56results about How to "Clear boundaries" patented technology

The invention relates to the technical field of optical fiber wireless communication and mobile forward transmission, in particular to a digital analog optical fiber wireless forward transmission transmitting end parameter self-adaptive optimization method and system based on reinforcement learning. Decomposing the normalized wireless signal into digital parts of all levels and a final analog residual error part through multi-level modulation, and transmitting the digital parts and the final analog residual error part in a time division multiplexing manner; a receiving end represents inter-symbol interference caused by link nonlinear distortion and bandwidth limitation, and the signal-to-noise ratio of a recovery signal is calculated based on the error vector amplitude. Parameters of a transmitting end are optimized and modeled into a Markov decision process, states comprise rounding factors, scaling factors, constellation forming factors and pre-equalization filter tap coefficients of all levels, and actions are discrete increments of the parameters; a reinforcement learning agent is constructed by using a dual deep action value network, and training is updated in combination with experience playback and a target network. The reward function comprises a threshold satisfaction item and a signal-to-noise ratio change item; parameters are interactively updated in the training stage, and online self-adaptive adjustment is performed in the reasoning stage.

The application relates to a boundary attention guiding-based encoder-decoder medical image segmentation system and method, and belongs to the technical field of medical image processing.The application comprises an encoder, a decoder, a feature enhancement module and a boundary attention guiding module.Different scale features are obtained in the encoder by inputting a medical image, and the features are output to the decoder; the decoder completes medical image segmentation according to the features; the feature enhancement module is further connected after the output of the different scale features of the encoder, and is used for enhancing the semantic features of the encoder, combining semantic information of different receptive fields, and inputting the feature output of the encoder into the decoder again after the feature output of the encoder passes through the feature enhancement module; the boundary attention guiding module is arranged between adjacent decoder modules in the decoder, and is used for strengthening boundary information of the medical image.The application is excellent in generalization ability and robustness, and the obtained medical image segmentation result is clear in boundary and obvious in small target.

This invention discloses a data processing method and system for government services, relating to the field of data transmission technology. The method includes: obtaining jitter-related USB events from USB events of a government self-service terminal; obtaining jitter-related government business events from government business events of the government self-service terminal based on the jitter-related USB events; obtaining a disturbance chain based on the jitter-related USB events and the jitter-related government business events, and determining the identifier of the disturbance chain; defining a target disturbance chain based on the identifier of the disturbance chain, and obtaining a disturbance chain pointing label; classifying all government data units according to the disturbance chain pointing label to obtain disturbance-related government data units and secure government data units. This invention improves the reliability and controllability of the government data processing process.

The present application relates to the technical field of stage lighting, and discloses an effect device for shaping stage light effect, comprising a lamp body, the lamp body comprising a connecting top seat and a lamp shell body, a light source is arranged in the lamp shell body, an adjusting device is connected to the lamp shell body, the adjusting device comprising: an extension shell body; two fixed pipes symmetrically connected to the extension shell body; an optical lens, both ends of the optical lens being connected to an upper cavity and a lower cavity respectively, the optical lens being a hollow structure, and the inner walls of the upper cavity and the lower cavity being fixedly connected to impellers; a connecting pipe, respectively connecting the upper cavity and the lower cavity; wherein, one side end face of the optical lens is a one-way perspective glass, the other side end face is ordinary glass, two fixed pipes are respectively connected to a liquid pipe, the two liquid pipes are both connected to an external water tank, a water pump is connected to the two liquid pipes, and the two liquid pipes are arranged on the inner wall of the lamp shell body, solving the problems of high energy consumption and single effect of the existing equipment.

The invention relates to the field of geophysics, and provides a magnetic anomaly inversion method, device, equipment and medium, and the method comprises the steps: obtaining magnetic anomaly observation data corresponding to an underground target region, and dividing an inversion region of the underground target region; calculating a smooth inversion result corresponding to the inversion region through a pre-constructed inversion expression based on the magnetic anomaly observation data, and calculating a sparse inversion result corresponding to the inversion region based on the smooth inversion result; updating the inversion model in the inversion expression based on the difference value between the sparse inversion result and the L0 norm of the preset smooth approximation function, and returning to the step of calculating the smooth inversion result and the sparse inversion result corresponding to the inversion region until the condition of stopping updating is reached; and predicting the magnetic body distribution of the to-be-observed area based on the updated inversion model. According to the scheme disclosed by the invention, accurate underground magnetic field intensity distribution and a clear boundary inversion result can be obtained, and the stability and the accuracy of a magnetic anomaly inversion result are improved.

