130 results about "Skeleton graph" patented technology

The invention discloses a sensing data fusion-based radiator abnormal state detection method, relates to the technical field of radiator state detection, and is used for solving the problem of poor radiator state detection efficiency. According to the method, multiple types of sensors are arranged in combination with a radiator structure, acquisition time alignment is completed, segmentation and de-noising processing is performed on multi-modal data based on a sliding time window, and a frequency domain texture feature is constructed by extracting a frequency spectrum information entropy and a high-frequency energy ratio; constructing the feature values into graph nodes, establishing a complete connection graph, compressing the complete connection graph into a skeleton graph through conditional independence test, generating a directed acyclic graph in combination with an intersection structure and topological sorting, and calculating a causal weight to realize structure updating; and extracting a current window node state, comparing the current window node state with a prediction state, identifying an abnormal node, executing path backtracking, calculating path cost, tracing to a root cause node, extracting path evaluation information, and generating a response signal, thereby improving the abnormality diagnosis precision and scheduling linkage responsivity of the radiator.

The invention discloses a geological map-oriented breakpoint repairing and closed curve reconstruction method and system and a medium. The method comprises the following steps of: preprocessing an original geological map to obtain a grayscale image; carrying out binarization and morphological closed operation processing to obtain a grayscale image of the edge of the smooth contour curve; performing pixel skeletonization to obtain a pixel-level skeleton diagram; performing secondary edge extraction, connected domain marking and endpoint statistics on the pixel-level skeleton diagram, and screening out a non-closed curve and an endpoint set; carrying out nearest neighbor retrieval by adopting KD-Tree to obtain a candidate pairing set of each end point; performing priority pairing in combination with the spatial distance and the direction smoothness; and generating a transition line segment according to a pairing result to supplement the fracture part of the non-closed curve. According to the method, non-closed breakpoints generated by scanning or drawing errors in the geological map can be automatically restored, and a closed curve with topological integrity and a smooth boundary is generated.

The invention discloses an internet consumption analysis method based on data analysis, and relates to the technical field of data analysis, and the method comprises the steps: a plurality of participation platforms generate local cause-effect sub-graphs according to user behavior data, convert the local cause-effect sub-graphs into Hash codes, upload the Hash codes to a central server, and aggregate the Hash codes to generate a global cause-effect skeleton graph; injecting a predefined sequential logic rule based on the global causal skeleton graph, dynamically adjusting a causal edge weight according to a user real-time behavior event, and incrementally updating nodes and edges in the dynamic knowledge graph; pre-training the policy network parameters by using each participation platform, and uploading the pre-trained policy network parameters to a central server for aggregation; and issuing the aggregated policy network parameters to each participation platform, executing an interpretable policy according to the user real-time behavior event, and returning user feedback behavior data to each participation platform. According to the method, collaborative optimization of privacy security, real-time response and causal interpretability is realized through federal causal discovery and a dynamic knowledge graph evolution architecture.

The invention provides a sports event real-time score statistical method and system based on AI visual identification, and relates to the technical field of intelligent scoring of sports events, and the method comprises the steps: collecting images through multiple cameras, and carrying out target detection to obtain data of athletes, equipment and scoring areas; generating a joint thermodynamic diagram by using the feature pyramid network, and constructing a dynamic skeleton diagram structure to extract action features; performing periodic detection and abnormity correction on the equipment; and performing feature matching and score calculation through a double-flow feature projection network and a multi-head attention mechanism, generating a score data packet including time, athlete numbers, types and video clips, and sending the score data packet to a scoring system. According to the invention, intelligent real-time scoring of sports events is realized, and the scoring accuracy and efficiency are improved.

The invention relates to the technical field of knee joint rehabilitation evaluation, in particular to a knee joint rehabilitation evaluation method based on a human skeleton by using a space-time diagram convolutional network, and the method comprises the following steps: obtaining a rehabilitation evaluation gait video set; extracting human body key point skeleton data, and slicing the human body key point skeleton data into gait samples according to the number of time frames; preprocessing the gait sample, and extracting joint, skeleton and joint angle features; a gait-link method is adopted to divide a space gait skeleton diagram, a space-time attention mechanism is added, and an improved CTR-GCN network model is constructed; training the improved CTR-GCN network model by using the gait sample to obtain a gait evaluation model; and processing a gait video to be evaluated, and inputting the processed gait video into the gait evaluation model to obtain a rehabilitation evaluation result of the patient. According to the method, the human body key point skeleton in the walking video of the patient is extracted and input into the space-time diagram convolutional network for training, the rehabilitation evaluation model is obtained, the rehabilitation condition of the patient is evaluated by using the model, and the requirement of rehabilitation evaluation is met.

