Intelligent personalized teaching system and method for mass art based on school-society linkage

Abstract

The application discloses a kind of mass art intelligent individualized teaching system and method based on school-society linkage, it is related to mass art teaching field.Method includes: extracting the time overhead and touch precision parameter of art teaching plan, generates process dependence topological graph;Obtain community terminal hardware metadata, calculate maximum coordinate drift threshold to construct kinematics error tolerance envelope space;Search candidate connected subgraph in topological graph based on expected fragmentation residence length;Compare subgraph touch precision and envelope space lower limit, if not exceed, then dispatch first microtask instruction, if exceed, then replace as visual appreciation instruction and dispatch;Finally, discrete coordinate stream uploaded by terminal is mapped to envelope space and is interpolated and smoothed processing.The application effectively solves the teaching fracture caused by learning time fragmentation in community scene, and the drawing skill misjudgment problem caused by cheap heterogeneous hardware.

Description

Technical Field

[0001] This invention relates to the field of art education for the general public, and in particular to an intelligent and personalized art education system and method based on school-community collaboration. Background Technology

[0002] In recent years, school-community collaboration has become an important way to promote art education among the general public. In a typical school-community collaboration scenario, schools are usually responsible for providing standardized, continuous, and systematic professional art lesson plans (such as requiring the completion of a still life watercolor painting with complete light and shadow relationships in two consecutive hours) and providing intelligent painting skill evaluation standards; while communities serve as venues for the public to practice art.

[0003] However, when standardized lesson plans from schools are disseminated to community members, existing intelligent teaching systems face two extremely difficult-to-reconcile physical and objective constraints: First, the extreme fragmentation of drawing time. Community members (such as the elderly, children, and working professionals) often can only practice during fragmented time while waiting or resting. Existing teaching systems typically employ tightly coupled streaming playback or mechanical, fixed steps. If a user exits midway, the original coherent teaching logic breaks down, leading to a situation where the user faces a reversal of the learning process or a complete inability to continue the lesson upon their next login.

[0004] Secondly, there is a severe heterogeneity in terminal hardware. Schools typically use professional digital drawing tablets with high sampling rates and high resolutions; while community members often use inexpensive tablets or even old smartphones. When drawing artistic lines, low touch sampling rates and low resolutions lead to severe coordinate drift and jagged edges. Traditional intelligent evaluation systems cannot address this underlying hardware defect and mechanically misjudge "line breaks caused by device lag" as "poor hand muscle control," causing school evaluation systems to completely fail in community settings. Summary of the Invention

[0005] This invention provides a school-community collaborative intelligent personalized art teaching system and method. The technical problem to be solved is that the existing technology lacks a digital scheduling and mapping mechanism that can deterministically break down continuous teaching processes and adaptively remove underlying hardware errors in scenarios where time fragmentation and hardware heterogeneity are both limiting.

[0006] To achieve the aforementioned objectives, the technical solution adopted by this invention is: a personalized and intelligent art teaching method based on school-community collaboration, wherein the method is executed by a server and includes:

[0007] S1. Obtain the target art lesson plan, extract the time cost parameters and touch accuracy parameters of the rendering nodes, and generate a process dependency topology diagram of the rendering nodes according to the prerequisite association relationship.

[0008] S2. Obtain the current device hardware metadata and the current user's expected fragmented dwell time sent by the community terminal, extract the screen parameters and touch sampling frequency from the current device hardware metadata, calculate the maximum coordinate drift threshold, and construct the kinematic error tolerance envelope space.

[0009] S3. Based on the current user's expected fragmented dwell time, perform a traversal search in the process dependency topology graph to extract candidate connected subgraphs that meet the time constraints.

[0010] S4. Determine whether the touch accuracy parameter in the candidate connected subgraph exceeds the lower limit of the kinematic error tolerance envelope space, and generate the task accuracy comparison result;

[0011] S5. If the task accuracy comparison result is not exceeded, the candidate connected subgraph is encapsulated to generate a first micro-task dispatch instruction and sent to the community terminal.

[0012] S6. If the task accuracy comparison result is excessive, perform visual appreciation downgrade replacement on the candidate connected subgraph to generate a second micro-task dispatch instruction and send it to the community terminal.

[0013] S7. Receive the discrete coordinate stream uploaded by the community terminal based on the first micro-task dispatch instruction, map the discrete coordinate stream to the kinematic error tolerance envelope space for interpolation processing, and generate a smooth compensation coordinate sequence.

[0014] Furthermore, step S1 specifically includes:

[0015] S11. Analyze the brush pixel area and historical average smearing rate of each layer in the target art teaching plan, and calculate the time overhead parameters of each rendering node based on the processor logic unit.

[0016] S12. Extract the brush edge sharpness threshold and pressure sensitivity response level of each layer, perform feature weighting on the brush edge sharpness threshold and the pressure sensitivity response level, and calculate the touch accuracy parameters of each rendering node.

[0017] S13. Extract the multiply and clipping mask identifiers of each layer, and establish a two-dimensional matrix of pre-constraint association between the rendering nodes based on the multiply and clipping mask identifiers.

[0018] S14. The rendering node is used as a directed graph vertex, and the constraint relationship in the pre-constraint association two-dimensional matrix is ​​used as a directed edge to compile the graph structure and output the process dependency topology graph.

[0019] Further, in step S2, obtaining the current user's expected fragmented dwell time sent by the community terminal includes:

[0020] S21. Extract the current geographical location data in the GPS coordinate system from the login request signaling of the community terminal, and extract the matching historical active logs from the historical behavior database;

[0021] S22. Analyze the login timestamp and logout timestamp in the historical active log, calculate the time difference and extract multiple historical dwell time period parameters to construct a hidden Markov time series prediction model.

[0022] S23. Input the multiple historical dwell time period parameters as observation sequences into the Hidden Markov Time Series Prediction Model, and use the forward-backward algorithm to solve for the convergence interval that maximizes the probability of the observation sequence.

[0023] S24. Extract the calculated maximum convergence interval value, and assign the maximum convergence interval value to the expected fragmented dwell time of the current user allocated for this login and write it into the cache.

[0024] Further, in step S2, the step of extracting screen parameters and touch sampling frequency from the current device hardware metadata, calculating the maximum coordinate drift threshold, and constructing the kinematic error tolerance envelope space includes:

[0025] S25. Extract the screen resolution parameter from the screen parameters, and read the system's preset high-frequency response attenuation parameter and interpolation tomography attenuation parameter.