The utility model discloses a Mark point structure, which adopts at least two rectangular parts which are arranged in a crossed manner to form a bonding pad area, and compared with a traditional circular structure, the composite geometric shape of the bonding pad area can provide higher feature vector dimension in a machine vision system, the HOG feature value is improved by at least 30%, the algorithm identification accuracy is greatly improved, and the reliability of the structure is improved. Especially in a high-density BGA packaging area, the positioning precision of the center coordinates of the cross points can reach + / -0.01 mm, the identification degree of the Mark points on the PCB is greatly improved, and the problem that the Mark points are difficult to distinguish due to the low identification degree is effectively solved.

This invention discloses a multifunctional energy-saving block forming device and method, relating to the field of block forming technology. The invention includes a conveyor line with a base plate at its top and a frame mounted above it. Inside the frame is a lifting and lowering pressing head, with several inner pressing blocks fixedly installed at its bottom. Lifting and lowering outer pressing blocks are located on both sides of each inner pressing block. This invention, through the novel design of inner and outer pressing blocks, creates a trapezoidal sawtooth pattern on the joint surface of the composite brick. Compared to a flat interface that relies entirely on chemical bonding, the trapezoidal sawtooth structure forms a regular tenon-and-mortise interlocking mechanism at the interface, resulting in clear boundaries, reduced separation, and high composite pressing strength. The alternating lifting and lowering design of the inner and outer pressing blocks, combined with the longitudinal conveying and transverse spreading of the material frame, ensures five-sided isolation after the outer layer is pressed, minimizing interlayer mixing and maintaining clear boundaries.

The invention relates to an optical target based on a three-dimensional cavity type structure. The optical target comprises a black box body and a white plate, a round hole is formed in one side surface of the black box body, and a light absorption cavity is formed in the box body; the white plate is connected with the black box body and arranged in the round hole, and an annular area communicated with the light absorption cavity is formed between the white plate and the black box body. When light irradiates the optical target, one part of the light is reflected by the white plate, and the other part of the light enters the annular area around the white plate and then enters the light absorption cavity, and the light is gradually attenuated after being reflected for multiple times in the cavity, so that an extremely low gray value close to a theoretical black body is formed in the annular area; the boundary between the light spot formed on the white plate and the annular area is distinct, and a good gray difference is formed between the light spot and the annular area.

This invention discloses a virtual immunohistochemical staining method and system based on relation generative adversarial networks (GANs) and sequence slide data, comprising the following steps: Step S1, constructing a sequence slide dataset, wherein each H&E-stained slide is followed by three consecutive Ki67-stained slides; Step S2, constructing a relation generative adversarial network (GAN) model, the model comprising two ResNet generators and two relation discriminators, used to train and evaluate the credibility of the sequence relationship between image pairs; Step S3, performing two-stage training on the relation generative adversarial network model, namely, basic style pre-training and Z-axis alignment relation adaptation. This invention fundamentally solves the Z-axis alignment problem, resulting in higher image fidelity and pathological accuracy, significantly improving clinical application value, and providing new research resources and evaluation paradigms.

This invention discloses a method for preserving cherry blossom germplasm, belonging to the field of forestry breeding technology. The method includes: (1) cultivating rootstocks with 6-10 primary branches, each primary branch containing 3-4 secondary branches; (2) selectively pruning secondary branches before grafting, retaining 1-2 weak outer branches as nutrient branches; (3) grafting in batches, first grafting strong inner branches, and then grafting reserved nutrient branches after survival; (4) determining the number of scions according to the branch diameter gradient; (5) supporting field management measures. This invention, through the core technology combination of "graded branching + batch grafting + retention of nutrient branches", can preserve 6-10 germplasm from a single rootstock, with a germplasm preservation rate of over 95%, increasing land utilization by 7 times, and reducing rootstock cost by 87.5%. It solves the technical problems of traditional cherry blossom germplasm preservation, such as large land occupation, high cost, and unstable survival rate, and is particularly suitable for the construction of large-scale cherry blossom germplasm resource banks.

