digital media content. Background Technology

[0002] With the rapid development of the digital economy and content creation industry, image content is widely used in numerous fields such as design, media, education, advertising, and documentation. Along with the popularization of new models such as remote work, cloud processing, and digital asset management, image generation, design, and the efficient editing, flexible management, and intelligent innovation of digital image content have become crucial foundations for content production, digital media operation, and management services. The widespread adoption and deepening application of digital images are driving the continuous upgrading of content production methods, organizational management models, and the industrial innovation ecosystem.

[0003] For example, the invention patent with publication number CN120707790A discloses a multi-view consistency 3D object editing method and system based on a diffusion model. The method includes: an initial frame style editing module that edits the initial viewpoint and consistently propagates the edited features to the multi-view sequence, alleviating the drift problem in traditional methods; a multi-view differential style injection module that, based on CLIP semantic embedding, calculates the difference features between the original and stylized images, and dynamically and accurately injects style differences into the video sequence of each viewpoint through an attention mechanism, effectively avoiding semantic illusion; and a structure-aware adaptive fusion module that employs a gradient-guided convolutional fusion strategy, combined with a spatial attention mechanism, selectively fuses stylized features into the video latent variables, ensuring that stylistic creativity is expressed while maintaining structural integrity. This invention is applicable to high-quality 3D content editing in the fields of virtual reality, augmented reality, and digital entertainment, and can significantly improve the multi-view consistency, structural fidelity, and semantic accuracy of 3D objects.

[0004] For example, invention patent CN108053495A discloses a method and system for converting 2D digital resources into dynamically changeable 3D digital resources. The method involves: extracting images from the 2D digital resources; placing the extracted results into a 3D coordinate system through texture mapping; performing motion processing and rendering in the 3D coordinate system to achieve dynamic changes; and then forming a 2D image through projection calculation. By calculating parallax in real time, the multimedia content of the 2D digital resources is integrated into a 3D scene through a movable plane, allowing it to dynamically change into a 3D image in a predetermined manner within the 3D virtual scene. This invention's terminal processing method eliminates the need for systems treating amblyopia or myopia to recreate specialized videos; existing VR images, videos, or films can be used to relax the ciliary muscles of the eyes and improve vision.

[0005] However, in collaborative design scenarios for digital media content, multiple users often design the same media content simultaneously in real time through different terminals. While existing collaborative design technologies can ensure consistency of multi-user operations at the data structure level, they can only resolve operational-level conflicts and cannot identify and handle semantic-level conflicts.

[0006] Therefore, in order to address the above issues, there is an urgent need for an intelligent, lightweight collaborative design method and system for digital media content. Summary of the Invention

[0007] Technical problems to be solved To address the shortcomings of existing technologies, this invention provides an intelligent, lightweight collaborative design method and system for digital media content. It solves the problem that existing collaborative design technologies can only resolve operational-level conflicts in digital media content collaborative design scenarios, and are unable to identify and handle semantic-level conflicts when multiple users are simultaneously editing the same media content.

[0008] Technical solution To achieve the above objectives, the present invention provides the following technical solution: an intelligent lightweight collaborative design method for digital media content, comprising the following steps: S1, real-time collection of collaborative operation data, data preprocessing and standardized storage of the collaborative operation data; S2, analysis of multi-user collaborative operation data to form operation pairs, quantification of operation conflicts in each operation pair, and selection of conflicting operation pairs based on the operation conflict quantification results; S3, for conflicting operation pairs, generation of content feature vectors after each operation using collaborative operation data, fusion of collaborative operation data and content feature vectors to evaluate the fusion similarity of operation pairs, and merging and branching of operation pairs based on the fusion similarity; S4, comprehensive evaluation of collaborative processing by combining operation conflict quantification results, fusion similarity, and merging and branching results of operation pairs, identification of collaborative health status, and adjustment of collaborative strategies, allocation of resources, and optimization of fusion schemes.

[0009] Furthermore, the specific process of real-time acquisition of collaborative operation data, data preprocessing, and standardized storage of collaborative operation data is as follows: Real-time acquisition of collaborative operation data specifically involves: real-time monitoring of all user operation events on media objects, acquiring timestamps, user IDs, operation types, and operation object IDs; simultaneously acquiring global attributes, regional attributes, and layer structure information of the operation object; where the regional attributes of the operation object include the image space coordinates involved in the operation; synchronously acquiring pixel values ​​before and after the operation within the area affected by each operation, calculating the difference in pixel values ​​before and after each operation, obtaining the average absolute change, and performing normalization processing as internal... The system incorporates several key features: First, it standardizes and transforms collaborative operation data collected from multiple terminals according to a unified data structure format, assigning a globally unique identifier to each piece of collaborative operation data. Second, it aggregates continuous operations performed by the same user within a fixed sliding time window, removing duplicate and invalid collaborative operation data. Third, it performs global time-series synchronization and deduplication of multi-terminal collaborative operation data based on unified operation timestamps and logical order, adjusting for operation disorder caused by clock errors and network latency. Fourth, it detects and marks concurrent conflicts between multiple users on the same operation object during time-series synchronization. Finally, it establishes a media collaboration database, storing the preprocessed collaborative operation data within it.

[0010] Furthermore, the specific process of analyzing multi-user collaborative operation data and grouping it into operation pairs is as follows: obtain the collaborative operation data for each operation event; using the operation object ID as the primary key, group all collaborative operations for the same operation object within a fixed sliding time window according to time order and spatial overlap, specifically satisfying: the time window interval of the operation is less than the time threshold and the regional attributes of the operation objects have intersection; for each group of collaborative operations, traverse and generate all possible pairwise operation pairs, and record the number of operation pairs.

[0011] Furthermore, the specific process for quantifying the operational conflict of each operation pair is as follows: For each operation pair, the operation types of the two operations are extracted and compared to obtain the operation type difference value. If the operation types are the same, the operation type difference value is assigned to 0, otherwise it is assigned to 1; the operation object region attributes of the two operations are extracted, and the intersection and union of the two operation object region attributes are calculated. The intersection is divided by the union to obtain the operation region overlap; the operation object layer structure of the two operations is extracted and compared to obtain the layer consistency value. If the operation object layer structure is the same, the layer consistency value is assigned to 1, otherwise it is assigned to 0; the content intensity of the two operations is extracted, and the smaller content intensity is divided by the larger content intensity to obtain the operation intensity normalization ratio; the operation type difference value, operation region overlap, layer consistency value, and operation intensity normalization ratio are multiplied to obtain the conflict degree value of the operation pair; based on the number of operation pairs, the conflict degree values ​​of all operation pairs are summed and averaged to obtain the collaborative operation conflict degree value.

[0012] Furthermore, the specific process of filtering out conflicting operation pairs based on the operation conflict quantification results is as follows: when multiple users perform collaborative operations, the collaborative operation conflict degree value is calculated in real time, and the collaborative operation data, operation pairs, conflict degree value of operation pairs and collaborative operation conflict degree value in each window are written into the media collaboration database. Collaborative operations with a collaborative operation conflict degree value greater than the conflict threshold are filtered out, marked as conflicting operations, and enter the semantic conflict reconciliation process.