[0026] S26. Obtain the instantaneous pen stroke speed parameter, divide the instantaneous pen stroke speed parameter by the touch sampling frequency, and then multiply it with the high-frequency response attenuation parameter to generate a speed error term;

[0027] S27. Divide the interpolation fault attenuation parameter by the screen resolution parameter to generate a basic error term, and add the basic error term to the velocity error term to obtain the maximum coordinate drift threshold.

[0028] S28. According to the formula Construct envelope features, where The maximum coordinate drift threshold. R is the touch sampling frequency, and R is the screen resolution parameter; where For instantaneous pen stroke speed parameter, The high-frequency response attenuation parameter is... The interpolation fault attenuation parameters are used to generate the kinematic error tolerance envelope space based on the envelope characteristics.

[0029] Furthermore, step S3 specifically includes:

[0030] S31. Using the currently completed node in the process dependency topology graph as the cursor start point, obtain all first-layer downstream connection nodes that are immediately adjacent to the cursor start point;

[0031] S32. Calculate the total time cost corresponding to the current traversal path, apply logical operations to determine whether the total time cost is less than the current user's expected fragmented dwell time, and generate coarse pruning results.

[0032] S33. If the coarse screening and pruning result is less than, perform a depth-first search to the deep nodes and extract the current subgraph node set, and extract all in-degree edge constraints of the current subgraph node set.

[0033] S34. Compare whether all in-degree edge constraints are included in the current completed node set. If the constraints are met, extract all nodes under the path and merge them to generate the candidate connected subgraph.

[0034] Furthermore, after step S5, the following steps are also included:

[0035] S51. Receive the micro-task completion confirmation signal sent by the community terminal, and trigger an interrupt signal to freeze the global multi-threaded rendering logic queue in the current digital canvas system.

[0036] S52. Extract the current layer overlay state matrix, real-time color palette parameter set, and global transparency mask information stream mounted in the global multi-threaded rendering logic queue;

[0037] S53. Perform feature dimensionality reduction on the current layer overlay state matrix and the real-time color palette parameter set, and apply a hash algorithm to generate a unique identifier string for the dimensionality-reduced data.

[0038] S54. Combining the unique identifier string, serialize and reassemble all extracted system operation parameters into a breakpoint context snapshot package carrying a precise timestamp, and encrypt and push it to the cloud database.

[0039] Furthermore, step S6 specifically includes:

[0040] S61. Locate high-frequency cable nodes that exceed the lower limit of the kinematic error tolerance envelope space from the candidate connected subgraph, and mark the high-frequency cable nodes as nodes to be replaced.

[0041] S62. In the system's preset basic appreciation task library, search for visual task modules whose time overhead parameter is strictly equal to the node to be replaced and whose touch accuracy parameter is equal to zero.

[0042] S63. Calculate the Euclidean distance between the node to be replaced and the visual task module in the multidimensional feature space, and extract the target degradation task corresponding to the minimum Euclidean distance.

[0043] S64. Use the target degradation task to forcibly overwrite the node to be replaced in the candidate connected subgraph, and recompile the data structure to generate a degradation task topology tree to encapsulate the second microtask dispatch instruction.

[0044] Further, in step S7, the discrete coordinate stream is mapped to the kinematic error tolerance envelope space for interpolation processing to generate a smoothed compensated coordinate sequence, including:

[0045] S71. Extract the set of discrete coordinate points with adjacent timestamps in the discrete coordinate stream, and calculate the spatial coordinate curvature variance parameter between the discrete coordinate point sets.

[0046] S72. Compare the spatial coordinate curvature variance parameter with the upper limit of the tolerance of the kinematic error tolerance envelope space, and determine that the spatial coordinate curvature variance parameter completely falls within the upper limit of the tolerance.

[0047] S73. Based on the difference between the spatial coordinate curvature variance parameter and the upper limit of tolerance, the adaptive smoothing weight coefficient is calculated using the negative exponential decay formula of the natural constant base.

[0048] S74. Request the allocation of dynamic interpolation cache space from the server processor, and substitute the discrete coordinate point set and the adaptive smoothing weight coefficient into the third-order Bessel equation to reconstruct the smoothing compensation coordinate sequence.

[0049] Furthermore, after comparing the spatial coordinate curvature variance parameter with the upper tolerance limit of the kinematic error tolerance envelope space, the method further includes:

[0050] S721. If it is determined that the spatial coordinate curvature variance parameter exceeds the tolerance upper limit, a subjective force control failure error flag is generated, and the resampling process of the third-order Bessel equation is blocked.

[0051] S722. Extract the brushstroke features of the current node corresponding to the subjective force control error identifier, and retrieve the correction guidance video stream that matches the brushstroke features of the current node in the preset technique error correction knowledge base;

[0052] S723. Align the corrected guiding video stream with the real-time screen coordinates at the pixel level to generate an anti-shake power correction command with local flickering prompts;

[0053] S724. The anti-shake force correction command is encapsulated into downlink control signaling and sent to the community terminal to drive the digital canvas interface to display standard pen stroke auxiliary guide lines superimposed on the discrete coordinate stream.

[0054] Another objective of this invention is to provide a personalized, intelligent art teaching system for the public based on school-community collaboration. The system runs on a server and includes:

[0055] The parsing module obtains the target art lesson plan, extracts the time cost parameters and touch accuracy parameters of the rendering nodes, and generates a process dependency topology graph of the rendering nodes according to the prerequisite association relationship.

[0056] The envelope construction module obtains the current device hardware metadata and the current user's expected fragmented dwell time sent by the community terminal;

[0057] The envelope construction module also extracts screen parameters and touch sampling frequency from the current device hardware metadata, calculates the maximum coordinate drift threshold, and constructs a kinematic error tolerance envelope space.

[0058] The search module performs a traversal search on the process dependency topology graph based on the current user's expected fragmented dwell time, and extracts candidate connected subgraphs that meet the time constraints.

[0059] The scheduling module determines whether the touch accuracy parameter within the candidate connected subgraph exceeds the lower limit of the kinematic error tolerance envelope space, and generates a task accuracy comparison result.

[0060] If the scheduling module does not exceed the limit, it encapsulates the candidate connected subgraph to generate a first microtask dispatch instruction and sends it to the terminal; if it exceeds the limit, it performs a downgrade replacement to generate a second microtask dispatch instruction.

[0061] The interpolation module receives the discrete coordinate stream uploaded based on the first micro-task dispatch instruction, maps the discrete coordinate stream to the kinematic error tolerance envelope space, and performs interpolation processing to generate a sequence.

[0062] The beneficial effects of this invention are as follows:

[0063] 1. This invention establishes a process-dependent topology graph and applies time constraints to perform subgraph search, which intelligently cuts the macro-level teaching tasks of the school into micro-tasks that are suitable for the fragmented time of the masses. Moreover, it ensures the absolute correctness of the art prerequisite logic from the bottom layer of graph theory algorithm, thus avoiding the interruption of the teaching process.

