An animation resource management system based on segment tagging

Abstract

The application relates to the technical field of resource management, in particular to an animation resource management system based on section tagging, wherein the AC-3 algorithm is applied to the constraint set to perform arc consistency checking on adjacent sections, sections without continuous connection capability are removed according to connection restrictions, sections with too short or too long time lengths are removed according to frame length intervals, sections with coding conflicts are removed according to condition restrictions, the remaining candidate quantity is adjusted to adjust the placement order and the back-off position, the behavior consistent nodes are folded into a single node combination through the Brzozowski algorithm, the folded path is reordered and spliced, a plurality of section paths originally needing to be maintained separately are compressed into a small-scale node structure, the candidate state quantity and the storage cost are reduced under the same animation resource scale, the tag matching link and the path traversal length are shortened, the conflict processing times are reduced, and the sequence generation stability and the response speed in the section combination and the reuse link are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of resource management, and in particular, to an animation resource management system based on section tagging. Background Art

[0002] The technical field of resource management aims to perform storage control, index construction, retrieval rule setting, scheduling strategy formulation on multiple digital resources through a structured organization method, and form a reusable resource set, establish a resource management system with a determinable access path, quantifiable retrieval efficiency, and controllable call process, so that resources have positionability, combinability, and schedulability in multi-scene usage, reduce the time-consuming of resource search, reduce the number of repeated constructions, and ensure that the resources can stably output the target data sequence or processing result according to the set rules during the call stage.

[0003] An animation resource management system based on section tagging aims to construct a management system that can perform section-level splitting, tagging annotation, indexed retrieval, and serial combination on animation resources, so that animation resources have positionability, quantifiable retrieval efficiency, and can perform deterministic combination according to the tag rules during the editing, call, and reuse processes. By establishing action category codes, timing position codes, and application condition codes for multiple animation sections, the sections can form a structured index set in the library, reduce the number of repeated productions, and ensure that the animation quickly generates the target section sequence according to the tag matching rules during the call stage.

[0004] In the prior art, the technical field of resource management generally focuses on resource storage control, index construction, retrieval rule setting, and scheduling strategy formulation. Although it can provide a structured resource set and a determinable access path, when it comes to the management of animation sections, it mostly focuses on the static organization of resource numbers, retrieval keys, and call strategies. The association relationships between sections mostly rely on simple serial numbers, single-tag matching, or preset scene scripts, lacking the联动 control of action categories, timing positions, and application conditions on the same sequence dimension. When there are multiple groups of sections with similar actions but different conditions in the library, single-tag or a small number of tag matches return a relatively wide range of section sets. During multi-scene calls, frame length mismatches, action category jumps, or usage condition conflicts between adjacent sections often require repeated adjustment of the time axis, replacement of materials, and modification of the call order, increasing the editing and scheduling burden. The number of state nodes in the resource library increases linearly or even super-linearly with the number of sections. Each time a combination is made, matching and checking are required on a large number of candidate sections. When there are shared resources in different scenes, the same logical conflicts are repeatedly exposed in multiple scenes. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and propose an animation resource management system based on section tagging.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an animation resource management system based on segment tagging includes: Tag parsing module: Based on segment tags, extract action category numbers in sequence, arrange temporal position numbers in frame order, match application condition numbers with occupancy identifiers item by item, and integrate numbers according to sequence position to form a basic tag set for the segment; Constraint Construction Module: Based on the segment basic tag set, compare the action category number with adjacent records one by one and organize the adjacent numbers, compare the frame length value with the interval limit and organize the frame length segment, compare the application condition number with the limit list one by one and organize the condition number, and process the three types of numbers in sequence to form a segment constraint item set; Segment Combination Module: Based on the segment constraint entry set, the AC-3 algorithm is used to remove inapplicable segments according to the connection constraints, filter out out-of-range segments according to the frame length interval, exclude conflicting segments according to the condition constraints, and adjust the placement order and back-off position according to the number of candidates to form a segment splicing sequence set. Sequence arrangement module: Based on the segment splicing sequence set, split mutually exclusive label groups, merge non-conflicting labels in sequence order, and extract each group in sequence according to the in-degree record and adjust the in-degree of subsequent groups to obtain the label sequence structure set; State compression module: Based on the label sequence structure set, label nodes are established in sequence, the differences in node behavior are identified according to the transformation record, nodes with different behaviors are grouped, and the Brzozowski algorithm is used to fold nodes with the same behavior into a single structure, while reorganizing the overall sequence to form a compressed segment structure set.

[0007] As a further embodiment of the present invention, the segment basic tag set includes an action category number group, a timing position number group, and an application condition number group; the segment constraint entry set includes an action connection restriction group, a frame length restriction group, and a condition restriction group; the segment splicing sequence set includes a segment sequence group, a sequence placement record, and a backtracking position group; the tag order structure set includes a tag sequence group, an in-degree record group, and a conflict splitting group; and the compressed segment structure set includes a folded node group, a node relationship group, and a sequence structure group.

[0008] As a further aspect of the present invention, the tag parsing module includes: Number Extraction Submodule: Based on segment labels, read action category numbers and arrange them in order, extract timing position numbers item by item and complete frame sequence labeling, check application condition numbers and occupancy identifiers item by item and complete number recording, and generate label number groups; Numbering Integration Submodule: Based on the label numbering group, align the action category numbering records and the time sequence position numbering records one by one and complete the numbering arrangement; arrange the application condition numbering records and the sequence position one by one and complete the numbering parallel; merge the three types of numbering records one by one and complete the sequence integration, and generate the segment basic label set.