The invention relates to the technical field of consumer behavior analysis, and discloses a network user shopping behavior causal analysis method and system, and the method comprises the following specific steps: collecting multi-dimensional time series data affecting a shopping decision; constructing an observation variable set V with time lag; initializing a completely undirected graph structure G; a tensor rank-based conditional independence test method is adopted to obtain a skeleton graph structure G'and a potential variable candidate pair set Lc after conditional independence judgment pruning; potential variables are grouped, and a graph structure GL containing the potential variables is obtained; and performing causal direction judgment operation on the GL and performing direction control on the potential variable nodes to generate a directed acyclic graph structure Gfinal. According to the method, the problem of difficulty in causal structure analysis in the prior art is solved, and the method has the characteristics of high interpretability and robustness.

The invention discloses a crack width measurement method and system based on dynamic topological structure analysis. A system hardware basis comprises an image acquisition module, an image processing unit, a data storage module and a result display module. According to the method, the working process of an image processing unit is divided into two stages according to a time sequence: the first stage is used for realizing data processing and crack segmentation, and the second stage is used for completing dynamic topology analysis and width measurement of cracks. In the second stage, the steps of form closed operation, skeleton extraction and enhancement, end point bridging, topological statistics and segment width measurement are executed in sequence, and finally a visual result and quantized data are output. The method comprises the following steps: processing an initial skeleton image by adopting an independently researched and developed topology enhancement algorithm; the method is especially good at processing complex net-shaped cracks, and the defect that measurement is inaccurate at crossed and bent cracks in a traditional method is effectively overcome through a dynamic topology analysis method.

The invention discloses a pedestrian re-recognition method based on a space-time hybrid Transform and a skeleton diagram attention mechanism, and belongs to the field of computer vision and pattern recognition. The invention aims to solve the problem that the existing pedestrian re-identification method based on a skeleton cannot fully utilize spatial-temporal characteristics to represent pedestrian identity information. Comprising the following steps: processing 3D coordinates of each articulation point in a skeleton diagram by adopting a track branch to obtain track characteristics of all articulation points of each frame of skeleton diagram; an adjacent matrix is determined based on the hinging relation of all the joint points by adopting a structure branch, and then processing is carried out to obtain an integrated structure feature corresponding to each frame of skeleton diagram; fusing the trajectory features with the integrated structure features to obtain space-time mixed features; and processing the space-time mixed feature to obtain an identity feature vector so as to realize pedestrian re-identification. The method is used for pedestrian re-identification.

The invention discloses a root system point cloud skeleton extraction method, system and device, and the method comprises the steps: obtaining a plant root system point cloud, carrying out the point cloud division of the plant root system point cloud, and forming a root system voxel grid set; a root system voxel graph is constructed, connection is performed based on the spatial position of the current root system voxel grid and the spatial position of the adjacent root system voxel grid to obtain the edge of the root system voxel graph, and the weight of the root system voxel graph is obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space; based on the root system voxel graph, constructing a root system voxel spanning tree, judging noise in the root system voxel graph according to the root system voxel spanning tree, and filtering to obtain an initial root system skeleton graph; acquiring root system connection data of the current root system voxel grid, and filtering the root system voxel grid corresponding to the burrs through judgment to obtain a root system skeleton diagram. According to the method, skeleton extraction is performed from the point cloud data, skeleton misjudgment is reduced, and the problems of smear, noisy points, burrs and the like on the surface of a root system are solved.

The invention discloses a tubular structure segmentation method based on frequency domain Mama and topology enhancement, and the method comprises the steps: carrying out the preprocessing of training data, extracting an initial skeleton and a distance map based on the label or structural features of the training data, generating an extended skeleton map based on inverse distance map weighting, and carrying out the subsequent topological constraint modeling and the subsequent supervision signal generation. The expression capability on the connectivity of the tiny branches and the structure is enhanced; constructing a segmentation network fusing a frequency domain Mama module and a difference perception gating attention module, and realizing cross-frequency feature modeling and multilayer semantic alignment; and step-by-step optimization strategy training is adopted, basic segmentation training is firstly carried out, then topology guidance supervision is introduced, and the overall structure consistency and branch connectivity are improved. Compared with the prior art, the method has the advantages that the high efficiency of the model is kept, the segmentation integrity and topological fidelity of the slender structure are remarkably improved, and more reliable technical support is provided for clinical diagnosis and preoperative planning. And constructing a segmentation network fusing a frequency domain Mama module and a difference perception attention mechanism.