[0064] 2. This invention does not rely on complex image recognition algorithms, but directly reads hardware metadata and uses kinematic principles to dynamically calculate and construct the error tolerance envelope space; this enables the system to accurately intercept tasks destined to fail due to poor equipment, thus protecting the creative enthusiasm of the public.

[0065] 3. When it is determined that the hardware cannot support high-precision painting tasks, the system of the present invention adaptively issues equivalent pure visual appreciation instructions; when the hardware is qualified, the system uses interpolation processing to remove hardware defects from the uploaded coordinate stream and provides the school with a smoothed and compensated real painting skill coordinate sequence, which fundamentally ensures the objectivity and fairness of the school-community joint evaluation system. Attached Figure Description

[0066] Figure 1 A schematic diagram illustrating the processing flow of the intelligent personalized art teaching method for the masses based on school-community collaboration, provided in an embodiment of the present invention.

[0067] Figure 2 This is a schematic diagram illustrating the principle framework for constructing the kinematic error tolerance envelope space and the process-dependent topology graph decomposition and mapping in this embodiment of the invention.

[0068] Figure 3 A block diagram of the module structure of the intelligent personalized art teaching system based on school-community collaboration provided in an embodiment of the present invention. Detailed Implementation

[0069] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0070] This invention provides a personalized, intelligent art teaching method for the public based on school-community collaboration. The method is executed by a server and includes the following steps:

[0071] S1. Obtain the target art lesson plan, extract the time overhead parameters and touch precision parameters of the rendering nodes, and generate a process dependency topology graph of the rendering nodes according to the prerequisite association relationship.

[0072] This step involves reconstructing the original lengthy and indivisible standard school art curriculum using graph theory to restructure its underlying data structure. The system analyzes the layer attributes of the digital lesson plan file to obtain the time cost parameters (approximately how long it takes to draw this stroke) and touch precision parameters (how detailed this stroke needs to be). Subsequently, based on the objective physical laws of art creation (such as the requirement to draw the line art before coloring), the system constructs a directed acyclic graph (DAG) from these nodes through prerequisite relationships, i.e., a process-dependent topology graph. This step provides a quantifiable and decomposable underlying mathematical model foundation for subsequent micro-task scheduling within fragmented timeframes.

[0073] Example: Xiaoming, a resident of the community, uses an old, low-sampling-rate tablet (60Hz, 1080P) in the community activity room to log into the system to study a 120-minute lesson plan on still life painting of apples with watercolors, provided by the school. The system parses this into three rendering nodes: Node A (basic sketching, 15 minutes, low precision requirement), Node B (laying down the base color over a large area, 30 minutes, extremely low precision requirement), and Node C (depicting the highlights on the fruit stem, 10 minutes, extremely high precision requirement). These three nodes form a process dependency topology graph where A points to B, and B points to C.

[0074] S2. Obtain the current device hardware metadata and the current user's expected fragmented dwell time sent by the community terminal, extract the screen parameters and touch sampling frequency from the current device hardware metadata, calculate the maximum coordinate drift threshold, and construct the kinematic error tolerance envelope space.

[0075] This step establishes an objective physical evaluation benchmark to address evaluation errors caused by heterogeneous and inexpensive hardware. Devices in the community environment vary greatly in quality; low-end screens can exhibit severe jagged edges or broken lines during rapid drawing. The system directly extracts the physical parameters of the current device (such as screen parameters and touch sampling frequency) by calling the operating system's low-level application programming interface. Combining kinematic principles, the system can calculate the coordinate distance that the system inevitably misses when the pen tip rapidly traverses the screen at the current sampling rate—the maximum coordinate drift threshold. The area defined by this threshold is the kinematic error tolerance envelope space. This step establishes an objective physical benchmark for subsequent determination of whether "line jitter is caused by poor user drawing skills or by machine lag."

[0076] Example: After obtaining the current device hardware metadata of Xiaoming's tablet, the system calculates that on this hardware, as long as the pen stroke speed exceeds a certain value, the line will inevitably drift by ±3 pixels. The range of these 3 pixels constitutes the kinematic error tolerance envelope space. Simultaneously, the system obtains the user's expected fragmented dwell time as 20 minutes through prediction or user input.

[0077] S3. Based on the current user's expected fragmented dwell time, perform a traversal search in the process dependency topology graph to extract candidate connected subgraphs that meet the time constraints.

[0078] This step is the core logic of time-fragmented scheduling, closely linked to the graph structure generated in the previous step and the user's time. The system applies a depth-first search algorithm from graph theory, starting from the current progress in the process dependency topology graph, searching for the next executable node combination, where the sum of the time cost parameters of all nodes in the combination must be less than the user's expected fragmented dwell time. The extracted graph structure is the candidate connected subgraph. This step ensures that the tasks assigned to the user can be completed within the current free time, preventing interruptions that could lead to unfinished work and ensuring the integrity of the teaching process.

[0079] Example: Xiaoming's current expected fragmented dwell time is 20 minutes. The system searches the process dependency topology graph and finds that node A requires 15 minutes, which is less than 20 minutes; however, if node B (30 minutes) is added, the time becomes 45 minutes, exceeding the limit. Therefore, the system only extracts node A as a candidate connected subgraph.

[0080] S4. Determine whether the touch accuracy parameters within the candidate connected subgraph exceed the lower limit of the kinematic error tolerance envelope space, and generate the task accuracy comparison results.

[0081] This step logically compares the required precision of the task extracted in the previous step with the established hardware tolerance. If a task requires extremely high precision (e.g., an error within 1 pixel), but the hardware itself introduces an error of 3 pixels, then the hardware is inadequate. The output is the task precision comparison result. This step can proactively intercept drawing tasks destined to fail due to poor hardware before the front-end operation occurs, avoiding unfair skill evaluations for users.

[0082] Example: The system determines that the touch accuracy parameter (allowing a deviation of 5 pixels) of node A (drafting) does not exceed the lower limit of the kinematic error tolerance envelope space of Xiaoming's old tablet (maximum tolerance of 3 pixels). Therefore, the generated task accuracy comparison result is within the limit.

[0083] S5. If the task accuracy comparison result is within the limit, encapsulate the candidate connected subgraph to generate the first micro-task dispatch instruction and send it to the community terminal.

[0084] When the hardware and software conditions are compatible and time permits, the system packages the candidate connected subgraphs into executable digital task interaction instructions, i.e., the first micro-task dispatch instructions, and sends them to the front-end device. This step enables the adaptive and precise distribution of complex art lesson plans to fragmented community scenarios.