[0009] As a further aspect of the present invention, the constraint construction module includes: Restriction comparison submodule: Based on the segment basic tag set, compare the action category number record with the adjacent record and complete the adjacent arrangement, compare the frame length value with the interval restriction and complete the segment value arrangement, compare the application condition number record with the restriction list and complete the condition placement, and generate a restriction record group; The constraint summary submodule: Based on the constraint record group, it organizes adjacent numbered records and frame length segment records and completes the numbering collection, classifies condition numbered records and sequence positions and completes the numbering collection, lists the three types of records one by one and completes the sequence compilation, and generates a segment constraint item set.

[0010] As a further aspect of the present invention, the three types of records are listed side by side and the sequence is completed. The adjacent numbered records, frame length segment records, and condition numbered records in the same sequence position are arranged in a fixed order to form a combined record. Then, all combined records are rearranged into a continuous structure according to the sequence position, so that the action restrictions, frame length restrictions, and condition restrictions corresponding to each position of the segment can be presented in a unified manner, forming a complete and usable serialized restriction structure.

[0011] As a further aspect of the present invention, the segment combination module includes: The connection filtering submodule: Based on the set of segment constraint entries, the AC-3 algorithm is used to read the action category number of each segment and determine the segments that can be retained by comparing with the adjacent number groups. Segments without adjacent numbers are removed from the sequence, and the remaining segments are rearranged to the corresponding sequence positions to generate a connection sequence group. Frame length filtering submodule: Based on the connection sequence group, collect the frame length value of each segment and confirm the interval correspondence with the frame length segment group. Remove the segments outside the interval and reorder them. Re-merge the segments within the interval into a continuous sequence to generate a frame length sequence group. Condition filtering submodule: Based on the frame length sequence group, read the application condition number of each segment and screen the source of conflict by comparing with the condition number group. After removing the conflicting segments, rearrange the placement order according to the number of remaining segments, and record the necessary back-off positions to generate a segment splicing sequence set.

[0012] As a further aspect of the present invention, the AC-3 algorithm uses the adjacent positions of a segment in the sequence as arcs, takes the action category numbers that each segment can take as a set of values, takes any arc in turn, and performs adjacent matching checks on the value sets of the segments at both ends of the arc. Values ​​that do not have a valid adjacent number relationship are deleted from the set. Then, the end that was deleted is added back to the list to be checked with the remaining adjacent positions and the comparison is continued. The arc extraction, adjacent check and value deletion operations are repeated until no new deletion items are generated for all arcs, resulting in a set of values ​​that only retains the value set that can form a legal connection relationship with the adjacent segments.

[0013] As a further aspect of the present invention, the sequence arrangement module includes: Exclusion splitting submodule: Based on the segment splicing sequence set, it identifies mutually exclusive label pairs in the label group and completes the splitting of exclusion items, sorts out non-exclusive labels and completes the sequence order marking, associates the positional relationship between exclusion items and non-exclusive items, completes the sequence distribution record within the group, and generates the label distribution group; Sequence straightening submodule: Based on the label distribution group, extract non-rejection labels and complete sequence merging, extract the in-degree records of each label group and complete the in-degree order arrangement, extract the label groups according to the in-degree order and complete the in-degree adjustment of subsequent groups, and generate a label order structure set.

[0014] As a further aspect of the present invention, the state compression module includes: Node generation submodule: Based on the tag sequence structure set, read the tag number one by one and establish the corresponding node, register the sequence relationship of adjacent tags as node connection, and then gather all nodes into a node mapping structure according to the reading order to generate tag node group; Node partitioning submodule: Based on the labeled node group, extract the record content between nodes and compare the behavioral differences of each node. Nodes with obvious differences are grouped into independent sets, and nodes with consistent behaviors are grouped into the same set. These sets are then formed into a collapsible structure to generate node collapsible groups. Sequence Recombination Submodule: Based on the node folding group, the Brzozowski algorithm is used to read the connection relationship of the folded nodes and connect each node in the order of connection. The connected structure is reordered into a continuous path, and all node structures are combined into a complete sequence set to generate a compressed segment structure set.

[0015] As a further aspect of the present invention, the Brzozowski algorithm, based on the connection relationships recorded in the node folding group, performs a complete reversal of the original node structure according to the connection direction and regenerates the reachability relationships between nodes in the reversed structure. Then, a new node set is constructed using the reversed node structure as input and the corresponding connection path is generated. Subsequently, the reversal operation is performed again on the new path structure to restore the connection direction, and the node set and connection order are regenerated on the restored structure. Through the loop of two reversals and two structure reconstructions, redundant nodes and redundant connections are eliminated, so that the remaining nodes and connection paths form the simplest continuous structure, resulting in a compressed segment structure set that can be used for the output sequence.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the action category number, timing position number, and application condition number in the segment label are extracted and integrated simultaneously. The three types of numbers are encoded into a unified label structure on the same sequence dimension, so that each segment has a multi-dimensional identifier that can be directly compared in the library. By comparing the action category number with adjacent records, the frame length value with the interval limit, and the application condition number with the limit list item by item, each segment is given clear connection constraints, duration constraints, and usage condition constraints before entering the combination stage. The constraint information is centrally recorded as an item set, providing a directly calculable judgment basis for the subsequent screening of inapplicable segments. In this invention, the AC-3 algorithm is applied to the constraint set to perform arc consistency checks on adjacent segments. Segments that do not have continuous connection capabilities are eliminated based on connection constraints. Segments with excessively short or long durations are excluded based on frame length intervals. Segments with conflicting encodings are removed based on condition constraints. The placement order and backoff position are adjusted according to the remaining number of candidates, so that the candidate segment sequence has both forward selection path and backoff path during the generation process. This is suitable for multi-round trial combination and rollback operations. After obtaining the candidate sequence, mutually exclusive label groups are split, and non-conflicting labels are merged according to sequence position. In-degree records are introduced to statistically analyze and sort the entry relationship of each label group. High-in-degree label groups obtain a more forward fixed position in sequence construction, reducing the number of subsequent structural adjustments. In this invention, nodes with consistent behavior are folded into a single node combination using the Brzozowski algorithm. The folded paths are then reordered and spliced, compressing multiple segment paths that originally needed to be maintained separately into a smaller node structure. This reduces the number of candidate states and storage overhead under the same animation resource scale, shortens the label matching link and path traversal length, reduces the number of conflict handling operations, and improves the stability and response speed of sequence generation in the segment combination and reuse process. Attached Figure Description