[0013] Furthermore, for conflicting operation pairs, the specific process of generating content feature vectors after each operation through collaborative operation data and evaluating the fusion similarity of operation pairs by fusing collaborative operation data and content feature vectors is as follows: The collaborative operation data, operation pairs, conflict degree values ​​of the operation pairs, and collaborative operation conflict degree values ​​corresponding to conflicting operations are received. A semantic conflict reconciliation process is executed: For each operation pair in a conflicting operation, the intersection of the regional attributes of the two operation objects is obtained as the overlapping region. Based on the operation object ID and the operation object regional attributes, a region pixel histogram is extracted within the overlapping region as a content feature vector, and the cosine similarity of the content feature vectors of the two operations is calculated. The normalized ratio of operation intensity, operation region overlap, and hierarchical consistency value of the operation pair in each conflicting operation are obtained. The normalized ratio of operation intensity, operation region overlap, hierarchical consistency value, and cosine similarity of the content feature vectors of the operation pair are multiplied to obtain the operation fusion similarity value.

[0014] Furthermore, the specific process of merging and branching operation pairs based on fusion similarity is as follows: Calculate the operation fusion similarity value of all operation pairs in the conflicting operation and compare it with the fusion threshold to determine operation fusion: When the operation fusion similarity value is greater than or equal to the fusion threshold, the current operation pair is determined to be similar, the operation results are directly merged, all operation content, content feature vectors and collaborative operation data are integrated, and a unique synthesized content version is output; when the operation fusion similarity value is less than the fusion threshold, the current operation pair is determined to be different, branches are created for different operation paths of the operation pair, independent version IDs are generated for each, multiple operation results are saved by version, and the operation intent of each user is fully preserved; the operation fusion similarity value of each operation pair and the operation fusion determination result are written into the media collaboration database.

[0015] Furthermore, the specific process for comprehensively evaluating collaborative processing, combining the results of operation conflict quantification, fusion similarity, and the results of operation pair merging and branching, is as follows: Based on a sliding time window, the number of operation pairs in conflicting operations is counted, and the conflict degree value, operation fusion similarity value, and operation fusion judgment result of each operation pair are obtained. The operation fusion judgment result of the operation pair is analyzed to obtain the conflict processing value. If the operation fusion judgment result is merging, the conflict processing value is assigned to 1; if the operation fusion judgment result is branching, the conflict processing value is assigned to 0.5. The conflict degree value of the operation pair is subtracted from the constant to obtain the conflict degree inverse value. For each operation pair in a conflicting operation, the conflict degree inverse value, operation fusion similarity value, and conflict processing value are added to obtain a single set of collaborative processing evaluation value. Based on the number of operation pairs in a conflicting operation, the single set of collaborative processing evaluation values ​​of all operation pairs are summed and divided by three times the number of operation pairs to obtain the total collaborative processing evaluation value.

[0016] Furthermore, the specific process for identifying the collaborative health status and adjusting collaborative strategies, allocating resources, and optimizing fusion schemes is as follows: The total collaborative processing evaluation value is periodically calculated and compared with the optimization threshold. When the total collaborative processing evaluation value is greater than or equal to the optimization threshold, the current collaborative health status is determined to be good, existing strategies and parameter settings are maintained, concurrent collaboration and elastic resource allocation are allowed, and collaborative design permissions are continuously opened. When the total collaborative processing evaluation value is less than the optimization threshold, the current collaborative health status is determined to be at risk, triggering an optimization mechanism: Based on the operation object ID and operation object region attributes of the collaborative operation data, the operation pair grouping is optimized, and the conflict threshold is reduced; regions in media content where the number of edits exceeds the frequency threshold are restricted, and collaborative design permissions and concurrent editing numbers are temporarily adjusted; high conflict incidence and collaborative risk warnings are pushed to relevant users and management terminals, along with off-peak collaboration prompts; a content feature vector matching algorithm based on the Transformer architecture is scheduled to analyze the similarity of cross-user multi-version content and optimize merging decisions for operation pairs that need to be merged, and output fusion suggestions and automatic synthesis results; simultaneously, reinforcement learning algorithms are periodically used to analyze historical total collaborative processing evaluation values ​​and actual operation fusion judgment results, and to optimize and adjust the conflict threshold, fusion threshold, and optimization threshold.

[0017] The second aspect of this invention provides an intelligent lightweight collaborative design system for digital media content, comprising: a collaborative operation data acquisition and preprocessing module for real-time acquisition of collaborative operation data, preprocessing the collaborative operation data, and standardizing and storing it; a collaborative intent recognition and conflict perception module for analyzing multi-user collaborative operation data, grouping it into operation pairs, quantifying the operational conflicts of each operation pair, and filtering out conflicting operation pairs based on the operation conflict quantification results; a semantic conflict reconciliation and fusion judgment module for generating content feature vectors after each operation based on the collaborative operation data for conflicting operation pairs, fusing the collaborative operation data and content feature vectors to evaluate the fusion similarity of the operation pairs, and merging and branching the operation pairs based on the fusion similarity; and a collaborative feedback and optimization module for comprehensively evaluating collaborative processing by combining the operation conflict quantification results, fusion similarity, and the merging and branching results of the operation pairs, identifying the collaborative health status, and adjusting collaborative strategies, allocating resources, and optimizing fusion schemes.

[0018] Beneficial effects The present invention has the following beneficial effects: (1) This invention achieves comprehensive extraction of multi-dimensional features such as operation type, regional overlap, layer consistency and content intensity by standardizing the collection and preprocessing of collaborative operation data. Based on a unified mechanism, it judges and filters conflicting operations in the multi-user operation process in real time, ensuring the objectivity, automation and real-time nature of conflict detection, and effectively supporting the efficient collaborative design of large-scale distributed media content.

[0019] (2) The present invention evaluates the fusion similarity of operation pairs by fusing the content feature vector of the region pixel histogram with operation intensity, region overlap, and hierarchical consistency collaborative operation data, which significantly improves the automatic identification and refined processing capability of semantic layer conflicts, supports intelligent merging or branching decision-making of multiple users and multiple paths in the same image region, and significantly improves the rationality of the design results and the complete preservation of user intent.

[0020] (3) This invention achieves comprehensive quantification and hierarchical feedback on the collaborative health status of the platform by periodically calculating the total evaluation value of collaborative processing. It can automatically identify high-risk areas, dynamically adjust conflict detection and editing permission parameters, and combine AI-assisted merging strategies to perform adaptive resource allocation and collaborative process optimization, effectively improving collaborative efficiency and reducing the difficulty of manual intervention and conflict management.

[0021] (4) This invention, through the standardization, labeling and database storage of data in the entire collaborative operation process, enables all conflict identification, fusion criteria and decision results to be traceable and automatically archived, which facilitates subsequent intelligent analysis, model training and platform self-learning evolution, and lays a solid foundation for continuously improving the level of intelligence and management accuracy.