[0085] Example: Since the hardware meets the requirements and there is enough time, the system generates the first micro-task dispatch instruction, and a drawing interface pops up on Xiaoming's tablet that says "Please complete the Apple pencil sketch in the next 15 minutes".

[0086] S6. If the task accuracy comparison result is excessive, perform visual appreciation downgrade replacement on the candidate connected subgraph to generate a second micro-task dispatch instruction and send it to the community terminal.

[0087] When the system determines that the current hardware is unable to handle the fine details of drawing, it will not force the user to draw. Instead, it will automatically find a purely visual task (such as finding errors in an image or a color theory multiple-choice question) from the background database that takes the same amount of time but does not depend on the hardware sampling rate to replace the original drawing task, and then generate a second micro-task dispatch instruction. This step ensures the continuity of the teaching logic even in the community's harsh heterogeneous hardware environment, ensuring that users are not passively interrupted in their learning process due to poor equipment.

[0088] Example: Suppose Xiaoming's current progress has reached node C (fruit stem highlight, requiring 1 pixel precision), and the system finds that Xiaoming's tablet (tolerance 3 pixels) cannot complete the task. At this point, a downgrade replacement is triggered, and the system generates a second microtask dispatch instruction. Xiaoming's tablet is no longer required to draw, but instead a 10-minute video explaining highlight theory and two multiple-choice questions pop up.

[0089] S7. Receive the discrete coordinate stream uploaded by the community terminal based on the first micro-task dispatch instruction, map the discrete coordinate stream to the kinematic error tolerance envelope space for interpolation processing, and generate a smooth compensation coordinate sequence.

[0090] This step is the final action in the closed-loop error correction process. For the underlying touch points (discrete coordinate stream) that have been correctly assigned to a drawing task and uploaded by the user, the system combines the tolerance envelope established in the previous step to filter out random jump points caused solely by poor screen quality. It then uses mathematical interpolation algorithms to smooth out the uneven lines, ultimately deriving a smooth, compensated coordinate sequence that reflects the user's true intention in applying force. This step completely eliminates the interference of hardware defects on art scoring, providing genuine and pure user drawing skill data for subsequent intelligent teaching evaluation.

[0091] Example: Xiaoming drew the outline of an apple on an old tablet. Due to the age of the device, the uploaded discrete coordinate stream was jagged. The system compared it with the kinematic error tolerance envelope space, confirming that the jaggedness was due to hardware-induced fluctuations, and then started the interpolation algorithm. Finally, the system rendered a smooth apple outline (smoothed compensation coordinate sequence) in the cloud, reflecting Xiaoming's true skill level.

[0092] As an embodiment of the present invention, step S1 specifically includes:

[0093] S11. Analyze the brush pixel area and historical average smearing rate of each layer in the target art teaching plan, and calculate the time overhead parameters of each rendering node based on the processor logic unit.

[0094] S12. Extract the brush edge sharpness threshold and pressure sensitivity level of each layer, perform feature weighting on the brush edge sharpness threshold and pressure sensitivity level, and calculate the touch accuracy parameters of each rendering node.

[0095] S13. Extract the multiply and clipping mask identifiers of each layer, and establish a two-dimensional matrix of pre-constraint associations between rendering nodes based on the multiply and clipping mask identifiers.

[0096] S14. Use the rendered nodes as vertices of the directed graph and the constraint relationships in the pre-constraint association two-dimensional matrix as directed edges to compile the graph structure and output the generated process-dependent topology graph.

[0097] The files in digital painting software are essentially multi-dimensional data matrices. By analyzing the transparency masks and blending modes of digital images (multiply or clipping masks must rely on the underlying graphics for physical rendering), the system can naturally form a logical lock for the painting sequence, thereby extracting the pre-constraint associative two-dimensional matrix. Simultaneously, by analyzing the target smear area and brush parameters, time overhead and touch precision parameters can be scientifically estimated. This step automatically transforms unstructured art lesson plan image files into graph-structured data that can be precisely scheduled and allocated by computer algorithms.

[0098] Example: When the system parses the "Apple Watercolor Still Life" file, it finds that the shadow layer has a multiply blending mode and is bound above the background layer. As a result, the system automatically creates a directed edge in the pre-constraint association matrix: the background layer must be completed before the shadow layer.

[0099] As an embodiment of the present invention, step S2, obtaining the current user's expected fragmented dwell time sent by the community terminal, includes:

[0100] S21. Extract the current geographical location data in the GPS coordinate system from the login request signaling of the community terminal, and extract the matching historical active logs from the historical behavior database.

[0101] S22. Analyze the login timestamp and logout timestamp in the historical active logs, calculate the time difference and extract multiple historical dwell time period parameters to construct a hidden Markov time series prediction model.

[0102] S23. Input multiple historical residence time parameters as observation sequences into the Hidden Markov Time Series Prediction Model, and use the forward-backward algorithm to solve for the convergence interval that maximizes the probability of the observation sequence.

[0103] S24. Extract the calculated maximum convergence interval value, assign the maximum convergence interval value to the expected fragmented dwell time of the current user used in this login, and write it into the cache.

[0104] This step explains how to intelligently predict time parameters in scenarios where users do not actively input the time. Hidden Markov Models (HMMs) are classic statistical models where the hidden state is set as the user's intended life scenario (e.g., "waiting for the bus" or "resting in a coffee shop"), and the observation sequence consists of multiple historical dwell time parameters generated by the user. Through a forward-backward algorithm, the model can predict the most likely time the user will stay based on their current geographical location. This step enables automated time perception without active user input, lowering the barrier to entry for community users.

[0105] Example: Xiaoming opens the system on a bench in the community square. The system extracts this location and finds that Xiaoming's dwell time at this location in historical activity logs is typically 10, 12, or 8 minutes. The Hidden Markov Model for Time Series Prediction uses these parameters as input to calculate that Xiaoming is most likely only taking a short break, thus determining that the maximum convergence interval is 10 minutes. The system automatically assigns 10 minutes as the current user's expected fragmented dwell time.

[0106] As an embodiment of the present invention, step S2 involves extracting screen parameters and touch sampling frequency from the current device hardware metadata, calculating the maximum coordinate drift threshold, and constructing a kinematic error tolerance envelope space, including:

[0107] S25. Extract the screen resolution parameter from the screen parameters, and read the system's preset high-frequency response attenuation parameter and interpolation tomographic attenuation parameter.

[0108] S26. Obtain the instantaneous pen stroke speed parameter, divide the instantaneous pen stroke speed parameter by the touch sampling frequency, and multiply it with the high-frequency response attenuation parameter to generate a speed error term.