[0017] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Example

[0019] Please see Figure 1 This invention provides a technical solution: an animation resource management system based on segment tagging, comprising: Tag parsing module: Based on segment tags, extract action category numbers in sequence, arrange temporal position numbers in frame order, match application condition numbers with occupancy identifiers item by item, and integrate numbers according to sequence position to form a basic tag set for the segment; Constraint Construction Module: Based on the segment basic tag set, it compares the action category number with adjacent records one by one and organizes the adjacent numbers, compares the frame length value with the interval limit and organizes the frame length segment, compares the application condition number with the limit list one by one and organizes the condition number, and processes the three types of numbers in sequence to form a segment constraint item set; Segment Combination Module: Based on the segment constraint entry set, the AC-3 algorithm is used to remove inapplicable segments according to the connection constraints, filter out out-of-range segments according to the frame length interval, exclude conflicting segments according to the condition constraints, and adjust the placement order and back-off position according to the number of candidates to form a segment splicing sequence set. Sequence arrangement module: Based on the segment splicing sequence set, split mutually exclusive label groups, merge non-conflicting labels in sequence order, and extract each group in sequence according to the in-degree record and adjust the in-degree of subsequent groups to obtain the label order structure set; State compression module: Based on the label sequence structure set, label nodes are established in sequence. The differences in node behavior are identified according to the transformation record. Nodes with different behaviors are grouped. The Brzozowski algorithm is used to fold nodes with the same behavior into a single structure. At the same time, the overall sequence is reorganized to form a compressed segment structure set.

[0020] The segment basic label set includes action category number group, timing position number group, and application condition number group; the segment constraint entry set includes action connection restriction group, frame length restriction group, and condition restriction group; the segment splicing sequence set includes segment sequence group, sequence placement record, and backtracking position group; the label order structure set includes label sequence group, in-degree record group, and conflict splitting group; and the compressed segment structure set includes folded node group, node relationship group, and sequence structure group.

[0021] The tag parsing module includes: Number Extraction Submodule: Based on segment labels, read action category numbers and arrange them in order, extract timing position numbers item by item and complete frame sequence labeling, check application condition numbers and occupancy identifiers item by item and complete number recording, and generate label number groups; Numbering Integration Submodule: Based on the tag numbering group, it aligns the action category numbering records with the time sequence position numbering records one by one and completes the numbering arrangement; it arranges the application condition numbering records with the sequence position one by one and completes the numbering parallelism; it merges the three types of numbering records one by one and completes the sequence integration, generating the segment basic tag set; Based on segment tags, a dynamic programming sequence parsing algorithm is used to read and process action category numbers. A fixed command set is used as the execution unit, including the INIT, LOAD, SCAN, and WRITE commands. Parameters include a start pointer of 0, a step size of 1, a maximum index parameter limited by the number of characters in the segment tag, and a buffer length of 64. During execution, the INIT command initializes the parsing buffer and sets all buffer units to 0. The LOAD command loads the character numbers from the segment tag into the corresponding index positions in the buffer with a step size of 1. The SCAN command compares the numbered characters with a fixed scanning step and records the action category number upon successful comparison. The WRITE command writes the recorded number into the action category number recording unit and arranges them in index order. The temporal position numbers contained in the segment tags are extracted and frame sequence labeling is completed using a finite state machine. The index positioning algorithm is executed. The algorithm includes states S0, S1, and S2, and condition codes A and B. During execution, in state S0, a pointer is used to detect whether the character is a number. If the number condition is met, the algorithm transitions to state S1 and records the starting index value. In state S1, consecutive numbers are read with a step size of 1. If a non-numeric character appears, the algorithm transitions to state S2 and records the ending index value. In state S2, a write command is used to write the recorded index to the time sequence position number recording unit and the algorithm returns to state S0 to continue scanning. The segment condition comparison algorithm is used to verify the application condition number and the occupancy identifier. It includes a fixed comparison matrix M. The rows of the matrix index correspond to the application condition number, and the columns of the matrix index correspond to the occupancy identifier. The matrix unit value is fixed to 0 or 1. During execution, the application condition number is loaded by a pointer and the corresponding occupancy identifier unit value is directly read by the matrix index and written to the number recording unit. The numbered records are arranged in the scanning order to generate a tag number group. Based on the tag number group, a three-way number alignment and merging algorithm is used to align action category number records and timing position number records. The command sequences ALIGN, SHIFT, and MERGE are used as execution units, and parameter group P is used as control conditions. Parameter group P includes the action category number length L1, the timing position number length L2, and the buffer sequence length L3. When the ALIGN command is executed, the index pointer synchronously reads and writes the alignment buffer sequence index position from 0 to the minimum value of L1 and L2. When the SHIFT command is executed, all timing position numbers are uniformly shifted with an offset of 2. When the MERGE command is executed, the alignment buffer sequence is used as the input source to sequentially write the action category number and timing position number into the merging sequence unit. Finally, a conditional numbering sequence arrangement algorithm is used to apply the conditional numbering. The records are processed by LOOP, STORE, and INC commands. The parameter K represents the number of application condition numbers, and the parameter S represents the sequence start position 0. When the LOOP command is executed, the index is traversed from 0 to K. When the STORE command is executed, the application condition number is written to the sequence position S. When the INC command is executed, S is incremented by 1 and written back to the control unit to form a parallel sequence. The number sequence integration algorithm is used to merge the three types of number records. A fixed-length sequence table T is used as the carrying sequence, and the TRAVERSE command is used to load the action category number record, the timing position number record, and the application condition number record in sequence. The PUT command is used to write the numbers to the continuous cells of T. The END command is used to mark the completion of the sequence table writing and output the segment basic label set to generate the segment basic label set.