The invention relates to an austenite structure display and grain size rating method, and belongs to the technical field of hot continuous rolling finishing mill production methods. According to the technical scheme, the method comprises the steps that an eroded sample is manufactured, and an austenite structure display graph is obtained; carrying out deep learning network training and grain boundary processing to obtain a boundary skeleton diagram; and grading the grain size by adopting a straight line intercept point method. The method has the beneficial effects that the Inconel 625 austenite structure with a uniform erosion surface and a clear and complete austenite grain boundary can be obtained, grain size rating is carried out on the austenite structure image through a deep learning algorithm, and compared with manual rating, the efficiency and accuracy of grain size rating are greatly improved.

The invention discloses an abnormal event detection method and device, equipment and a medium, and relates to the technical field of deep learning, and the method comprises the steps: determining a plurality of skeleton sequences corresponding to each human skeleton track of all pedestrians in a to-be-detected video, and constructing a space-time skeleton diagram based on the plurality of skeleton sequences; determining a global feature component through the feature vector and the length and width of the detection bounding box, and determining a local feature component based on the global feature component; obtaining a reconstruction sequence and a prediction sequence of the local feature component by using the trained space-time diagram convolution auto-encoder network, and determining a target reconstruction anomaly score and a target prediction anomaly score of the video frame at the target moment based on the reconstruction sequence and the prediction sequence; and determining a target abnormal score based on the target reconstruction abnormal score and the target prediction abnormal score, and detecting whether an abnormal event exists in the to-be-detected video according to the target abnormal score. Therefore, the abnormal event in the video can be accurately detected.

A method, apparatus, non-transitory computer readable medium, and system for image generation includes obtaining an input image depicting an entity and a skeleton map depicting a pose of the entity and performing a cross-attention mechanism between image features of the input image and entity features representing the pose to obtain modified image features. An output image is generated based on the modified image features that depicts the entity with the pose.

The invention relates to a multi-modal gait recognition method based on space-time semantic modeling and cross-modal cooperation, and the method comprises the steps: inputting a skeleton energy diagram, a contour energy diagram, a contour diagram and a skeleton diagram into four feature extraction branches with the same structure, and obtaining four modal features and two stage features of each modal; inputting the two stage features of each group of cross-modal combination into a double-stage feature interaction enhancement module to obtain a cross-modal fusion feature; all the modal features and the cross-modal fusion features are mapped and pooled and then input into a classification head, and a multi-modal gait recognition result is obtained. According to the method, key features under different view angles and walking conditions are effectively focused, the integrity of local channel information is reserved, modeling is performed on spatio-temporal information through multiple branches, high-level semantic features are extracted, spatio-temporal changes of gaits under different view angles and walking conditions are accurately captured, features of different modal data are fused, and the gait time-space information is obtained. And the gait recognition performance under multi-view and different walking conditions is improved.

The invention provides an ancient character defect font reconstruction method and device based on a graph structure and deep learning, and relates to the technical field of image processing, and the method comprises the steps: carrying out the graph structure representation extraction of a defective ancient character image, and obtaining a target contour graph and a target skeleton graph corresponding to the defective ancient character image; according to the target contour map and the target skeleton map, determining a defect type corresponding to the defective ancient character image, determining a font reconstruction strategy based on the defect type, and determining a map neural network corresponding to the defective ancient character image based on the font reconstruction strategy; the graph neural network comprises a contour graph reconstruction network, a skeleton graph reconstruction network or a cross-modal reconstruction network; and inputting the defective ancient character image into a pre-trained graph neural network based on a font reconstruction strategy to obtain a reconstruction result. According to the method, high-quality ancient character defect font reconstruction is realized by using the graph neural network.

