66results about How to "Improve certainty" patented technology

The invention discloses a fermentation strain yield prediction method based on artificial intelligence, particularly relates to the field of biological fermentation production, and is used for solving the problems that in the prior art, due to the fact that the yield is difficult to directly map in concentration prediction, monitoring indexes are normal, the final yield deviates from expectation, and reliable strain comparison and production management decision cannot be supported. A historical batch process observation sequence and a material account sequence are aligned to generate a conversion efficiency characterization sequence to form a yield learning sample training yield prediction model, and the conversion efficiency characterization sequence is generated on batches and input into the model to obtain a yield prediction value. And calculating an accounting volume trajectory consistency parameter and a conversion contribution steady-state proportion parameter, inputting the parameters into a coefficient analysis model to obtain a decision credible coefficient, generating a yield risk judgment result for a yield prediction value based on the decision credible coefficient, and realizing a new-batch online prediction record.

The invention is suitable for the technical field of vehicle dynamics modeling and control, and provides a preview active suspension stiffness control method based on road surface grading and MAP mapping, and the method comprises the steps: collecting a road image through a vehicle-mounted visual sensor, dividing regions, and recognizing the road surface unevenness grade of each region through a convolutional neural network; obtaining the vehicle speed; and according to the current pavement grade and the vehicle speed, a vehicle speed-pavement grade-rigidity mapping MAP table generated in advance is inquired, and the target suspension rigidity is rapidly obtained and adjusted. According to the invention, the dependence on a high-precision sensor and a high-computing-power processor is reduced, and low-cost and fast-response rigidity optimization control of the preview active suspension is realized.

The invention relates to the technical field of time-triggered network communication, in particular to a time-triggered Ethernet scheduling generation and synchronization system based on a hardware time axis construction engine. According to the system, a time axis structure consistent with global synchronization time is constructed on a switch side through hardware, sending time and a sending sequence of a time trigger frame configured by a user are directly mapped into a resource occupation relation on a time axis, and scheduling window mapping, conflict detection and window shaping are completed on a hardware level; therefore, a deterministic scheduling table is generated; through a double-table storage and time alignment switching mechanism, undisturbed updating and synchronous effectiveness of the dispatch table between the switch and the end system are realized. Compared with a traditional scheme of generating a scheduling table depending on a software algorithm, the method does not need a complex scheduling solution process, supports manual controllable configuration of time-triggered frame scheduling, and improves the certainty, autonomy and engineering realizability of the system.

The invention discloses a network dynamic flow scheduling method, device, equipment and product, and relates to the technical field of network scheduling, and the method comprises the steps: monitoring the change of target parameters of the equipment in real time through obtaining the state data and historical data of each piece of equipment in real time, and analyzing the behavior and flow trend of the equipment. The method comprises the following steps: calculating a Q value of each device through a Q learning algorithm to obtain an optimal Q value of each device under a current task, calculating an optimal path and a time slot allocation result of each device under the current task according to the optimal Q value, and scheduling a network dynamic flow according to the optimal path and the time slot allocation result to ensure that the path is shortest and the resource allocation is most reasonable, thereby improving the network transmission efficiency. And the risks of time delay and network congestion are reduced. The dynamic flow scheduling mechanism can flexibly adjust the flow load of each path in the network according to the real-time data and the historical trend, thereby ensuring the efficient utilization of network resources, and improving the certainty of equipment in the dynamic network.

The invention provides a processor synchronization system and electronic equipment, and relates to the field of digital integrated circuits, the processor synchronization system comprises a plurality of execution engines, each execution engine is configured to execute an instruction stream, and the instruction stream comprises a synchronization mark instruction used for indicating a dependency relationship across the execution engines; each synchronous control module is embedded in the corresponding execution engine, and each synchronous control module is configured to generate a synchronous request signal or a synchronous event signal according to the current state of the execution engine in response to the execution of the execution engine to the synchronous mark instruction; and the inter-engine synchronization module is coupled with the plurality of execution engines to receive the synchronization request signals and the synchronization event signals and return synchronization completion signals to the corresponding execution engines when the dependency relationship is detected to be met, so that all explicit state access operations can be eliminated, and the performance of the execution engines is improved. And conversion of a synchronization mechanism from software explicit scheduling to hardware implicit response is realized.