[0022] The constraint building module includes: Restriction Comparison Submodule: Based on the segment basic tag set, it compares the action category number record with adjacent records and completes the adjacent arrangement, compares the frame length value with the interval restriction and completes the segment value arrangement, compares the application condition number record with the restriction list and completes the condition placement, and generates a restriction record group. The constraint summary submodule: Based on the constraint record group, it organizes adjacent numbered records and frame length segment records and completes the numbering collection, classifies condition numbered records and sequence positions and completes the numbering collection, lists the three types of records item by item and completes the sequence compilation, and generates a set of segment constraint entries; Based on the segment-based tag set, an adjacency number comparison and segment organization algorithm is used to compare action category number records with adjacent records. The command sequence LOAD, COMPARE, and WRITE are used as execution steps, with parameters including a read pointer of 0, a step size of 1, and a record buffer length of 128. When the LOAD command is executed, it loads action category number records from the segment-based tag set item by item, starting from pointer 0, and writes them to the comparison buffer. When the COMPARE command is executed, it performs a fixed-length character comparison between the current number and the next number with a step size of 1, and writes the comparison result to the adjacency state unit. When the WRITE command is executed, it writes the adjacency state unit and the action category number record into the organization sequence unit and arranges them in order. A frame length interval verification algorithm is then used to compare the frame length value with the interval limit, with the command sequence READ, CHECK, and STORE being used as execution steps. Parameters include a lower interval limit of 1 and an upper interval limit of 1. 999. The frame length buffer length is 64. When the READ command is executed, the frame length value is loaded into the frame length buffer unit. When the CHECK command is executed, the frame length value is compared with the upper and lower limits of the interval based on fixed conditions, and the check result is written into the segment value status unit. When the STORE command is executed, the segment value status unit and the frame length value are written into the segment value sorting sequence. The application condition number matching algorithm is used to compare the application condition number record with the restriction list. The command sequence FETCH command, MATCH command, and PUT command are used as execution steps. The parameters include the restriction list length M, the condition number length N, and the initial value of the matching cursor 0. When the FETCH command is executed, the cursor is used to load the target number from the application condition number record item by item. When the MATCH command is executed, the cursor value is used to perform item-by-item character comparison in the restriction list and write the matching status into the matching unit. When the PUT command is executed, the content of the matching unit is written into the condition sorting sequence and arranged according to the record order to generate a restriction record group. Based on the restricted record group, an adjacent numbered record and frame length segment record merging algorithm is used to organize adjacent numbered records and frame length segment records. The main steps are the command sequence SCAN, GROUP, and MERGE. Parameters include the numbered record length X, frame length segment record length Y, and merge buffer length Z. During the SCAN command execution, the numbered records and frame length segment records in the restricted record group are scanned with a step size of 1 and written to the scan buffer. During the GROUP command execution, numbered records and frame length segment records at the same index position are written to the merging unit according to a fixed merging rule. During the MERGE command execution, the merging unit is written sequentially into the numbered merging sequence. A conditional numbering sequence classification algorithm is then used to classify conditional numbered records and sequence positions. The execution steps are the command sequence ITER, CLASS, and APPEND. Parameters include the total number of conditional numbers K, the sequence length K, and the number of conditional numbers K. The position reference value is 0. When the ITER command is executed, condition number records are loaded with indices from 0 to K. When the CLASS command is executed, the loaded number records are written into category units according to fixed classification rules. When the APPEND command is executed, the category units are written into the number set arrangement sequence and arranged in index order. The three types of records are processed item by item in parallel using a three-class record parallel arrangement algorithm. It includes a sequence mapping table R. The index positions in the table include the adjacent number record sequence position, the frame length segment record sequence position, and the condition number record sequence position. During the execution, the LOAD3 command is used to write the three types of records into a three-slot buffer. The ORDER command arranges the contents of the three-slot buffer into a combined record in a fixed order. The WRITE3 command writes the combined record into the corresponding index position of the continuous sequence structure. The ENDSEQ command marks the end of writing for all indices and generates a segment constraint entry set.

[0023] The three types of records are listed side by side and the sequence is completed. Adjacent numbered records, frame length segment records, and condition numbered records in the same sequence position are arranged into a combined record in a fixed order. Then, all combined records are rearranged into a continuous structure according to the sequence position, so that the action restrictions, frame length restrictions, and condition restrictions corresponding to each position of the segment can be presented in a unified manner, forming a complete and usable serialized restriction structure.