The invention provides an attack scene reconstruction method based on network threat clue causal mining, and relates to the technical field of network security. The method comprises the following steps: performing feature extraction and representation on collected network data to obtain multi-source data; extracting open source data based on an open source threat intelligence platform to establish a marginal causal relationship of multi-source data; the method comprises the following steps: constructing a completely undirected graph based on multi-source data, initializing a condition set, testing whether adjacent node pairs are independent or not based on the condition set, removing edges without a direct causal relationship when the adjacent node pairs are independent, obtaining a skeleton graph, and expanding a marginal causal relationship based on the condition set at the moment; and performing edge orientation processing on the skeleton diagram based on the causal priori knowledge base to obtain a causal diagram, and obtaining a causal relationship between the network data based on the causal diagram and reconstructing an attack path. According to the method, a causal discovery process based on marginal causal priori knowledge extension is introduced, so that the generalization ability of the model is enhanced, and the adaptability to unknown attacks is improved.

The invention discloses a railway driver distraction behavior identification method based on a visual large model, and belongs to the technical field of railway safety monitoring. Comprises: constructing a data set; based on the data set, training the visual big language model by adopting a LoRA method to obtain a fine-tuning visual big language model; obtaining human body key point information on the original video data frame, and generating a skeleton diagram; and inputting the skeleton graph into the fine-tuning visual large language model, and outputting description information of the image and a corresponding cue word. According to the method, the technical route of lightweight posture recognition and visual large language model analysis is fused, and the posture of the driver is captured by adopting a character key point detection tool; meanwhile, a visual large language model is introduced, and correlation analysis of operation behaviors and environment equipment is achieved by constructing a driving scene knowledge graph; the performance bottleneck of a traditional monitoring system is broken through, a technical template capable of being copied and popularized is formed, and the method has important practical significance for improving the intrinsic safety level of railway transportation.

The application provides a tab folding detection method and device, electronic equipment and storage medium. The method comprises: performing skeleton extraction processing on a cross-sectional image of a multi-layer tab to obtain a skeleton image of the multi-layer tab; performing merging processing on a broken connected domain in the skeleton image to obtain a merged connected domain, the broken connected domain being a connected domain that is broken in the same tab cross section; according to the merged connected domain and a non-broken connected domain, counting a target number of the multi-layer tab; and according to the target number and a preset number, detecting whether a tab in the multi-layer tab is in a folded state. The broken connected domain is merged to obtain the merged connected domain, and according to the merged connected domain and the non-broken connected domain, the target number of the multi-layer tab is more accurate, thereby making the tab folding detection more accurate.

The application provides a human motion sequence prediction method and system based on skeleton-enhanced Transformer, comprising the following steps: S1: sampling human motion to obtain original motion sequence data; S2: generating virtual joints based on real joints, adding the virtual joints to the sequence, constructing enhanced motion sequence data, and generating an enhanced skeleton graph; S3: inputting the enhanced skeleton graph into a graph convolution network, establishing the skeletal correlation between virtual joints and real joints, and optimizing feature extraction through a spatiotemporal attention mechanism; S4: decomposing the input motion sequence data and dividing it into a trend part and a residual part, wherein the trend part is used to capture the trend information of the overall motion, and the residual part is used to retain local details; S5: combining the extracted features with the trend part and the residual part to generate future motion sequence data to obtain a prediction result; the skeleton features are enhanced by simulating joint combinations and introducing virtual joints.

This application discloses a method, system, and device for detecting malicious code variants using control flow graph semantic matching, belonging to the field of computer network security technology. It includes: disassembling the target executable file, constructing a control flow graph, and extracting basic block semantic features to generate semantic vectors; obtaining semantic equivalence classes based on semantic vector clustering, identifying main anchor points, and constructing a main anchor point skeleton graph; performing semantic compression and noise processing with the anchor point skeleton as constraints, and obtaining a semantic core graph through topological verification; hierarchically normalizing the semantic core graph and the main anchor point skeleton graph to generate a composite hash signature; and matching the composite hash signature with a malicious code family signature library to complete variant determination. The above scheme, through the main anchor point constraint compression and normalization process, resists structural obfuscation disturbances, improves the consistency of homologous variant identification, and has advantages such as strong robustness, high efficiency, and good interpretability. It is suitable for detecting malicious code variants using control flow graph semantic matching in complex obfuscated environments.

The invention discloses an epidemic prevention robot path planning method fused with a Voronoi skeleton diagram RRT algorithm. The method comprises the following specific implementation steps: step 1, acquiring and connecting equidistant points among map obstacles to construct a Voronoi skeleton diagram, and storing the Voronoi skeleton diagram locally; 2, acquiring an initial heuristic path on the skeleton diagram by adopting an A * algorithm; step 3, based on the initial path node, subdividing the global path into a plurality of sub-paths, and guiding the RRT algorithm to extend to sub-target nodes one by one; step 4, constructing an elliptical constraint channel for each sub-path to perform non-uniform sampling; 5, redundant exploration of an RRT algorithm is avoided in combination with a gravitational field bias strategy; and 6, after the planning is completed, sequentially connecting each sub-path as a result path, and smoothing a final path by using a multi-segment pruning strategy. Compared with an existing algorithm, the method has the advantages of path smoothness, high operation efficiency and environmental adaptability, and the epidemic prevention robot can stably and quickly execute a planning task.