This application relates to the field of integrated circuit design technology and discloses an on-chip network deadlock prevention method and system based on path reservation and credit backhaul. The core of the method lies in proactively locking virtual channel resources along the entire path via an independent reservation request packet along an adaptive route before data packet transmission. Combined with a "credit backhaul confirmation" mechanism, data packets are only allowed to be injected into the network after the source node receives an confirmation packet, thus transforming resource dependencies into unidirectional credit flow control. Simultaneously, reservation resource management is achieved by setting a distributed reservation state table in each router, and soft reservation and multi-path exploration mechanisms are introduced to improve resource utilization efficiency and path selection flexibility. This scheme fundamentally breaks the conditions for circular waiting, completely avoids deadlock, and eliminates the performance bottleneck of traditional methods when virtual channel resources are limited, achieving high determinism in data transmission, low blocking risk, and high efficiency in resource utilization.

The application relates to the field of heterogeneous bus communication and discloses a heterogeneous bus fault self-adaptive retransmission decision method and system, which solves the problem that secondary congestion and semantic fault are caused by the incapability of existing static retransmission to perceive cross-domain states. The method comprises the following steps: performing polymorphic fault fusion to output a comprehensive fault judgment signal; realizing standard scalarization and consistency checking of heterogeneous parameters through cross-domain state synchronization and a mapping table; calculating a real-time communication quality index, using clustering decision to determine an adaptive retransmission interval, number and path; performing semantic disassembly on cross-protocol domain data to build a data semantic structure tree, generating a compensation retransmission instruction and interweaving the instruction with a time sequence to issue the instruction. The application realizes a closed loop from perception to semantic compensation, improves communication reliability and can be applied to communication system equipment manufacturing.

The invention discloses a time efficiency label generation method and device and a storage medium, and relates to the technical field of logistics time efficiency, and the time efficiency label generation method comprises the steps: responding to a time efficiency label obtaining request, and determining whether the client time corresponding to the time efficiency label obtaining request is before the preset time; and if the client time is before the preset time, constructing query input based on a commodity identifier corresponding to the timeliness label acquisition request, and querying the white list data in the memory cache. When the subitems in the white list data are hit, the shipping warehouse information and the receiving place information corresponding to the timeliness label acquisition request are assembled into a key, and the precomputation state of the target route is inquired in the Redis cache based on the key assembled by the shipping warehouse information and the receiving place information. And if the pre-calculation state is normal, the next day arrival tag is used as a response of the timeliness tag acquisition request and is sent to the client. According to the method, the technical effect of high-robustness and high-certainty time-efficiency label calculation capability of the million-level QPS can be realized.

The application relates to the technical field of industrial automation, and provides a device testing method and system, an electronic device and a computer readable storage medium, the method comprises the following steps: being applied to a device testing system, the device testing system comprises a plurality of devices to be tested, each device to be tested is connected with a physical port of a switch, the switch is connected with a test server, a structured configuration file is analyzed to obtain a device mapping relationship. Based on the device mapping relationship, a plurality of virtual network interfaces are created, and a corresponding network protocol address is allocated to each virtual network interface. An independent routing table is established for each virtual network interface, and a source network protocol address of the virtual network interface is bound to the routing table. Address resolution protocol response control parameters are set for each virtual network interface. A plurality of logically isolated virtual network interfaces can be automatically created, an automatic configuration driven testing environment can be built, and the testing efficiency is improved.

The invention provides a method for highlighting texture features based on multi-focus images. The method comprises the following steps: step 1, collecting multi-layer focus sequence images of fiber fragments; step 2, carrying out corresponding position pixel accumulation on the acquired multi-layer focusing sequence image and then averaging to obtain an average image; 3, calculating the definition of each layer of image; 4, calculating the sum of the sharpness of each multi-layer graph interval, and finding out the optimal multi-layer graph interval with the maximum value; step 5, respectively subtracting pixels at corresponding positions of the average image from each layer of image in the multi-layer image interval, and taking absolute values to obtain a difference image sequence of each image to the average image; and step 6, adding and normalizing all the difference images to obtain a final texture enhancement image. The method has the technical effects of good texture highlighting effect, good certainty, high efficiency and good robustness for the fiber image in the microscopic imaging state, and can be used for an image preprocessing stage of a qualitative and quantitative analysis system for textile fiber types and components.