[0024] The section combination module includes: The connection filtering submodule: Based on the segment constraint item set, the AC-3 algorithm is used to read the action category number of each segment and determine the segments that can be retained by comparing with the adjacent number groups. Segments without adjacent numbers are removed from the sequence, and the remaining segments are rearranged to the corresponding sequence positions to generate a connection sequence group. Frame length filtering submodule: Based on the connection sequence group, collect the frame length value of each segment and confirm the interval correspondence with the frame length segment group. After removing the segments outside the interval, reorder them and merge the segments within the interval into a continuous sequence to generate a frame length sequence group. Condition filtering submodule: Based on frame length sequence group, read the application condition number of each segment and screen the source of conflict by comparing with the condition number group. After removing the conflicting segments, rearrange the placement order according to the number of remaining segments, and record the necessary back-off positions to generate segment splicing sequence set. Based on the segment constraint entry set, the AC-3 algorithm is used to read and process the action category numbers of each segment. The execution steps are the command sequence INIT, ENQUEUE, DEQUEUE, CHECK, REDUCE, and WRITE. Parameters include the maximum arc queue length of 256, the maximum number of segment numbers of 128, the initial index value of 0, the scan step size of 1, and the record buffer length of 64. When the INIT command is executed, the arc queue is initialized and all arc queue cells are set to integer values ​​of 0. When the ENQUEUE command is executed, the segment constraint entry set is... Adjacent segments are pushed into the arc queue in the form of segment indexes. When the DEQUEUE command is executed, a segment arc is taken out from the arc queue and written into the active arc unit. When the CHECK command is executed, the action category numbers of the segments at both ends of the active arc are read with a step size of 1, and the character sequence is compared and the comparison result is written into the check unit. When the REDUCE command is executed, the segment indexes that are determined to have no adjacent numbers in the check unit are written into the remove unit. When the WRITE command is executed, the segment numbers that have not been written into the remove unit are written into the reserved segment sequence and rearranged into the corresponding sequence positions according to the original index order to generate the connecting sequence group. Based on the concatenation sequence group, a frame length interval filtering algorithm is used to read and process the frame length value of each segment. The execution steps are the command sequence LOAD, COMPARE, REMOVE, REORDER, and MERGE. The parameters include a minimum frame length of 1, a maximum frame length of 999, an upper limit of the number of segments of 128, an array step size of 1, and a frame length record buffer length of 64. When the LOAD command is executed, the segment frame length value is loaded from the concatenation sequence group with a step size of 1 and written to the frame length buffer unit. When the COMPARE command is executed, the content of the frame length buffer unit is compared with the given lower and upper limits of the frame length segment group and the comparison result is written to the interval status unit. When the REMOVE command is executed, the index of the segment marked as outside the interval in the interval status unit is written to the elimination unit and its index position is deleted from the cache structure. When the REORDER command is executed, the sequence position of the remaining segments is rearranged in a continuous index manner according to the number of remaining segments. When the MERGE command is executed, the rearranged segment set is written to the continuous sequence unit and combined into a continuous structure to generate a frame length sequence group. Based on frame length sequence groups, a condition numbering conflict screening algorithm is used to read and process the condition numbers applied to each segment. The execution steps are the command sequence READ, MATCH, FLAG, REMOVE, RELOC, and RECORD. The parameters include a maximum number of condition numbers (64), a maximum number of segments (128), a conflict record buffer length (32), an initial sequence position value of 0, and a step size of 1. When the READ command is executed, the segment application condition numbers are loaded with a step size of 1 and written to the condition check unit. When the MATCH command is executed, the contents of the condition check unit are compared with the condition number group item by item, and the condition identifiers with conflicting characters are written to the conflict unit. When the FLAG command is executed, the index of the segment marked by the conflict unit is written to the removal list. When the REMOVE command is executed, the segment in the removal list is removed from the sequence and its index position is deleted. When the RELOC command is executed, the sequence position is rewritten in a continuous order according to the number of remaining segments, and a new position mapping unit is generated. When the RECORD command is executed, the backtrack position in the position mapping unit is written to the backtrack record structure and arranged by index, generating a segment splicing sequence set.

[0025] The AC-3 algorithm uses the adjacent positions of a segment in the sequence as arcs. It takes the possible action category numbers for each segment as a set of values, sequentially extracts any arc, and performs an adjacency matching check on the value sets of the segments at both ends of the arc. Values ​​without a valid adjacent number relationship are deleted from the set. The deleted end is then added back to the check list with the remaining adjacent positions for further comparison. This process of arc extraction, adjacency checking, and value deletion is repeated until no new deletions are generated for any arcs, resulting in a set of values ​​that only have a valid connection with adjacent segments. The AC-3 algorithm follows the formula:

[0026] in: Indicates a section In the segment-based animation resource management system, the corresponding action category number. Indicates a section In the segment-based animation resource management system, the corresponding action category number. This indicates the index of the target segment being examined in the animation resource management process. Indicates the segment Segment indexes that have adjacent relationships in a time series or logical sequence. Indicates a section The set of adjacent segment indices in the animation resource sequence This represents the set of adjacency relationships for allowed action categories registered in the animation resource constraint library. This refers to the segment obtained from the historical tag sequence statistics in animation resource management. and section The adjacency reliability weighting coefficient, Indicates a section and section The sequence position proximity coefficient is calculated based on the frame difference on the animation timeline. This represents the positive smoothing constant used in calculating the stable adjacency ratio. Indicates a section The retention threshold coefficient, Indicates a section The retention confidence flag value under the adjacency calculation rule is used to guide the segment retention or deletion process in the animation resource management system; Execution process: First, read the segment indices in the segment index set sequentially. Extracting segments from segment label sequences Action category number At the same time, the section is determined. Adjacent segment index set and for the set Index of each adjacent segment Read their action category numbers sequentially Then, match the action category number to... Adjacency constraint set Matching is performed to determine the value of the indicator function, and then the adjacency reliability weighting coefficient is calculated. Based on the action category in the historical label sequence The frequency of occurrence is obtained through a normalization step, and then the sequence position proximity coefficient is calculated. , by section and section The difference in the starting frame index is obtained by applying an exponential decay function, which is used to construct the product of multiple terms in the numerator of the formula. The summation value is calculated, and the product of multiple terms in the denominator of the formula is constructed simultaneously. Sum of the values ​​and add the smoothing constant To ensure the ratio calculation is stable, the ratio of the numerator to the denominator is then calculated to obtain the segment. The adjacent matching ratio is then used to calculate the segment from the historical label sequence. The average matching ratio is combined with the adjustment coefficient to obtain the retention threshold. Finally, the adjacency matching ratio is compared with the threshold. If the ratio is greater than or equal to the given ratio, then... Then let Indicates the reserved section Otherwise Indicates the deletion of the segment. This ensures that the animation resource management system can stably filter reusable segments based on the adjacency constraints of segment labels.