The invention discloses a construction settlement dynamic monitoring method based on data analysis, and relates to the technical field of construction engineering monitoring and data-driven modeling, and the method comprises the steps: combining a causal diagram modeling method with a multi-source data analysis mechanism, achieving the structural modeling and dynamic correlation recognition of multi-level observation variables in a construction settlement region, and achieving the dynamic monitoring of the construction settlement. According to the method, a causal structure is updated in real time under multi-source heterogeneous data flow through a fast conditional independence test algorithm in combination with a modular skeleton diagram updating mechanism, logic consistency of a monitoring model is kept, and a multi-objective function optimization algorithm is combined with DAG constraint and a direction confidence coefficient matrix, so that a multi-source heterogeneous data flow is optimized. According to the method, automatic judgment and global optimal causal direction reasoning of conflict causal relationships are achieved, a priori knowledge attenuation function is combined with a data-driven confidence fusion model, time sequence dynamic evaluation of causal edge reliability is achieved, and self-evolution of a knowledge system is achieved.

The invention relates to the technical field of automatic fruit picking, and particularly discloses a skeleton-guided fine-tuning-based longan picking point identification system and method, and the method comprises the steps: introducing a region attention mechanism into a C3K2 module of YOLOv11 to form a C3K2AAttn feature extraction module, inputting a feature map into a YOLOv11 network, carrying out the deep feature extraction through the C3K2AAttn feature extraction module, and carrying out the recognition of a longan picking point. Performing multi-direction region division on the feature map, calculating the attention weight of each region, integrating multi-direction features through a weight fusion layer, outputting an enhanced feature map, and on this basis, taking a branch binary skeleton map as geometric prior information, splicing high-level semantic features of the backbone network and the binary skeleton map in a channel dimension, and obtaining a multi-directional feature map; after region attention enhancement feature expression, a 3 * 3 convolutional layer is input to generate a biased field, and the biased field is utilized to dynamically calibrate an initial prediction picking point, so that a prediction point returns to a real branch communication region, and the accuracy of automatic longan picking is improved.

The application discloses a photovoltaic power station monitoring method and system based on three-dimensional image modeling and virtual simulation fusion, relates to the technical field of image modeling, and comprises the following steps: generating a reflection response graph according to an original inspection graph of a photovoltaic power station, sequentially performing adaptive threshold segmentation and morphological processing on the reflection response graph to obtain a candidate reflection graph of a screen; performing skeleton extraction on the candidate reflection graph of the screen to obtain a candidate highlight skeleton graph, screening out attached edge pixel points according to a skeleton pixel point set in the candidate highlight skeleton graph, connecting the attached edge pixel points to obtain an edge traction sliding graph; generating a basic outer adjacent band according to a three-dimensional model of a station of the photovoltaic power station, performing region merging on the basic outer adjacent band to obtain a grounding conductor layout band, and performing projection on the grounding conductor layout band to obtain a virtual screen corridor projection graph; and the application improves the reliability of a photovoltaic power station monitoring result.

The invention discloses a real ink mark driven facsimile authentic method based on submerged space diffusion, which comprises the following steps: firstly, collecting a target facsimile image and a large number of real ink mark sample images, carrying out preprocessing operation on the images, and constructing a standardized training set and a test set; secondly, extracting a contour diagram and a skeleton diagram from the target copy image, and mixing the contour diagram and the skeleton diagram in a linear mode to generate a continuous structural feature map; constructing an ink mark encoder, encoding the authentic mark sample, and extracting the pen and ink characteristics of the authentic ink mark sample; and finally, constructing a reinking network based on a latent space diffusion model, and finally converting the optimized data representation into a pixel image through a decoder to obtain a calligraphy image with natural ink marks and real strokes. The problems that in the prior art, copy images are inflexible in ink color and lack of natural ink color change are solved, meanwhile, stability of a font skeleton and refinement of the stroke of the to-be-restored area are guaranteed, and authentic transformation work is achieved.