[0022] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0023] Figure 1 A flowchart of an intelligent, lightweight collaborative design method for digital media content; Figure 2 This is a structural diagram of an intelligent, lightweight, collaborative design system for digital media content. Figure 3 To perform operation on multi-dimensional feature radar maps based on fusion similarity; Figure 4 This is a distribution map of the fusion similarity values. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. As those skilled in the art will understand, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Please see Figures 1-4 This invention provides a technical solution: an intelligent lightweight collaborative design method and system for digital media content, such as... Figure 1 As shown, the process includes the following steps: S1, real-time collection of collaborative operation data, data preprocessing and standardized storage of the collaborative operation data; S2, analysis of multi-user collaborative operation data to group into operation pairs, quantification of operation conflicts for each operation pair, and selection of conflicting operation pairs based on the operation conflict quantification results; S3, for conflicting operation pairs, generation of content feature vectors after each operation using collaborative operation data, fusion of collaborative operation data and content feature vectors to evaluate the fusion similarity of operation pairs, and merging and branching of operation pairs based on the fusion similarity; S4, comprehensive evaluation of collaborative processing by combining operation conflict quantification results, fusion similarity, and merging and branching results of operation pairs, identification of collaborative health status, and adjustment of collaborative strategies, allocation of resources, and optimization of fusion schemes.

[0026] Specifically, the process of real-time acquisition of collaborative operation data, data preprocessing, and standardized storage of collaborative operation data is as follows: Real-time acquisition of collaborative operation data specifically involves: real-time monitoring of all user operation events on media objects, acquiring timestamps, user IDs, operation types, and operation object IDs; simultaneously acquiring global attributes, regional attributes, and layer structure information of the operation object; where media objects include static images and image frames, and operation types include but are not limited to drawing, smearing, cropping, moving, scaling, region selection, layer adjustment, and pixel-level modification, covering common image design behaviors; the regional attributes of the operation object include the image space coordinates involved in the operation; where global attributes of the operation object are used to identify the uniqueness of the image file and metadata, such as resolution and color mode, and the regional attributes of the operation object use a two-dimensional coordinate system to accurately mark the range of the image area edited by the user; the layer structure information of the operation object is used to record the image layering relationship, such as the level and visibility of each layer, facilitating the identification and analysis of different layers in subsequent operations. The system synchronously collects pixel values ​​before and after each operation within the affected area, calculates the difference in pixel values ​​before and after each operation, and obtains the average absolute change. This average change is then normalized to represent the content intensity. Specifically, it iterates through all pixels within the operation area, recording pixel values ​​(e.g., RGB channel intensity) before and after the operation. The difference is calculated per pixel, and the sum of these differences is taken as the average absolute value within the area. After normalization, this average absolute value is used as a quantitative indicator of the content intensity of the image editing operation, ranging from 0 to 1, to quantitatively reflect the editing scope and impact. The collaborative operation data collected from multiple terminals is standardized using a unified JSON data structure format. Each piece of collaborative operation data is assigned a globally unique identifier, along with a composite unique identifier consisting of a timestamp and a user ID. Continuous operations by the same user within a fixed sliding time window are aggregated, and duplicate and invalid collaborative operation data are removed. The sliding time window is parameterizable, preferably set between 0.5 and 5 seconds, and can be flexibly adjusted according to specific collaborative editing scenarios and user operating habits. For example, in high-interaction-frequency scenarios, the window length can be set to 1 second to efficiently aggregate consecutive operations within the same area in a short period, reducing fragmented data. In low-frequency scenarios, the window length can be appropriately increased to ensure the completeness and rationality of operation aggregation. The system can automatically recommend the optimal window length parameter based on operation frequency and conflict rate, supporting dynamic adaptive adjustment. Deduplication rules include: operations by the same user within the same time window and area where the content has not actually changed are directly removed and judged as duplicate operations; operations with completely identical content but different user IDs or timestamps are retained only the latest one; invalid operations that have been undone or reworked but have not resulted in actual content changes are excluded, improving the effectiveness of the dataset.Based on UTC standard time, and using a unified operation timestamp and logical order, global time synchronization and deduplication are performed on multi-terminal collaborative operation data to correct operation disorder caused by clock errors and network latency. Specifically, the system integrates network time protocols to automatically align the clocks of multiple systems, employs a timestamp correction algorithm, and allows a maximum error of no more than 50 milliseconds. In case of extreme network fluctuations, a secondary sorting correction can be performed based on the logical order and content change relationship to ensure the authenticity and reliability of the sequential relationship of all operations on the global timeline, providing a stable foundation for subsequent conflict determination. During time synchronization, concurrent conflicts of multiple users on the same operation object are detected and marked. Concurrent conflicts refer to multiple users performing incompatible operations on the same area or layer of an image within the same time period, such as simultaneously modifying pixels in the same area or different users editing elements on the same layer simultaneously. Such concurrency is detected in real time, and related operations are automatically marked with concurrent conflict flags as a structural pre-screening. A media collaboration database is established, and the preprocessed collaborative operation data is stored in the media collaboration database.

[0027] This implementation scheme, through standardized acquisition and processing of collaborative image operation data throughout the entire process, meticulously records every specific operation performed by the user during image editing, including spatial coordinates, layer structure, and multi-dimensional information on content intensity. Content intensity is quantified using average absolute pixel change normalization, effectively reflecting the actual impact of each edit on the image content. A sliding time window aggregates continuous operations and eliminates invalid and duplicate operations, improving data processing efficiency and effectiveness. A unified temporal synchronization and out-of-order correction mechanism ensures global consistency across multiple terminals and users. Furthermore, concurrent conflicts within the same region or layer are automatically detected and marked, providing high-quality data support for subsequent conflict quantification analysis and intelligent fusion decision-making. Overall, this significantly enhances the traceability, data consistency, and conflict control capabilities of collaborative image editing, meeting the high standards required by intelligent digital image collaborative processing systems.

[0028] Specifically, the process of analyzing and grouping multi-user collaborative operation data into operation pairs is as follows: First, acquire the collaborative operation data for each operation event. Second, using the operation object ID as the primary key, group all collaborative operations for the same operation object within a fixed sliding time window according to time order and spatial overlap, specifically satisfying the following conditions: the time window interval of the operations is less than a time threshold, and the operation object's regional attributes overlap. The width of the fixed sliding time window can be flexibly configured according to actual needs, typically ranging from a few seconds to tens of seconds, ensuring the timeliness of operation grouping. Spatial overlap is determined by the intersection of the operation object's regional attributes; if the coordinate intervals of two operations overlap, they are considered to have spatial overlap. This ensures that editing operations within the same time period and area are accurately grouped, facilitating subsequent conflict analysis. For each group of collaborative operations, iterate through and generate all possible pairwise operation pairs, recording the number of operation pairs: combine all collaborative operation events within the same group pairwise; for example, n operation events will generate n(n-1) / 2 operation pairs, each containing complete operation metadata.