[0109] S27. Divide the interpolation fault attenuation parameter by the screen resolution parameter to generate the basic error term. Add the basic error term to the velocity error term to obtain the maximum coordinate drift threshold.

[0110] S28. According to the formula Construct envelope features, where The maximum coordinate drift threshold. R is the touch sampling frequency, R is the screen resolution parameter, and v(t) is the instantaneous pen stroke speed parameter. For high-frequency response attenuation parameters, The interpolation fault attenuation parameters are used to generate a kinematic error tolerance envelope space based on the envelope characteristics.

[0111] This step introduces mathematical physics formulas to support the construction of the hardware envelope space. The principle is that the electronic screen's signal acquisition is a discrete sampling process. According to the Nyquist sampling theorem, the faster the speed and the lower the sampling frequency, the greater the lost physical distance between two points (speed error term); simultaneously, the lower the resolution, the larger the pixel size, and the greater the resulting discontinuity (fundamental error term). The sum of these two factors constitutes the total tolerance. This step establishes the boundary of hardware evaluation error from a mathematical and logical perspective, making the system's tolerance determination mathematically rigorous and repeatable.

[0112] Example of continuity: Xiaoming's tablet touch sampling frequency The screen refresh rate is 60Hz, and the screen resolution parameter R is 1080. Xiaoming's instantaneous pen stroke speed parameter v(t) is extremely fast when drawing lines. The system substitutes the formula... By combining the preset high-frequency response attenuation parameter α and the interpolation tomography attenuation parameter β, the maximum coordinate drift threshold of the system at the instant the screen is swiped is accurately calculated. It reached a width of 4 pixels.

[0113] As an embodiment of the present invention, step S3 specifically includes:

[0114] S31. Using the currently completed node in the process dependency topology graph as the cursor starting point, obtain all first-layer downstream connection nodes immediately adjacent to the cursor starting point.

[0115] S32. Calculate the total time cost corresponding to the current traversal path, apply logical operations to determine whether the total time cost is less than the current user's expected fragmented dwell time, and generate coarse pruning results.

[0116] S33. If the coarse screening and pruning result is less than, perform a depth-first search to the deep nodes and extract the current subgraph node set, and extract all in-degree edge constraints of the current subgraph node set.

[0117] S34. Compare whether all in-degree edge constraints are included in the current set of completed nodes. If the constraints are met, extract all nodes under the path and merge them to generate a candidate connected subgraph.

[0118] This step further refines the logical protection mechanism in the graph theory search algorithm. In-degree edge constraints are edges pointing to a node, representing all the preconditions that must be satisfied to execute that node. The system not only calculates whether the total time exceeds the limit, but also verifies whether these preconditions (all in-degree edge constraints) have already been executed in previous fragmented time (included in the currently completed node set). This step thoroughly prevents errors caused by forced time segmentation due to reversed process order at the underlying graph data structure level, ensuring the strict progression of the art teaching logic.

[0119] Example: The algorithm found the "apple shadow" node, which takes 5 minutes, meeting the time requirement. However, extracting its in-degree edge constraints reveals that drawing the shadow requires first completing the light source setting node. Upon comparison, the light source setting is not in the currently completed node set. Therefore, the system determines that the constraint condition is not met, abandons this path, and ensures that Xiaoming's learning order is absolutely correct.

[0120] As an embodiment of the present invention, step S6 specifically includes:

[0121] S61. Locate high-frequency cabling nodes that exceed the lower limit of the kinematic error tolerance envelope space from the candidate connected subgraph, and mark and isolate the high-frequency cabling nodes as nodes to be replaced.

[0122] S62. In the system's preset basic appreciation task library, search for visual task modules whose time overhead parameter is strictly equal to the node to be replaced and whose touch accuracy parameter is equal to zero.

[0123] S63. Calculate the Euclidean distance between the node to be replaced and the visual task module in the multidimensional feature space, and extract the target degradation task corresponding to the minimum Euclidean distance.

[0124] S64. Use the target degradation task to forcibly overwrite the nodes to be replaced in the candidate connected subgraph, and recompile the data structure to generate a degradation task topology tree to encapsulate the second microtask dispatch instruction.

[0125] This step clarifies the substitution algorithm logic during downgrading and replacement. To ensure a high degree of consistency between the replacement task and the original painting task in terms of teaching objectives, the algorithm vectorizes the artistic tags (such as color, light and shadow, and form) of different tasks and maps them to a high-dimensional feature space. By calculating the Euclidean distance, the smaller the distance value, the closer the teaching knowledge points of the two tasks are. This step enables the intelligent teaching system to flexibly downgrade under extreme hardware conditions while ensuring that the core teaching knowledge points do not shift before and after downgrading.

[0126] Example: The core teaching tag for the high-frequency ribbon cable node that needs to be replaced is "the boundary between light and shadow". The system calculates the Euclidean distance in the feature space and finds the visual task module with the smallest Euclidean distance value—Masterpiece Appreciation: Find the boundary between light and shadow in this Rembrandt painting. The system forcibly overwrites the former with the latter and generates a second microtask dispatch instruction, which is then sent to Xiaoming.

[0127] As an embodiment of the present invention, in step S7, the discrete coordinate stream is mapped to the kinematic error-tolerant envelope space for interpolation processing to generate a smoothed compensated coordinate sequence, including:

[0128] S71. Extract the set of discrete coordinate points with adjacent timestamps in the discrete coordinate stream, and calculate the spatial coordinate curvature variance parameter between the discrete coordinate point sets.

[0129] S72. Compare the spatial coordinate curvature variance parameter with the upper limit of the tolerance of the kinematic error tolerance envelope space, and determine that the spatial coordinate curvature variance parameter completely falls within the upper limit of the tolerance.

[0130] S73. Based on the difference between the spatial coordinate curvature variance parameter and the upper limit of the tolerance, the adaptive smoothing weight coefficient is calculated using the negative exponential decay formula of the natural constant base.

[0131] S74. Request the allocation of dynamic interpolation cache space from the server processor, substitute the discrete coordinate point set and adaptive smoothing weight coefficients into the third-order Bessel equation to reconstruct and generate a smooth compensation coordinate sequence.

[0132] This step defines the dynamic mathematical implementation of smoothing compensation. Its principle lies in applying different correction levels to different degrees of jitter. The closer the spatial coordinate curvature variance parameter (reflecting the degree of abrupt changes in line curvature) is to the hardware tolerance limit, the greater the influence of the hardware, requiring a higher smoothing weight. Using a negative exponential decay formula to calculate the weights enables a smooth transition of correction levels, avoiding over-editing. Subsequently, substituting the discrete coordinate point set into the third-order Bessel equation generates a smooth computer graphics curve. This step achieves pixel-level adaptive repair for heterogeneous hardware errors, avoiding image distortion caused by using fixed smoothing parameters.