The invention relates to the field of paint spraying, in particular to a curved surface paint spraying boundary technology which comprises the following steps that S1, a first color layer is sprayed to a curved surface structure; s2, removing the first color layer on the second side of the color block boundary; s3, spraying a release interlayer on the curved surface structure, wherein the release interlayer is sprayed on the second side of the color block boundary and the surface of the first color layer; s4, removing the release interlayer on the second side of the color block boundary; s5, spraying a second color layer on the curved surface structure, wherein the second color layer is sprayed on the second side of the color block boundary and the surface of the release interlayer; and S6, removing the release interlayer so as to remove the second color layer sprayed on the surface of the release interlayer, and respectively forming a first color block and a second color block on the first side and the second side of the color block boundary. The multi-color paint spraying device has the effect of conveniently carrying out boundary multi-color paint spraying on the curved surface structure.

The application discloses a kind of diene extraction process saponification fat removal methods, first to the crude diene raw material is pretreated to remove low-boiling component, then carry out low-temperature weak base pre-saponification break down gum and waxy, middle-temperature compound saponification removes grease and fatty acid ester impurity in depth, after constant temperature stationary stratification, utilize compound solvent and carry out multistage countercurrent extraction, again by washing, reduced pressure desolventization, vacuum drying and obtain finished product.The application is by grading temperature control, step by step, effectively avoids emulsification, impurity removal effect is good, finished product purity and material yield significantly improve, reagent and three waste emissions are reduced.The process can be used for batch production, and can be matched with automatic equipment to realize continuous operation, process adaptability is strong, suitable for industrial scale popularization and application.

The application relates to a digital wave trap, a digital low-pass filter and a design method, a medium and an electronic device. The digital wave trap comprises a wave limit boundary anchoring unit and a wave limit zero point modulation unit which are cascaded in sequence, the wave limit boundary anchoring unit adopts a cascaded integral comb filter structure and is used for generating a first attenuation zero point at a boundary frequency of a target wave trap frequency band; the wave limit zero point modulation unit adopts a finite-length unit impulse response filter structure and is used for generating at least one second attenuation zero point in the target wave trap frequency band; and a continuous suppression band for the target wave trap frequency band is formed in the amplitude-frequency response of an output signal of the wave limit zero point modulation unit by the first attenuation zero point and the at least one second attenuation zero point. By using the digital wave trap provided by the application, a digital low-pass filter which can deeply and widely suppress power frequency interference and maintain strict linear phase and excellent low-pass characteristics can be designed without significantly increasing hardware resources and power consumption.

The utility model is applicable to the technical field of construction environment governance, and provides a spraying device for highway construction, which comprises an isolation frame for isolating a highway construction area, and an anti-slip mat for increasing friction force is arranged at the bottom of the isolation frame; the baffles are installed on the two sides of the isolation frame, and a storage bin used for storing water is arranged between the baffles; the liquid discharge pipe is fixed on the isolation frame and extends into the water storage bin; the water pump is arranged in the water storage bin and is used for pumping water for dust falling; and a connecting pipe is arranged between the liquid discharge end of the water pump and the liquid inlet pipe of the liquid discharge pipe. According to the scheme, the spraying device for highway construction is large in liquid storage range, rainwater can be collected for dust falling while frequent water supplementing by personnel is not needed, the spraying device can also serve as an isolation fence of a construction site, and the influence on the construction site is small.

The present application relates to the technical field of phenotype measurement in aquatic genetic breeding, and particularly relates to a high-throughput phenotype measurement method and system for small-sized Macrobrachium rosenbergii. In the measurement box, a dark box environment is constructed by a white shooting bottom plate, dark light-absorbing inner walls and an LED surface array light source, combined with preview graph rough segmentation and local gray histogram analysis, to adaptively adjust the light to make the shrimp body and the background form a stable bimodal gray distribution. On this basis, a segmentation-key point joint deep learning model is used to simultaneously output the shrimp body mask and standard key points, and the topological consistency loss constraint of the skeleton side length ratio and the included angle is introduced in the training to significantly reduce the key point mislabeling and drift caused by transparent body color. The present application realizes the rapid, non-contact, high-precision and high-throughput determination of the phenotype parameters and weight of small-sized Macrobrachium rosenbergii individuals, and provides a high-quality structured phenotype database for genetic evaluation and multi-trait joint selection.