[0027] The sorting module includes: Exclusion splitting submodule: Based on the segment splicing sequence set, it identifies mutually exclusive label pairs in the label group and completes the splitting of exclusion items, sorts out non-exclusive labels and completes the sequence order marking, associates the positional relationship between exclusion items and non-exclusive items, completes the sequence distribution record within the group, and generates the label distribution group; Sequence straightening submodule: Based on the label distribution group, extract non-rejection labels and complete sequence merging, extract the in-degree records of each label group and complete the in-degree order arrangement, extract the label groups according to the in-degree order and complete the in-degree adjustment of subsequent groups, and generate a label order structure set. Based on the segment splicing sequence set, an exclusion tag identification and splitting algorithm is used to identify mutually exclusive tag pairs in the tag group. The command sequence INIT, LOAD, SCAN, CHECKPAIR, SPLIT, and WRITE are used as execution steps. Parameters include a maximum number of tags (128), a maximum number of exclusion pairs (64), a scan step size of 1, a tag buffer length of 64, and an exclusion record buffer length of 32. When the INIT command is executed, all exclusion record buffer entries are set to 0, and the scan pointer is initialized to 0. When the LOAD command is executed, tags from the segment splicing sequence set are written to the tag buffer unit item by item at a step size of 1. When the SCAN command is executed, the scanning pointer reads two adjacent tags in the tag buffer unit and writes them to the comparison register. When the CHECKPAIR command is executed, the two tags in the comparison register are compared character by character in the form of a character sequence and the comparison result is written to the rejection state unit. When the SPLIT command is executed, the contents of the rejection state unit are read and the corresponding tag record is written to the rejection split sequence when the state value is 1 and the corresponding tag is written to the non-rejection sequence when the state value is 0. When the WRITE command is executed, the non-rejection sequence is indexed and the index positions of the rejection item and the non-rejection item are mapped and written to the sequence distribution record unit in the group to generate the tag distribution group. Based on the label distribution group, a non-rejection label merging algorithm is used to extract non-rejection labels and write them into the merge sequence. The command sequence used is FETCH, GROUP, and APPEND. Parameters include the number of non-rejection labels N, the initial value of the group mapping index (0), the step size (1), and the merge buffer length (64). When the FETCH command is executed, non-rejection labels are read from the label distribution group at a step size of 1 and written into the extraction unit. When the GROUP command is executed, the contents of the extraction unit are written into the merging unit according to the label group number given by the label distribution group. When the APPEND command is executed, the merging unit is written into the non-rejection label merge sequence and arranged in index order. An in-degree record ordering algorithm is then used to sort the in-degree records of each label group. The row extraction process uses the command sequence READ, RANK, SORT, and UPDATE. Parameters include a maximum of 64 in-degree records, a sorting buffer length of 64, and a comparison step size of 1. When the READ command is executed, it reads the in-degree records of each tag group from the tag distribution group and writes them to the in-degree buffer unit. When the RANK command is executed, it compares the values ​​of the in-degree buffer unit item by item with the comparison step size and writes the comparison results to the in-degree order unit. When the SORT command is executed, it sorts the tag group indexes according to the in-degree order unit and writes them to the sorting sequence. When the UPDATE command is executed, it writes the in-degree values ​​of the tag groups in descending order according to the sorting sequence and completes the subsequent group in-degree adjustment, generating a tag order structure set.

[0028] The state compression module includes: Node generation submodule: Based on the tag order structure set, it reads the tag number one by one and establishes the corresponding node, registers the order relationship of adjacent tags as node connection, and then gathers all nodes into a node mapping structure according to the reading order to generate tag node group; Node partitioning submodule: Based on the tag node group, extract the record content between nodes and compare the behavioral differences of each node. Nodes with obvious differences are grouped into independent sets, and nodes with consistent behaviors are grouped into the same set. These sets are then formed into a collapsible structure to generate node collapsible groups. Sequence Recombination Submodule: Based on node folding groups, using the Brzozowski algorithm, it reads the connection relationship of folded nodes and concatenates each node according to the connection order, reorders the concatenated structure into a continuous path, and combines all node structures into a complete sequence set to generate a compressed segment structure set. Based on the tag sequence structure set, a node construction and mapping algorithm is used to read the tag numbers one by one. This process uses the command sequence INIT, LOAD, NEWNODE, LINK, and WRITE as execution steps. Parameters include the maximum number of tags (128), the node array length (128), the connection matrix size (128×128), the initial value of the scan pointer (0), the step size (1), and the default value of empty cells (0). When the INIT command is executed, all cells in the node array are written to the default value of empty cells, all cells in the connection matrix are written to the default value of empty cells, and the scan pointer is set to 0. When the LOAD command is executed... During row execution, a single tag number is written from the tag sequence structure set to the tag cache unit with a step size of 1. When the NEWNODE command is executed, the contents of the tag cache unit are read and a corresponding node identifier is created in the node array at the current scan pointer position, and the node identifier value is written. When the LINK command is executed, the previous node identifier and the current node identifier are written to the corresponding row and column positions of the connection matrix and a fixed value of 1 is used to indicate that there is an ordered connection relationship. When the WRITE command is executed, the contents of the node array and the contents of the connection matrix are written to the node mapping structure according to the scan order, and a continuous node record structure is formed after all tag numbers are processed, generating a tag node group. Based on the tag node group, a node difference comparison partitioning algorithm is used to read and process the record content between nodes. This process uses the command sequence SCAN, COMPARE, SEPARATE, MERGE, and FOLD as execution steps. The parameters include the upper limit of the number of nodes (128), the node record length (64), the difference detection step size (1), the set cache length (64), and the initial value of the fold index (0). When the SCAN command is executed, the record content of two nodes is read at the difference detection step size and written to the comparison cache unit. When the COMPARE command is executed, the content of the comparison cache unit is compared character by character according to the character sequence and the difference result is written to the difference status unit. When the SEPARATE command is executed, when the value of the difference status unit is 1, the corresponding node identifier is written to the independent set and then to the set cache unit. When the MERGE command is executed, when the value of the difference status unit is 0, the node identifier is written to the consistent set and then to the set cache unit. When the FOLD command is executed, the independent set and the consistent set are combined in the order of the fold index and written to the fold structure array and then to the fold record unit to generate a node fold group. Based on node folding groups, the Brzozowski algorithm is used to read and process the connection relationships of folded nodes. This process uses the command sequence REVERSE, DETERM, REBUILD, ORDER, and COMBINE as execution steps. The parameters include the maximum number of folded nodes (128), the connection matrix size (128×128), the traversal step size (1), the path buffer length (128), and the sequence buffer length (128). When the REVERSE command is executed, the row and column positions of the folded node connection matrix are read according to the traversal step size and written into the inverted connection buffer structure. When the DETERM command is executed, the inverted connection buffer structure is read row by row and the connection content of each row is written into the deterministic structure unit. When the REBUILD command is executed, the deterministic structure unit is used as input and written into the reconstructed node sequence according to the row and column index order and written into the sequence reconstruction buffer. When the ORDER command is executed, the sequence reconstruction buffer is indexed and sorted and the sorting result is written into the continuous path unit. When the COMBINE command is executed, the content of the continuous path unit and the folded node information are merged according to the index and written into the complete sequence structure array to generate a compressed segment structure set.