[0029] This implementation scheme, through temporal and spatial dual grouping of multi-user image editing operations, can accurately extract and organize collaborative operation pairs within the same time period and region, ensuring that all potentially conflicting editing behaviors are effectively captured. Combining sliding time windows and spatial intersection determination ensures both the timeliness of grouping and improves the accuracy of conflict detection. Generating operation pairs through pairwise traversal provides a sufficient data foundation for subsequent quantitative analysis of operation conflict degree and fusion similarity. Overall, it enhances the conflict identification capability and operation management efficiency in a collaborative design environment.

[0030] Specifically, the process for quantifying the operational conflict of each operation pair is as follows: For each operation pair, the operation types of the two operations are extracted and compared to obtain the operation type difference value. If the operation types are the same, the operation type difference value is assigned to 0; otherwise, it is assigned to 1. Different types of operations may have essential differences in operational intent. The operation object region attributes of the two operations are extracted, and the intersection and union of the two operation object region attributes are calculated. The intersection is divided by the union to obtain the operation region overlap, with a value ranging from 0 to 1. The operation region overlap refers to the actual overlapping part of the two user-edited regions on the image and the overlap between the two regions. The proportion of the total coverage area; for example, if user A's operation area is the top left corner (10,20) and bottom right corner (60,80), and user B's is the top left corner (40,50) and bottom right corner (90,100), then the overlap area of ​​the two areas is 600 pixels, the union area is 4900 pixels, and the calculated operation area overlap is 0.122, which quantitatively reflects the degree of spatial intersection of multi-user operations, providing basic data for subsequent conflict analysis and content fusion; the operation area overlap reflects the degree of spatial overlap between two operations, the larger the value, the closer the operation areas are, and the higher the potential conflict risk. Extract the layer structure of the operation objects of the two operations and compare them to obtain the layer consistency value. If the layer structure of the operation objects is the same, the layer consistency value is assigned to 1, otherwise it is assigned to 0; the layer structure information is used to distinguish whether the operation occurs at the same image layer, such as the main layer, mask layer, and annotation layer. When the layer consistency is 1, it means that the two operations are at the same layer, which is prone to direct content coverage conflict; otherwise, it is 0, which means there is no direct conflict. The content intensity of two operations is extracted, and the smaller content intensity is divided by the larger content intensity to obtain the normalized ratio of operation intensity. The conflict degree of the operation pair is obtained by multiplying the operation type difference value, operation area overlap degree, hierarchical consistency value, and operation intensity normalized ratio. This comprehensively reflects the role of multiple factors such as operation type, spatial area, hierarchical consistency, and operation amplitude in conflict determination. The higher the value, the greater the probability of a substantial conflict between the two operations. Based on the number of operation pairs, the conflict degree values ​​of all operation pairs are summed and averaged to obtain the collaborative operation conflict degree value, which serves as the core indicator for quantifying the conflict intensity of multi-person collaborative design.

[0031] The specific formula for the collaborative operation conflict degree value is as follows: ; In the formula, It represents the degree of conflict in collaborative operations, quantifying the actual degree of conflict between different user operations when multiple people collaboratively design the same digital media content; Indicates the number of operation pairs; This indicates the difference in operation type, determines whether two operation types are different, reflects the fundamental difference in operation intent, filters out invalid conflicts with the same type, and highlights the collaborative contradictions caused by different types. This indicates the degree of overlap between operation regions, calculating the actual spatial overlap ratio of two operations. It highlights operations that truly intersect in space; no overlap means no substantial impact, while significant overlap may lead to conflict. This indicates a region attribute of one of the operands in an operation pair. This indicates the region attribute of another operand in the operand pair; It represents the layer consistency value, determines whether two operations occur at the same layer level, effectively distinguishes between direct conflicts at the same layer and false conflicts at different layers, and enhances semantic discriminative power; This represents the normalized ratio of operational intensity, measuring whether the actual changes to the content caused by two operations are similar. It prevents the overestimation of conflicts between minor and major operations, improving the objectivity of conflict assessment and its practical engineering guidance significance. This indicates the intensity of one of the contents in the operation pair. This indicates the intensity of another content in the operation pair.

[0032] This implementation plan, through quantitative analysis of multiple factors including operation type, spatial region, layer level, and content intensity for each operation pair, can accurately identify actual conflict relationships in multi-user collaborative design processes. The construction of indicators such as operation type differences, regional overlap, and layer consistency effectively improves the ability to identify conflicts in complex editing scenarios. In particular, the calculation of spatial overlap provides a scientific basis for risk assessment of operations that intersect in space; the normalized ratio of content intensity further reflects the proximity of the influence of different operations. By averaging the conflict intensity values ​​of all operation pairs, the overall conflict intensity in the current collaborative design state can be intuitively reflected, facilitating the automatic identification of high-risk collaborative areas and providing solid data support for subsequent intelligent fusion and optimization.

[0033] Specifically, the process of filtering conflicting operation pairs based on the quantitative results of operation conflicts is as follows: When multiple users perform collaborative operations, the collaboration operation conflict degree value is calculated in real time. The collaborative operation data, operation pairs, conflict degree values ​​of operation pairs, and collaboration operation conflict degree values ​​in each window are written into the media collaboration database to ensure that all key intermediate data has traceability and subsequent retrieval and analysis capabilities. Collaborative operations with a collaboration operation conflict degree value greater than the conflict threshold are filtered out and marked as conflicting operations, entering the semantic conflict reconciliation process. For collaborative operations with a collaboration operation conflict degree value less than or equal to the conflict threshold, they are determined to have no substantial conflict and do not need to enter the subsequent semantic conflict reconciliation process. They can be directly archived according to the original operation order to ensure operational efficiency and the smoothness of collaborative design.

[0034] This implementation scheme achieves efficient conflict screening and automatic classification in multi-user editing processes by dynamically determining the conflict level of collaborative operations. Operations exceeding the threshold are promptly identified and marked as conflicting, entering the subsequent semantic reconciliation process to ensure intelligent handling of potential risks. Operations below or equal to the threshold are automatically determined to have no substantial conflict and are directly archived, avoiding unnecessary resource waste and effectively improving the smoothness and efficiency of collaborative design. Overall, it achieves hierarchical response and refined management of conflict handling, ensuring the traceability and stable operation of operational data in a collaborative environment.

[0035] Specifically, for conflicting operation pairs, the process of generating content feature vectors after each operation using collaborative operation data and evaluating the fusion similarity of operation pairs by fusing collaborative operation data and content feature vectors is as follows: The process involves receiving collaborative operation data, operation pairs, conflict degree values ​​of the operation pairs, and collaborative operation conflict degree values ​​corresponding to conflicting operations, and executing a semantic conflict reconciliation process. For each operation pair in a conflicting operation, the intersection of the regional attributes of the two operation objects is obtained as the overlapping region. Based on the operation object ID and the operation object regional attributes, all pixels within the overlapping region are traversed, and the distribution frequency of each pixel value is statistically analyzed. A multi-dimensional regional pixel histogram is extracted within the overlapping region, and a high-dimensional vector representation of the regional content features is constructed as the content feature vector. The content feature vector effectively characterizes the fine-grained differences in content between different operation results. The cosine similarity of the content feature vectors of the two operations is then calculated. Cosine similarity is a commonly used algorithm to measure the angle between two vectors; a value closer to 1 indicates greater content similarity, while a value closer to 0 indicates greater content difference, used to quantitatively express the consistency of content changes. Obtain the normalized ratio of operation intensity, overlap of operation regions, and hierarchical consistency value of operation pairs in each conflicting operation; multiply the normalized ratio of operation intensity, overlap of operation regions, hierarchical consistency value, and cosine similarity of content feature vectors of operation pairs to obtain the operation fusion similarity value, which serves as the core basis for subsequent automatic fusion and branch determination. The higher the value, the more similar the two operations are in terms of content, space, hierarchy, and magnitude, and the higher the feasibility of automatic fusion.