The present invention provides a multilayer milk skin layered yogurt and a preparation method thereof, and relates to the technical field of yogurt, the multilayer milk skin layered yogurt comprises 80-90 parts of raw material milk, 1-10 parts of cream, 0-1.2 parts of a compound stabilizer, 0.01-0.05 part of a strain, 3-8 parts of an anti-acid fruit and vegetable sauce, and 1-8 parts of sugar, the yogurt has at least three layers, the milk skin is located on the upper layer, the middle layer is a yogurt base, and the bottom layer is the fruit and vegetable sauce. The anti-acid fruit and vegetable sauce and the compound stabilizer cooperate to realize fruit and milk co-fermentation fermented milk, multilayer milk skin layering weaning can be formed, the multilayer milk skin layering weaning layering effect is remarkable, the boundary is clear, the product taste is fine, smooth and smooth, the product has natural fruit fragrance and milk fragrance, and the product is visually bright in color and distinct in gradation, and has a good market prospect. The appetite feeling and the eating interestingness are greatly enhanced.

The invention discloses a multi-style differentiation graph lightening system and device of a sighting telescope, and relates to the technical field of sighting telescopes. Comprising a prism, a differentiation plate, a first LED lamp, a second LED lamp and a lens, the differentiation plate is provided with a first pattern and a second pattern which are concentric, and the surface of the second pattern is coated with a black shading coating. By independently controlling on-off and color switching of the two LED lamps, free switching of independent display of the first graph, independent display of the second graph and combined display of the first graph and the second graph can be achieved, and multi-color adaptation is supported. The device ensures light conduction and imaging effects by accurately assembling all parts. The dividers can be replaced to expand and differentiate pattern styles, the problem that a traditional sighting telescope is fixed in dividers and colors is solved, the dividers are adaptive to different sighting scenes, the sighting flexibility and accuracy are improved, and the dividers are suitable for various sighting telescope devices.

This invention discloses a semi-supervised remote sensing image change detection method and system based on neighborhood relationship propagation, belonging to the field of computer vision and remote sensing image processing technology. It includes: acquiring dual-temporal remote sensing image pairs and dividing them into labeled and unlabeled datasets; constructing a teacher-student framework semi-supervised model, calculating the squared Euclidean distance between the center pixel and its neighboring pixels, converting it into a relevance weight matrix using Softmax, and performing neighborhood-weighted aggregation on the initial prediction map to obtain a corrected prediction map; calculating the supervised loss using labeled samples, and calculating the unsupervised loss using unlabeled samples through weak perturbation to generate pseudo-labels and strong perturbation for consistency learning; jointly optimizing the student model parameters and updating the teacher model using exponential moving average; inputting the dual-temporal images into the trained student model, outputting a probability map, and generating a binary change detection mask. This invention enhances spatial consistency through neighborhood relationship propagation, improving detection performance when labels are limited.

This application relates to the field of zero-carbon determination technology, and provides a method for determining a zero-carbon watershed. The method includes: obtaining the annual average carbon emissions, annual average carbon sink, and annual average clean energy carbon neutrality contribution of a target watershed; and determining the target watershed as a zero-carbon watershed based on the annual average carbon emissions, annual average carbon sink, and annual average clean energy carbon neutrality contribution. This application enables the scientific determination of zero-carbon watersheds in a manner adapted to the characteristics of the watershed, thereby helping to clarify the boundaries and connotations of zero-carbon construction and realizing the systematic advancement and practical exploration of zero-carbon construction.

The invention relates to the technical field of biomedical engineering, and discloses an electromyographic signal simulation display method and system, and the method comprises the steps: carrying out the channel analysis and synchronous verification of a mixed signal sequence, and constructing a shunt signal which retains the time sequence correlation; extracting envelope features of the signals and performing trend fitting correction to generate initial intensity distribution; constructing an interference distribution diagram based on cross-correlation analysis, and determining an intensity correction value by using adaptive filtering and residual analysis; carrying out reverse reconstruction on the signal in combination with a tissue impedance model, and restoring real muscle activity distribution; and finally, mapping the distribution data to a visual model, and generating a muscle activity image through interpolation and sharpening processing. According to the method, the problem of low visual image accuracy in the prior art can be solved.