[0029] The Brzozowski algorithm, based on the connection relationships recorded in the node folding group, reverses the original node structure in the connection direction and regenerates the reachability relationships between nodes in the reversed structure. Then, it uses the reversed node structure as input to construct a new set of nodes and generate the corresponding connection paths. Subsequently, it reverses the new path structure again to restore the connection direction and regenerates the node set and connection order on the restored structure. Through the loop of two reversals and two structure reconstructions, redundant nodes and redundant connections are eliminated, so that the remaining nodes and connection paths form the simplest continuous structure, resulting in a compressed segment structure set that can be used for the output sequence. Brzozowski algorithm, according to the formula:

[0030] in: This represents the weighted average value of the compressed paths obtained after applying a comprehensive weighting to all compressed paths in an animation resource management system based on segment tags. This indicates the path used to identify the compressed path in the animation resource management set. The path index of the compressed path. This indicates the number of compressed paths included in the set of compressed paths. Indicates the compressed path The middle is used to identify the first The node index of the position of each folded node. Indicates the compressed path The number of folded nodes included. Show compression path Upper A unique identifier for a collapsed node within the set of collapsed nodes. Indicates the compressed path The middle is composed of folded nodes To fold node The corresponding basic segment weight value, Show compression path The narrative coherence coefficient, Represents a folded node and The style consistency coefficient between corresponding segments This represents the reuse frequency weighting coefficient for segments between folded nodes. This represents the user preference matching coefficient between segments of folded nodes. This represents the transition penalty factor for segments between folded nodes; Execution process: The entire compressed path output by the sequence recombination submodule is traversed sequentially. During the traversal, each compressed path is first processed... Read the node sequence consisting of all folded nodes in the path and determine the number of nodes. Subsequently, the narrative coherence coefficient corresponding to the path is calculated based on the path narrative score obtained by the system in the narrative segment annotation process. And when entering node-level computation, it is based on the node index. Read adjacent folded node pairs in the path The node pairs read the weights of the basic segments generated during the compression structure construction phase. Then, based on the system's rendering parameter difference analysis module, the style difference is calculated and converted into a style consistency coefficient. The animation resource reuse statistics module is further invoked to count the number of times a segment is reused in the historical sequence and normalized to generate a reuse frequency weighting coefficient. Then, the preference scores for each segment are extracted from the user behavior analysis module and mapped to user preference matching coefficients. The system's lens detection and audio envelope analysis modules then determine whether there are visual or audio abrupt changes in the connection and generate a transition penalty factor accordingly. After obtaining all the above parameters, calculate the product for each node pair. and divide by Obtain the weighted contribution value of the connection, and finally perform path indexing. and node index Perform a double summation to obtain the weighted composite value of the compressed path. This is a comprehensive quality indicator used to characterize the compressed segment structure set under the combined effects of narrative coherence, stylistic continuity, reusability, user preference adaptation, and visual and audio transition quality.

[0031] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An animation resource management system based on segment tagging, characterized in that, The system includes: Tag parsing module: Based on segment tags, extract action category numbers in sequence, arrange temporal position numbers in frame order, match application condition numbers with occupancy identifiers item by item, and integrate numbers according to sequence position to form a basic tag set for the segment; Constraint Construction Module: Based on the segment basic tag set, compare the action category number with adjacent records one by one and organize the adjacent numbers, compare the frame length value with the interval limit and organize the frame length segment, compare the application condition number with the limit list one by one and organize the condition number, and process the three types of numbers in sequence to form a segment constraint item set; Segment Combination Module: Based on the segment constraint entry set, the AC-3 algorithm is used to remove inapplicable segments according to the connection constraints, filter out out-of-range segments according to the frame length interval, exclude conflicting segments according to the condition constraints, and adjust the placement order and back-off position according to the number of candidates to form a segment splicing sequence set. Sequence arrangement module: Based on the segment splicing sequence set, split mutually exclusive label groups, merge non-conflicting labels in sequence order, and extract each group in sequence according to the in-degree record and adjust the in-degree of subsequent groups to obtain the label sequence structure set; State compression module: Based on the label sequence structure set, label nodes are established in sequence, the differences in node behavior are identified according to the transformation record, nodes with different behaviors are grouped, and the Brzozowski algorithm is used to fold nodes with the same behavior into a single structure, while reorganizing the overall sequence to form a compressed segment structure set.