[0036] The specific formula for the operation fusion similarity value is as follows: ; In the formula, This represents the operation fusion similarity value, which measures the content similarity of each pair of conflicting operations in the overlapping area. The larger the value, the more similar the two operations are, and the higher the feasibility of automatic merging. This represents the normalized ratio of operation intensity, which measures the similarity of the actual changes to the content caused by two operations. The larger the value, the closer the intensity of the two operations are, the higher the similarity, and the more significant it is to be merged. This indicates the degree of overlap between the operation areas, measuring the proportion of spatial overlap between two operations. The larger the value, the closer the areas of action are, the higher the similarity, and the more necessary it is to consider merging. This represents the layer consistency value, which determines whether two operations occur at the same layer level. 1 indicates the same layer and 0 indicates different layers. Only operations at the same layer have actual merging significance. This represents one of the content feature vectors in the operation pair. This represents another content feature vector in the operation pair. The cosine similarity of the content feature vectors is represented by the value. The larger the value, the closer the two contents are in essence, and the more suitable they are to be merged. Conversely, when the difference is large, the value is close to 0, and the contents should not be merged.

[0037] In this embodiment, Table 1 is a data table of operation fusion similarity values. The table details the content intensity 1, content intensity 2, intersection of operation object region attributes, union of operation object region attributes, hierarchical consistency value, cosine similarity of content feature vectors, and operation fusion similarity value for five operation pairs. Specifically, for operation pair 1, the content intensity 1 is 0.87, content intensity 2 is 0.92, the intersection of operation object region attributes is 117, the union of operation object region attributes is 163, the hierarchical consistency value is 1, the cosine similarity of content feature vectors is 0.913, and the operation fusion similarity value is 0.619; for operation pair 2, the content intensity 1 is 0.69, content intensity 2 is 0.73, the intersection of operation object region attributes is 54, the union of operation object region attributes is 97, the hierarchical consistency value is 1, the cosine similarity of content feature vectors is 0.674, and the operation fusion similarity value is 0.354; the content intensity 2 of operation pair 3 is... For operation pair 1, the content intensity is 0.94, content intensity 2 is 0.81, the intersection of operation object region attributes is 49, the union of operation object region attributes is 92, the hierarchical consistency value is 0, the cosine similarity of content feature vectors is 0.431, and the operation fusion similarity value is 0.000. For operation pair 4, the content intensity 1 is 0.81, content intensity 2 is 0.83, the intersection of operation object region attributes is 98, the union of operation object region attributes is 143, the hierarchical consistency value is 1, the cosine similarity of content feature vectors is 0.867, and the operation fusion similarity value is 0.579. For operation pair 5, the content intensity 1 is 0.79, content intensity 2 is 0.86, the intersection of operation object region attributes is 74, the union of operation object region attributes is 110, the hierarchical consistency value is 1, the cosine similarity of content feature vectors is 0.712, and the operation fusion similarity value is 0.439.

[0038] Table 1. Operational Fusion Similarity Values ​​Data Table

[0039] like Figure 3As shown, this is a multi-dimensional feature radar chart of operation fusion similarity. It visualizes the data of five operation pairs across five dimensions: content intensity normalization ratio, region overlap, hierarchical consistency, feature similarity, and fusion similarity. Each line represents a specific operation pair, and the length of the line in different dimensions reflects the quantitative performance of the operation pair in the corresponding dimension. (Based on Table 1 and...) Figure 3 It can be seen that most operation pairs perform well in terms of content intensity normalization ratio and hierarchical consistency, indicating that the operation pairs generally have good consistency and comparability in content intensity and hierarchical structure. Significant differences exist in the feature similarity and fusion similarity dimensions. Some operation pairs, such as operation pair 3, have significantly lower values ​​in these dimensions, indicating large differences in content features and high fusion similarity. Operation pairs 1 and 4 perform relatively evenly across all dimensions, with fusion similarity at a moderate to low level, indicating that these operation pairs are relatively easy to merge. Operation pair 3 has a hierarchical consistency of 0 and extremely low fusion and feature similarity, meaning that the content of this operation pair has no practical merging significance. Operation pairs 2 and 5 have low values ​​in regional overlap and fusion similarity, indicating some differences in spatial regions or content features, requiring careful handling during merging.

[0040] like Figure 4 The figure shows the distribution of operation fusion similarity values. It displays the fusion similarity values ​​for five different operation pairs, with the fusion threshold marked by dashed lines. The horizontal axis represents the operation pair number, and the vertical axis represents the fusion similarity value. (Based on Table 1 and...) Figure 4 It can be seen that the fusion similarity values ​​of operation pair 1 and operation pair 4 exceed the fusion threshold, indicating that the content changes between these two operation pairs are relatively small, the regions overlap significantly, the layers are consistent, and the content features are highly similar, thus providing a basis for automatic merging. The fusion similarity values ​​of operation pair 2 and operation pair 5 are significantly lower than the threshold, indicating that these operation pairs have large content differences or low overlap, resulting in a high risk of automatic merging and making them unsuitable for automatic merging. The fusion similarity value of operation pair 3 is 0 because the layer consistency is 0, meaning that there is no substantial content conflict at different layers, and no fusion processing is required.

[0041] This implementation scheme accurately measures the actual fusion similarity of various conflicting operations under multi-user collaborative design by extracting pixel-level content features of conflicting operation pairs, calculating cosine similarity, and fusing multi-dimensional operation metadata. It effectively characterizes the fine-grained differences in content after different operations using pixel histograms of overlapping regions, and achieves a comprehensive quantitative analysis of fusion possibilities by combining multiple indicators such as operation intensity normalization, spatial overlap, and hierarchical consistency. Overall, it not only significantly improves the intelligence level of automatic conflict resolution and content merging in complex image editing scenarios, but also provides scientific and executable data support for subsequent fusion judgment and version management, enhancing collaborative consistency and editing efficiency.