The invention provides a large-scale generation method and system for interpretable question and answer pairs for the vertical field, and relates to the technical field of data processing, and the method comprises the steps: 1, constructing a cross-modal heterogeneous meteorological knowledge graph; heterogeneous graph neural network modeling is carried out on the cross-modal heterogeneous meteorological knowledge graph, an embedded vector of each node is learned, a knowledge graph with node embedded vectors is obtained, semantic community division is carried out on an entity based on the embedded vectors, and an optimized knowledge graph is generated. According to the method, through knowledge graph optimization, conical semantic space construction, structured evidence chain retrieval and pre-training large language model generation, large-scale generation of interpretable meteorological question and answer pairs in the vertical field is realized, and interpretability and credibility of question and answer results are improved.

The embodiment of the application discloses a boiler burner multi-layer flame temperature field distribution reconstruction method, which comprises the following steps: setting up a temperature measurement and fire detection hardware and collecting multi-layer flame radiation and full-quantity operation parameters, fusing output spectrum and working condition fusion basic databases through time alignment, using a pre-calibrated coupling coefficient matrix for online iterative decoupling, outputting corrected temperature measurement reference data of each layer, constructing a CFD reduced-order model, correcting simulation deviation through a neural network, outputting high-precision in-furnace temperature field calibration data, three-dimensionally reconstructing the temperature field after layering and smoothing the high-precision in-furnace temperature field calibration data, establishing a prediction model based on historical three-dimensional temperature fields, and outputting final reconstruction data of the multi-layer flame temperature field; early abnormalities are identified, and the boiler is output through a standard interface to generate and execute coal quantity and air quantity control instructions. The problems that multi-layer flame coupling interference is not eliminated and temperature field reconstruction only depends on instantaneous static data are solved.

This invention proposes a global parameter configuration method based on parallel discrete event multi-process simulation. This method achieves bidirectional handshake interaction between the host configuration software and the simulation platform software through a Socket transmission interface, mutually verifying the initialization status and automatically restarting the corresponding software initialization process in case of initialization errors. The simulation platform software aggregates all process local parameters and reports them to the host configuration software, which then generates a unified format global parameter file. After the simulation platform software sends a distribution request, the host configuration software distributes the global parameter files sequentially according to the process number. The simulation platform software collects the file reception identifier and process parameter configuration identifier of each process for hierarchical judgment, distinguishing between file reception failure and parameter configuration abnormality, and performs targeted retries for repair. After all processes pass verification, the parallel discrete event multi-process simulation is started. This invention solves the problems of complex global parameter configuration, difficult synchronization, and cumbersome fault handling in existing parallel simulations.

The application discloses an underground intelligent early warning cable and a preparation method thereof, and belongs to the technical field of cables. The cable comprises, from inside to outside, a core assembly, a wrapping assembly and an outer sheath assembly. The core assembly comprises at least one conductive wire core and a communication optical cable unit. The conductive wire core is in a twisted distribution, and the communication optical cable unit is arranged outside the conductive wire core. Each conductive wire core comprises a conductor, a temperature sensing optical fiber conductor unit and an insulation layer. A copper tape shielding layer is arranged outside the conductive wire core. Mica tape is arranged in the gap of the core assembly. The outer sheath assembly is low-smoke halogen-free polyolefin. Through the application, the communication optical cable unit and the temperature sensing optical fiber conductor unit are arranged, and the communication optical cable unit is arranged in the gap surrounded by the conductive wire core, so that synchronous transmission and backhaul of monitoring information are realized. Operation data can be quickly transmitted through the communication optical cable unit, and the problems of dispersed monitoring signal link, unstable transmission and feedback lag of traditional cables are avoided.

The application relates to the technical field of industrial anomaly detection, and discloses a rotation-invariant multi-prototype industrial anomaly detection method, which comprises the following steps: S1: obtaining an industrial image to be processed and a corresponding defect region frame-level label, and constructing a foreground mask and a background mask based on the frame-level label; S2: inputting the industrial image into a pre-trained feature extraction network, obtaining an initial feature map, and performing channel mapping and normalization processing on the initial feature map to obtain a normalized feature map; and S3: rotating the normalized feature map at least two different angles to obtain a feature map in each rotation direction. The rotation-invariant multi-prototype industrial anomaly detection method is based on a feature extraction network pre-trained on a large-scale data set, can make a heat map generation module converge only through a small number of training rounds, does not need to train a complex model from the beginning, and greatly reduces data demand, computing resource and training time cost.