2. The animation resource management system based on segment tagging according to claim 1, characterized in that, The segment basic tag set includes an action category number group, a timing position number group, and an application condition number group. The segment constraint entry set includes an action connection restriction group, a frame length restriction group, and a condition restriction group. The segment splicing sequence set includes a segment sequence group, a sequence placement record, and a backtracking position group. The tag order structure set includes a tag sequence group, an in-degree record group, and a conflict splitting group. The compressed segment structure set includes a folded node group, a node relationship group, and a sequence structure group.

3. The animation resource management system based on segment tagging according to claim 1, characterized in that, The tag parsing module includes: Number Extraction Submodule: Based on segment labels, read action category numbers and arrange them in order, extract timing position numbers item by item and complete frame sequence labeling, check application condition numbers and occupancy identifiers item by item and complete number recording, and generate label number groups; Numbering Integration Submodule: Based on the label numbering group, align the action category numbering records and the time sequence position numbering records one by one and complete the numbering arrangement; arrange the application condition numbering records and the sequence position one by one and complete the numbering parallel; merge the three types of numbering records one by one and complete the sequence integration, and generate the segment basic label set.

4. The animation resource management system based on segment tagging according to claim 1, characterized in that, The constraint construction module includes: Restriction comparison submodule: Based on the segment basic tag set, compare the action category number record with the adjacent record and complete the adjacent arrangement, compare the frame length value with the interval restriction and complete the segment value arrangement, compare the application condition number record with the restriction list and complete the condition placement, and generate a restriction record group; The constraint summary submodule: Based on the constraint record group, it organizes adjacent numbered records and frame length segment records and completes the numbering collection, classifies condition numbered records and sequence positions and completes the numbering collection, lists the adjacent numbered records, frame length segment records and condition numbered records item by item and completes the sequence compilation, and generates a segment constraint item set.

5. The animation resource management system based on segment tagging according to claim 4, characterized in that, The process involves sequentially listing the three types of records and completing the sequence arrangement. The adjacent numbered records, frame length segment records, and condition numbered records in the same sequence position are arranged in a fixed order to form a combined record. Then, all combined records are rearranged into a continuous structure according to their sequence positions, so that the action restrictions, frame length restrictions, and condition restrictions corresponding to each position of the segment can be presented in a unified manner, forming a complete and usable serialized restriction structure.

6. The animation resource management system based on segment tagging according to claim 1, characterized in that, The segment combination module includes: The connection filtering submodule: Based on the set of segment constraint entries, the AC-3 algorithm is used to read the action category number of each segment and determine the segments that can be retained by comparing with the adjacent number groups. Segments without adjacent numbers are removed from the sequence, and the remaining segments are rearranged to the corresponding sequence positions to generate a connection sequence group. Frame length filtering submodule: Based on the connection sequence group, collect the frame length value of each segment and confirm the interval correspondence with the frame length segment group. Remove the segments outside the interval and reorder them. Re-merge the segments within the interval into a continuous sequence to generate a frame length sequence group. Condition filtering submodule: Based on the frame length sequence group, read the application condition number of each segment and screen the source of conflict by comparing with the condition number group. After removing the conflicting segments, rearrange the placement order according to the number of remaining segments, and record the necessary back-off positions to generate a segment splicing sequence set.

7. The animation resource management system based on segment tagging according to claim 6, characterized in that, The AC-3 algorithm uses the adjacent positions of a segment in the sequence as arcs, takes the action category numbers that each segment can take as the value set, takes any arc in turn, and performs an adjacent matching check on the value sets of the segments at both ends of the arc. Values ​​that do not have a valid adjacent number relationship are deleted from the set. Then, the end that was deleted is added back to the check list with the other adjacent positions and the comparison is continued. The arc extraction, adjacent check and value deletion operations are repeated until no new deletion items are generated for all arcs, and only the value set that can form a legal connection relationship with the adjacent segments is retained.

8. The animation resource management system based on segment tagging according to claim 1, characterized in that, The sequence arrangement module includes: Exclusion splitting submodule: Based on the segment splicing sequence set, it identifies mutually exclusive label pairs in the label group and completes the splitting of exclusion items, sorts out non-exclusive labels and completes the sequence order marking, associates the positional relationship between exclusion items and non-exclusive items, completes the sequence distribution record within the group, and generates the label distribution group; Sequence straightening submodule: Based on the label distribution group, extract non-rejection labels and complete sequence merging, extract the in-degree records of each label group and complete the in-degree order arrangement, extract the label groups according to the in-degree order and complete the in-degree adjustment of subsequent groups, and generate a label order structure set.

9. The animation resource management system based on segment tagging according to claim 1, characterized in that, The state compression module includes: Node generation submodule: Based on the tag sequence structure set, read the tag number one by one and establish the corresponding node, register the sequence relationship of adjacent tags as node connection, and then gather all nodes into a node mapping structure according to the reading order to generate tag node group; Node partitioning submodule: Based on the labeled node group, extract the record content between nodes and compare the behavioral differences of each node. Nodes with obvious differences are grouped into independent sets, and nodes with consistent behaviors are grouped into the same set. These sets are then formed into a collapsible structure to generate node collapsible groups. Sequence Recombination Submodule: Based on the node folding group, the Brzozowski algorithm is used to read the connection relationship of the folded nodes and connect each node in the order of connection. The connected structure is reordered into a continuous path, and all node structures are combined into a complete sequence set to generate a compressed segment structure set.

10. The animation resource management system based on segment tagging according to claim 9, characterized in that, The Brzozowski algorithm, based on the connection relationships recorded in the node folding group, reverses the original node structure in the connection direction and regenerates the reachability relationships between nodes in the reversed structure. Then, it uses the reversed node structure as input to construct a new node set and generate the corresponding connection path. Subsequently, it performs the reversal operation again on the new path structure to restore the connection direction and regenerates the node set and connection order on the restored structure. Through the loop of two reversals and two structure reconstructions, redundant nodes and redundant connections are eliminated, so that the remaining nodes and connection paths form the simplest continuous structure, resulting in a compressed segment structure set that can be used for the output sequence.

Images