[0133] Example: The system extracts the discrete coordinate point set corresponding to the jagged apple outline drawn by Xiaoming, and calculates that its spatial coordinate curvature variance parameter is within the hardware tolerance limit of the old tablet. The system uses the negative exponential decay formula to calculate an extremely high adaptive smoothing weight coefficient. After substituting it into the third-order Bessel equation, the jagged edges are re-rendered into a smooth, rounded, compensated coordinate sequence.

[0134] Furthermore, after comparing the spatial coordinate curvature variance parameter with the upper tolerance limit of the kinematic error tolerance envelope space, the method also includes:

[0135] S721. If the curvature variance parameter of the spatial coordinates is determined to exceed the upper tolerance limit, a subjective force control failure error flag is generated, blocking the resampling process of the third-order Bessel equation.

[0136] S722. Extract the brushstroke features of the current node corresponding to the subjective force control error identifier, and retrieve the correction guidance video stream that matches the brushstroke features of the current node from the preset technique error correction knowledge base.

[0137] S723. Align the corrected guide video stream with the real-time screen coordinates at the pixel level, and generate an anti-shake power correction command with local flicker indication.

[0138] S724 encapsulates the anti-shake force correction command into downlink control signaling and sends it to the community terminal, driving the digital canvas interface to display standard pen stroke auxiliary guide lines superimposed on the discrete coordinate stream.

[0139] This step establishes a complete business loop, transitioning from objective hardware correction to subjective drawing skill error correction. When previous steps calculate that the spatial coordinate curvature variance parameter exceeds the hardware's maximum allowable jitter, the system determines that this is not an objective equipment defect, but rather a result of the user's subjective muscle control (hand tremors). At this point, the system proactively blocks the interpolation and beautification process, instead activating the teaching error correction mechanism for real-world intervention. This step achieves precise localization of the user's actual drawing skill errors and provides highly interactive real-time guidance, preventing the teaching system from becoming merely an automated image editing software.

[0140] Example: When Xiaoming was drawing a long straight line, his hand trembled severely. The system calculated that the variance of the curvature parameter of his spatial coordinates far exceeded the tolerance limit of the old tablet. The system immediately generated a subjective force control error flag, stopped the interpolation process, and sent a shaky force correction command to Xiaoming's community terminal. A red standard pen stroke guide line flashed on the screen, right above Xiaoming's crooked line, prompting Xiaoming to stabilize his force again.

[0141] As an embodiment of the present invention, after step S5, the method further includes:

[0142] S51. Receive the micro-task completion confirmation signal sent by the community terminal and trigger an interrupt signal to freeze the global multi-threaded rendering logic queue in the current digital canvas system.

[0143] S52. Extract the current layer overlay state matrix, real-time color palette parameter set, and global transparency mask information stream mounted in the global multi-threaded rendering logic queue.

[0144] S53. Perform feature dimensionality reduction on the current layer overlay state matrix and the real-time color palette parameter set, and apply a hash algorithm to generate a unique identifier string for the dimensionality-reduced data.

[0145] S54. Combine the unique identifier string to serialize and reassemble all extracted system operation parameters into a breakpoint context snapshot package with a precise timestamp, and encrypt and push it to the cloud database.

[0146] This step establishes a cloud-based archiving mechanism to ensure a smooth learning experience during fragmented periods. The state of an artwork is composed of a vast number of parameters, including layers and brush attributes. The system momentarily pauses all actions by triggering an interrupt signal, serializing data such as colors and layer masks at that instant (converting the program's running state into a format that can be stored long-term). A hash algorithm is applied to generate a unique identifier string for each archive, ensuring data uniqueness and tamper-proof protection. This step achieves high-density environmental snapshot retention within extremely short, fragmented timeframes, ensuring that users can accurately resume their drawing progress the next time they log in across devices.

[0147] Example: Xiaoming finishes his initial sketch with Apple Pencil after 20 minutes of free time. The system immediately freezes the screen, extracts the colors and layer parameters Xiaoming just adjusted, packages them into a breakpoint context snapshot using a hash algorithm, and stores it in the cloud database. The next day, Xiaoming logs in on a different device and can instantly and perfectly resume his drawing from the previous day.

[0148] To further illustrate the above embodiments, the technical solutions in the above embodiments will be clearly and completely described below with reference to the accompanying drawings.

[0149] In a specific implementation scenario, schools distribute standard teaching resources via the cloud, while community members access these resources using mobile devices with varying performance. The method and system provided in this embodiment of the invention run on a teaching server connecting these two ends.

[0150] like Figure 1 and Figure 2 As shown in the figure, this embodiment of the invention provides a smart and personalized art teaching method for the public based on school-community collaboration, which is executed by the server. The specific processing procedure is detailed below:

[0151] First, the server retrieves the target art lesson plan. To adapt the large and coherent art lesson plan to fragmented scenarios, the system extracts the rendering node's time overhead and touch precision parameters, and generates a process dependency topology graph based on the rendering nodes' prerequisite relationships. This processing... Figure 2 Specifically, this is manifested in decoupling the multi-layered structure of the image into a directional acyclic graph structure, which ensures the logical validity of subsequent scheduling.

[0152] Subsequently, in response to community scenarios, the system obtains the current device hardware metadata and the user's expected fragmented dwell time sent by the community terminal 200. Addressing hardware differences, the system extracts screen parameters and touch sampling frequency from the current device hardware metadata, calculates the maximum coordinate drift threshold, and constructs a kinematic error tolerance envelope space. For example... Figure 2 As shown in the principle framework, this envelope space serves as a physical benchmark for subsequent comparisons, defining the limits of the inherent error of the equipment.

[0153] Furthermore, based on the current user's expected fragmented dwell time, the system performs a traversal search in the process dependency topology graph to extract candidate connected subgraphs that meet the time constraints. This enables the dynamic "on-demand segmentation" of lesson plans according to the actual dwell time of the users.

[0154] After completing the task segmentation, the system does not issue the task directly, but instead introduces security verification logic. The system determines whether the touch accuracy parameters within the candidate connected subgraph exceed the lower limit of the kinematic error tolerance envelope space, and generates task accuracy comparison results.

[0155] If the task accuracy comparison result is not exceeded, it indicates that the current device's hardware capability is sufficient to support the task. The system then encapsulates the candidate connected subgraph to generate the first micro-task dispatch instruction and sends it to the community terminal 200.