[0042] Specifically, the process of merging and branching operation pairs based on fusion similarity is as follows: Calculate the fusion similarity value of all operation pairs in the conflicting operation and compare it with the fusion threshold to determine the operation fusion: When the fusion similarity value is greater than or equal to the fusion threshold, the current operation pair is determined to be similar, and the operation results are directly merged. All operation content, content feature vectors, and collaborative operation data are integrated to output a unique synthesized content version. At this time, the edited content of two similar operations is automatically merged at the pixel level. Specifically, pixel-level merging rules are adopted, including: for each pixel in the overlapping area, the operation with the largest absolute value of pixel value change is selected as the final pixel value, and the corresponding collaborative operation data is written into the traceability information; for non-overlapping areas, their respective edited results are retained. During the merging process, priority is determined primarily by content intensity. When intensity is equal, the latest operation time takes precedence to ensure the uniqueness, engineering consistency, and traceability of the final value of each pixel. For example, if user A changes the value of pixel (20, 30) from 150 to 210, and user B changes the value of the same pixel from 150 to 170, then A's absolute change is 60, and B's is 20. The final pixel value is A's result of 210, and this decision is recorded as originating from A's collaborative operation data. If A and B have the same operation intensity, the one with the later update time is selected. All merging results are automatically archived in the merged content version for easy unified display, retrospection, and further processing. The merged content version contains data traceability information for all related operations and the final pixel content for easy unified tracing and display. When the operation fusion similarity value is less than the fusion threshold, the current operation pair is determined to be different. Branches are created for different operation paths of the operation pair, and independent version IDs are generated for each. Multiple operation results are saved by version, fully preserving the operation intent of each user. The branching mechanism ensures that differing operations can be retained separately and traced independently, and automatically generates a unique version identifier for each branch. Related edited content and metadata are archived by version, facilitating subsequent rollback, retrospection, and multi-version management. The operation fusion similarity value and operation fusion determination result of each operation pair are written to the media collaboration database.

[0043] This implementation scheme achieves intelligent merging and branch management in a collaborative design environment by dynamically determining the fusion similarity value of conflicting operation pairs. For operation pairs with high similarity, they can be automatically merged to generate a unified content version, improving editing consistency and collaboration efficiency. For operation pairs with differences, they are saved separately according to the branch method, fully recording the operation intentions of each user, supporting version backtracking and diversified content management. The linkage application of fusion similarity value and fusion threshold ensures both the consistency and continuity of content when facing complex conflicts, while also taking into account the traceability of personalized operations and the coexistence of multiple versions, greatly improving the intelligence and flexibility of collaborative design.

[0044] Specifically, the comprehensive evaluation process for collaborative processing, combining the quantitative results of operational conflicts, fusion similarity, and the merging and branching results of operational pairs, is as follows: Based on a sliding time window, the number of operational pairs in conflicting operations is counted. The length of the sliding time window can be flexibly set according to actual business needs to ensure that the data statistics can reflect the collaborative status in real time and have historical backtracking capabilities. The conflict degree value, operational fusion similarity value, and operational fusion judgment result of each operational pair are obtained. The operational fusion judgment result of the operational pair is analyzed to obtain the conflict processing value. If the operational fusion judgment result is merging, the conflict processing value is assigned to 1; if the operational fusion judgment result is branching, the conflict processing value is assigned to 0.5. The conflict processing value serves as a quantitative expression of the final content processing method. Merging operations indicate that the user's edited results have been successfully merged, while branching operations represent the coexistence of multiple versions. The conflict degree inverse value is obtained by subtracting the conflict degree value of the operational pair from a constant. The conflict degree inverse value reflects the compatibility and collaborative potential between operational pairs; the larger the value, the lower the actual conflict risk and the higher the probability of merging. For each pair of operations in a conflicting operation, the conflict degree inverse value, operation fusion similarity value, and conflict handling value are added together to obtain a single-group collaborative processing evaluation value. This single-group collaborative processing evaluation value comprehensively reflects the spatial conflict intensity, fusion feasibility, and final processing result, facilitating the quantitative assessment of the collaborative health status of each pair of conflicting operations. Based on the number of operation pairs in a conflicting operation, the single-group collaborative processing evaluation values ​​of all operation pairs are summed and divided by three times the number of operation pairs to obtain the total collaborative processing evaluation value. The purpose of dividing by three times the number of operation pairs is to normalize the evaluation values ​​of all conflicting operation pairs. Since each term theoretically ranges from 0 to 1, the maximum possible value after adding the three terms is 3. To ensure that the final total collaborative processing evaluation value is stable within the standard range of 0 to 1, so that the total collaborative processing evaluation value has clear physical and engineering significance and can be used for horizontal comparison and threshold determination in different scenarios.

[0045] The specific formula for the overall evaluation value of collaborative processing is as follows: ; In the formula, This represents the overall evaluation value of collaborative processing, which comprehensively quantifies the overall capability and optimization effectiveness of automatically handling all conflicting operations within a certain statistical period. The larger the value, the higher the quality of collaborative processing. This represents the conflict level of the operation pair, measuring the severity of conflict during the collaboration process. A higher value indicates a more severe conflict and a higher difficulty in merging. This is used to reflect that the less conflict there is, the more conducive it is to merger; This represents the operation fusion similarity value of the operation pair, which measures the feasibility of automatic fusion of the operation pair. The larger the value, the easier the fusion and the stronger the automatic processing capability. This represents the conflict handling value, used to quantify the quality of the operation on the final automatic processing. It is set to 1 when merging and 0.5 when branching, reflecting the uniformity and integration level of the actual output content. This indicates the number of operation pairs in a conflicting operation, used for normalizing the overall evaluation.

[0046] This implementation plan normalizes and quantifies multi-dimensional information such as conflict degree, fusion similarity, and processing method to construct a unified overall evaluation value for collaborative processing. This enables real-time and objective evaluation of the overall health status and processing effect of the multi-user image collaborative design process. A sliding time window mechanism ensures the timeliness and traceability of the evaluation, while the normalization of the three indicators ensures the comparability of evaluation results across different time periods and collaborative scenarios. It accurately reflects the intensity of collaborative conflicts, the feasibility of fusion, and the rationality of the final processing, providing a scientific basis for subsequent strategy adjustments, resource allocation, and system optimization. This significantly improves management efficiency and content consistency in the intelligent collaborative design environment.