This invention provides a method and system for cancer subtype identification based on multi-omics data, belonging to the field of bioinformatics processing technology. The method includes: constructing a DILCORE model that integrates a multi-branch variational autoencoder, contrastive learning, and cross-view attention mechanisms. The multi-branch variational autoencoder decomposes omics observation data into common components for subtype classification and view-specific noise components, achieving noise suppression; the InfoNCE contrastive loss is introduced to bring common representations of the same sample closer across different views, enhancing cross-omics consistency; the cross-view residual self-attention mechanism is used to adaptively weight and fuse common vectors; finally, a self-supervised clustering fine-tuning optimization strategy is introduced to jointly improve representation quality and clustering performance in the latent space. This achieves deep and effective integration of multi-omics data, significantly improving the accuracy of cancer subtype identification and providing a powerful tool for personalized cancer treatment and prognostic assessment.

This invention discloses a language model-driven zero-shot object detection method and system. The method includes: training a supervised detection model using visible class data from a dataset; extracting semantic vectors from a large language model based on data class names to generate external knowledge; extracting visual features of visible class images using the supervised detection model, and training a generative adversarial network (GAN) based on pseudo-visual features synthesized from external knowledge; using the GAN to synthesize pseudo-visual features for invisible class data, training an invisible visual feature classifier, and obtaining an updated supervised detection model through parameter fusion, thereby achieving zero-shot object detection of image data. Through the technical solution of this invention, visual features with rich semantics and discriminative power can be generated, improving the understanding and expression of visual content, better aligning the visual space with the semantic space, and solving the problem of semantic confusion in zero-shot object detection.

The invention discloses a microfilter combination automatic search heterogenous remote sensing image change detection method and system, and the method comprises the steps: S1, carrying out the feature fusion processing of two obtained heterogenous remote sensing images, and obtaining a fusion feature; s2, according to the fusion features, a convolution kernel in a discriminator is dynamically synthesized by adopting a microfilter combinatorial search strategy, and parameters of a generator and the discriminator are optimized through adversarial training; and S3, supervising the change atlas output by the generator based on geometric structure collaborative supervision loss to obtain a final change detection result. According to the method, the common texture and structural features in the optical and SAR images can be effectively extracted, the problems of large radiation difference and much noise of different-source images are solved, and the omission ratio and the false detection rate are remarkably reduced. Meanwhile, the pixel precision, the marginal definition and the global structure authenticity are coordinated, so that the detected change area is clear in boundary and complete in structure, and common salt and pepper noise and boundary blur phenomena are effectively eliminated.

The invention discloses an electrode reflection structure and a preparation method thereof, and the method comprises the steps: preparing a photoresist first chamfer and an undercut with a preset length on a transparent conductive layer, vacuumizing an operation environment to 10 <-6 > Torr or above, and sequentially carrying out the evaporation of a metal adhesion layer and a metal reflection layer on the transparent conductive layer; evaporating a Ti stress buffer layer on the metal reflecting layer, then evaporating a Pt barrier layer on the Ti stress buffer layer, and controlling the thickness of the Ti stress buffer layer to be at least 1.5 times greater than the thickness of the Pt barrier layer to form a Ti / Pt metal barrier layer; and evaporating a covering layer on the outer surface of the Ti / Pt metal barrier layer to complete the preparation of the electrode reflection structure. On the premise that the design size of the electrode is not changed, shielding and absorption of the electrode to emergent light of the chip are effectively reduced, the probability that photons escape from the active layer is improved, therefore, the brightness of the LED chip is improved, and meanwhile the stability and long-term reliability of the electrode structure are considered.

This application relates to a motion capture method, device, equipment, and medium for ice hockey passing and control training. The method includes: acquiring raw motion data from an inertial measurement unit (IMU), synchronizing and cleaning it to obtain a synchronization data matrix; constructing an initial spatiotemporal graph using IMUs as nodes and spatial relationships as edges; generating weight information using a LASSO feature selector, and weighting it to obtain an updated spatiotemporal graph; inputting the updated spatiotemporal graph and weights into a graph neural network for synchronous inference of motion segmentation, category, and quality assessment information; decoding the motion boundaries and category labels using the Viterbi algorithm; inputting the data into a scoring regression model to calculate a technical quality score, and outputting the motion analysis results. This method can accurately capture ice hockey passing and control motions, achieving integrated motion segmentation, classification, and quality assessment, thus improving training analysis efficiency.