[0156] Conversely, if the task accuracy comparison result exceeds the limit, the system will not force the user to perform micro-operations that are destined to fail. Instead, it will visually evaluate and downgrade the candidate connected subgraphs to generate a second micro-task dispatch instruction and send it to the community terminal 200, thus ensuring the continuity of learning.

[0157] Finally, for the device that performed the first micro-task, the server received the discrete coordinate stream uploaded by the community terminal 200. To completely eliminate the "jagged edges" or "jitter" caused by hardware defects, the system mapped the discrete coordinate stream to the kinematic error tolerance envelope space for interpolation processing, generating a smoothed and compensated coordinate sequence for fair review by the school's evaluation system.

[0158] like Figure 3 As shown, this invention also provides a personalized intelligent art teaching system for the public based on school-community collaboration. This system runs on the teaching server side and is used to execute the above-mentioned... Figure 1 and Figure 2 The method shown. The system specifically includes:

[0159] The parsing module 101 is used to obtain the target art lesson plan, extract the time cost parameters and touch accuracy parameters of the rendering nodes, and generate a process dependency topology diagram of the rendering nodes according to the prerequisite association relationship.

[0160] The envelope construction module 102 is communicatively connected to the parsing module 101. It is used to obtain the current device hardware metadata and the current user's expected fragmented dwell time sent by the community terminal 200, extract the screen parameters and touch sampling frequency from the current device hardware metadata, calculate the maximum coordinate drift threshold, and construct the kinematic error tolerance envelope space.

[0161] Search module 103 is used to perform a traversal search in the process dependency topology graph based on the current user's expected fragmented dwell time, and extract candidate connected subgraphs that meet the time constraints.

[0162] The scheduling module 104 interacts with the search module 103 and the envelope construction module 102 to determine whether the touch accuracy parameters in the candidate connected subgraph exceed the lower limit of the kinematic error tolerance envelope space and generate a task accuracy comparison result. If the task accuracy comparison result does not exceed the limit, the candidate connected subgraph is encapsulated to generate a first micro-task dispatch instruction and sent to the community terminal 200. If the task accuracy comparison result exceeds the limit, the candidate connected subgraph is visually downgraded and replaced to generate a second micro-task dispatch instruction and sent to the community terminal 200.

[0163] The interpolation module 105 is used to receive the discrete coordinate stream uploaded by the community terminal 200 based on the first micro-task dispatch instruction, map the discrete coordinate stream to the kinematic error tolerance envelope space for interpolation processing, and generate a smooth compensation coordinate sequence.

[0164] In the actual hardware deployment architecture, the teaching server is also equipped with a processor and a memory. The memory is used to store computer programs, and when the processor executes the computer programs, it can instantiate and run the parsing module 101, envelope construction module 102, search module 103, scheduling module 104, and interpolation module 105 to complete the concurrent processing of massive requests from the community terminals 200.

[0165] Those skilled in the art will understand that the systems and methods described in the above embodiments can be referred to in correspondence with each other. This invention is not limited to the specific embodiments described above, and any modifications or equivalent substitutions made without departing from the spirit and scope of this invention should be included within the protection scope of this invention.

[0166] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A personalized, intelligent teaching method for mass art education based on school-community collaboration, characterized in that: The method is executed by the server and includes: S1. Obtain the target art lesson plan, extract the time cost parameters and touch accuracy parameters of the rendering nodes, and generate a process dependency topology diagram of the rendering nodes according to the prerequisite association relationship. S2. Obtain the current device hardware metadata and the current user's expected fragmented dwell time sent by the community terminal, extract the screen parameters and touch sampling frequency from the current device hardware metadata, calculate the maximum coordinate drift threshold, and construct the kinematic error tolerance envelope space. S3. Based on the current user's expected fragmented dwell time, perform a traversal search in the process dependency topology graph to extract candidate connected subgraphs that meet the time constraints. S4. Determine whether the touch accuracy parameter in the candidate connected subgraph exceeds the lower limit of the kinematic error tolerance envelope space, and generate the task accuracy comparison result; S5. If the task accuracy comparison result is not exceeded, the candidate connected subgraph is encapsulated to generate a first micro-task dispatch instruction and sent to the community terminal. S6. If the task accuracy comparison result is excessive, perform visual appreciation downgrade replacement on the candidate connected subgraph to generate a second micro-task dispatch instruction and send it to the community terminal. S7. Receive the discrete coordinate stream uploaded by the community terminal based on the first micro-task dispatch instruction, map the discrete coordinate stream to the kinematic error tolerance envelope space for interpolation processing, and generate a smooth compensation coordinate sequence.

2. The intelligent personalized art teaching method based on school-community collaboration according to claim 1, characterized in that, Step S1 specifically includes: S11. Analyze the brush pixel area and historical average smearing rate of each layer in the target art teaching plan, and calculate the time overhead parameters of each rendering node based on the processor logic unit. S12. Extract the brush edge sharpness threshold and pressure sensitivity response level of each layer, perform feature weighting on the brush edge sharpness threshold and the pressure sensitivity response level, and calculate the touch accuracy parameters of each rendering node. S13. Extract the multiply and clipping mask identifiers of each layer, and establish a two-dimensional matrix of pre-constraint association between the rendering nodes based on the multiply and clipping mask identifiers. S14. The rendering node is used as a directed graph vertex, and the constraint relationship in the pre-constraint association two-dimensional matrix is ​​used as a directed edge to compile the graph structure and output the process dependency topology graph.

3. The intelligent personalized art teaching method based on school-community collaboration according to claim 1, characterized in that, In step S2, obtaining the current user's expected fragmented dwell time sent by the community terminal includes: S21. Extract the current geographical location data in the GPS coordinate system from the login request signaling of the community terminal, and extract the matching historical active logs from the historical behavior database; S22. Analyze the login timestamp and logout timestamp in the historical active log, calculate the time difference and extract multiple historical dwell time period parameters to construct a hidden Markov time series prediction model. S23. Input the multiple historical dwell time period parameters as observation sequences into the Hidden Markov Time Series Prediction Model, and use the forward-backward algorithm to solve for the convergence interval that maximizes the probability of the observation sequence. S24. Extract the calculated maximum convergence interval value, and assign the maximum convergence interval value to the expected fragmented dwell time of the current user used in this login and write it into the cache.