[0047] Specifically, the process of identifying collaborative health status and adjusting collaborative strategies, allocating resources, and optimizing fusion schemes is as follows: The total collaborative processing evaluation value is periodically calculated and compared with an optimization threshold. When the total collaborative processing evaluation value is greater than or equal to the optimization threshold, the current collaborative health status is considered good. Existing strategies and parameter settings are maintained, allowing concurrent collaboration and flexible resource allocation, and collaborative design permissions remain open. No additional tuning operations are needed at this time, ensuring user experience and editing smoothness. When the total collaborative processing evaluation value is less than the optimization threshold, the current collaborative health status is considered risky, triggering an optimization mechanism: Based on the operation object ID and operation object region attributes of the collaborative operation data, operation pair grouping is optimized, and the conflict threshold is reduced. Optimized operation pair grouping avoids excessive concentration of operations in high-conflict areas by re-clustering and partitioning, reducing overall pressure. Lowering the conflict threshold enhances sensitivity to potential risks, enabling early warning and proactive intervention. Areas in media content where the number of edits exceeds a frequency threshold are restricted, and collaborative design permissions and concurrent editing counts are temporarily adjusted. For frequently edited hotspot areas, automatic locking and downgrading of concurrent editing permissions can be implemented to prevent the continuous accumulation of editing conflicts in local areas, ensuring global content consistency and local stability. Pop-up windows push warnings of high-risk conflicts and collaborative risks to relevant users and management, along with staggered collaboration prompts, ensuring users are aware of high-risk areas and operational suggestions immediately, allowing them to take preventative measures and optimize the overall collaborative design rhythm. A content feature vector matching algorithm based on the Transformer architecture is used to efficiently model the feature representations and contextual relationships between different editing versions, enabling intelligent cross-user and cross-version fusion and high-quality automatic synthesis of complex image content. This improves the intelligence level of version management and content consistency, analyzes the similarity of cross-user multi-version content for operations requiring merging, and makes optimization merging decisions, outputting fusion suggestions and automatic synthesis results. Simultaneously, reinforcement learning algorithms are periodically used to automatically learn optimal parameter configuration strategies by analyzing long-term accumulated historical collaborative processing total evaluation values ​​and actual operation fusion judgment results. Conflict thresholds, fusion thresholds, and optimization thresholds are optimized and adjusted. Specifically, within each optimization cycle, using the current threshold configuration and historical data as input, parameters are automatically fine-tuned based on the improvement or decline of the platform's collaborative effect, with an improvement in evaluation value serving as a positive feedback signal. Through multiple iterations, adaptive convergence of threshold parameters is achieved, dynamically maintaining the health and optimal performance of the collaborative design platform. For example, the current conflict threshold is 0.7, the fusion threshold is 0.6, and the optimization threshold is 0.75. Through reinforcement learning, the platform discovers that when the conflict threshold is lowered to 0.65, automatic conflict detection becomes more sensitive, the branching rate decreases, and the overall collaborative processing evaluation value increases by 0.05. The platform then automatically solidifies this new parameter and continuously optimizes it online, ultimately ensuring that all thresholds adaptively stabilize within the optimal range, maintaining a consistently healthy platform.All threshold adjustments and evaluation processes are automatically archived to ensure that the optimization process is traceable and controllable, thereby improving the intelligence level and business robustness of the collaborative design platform.

[0048] This implementation scheme achieves intelligent identification and tiered response to health status of the collaborative design platform by real-time monitoring of the overall collaborative processing evaluation value and dynamic comparison and optimization thresholds. When the health status is good, it ensures high efficiency and resource elasticity, improving user experience; when the health status is at risk, it automatically triggers optimization mechanisms, effectively suppressing editing risks in high-conflict areas through multiple measures such as group reconstruction, hotspot area permission adjustment, and pop-up warning pushes. The introduction of Transformer-based content feature vector matching and reinforcement learning algorithms enables high-quality intelligent fusion of multiple versions of content and adaptive threshold optimization, providing continuous evolution capabilities. Overall, it significantly improves the operational efficiency, content consistency, and intelligence level of the collaborative design platform, ensuring highly reliable collaboration and efficient content management in large-scale complex scenarios.

[0049] Reference Figure 2 As shown, the second aspect of the present invention provides an intelligent lightweight collaborative design system for digital media content, applied to the aforementioned intelligent lightweight collaborative design method for digital media content, comprising: a collaborative operation data acquisition and preprocessing module, used to acquire collaborative operation data in real time, preprocess the collaborative operation data, and standardize and store it; a collaborative intent recognition and conflict perception module, used to analyze the collaborative operation data of multiple users, group them into operation pairs, quantify the operation conflicts of each operation pair, and filter out conflicting operation pairs based on the operation conflict quantification results; a semantic conflict reconciliation and fusion judgment module, used to generate content feature vectors after each operation for conflicting operation pairs through collaborative operation data, fuse collaborative operation data and content feature vectors to evaluate the fusion similarity of operation pairs, and perform merging and branching processing of operation pairs based on the fusion similarity; and a collaborative feedback and optimization module, used to comprehensively evaluate collaborative processing by combining the operation conflict quantification results, fusion similarity, and operation pair merging and branching processing results, identify the collaborative health status, and adjust collaborative strategies, allocate resources, and optimize fusion schemes.

[0050] This implementation scheme achieves efficient perception, refined analysis, and intelligent feedback throughout the entire process of multi-user collaborative image design through a multi-layered modular architecture. The collaborative operation data acquisition and preprocessing module ensures the comprehensiveness and standardization of raw data, providing a solid foundation for subsequent analysis; the collaborative intent recognition and conflict perception module can promptly and accurately identify actual conflicts in multi-user collaboration, improving the real-time risk identification capability of collaborative design; the semantic conflict reconciliation and fusion judgment module, relying on content feature vectors and fusion similarity quantification, achieves intelligent reconciliation of complex conflicts and high-quality content synthesis; and the collaborative feedback and optimization module, through comprehensive evaluation and dynamic adjustment, continuously optimizes collaborative strategies and resource allocation, significantly improving the platform's management efficiency, editing consistency, and intelligence level in large-scale complex scenarios.

[0051] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0052] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. As those skilled in the art will understand, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A smart, lightweight collaborative design method for digital media content, characterized in that, Includes the following steps: S1 collects collaborative operation data in real time, performs data preprocessing on the collaborative operation data, and stores it in a standardized manner; S2, analyze the collaborative operation data of multiple users, group them into operation pairs, quantify the operation conflicts of each operation pair, and filter out conflicting operation pairs based on the operation conflict quantification results; S3, for conflicting operation pairs, generates content feature vectors after each operation through collaborative operation data, evaluates the fusion similarity of operation pairs by fusing collaborative operation data and content feature vectors, and performs merging and branching processing of operation pairs based on fusion similarity; S4 combines the results of operation conflict quantification, fusion similarity, and operation pair merging and branching to comprehensively evaluate collaborative processing, identify collaborative health status, and adjust collaborative strategies, allocate resources, and optimize fusion schemes.

2. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process of real-time acquisition of collaborative operation data, data preprocessing of collaborative operation data, and standardized storage is as follows: Real-time acquisition of collaborative operation data specifically involves: real-time monitoring of all user operation events on media objects, collecting timestamps, user IDs, operation types, and operation object IDs; simultaneously collecting global attributes, regional attributes, and layer structure information of the operation object; where the regional attributes include the image space coordinates involved in the operation; synchronously collecting pixel values ​​before and after the operation within the area affected by each operation, calculating the difference in pixel values ​​before and after each operation, obtaining the average absolute change, and normalizing it as the content intensity; The collaborative operation data collected from multiple terminals is standardized and transformed according to a unified data structure format, and a globally unique identifier is assigned to each piece of collaborative operation data; continuous operations of the same user within a fixed sliding time window are aggregated, and duplicate and invalid collaborative operation data are removed; based on a unified operation timestamp and logical order, global time sequence synchronization and deduplication of multi-terminal collaborative operation data are performed to adjust the operation disorder caused by clock errors and network latency; During time synchronization, concurrent conflicts between multiple users on the same operation object are detected and marked. Establish a media collaboration database and store the pre-processed collaborative operation data into the media collaboration database.

3. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process of analyzing multi-user collaborative operation data and grouping it into operation pairs is as follows: Retrieve collaborative operation data for each operation event; using the operation object ID as the primary key, group all collaborative operations for the same operation object within a fixed sliding time window according to time order and spatial overlap, specifically satisfying: the time window interval of the operation is less than the time threshold and the regional attributes of the operation objects have intersection; for each group of collaborative operations, traverse and generate all possible pairwise operation pairs, and record the number of operation pairs.

4. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process of quantifying the operational conflicts of each operation pair is as follows: For each pair of operations, extract the operation types of the two operations and compare them to obtain the operation type difference value. If the operation types are the same, the operation type difference value is assigned to 0; otherwise, it is assigned to 1. Extract the operation object region attributes of the two operations, calculate the intersection and union of the two operation object region attributes, and divide the intersection by the union to obtain the operation region overlap. Extract the layer structure of the operation objects of the two operations and compare them to obtain the layer consistency value. If the layer structure of the operation objects is the same, the layer consistency value is assigned to 1, otherwise it is assigned to 0. Extract the content intensity of the two operations, and divide the smaller content intensity by the larger content intensity to obtain the normalized ratio of the operation intensity. The conflict degree of the operation pair is obtained by multiplying the operation type difference value, operation area overlap, hierarchical consistency value, and operation intensity normalization ratio. Based on the number of operation pairs, the conflict degree values ​​of all operation pairs are summed and averaged to obtain the collaborative operation conflict degree value.

5. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process for selecting conflicting operation pairs based on the operation conflict quantification results is as follows: When multiple users perform collaborative operations, the collaborative operation conflict degree value is calculated in real time. The collaborative operation data, operation pairs, conflict degree values ​​of operation pairs, and collaborative operation conflict degree values ​​in each window are written into the media collaboration database. Collaborative operations with a conflict degree value greater than the conflict threshold are filtered out, marked as conflict operations, and enter the semantic conflict reconciliation process.

6. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process for evaluating the fusion similarity of operation pairs by generating content feature vectors after each operation using collaborative operation data and fusing collaborative operation data and content feature vectors is as follows: Receive the collaborative operation data, operation pair, conflict degree value of operation pair and collaborative operation conflict degree value corresponding to the conflict operation, and execute the semantic conflict reconciliation process: For the operation pair in each conflict operation, obtain the intersection of the regional attributes of the two operation objects as the overlapping region. For the operation object ID and the regional attributes of the operation object, extract the regional pixel histogram in the overlapping region as the content feature vector, and calculate the cosine similarity of the content feature vectors of the two operations. Obtain the normalized ratio of operation intensity, overlap of operation regions, and hierarchical consistency value of operation pairs in each conflict operation; multiply the normalized ratio of operation intensity, overlap of operation regions, hierarchical consistency value, and cosine similarity of content feature vectors of operation pairs to obtain the operation fusion similarity value.

7. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process of merging and branching operation pairs based on fusion similarity is as follows: Calculate the operation fusion similarity value of all operation pairs in the conflict operation and compare it with the fusion threshold to determine the operation fusion: when the operation fusion similarity value is greater than or equal to the fusion threshold, the current operation pair is determined to be similar, the operation results are directly merged, all operation content, content feature vector and collaborative operation data are integrated, and a unique synthesized content version is output. When the operation fusion similarity value is less than the fusion threshold, it is determined that the current operation pair is different. Branches are created for different operation paths of the operation pair, and independent version IDs are generated for each. Multiple operation results are saved by version to fully preserve the operation intent of each user. Write the operation fusion similarity value and operation fusion determination result of each operation pair into the media collaboration database.

8. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process for comprehensively evaluating collaborative processing by combining the results of operation conflict quantification, fusion similarity, and the results of operation pair merging and branching is as follows: Based on the sliding time window, the number of operation pairs in conflicting operations is counted, and the conflict degree value, operation fusion similarity value, and operation fusion judgment result of each operation pair are obtained. The conflict handling value is obtained by analyzing the operation fusion judgment result of the operation pair. If the operation fusion judgment result is merging, the conflict handling value is assigned to 1; if the operation fusion judgment result is branching, the conflict handling value is assigned to 0.

5. The conflict inverse value is obtained by subtracting the conflict degree of the operation pair from the constant. For each operation pair in a conflicting operation, the conflict inverse value, the operation fusion similarity value, and the conflict handling value are added together to obtain a single-group collaborative processing evaluation value. Based on the number of operation pairs in a conflicting operation, the single-group collaborative processing evaluation values ​​of all operation pairs are summed and divided by three times the number of operation pairs to obtain the total collaborative processing evaluation value.

9. The intelligent lightweight collaborative design method for digital media content according to claim 1, characterized in that, The specific process of identifying the collaborative health status and adjusting the collaborative strategy, allocating resources, and optimizing the fusion scheme is as follows: Periodically calculate the total collaborative processing evaluation value and compare it with the optimization threshold; when the total collaborative processing evaluation value is greater than or equal to the optimization threshold, the current collaborative health status is determined to be good, the existing strategies and parameter settings are maintained, concurrent collaboration and elastic resource allocation are allowed, and collaborative design permissions are continuously opened. When the total collaborative processing evaluation value is less than the optimization threshold, the current collaborative health status is deemed to be at risk, triggering an optimization mechanism: Based on the operation object ID and operation object region attribute of the collaborative operation data, the operation pair grouping is optimized, and the conflict threshold is lowered; regions in media content where the number of edits exceeds the frequency threshold are restricted, and collaborative design permissions and concurrent editing numbers are temporarily adjusted; high conflict incidence and collaborative risk warnings are pushed to relevant users and management terminals, along with off-peak collaboration prompts; a content feature vector matching algorithm based on the Transformer architecture is scheduled to analyze the similarity of cross-user multi-version content for the operation pairs to be merged and make optimization merging decisions, outputting fusion suggestions and automatic synthesis results; Meanwhile, reinforcement learning algorithms are used regularly to analyze the historical collaborative processing total evaluation value and the actual operation fusion judgment results, and to optimize and adjust the conflict threshold, fusion threshold and optimization threshold.

10. A smart, lightweight collaborative design system for digital media content, characterized in that, include: The collaborative operation data acquisition and preprocessing module is used to acquire collaborative operation data in real time, preprocess the collaborative operation data, and store it in a standardized manner. The collaborative intent recognition and conflict perception module is used to analyze the collaborative operation data of multiple users, group them into operation pairs, quantify the operation conflicts of each operation pair, and filter out conflicting operation pairs based on the operation conflict quantification results. The semantic conflict reconciliation and fusion determination module is used to generate content feature vectors after each operation for conflicting operation pairs through collaborative operation data, fuse collaborative operation data and content feature vectors to evaluate the fusion similarity of operation pairs, and perform merging and branching processing of operation pairs based on fusion similarity. The collaborative feedback and optimization module is used to comprehensively evaluate collaborative processing by combining the quantitative results of operational conflict, fusion similarity, and the results of merging and branching of operation pairs, identify the collaborative health status, and adjust collaborative strategies, allocate resources, and optimize fusion schemes.

Images