The application discloses a kind of fruit tree image acquisition method and system based on two-dimensional skeleton and viewpoint planning, it is related to wisdom agricultural equipment technical field, method includes: obtaining crown image obtains branch mask and depth information;To mask thinning pruning obtains initial skeleton, identifies feature point and calculates growth direction after artifact removal;Based on feature point and direction vector, topological reconstruction is carried out to broken skeleton, and complete two-dimensional skeleton graph is obtained;Based on two-dimensional skeleton graph, viewpoint planning is carried out in reference depth plane, generates viewpoint sequence and is mapped into three-dimensional pose, realizes autonomous image acquisition.The application solves the problem of information loss caused by occlusion and three-dimensional reconstruction calculation lag, and realizes real-time, safe and high-resolution image acquisition.

The invention provides a multi-person whole body grid three-dimensional reconstruction method and system based on hierarchical query and graph convolution refinement. The method comprises the following steps: performing parallel prediction on parameters of each part of a human body from coarse to fine by using a hierarchical query decoder; aggregating all the preliminary attitude parameters into a human body global skeleton graph structure, and carrying out whole body attitude collaborative optimization through an attitude graph refinement device based on a graph convolution pyramid to obtain an optimized whole body fusion parameter set; and inputting the optimized whole body fusion parameter set, the shape features and the facial expression parameters into an SMPL-X human body model to carry out three-dimensional human body grid reconstruction. According to the invention, the natural dependency relationship of each part of the human body is explicitly captured from coarse to fine; the problem of space decoupling of body parts is solved while high efficiency is kept, and the structural consistency of attitude estimation is remarkably improved; and through explicit modeling of a dependency relationship between joints and fusion of multi-scale features, distortion of a local attitude is effectively corrected, and the method has excellent robustness and accuracy.

The application discloses a method, device and medium for establishing a Parkinson's disease gait impairment evaluation model. The method for establishing the model comprises the following steps: pre-processing a gait video, extracting a skeleton graph sequence and a multi-cycle gait energy graph; using the skeleton graph sequence and the multi-cycle gait energy graph, combining a gait impairment score label obtained by a doctor according to gait video diagnosis, training a neural network to obtain a Parkinson's disease gait impairment evaluation model, wherein the gait impairment score label is used to divide the gait impairment degree. The application uses the contour and joint extracted from the gait video to generate two different morphological data streams of the gait energy cycle graph and the skeleton graph, and combines neural network training to generate a corresponding neural network model. The neural network model can analyze the gait video of a Parkinson's disease patient to be evaluated to obtain a gait score. Based on the gait score, the gait impairment degree of the Parkinson's disease patient can be evaluated conveniently and quickly, and the accuracy is high.

The application discloses a multi-robot cooperative exploration method based on a real-time environment skeleton graph and related equipment, and the method comprises the following steps: each robot acquires sensing data of itself, constructs a grid map according to the sensing data, performs binary processing on the grid map to obtain a binary map, a server constructs a global binary environment map based on the binary maps uploaded by all the robots, extracts an environment skeleton graph based on the global binary environment map, performs structural optimization on the environment skeleton graph to obtain an optimized environment skeleton graph, and sends the optimized environment skeleton graph to each robot; each robot performs state detection based on the optimized environment skeleton graph, and if the robot falls into a skeleton dead end or is stuck, the robot starts an escape strategy; each robot performs path planning and outputs an optimal exploration target and a path; and each robot executes a next exploration action according to the optimal exploration target and the path until the exploration task is completed.

The application belongs to the technical field of behavior recognition, and particularly relates to a human behavior recognition method and system based on dynamic space-time modeling and semantic quantification, which comprises the following steps: acquiring human skeleton sequence data and converting the data into space-time feature representation; constructing a Transformer network based on dynamic space-time modeling and semantic quantification, which comprises a preprocessing layer, a progressive feature extraction layer and a classifier; the preprocessing layer generates initial space-time features; in the progressive feature extraction layer, the joint importance weight is adaptively assigned by a global dynamic joint weighting module in the shallow stage, and the global dynamic joint weighting and a semantic quantification module are simultaneously used in the deep stage, the continuous features are discretized into semantic indexes by using a learnable motion primitive codebook, a semantic graph is constructed based on hard assignment and is fused with a physical skeleton graph, and the differentiable feature reconstruction is performed based on soft assignment; finally, the semantic enhanced features are input into the classifier to complete behavior recognition. The application improves the precision, interpretability and generalization ability of behavior recognition.