4. The intelligent personalized art teaching method based on school-community collaboration according to claim 1, characterized in that, In step S2, the step of extracting screen parameters and touch sampling frequency from the current device hardware metadata, calculating the maximum coordinate drift threshold, and constructing the kinematic error tolerance envelope space includes: S25. Extract the screen resolution parameter from the screen parameters, and read the system's preset high-frequency response attenuation parameter and interpolation tomography attenuation parameter. S26. Obtain the instantaneous pen stroke speed parameter, divide the instantaneous pen stroke speed parameter by the touch sampling frequency, and then multiply it with the high-frequency response attenuation parameter to generate a speed error term; S27. Divide the interpolation fault attenuation parameter by the screen resolution parameter to generate a basic error term, and add the basic error term to the velocity error term to obtain the maximum coordinate drift threshold. S28. According to the formula Construct envelope features, where The maximum coordinate drift threshold. R is the touch sampling frequency, and R is the screen resolution parameter; where For instantaneous pen stroke speed parameter, The high-frequency response attenuation parameter is... The interpolation fault attenuation parameters are used to generate the kinematic error tolerance envelope space based on the envelope characteristics.

5. The intelligent personalized art teaching method based on school-community collaboration according to claim 1, characterized in that, Step S3 specifically includes: S31. Using the currently completed node in the process dependency topology graph as the cursor start point, obtain all first-layer downstream connection nodes that are immediately adjacent to the cursor start point; S32. Calculate the total time cost corresponding to the current traversal path, apply logical operations to determine whether the total time cost is less than the current user's expected fragmented dwell time, and generate coarse pruning results. S33. If the coarse screening and pruning result is less than, perform a depth-first search to the deep nodes and extract the current subgraph node set, and extract all in-degree edge constraints of the current subgraph node set. S34. Compare whether all in-degree edge constraints are included in the current completed node set. If the constraints are met, extract all nodes under the path and merge them to generate the candidate connected subgraph.

6. The intelligent personalized art teaching method based on school-community collaboration according to claim 1, characterized in that, After step S5, the following also includes: S51. Receive the micro-task completion confirmation signal sent by the community terminal, and trigger an interrupt signal to freeze the global multi-threaded rendering logic queue in the current digital canvas system. S52. Extract the current layer overlay state matrix, real-time color palette parameter set, and global transparency mask information stream mounted in the global multi-threaded rendering logic queue; S53. Perform feature dimensionality reduction on the current layer overlay state matrix and the real-time color palette parameter set, and apply a hash algorithm to generate a unique identifier string for the dimensionality-reduced data. S54. Combining the unique identifier string, serialize and reassemble all extracted system operation parameters into a breakpoint context snapshot package carrying a precise timestamp, and encrypt and push it to the cloud database.

7. The intelligent personalized art teaching method based on school-community collaboration according to claim 1, characterized in that, Step S6 specifically includes: S61. Locate high-frequency cable nodes that exceed the lower limit of the kinematic error tolerance envelope space from the candidate connected subgraph, and mark the high-frequency cable nodes as nodes to be replaced. S62. In the system's preset basic appreciation task library, search for visual task modules whose time overhead parameter is strictly equal to the node to be replaced and whose touch accuracy parameter is equal to zero. S63. Calculate the Euclidean distance between the node to be replaced and the visual task module in the multidimensional feature space, and extract the target degradation task corresponding to the minimum Euclidean distance. S64. Use the target degradation task to forcibly overwrite the node to be replaced in the candidate connected subgraph, and recompile the data structure to generate a degradation task topology tree to encapsulate the second microtask dispatch instruction.

8. The intelligent personalized art teaching method based on school-community collaboration according to claim 1, characterized in that, In step S7, the discrete coordinate stream is mapped to the kinematic error tolerance envelope space for interpolation processing to generate a smoothed compensated coordinate sequence, including: S71. Extract the set of discrete coordinate points with adjacent timestamps in the discrete coordinate stream, and calculate the spatial coordinate curvature variance parameter between the discrete coordinate point sets. S72. Compare the spatial coordinate curvature variance parameter with the upper limit of the tolerance of the kinematic error tolerance envelope space, and determine that the spatial coordinate curvature variance parameter completely falls within the upper limit of the tolerance. S73. Based on the difference between the spatial coordinate curvature variance parameter and the upper limit of tolerance, the adaptive smoothing weight coefficient is calculated using the negative exponential decay formula of the natural constant base. S74. Request the allocation of dynamic interpolation cache space from the server processor, and substitute the discrete coordinate point set and the adaptive smoothing weight coefficient into the third-order Bessel equation to reconstruct the smoothing compensation coordinate sequence.

9. The intelligent personalized art teaching method based on school-community collaboration according to claim 8, characterized in that, After comparing the spatial coordinate curvature variance parameter with the upper tolerance limit of the kinematic error tolerance envelope space, the method further includes: S721. If it is determined that the spatial coordinate curvature variance parameter exceeds the tolerance upper limit, a subjective force control failure error flag is generated, and the resampling process of the third-order Bessel equation is blocked. S722. Extract the brushstroke features of the current node corresponding to the subjective force control error identifier, and retrieve the correction guidance video stream that matches the brushstroke features of the current node in the preset technique error correction knowledge base; S723. Align the corrected guiding video stream with the real-time screen coordinates at the pixel level to generate an anti-shake power correction command with local flickering prompts; S724. The anti-shake force correction command is encapsulated into downlink control signaling and sent to the community terminal to drive the digital canvas interface to display standard pen stroke auxiliary guide lines superimposed on the discrete coordinate stream.

10. A system applied to the intelligent personalized art teaching method based on school-community collaboration as described in any one of claims 1 to 9, characterized in that, The system runs on a server and includes: The parsing module is used to obtain the target art teaching plan, extract the time cost parameters and touch accuracy parameters of the rendering nodes, and generate a process dependency topology graph of the rendering nodes according to the prerequisite association relationship. The envelope construction module is used to obtain the current device hardware metadata and the current user's expected fragmented dwell time sent by the community terminal, extract the screen parameters and touch sampling frequency from the current device hardware metadata, calculate the maximum coordinate drift threshold, and construct the kinematic error tolerance envelope space. The search module is used to traverse and search the process dependency topology graph based on the current user's expected fragmented dwell time and extract candidate connected subgraphs that meet the time constraints. The scheduling module is used to determine whether the touch accuracy parameter within the candidate connected subgraph exceeds the lower limit of the kinematic error tolerance envelope space, and generate a task accuracy comparison result; if the task accuracy comparison result is within the limit, the candidate connected subgraph is encapsulated to generate a first micro-task dispatch instruction and sent to the community terminal; if the task accuracy comparison result is outside the limit, the candidate connected subgraph is visually downgraded and replaced to generate a second micro-task dispatch instruction and sent to the community terminal. The interpolation module is used to receive the discrete coordinate stream uploaded by the community terminal based on the first micro-task dispatch instruction, map the discrete coordinate stream to the kinematic error tolerance envelope space for interpolation processing, and generate a smooth compensation coordinate sequence.

Images