This invention provides a method for predicting the dominant frequency of blasting vibration by integrating parameter fluctuation processing and the Dream Optimization Algorithm (DOA), belonging to the field of underground engineering safety control technology. The system includes a data acquisition and uncertainty modeling module, a data preprocessing and feature construction module, a hyperparameter optimization module, and a prediction model training and result output module. The method obtains relevant parameters through on-site investigation and monitoring, and generates an extended sample set using a combination of probabilistic perturbation modeling and fuzzy triangular modeling. The data is preprocessed and features are selected. The DOA algorithm is used to globally search the hyperparameters of the Support Vector Regression (SVR) model, and the optimal hyperparameter combination is selected by combining a robustness fitness function. The SVR model is trained, and prediction results and uncertainty prediction intervals are generated. This invention can explicitly characterize the uncertainty of geological parameters, achieve efficient global optimization of hyperparameters, improve prediction accuracy, robustness, and generalization ability, and is applicable to different geological conditions and blasting scenarios. It can be extended to various blasting dynamic response prediction tasks.

The invention relates to a cluster distributed virtual interrupt control method, device and system. The method comprises the following steps: under the condition that a first distributed virtual interrupt controller DVIC node receives a virtual interrupt request, determining interrupt target routing information corresponding to virtual interrupt; the first DVIC node constructs a cross-chip hardware-level direct injection message according to the interrupt target routing information; and the first DVIC node sends a hardware-level direct injection message to the second DVIC node through inter-chip interconnection, and the hardware-level direct injection message directly writes a virtual interrupt into an IMSIC interrupt file of an incoming message signal interrupt controller through hardware logic by the second DVIC node. According to the invention, the cross-chip virtual interrupt delivery delay can be obviously reduced, and the system certainty is improved.

The invention relates to the technical field of data processing of multi-source data fusion, and discloses an investigation data integration analysis method and system based on multi-source data fusion. According to the scheme, a multi-source sensor is deployed, original voltage is collected, and standardization is carried out in each physical quantity range; constructing temperature, pressure and vibration three-dimensional observation and states based on a unified time index, and setting initial values; recursive prediction is carried out by using a constant preserving model; counting a multi-measuring-point variance in a static window, determining observation noise in combination with the minimum quantization step size, and calculating fusion gain; linearly updating and correcting covariance, and then carrying out dimension-by-dimension amplitude limiting according to an abnormal upper limit; carrying out space averaging on the three-dimensional results of the multiple measuring points at each sampling moment, and carrying out rollback if necessary; and a fusion sequence and a statistical report are formed in the whole cycle. According to the method, the data scale is unified, subjective parameters and computing power burden are reduced, adaptive noise modeling along with the environment is achieved, stable noise suppression and fault resistance are achieved, and continuous output and comparable analysis are guaranteed.

The present application belongs to the technical field of geological engineering monitoring, and particularly relates to a geological stability monitoring device for salt cavern energy storage, aiming to solve the problems of single monitoring dimension, existing blind area and insufficient sensitivity in identifying creep. The monitoring device is internally provided with a geological stability monitoring system, and the system comprises: a surface multi-source deformation sensing unit, a shaft deep layer strain detection unit, a microseismic global activity monitoring unit and a cavity shape sensing unit, which are used for collecting surface and deep multi-dimensional data; a geomechanics digital twin and evolution analysis unit, which constructs a mechanics model and dynamically inverts real-time data as a constraint boundary; a stability discrimination and risk early warning scheduling unit, which outputs an evaluation report based on the inversion result and triggers an early warning. Through the deep coupling of multi-dimensional full-space perception and mechanics model, the present application effectively reduces the perception blind area and improves the accuracy of salt rock creep risk prediction, thereby providing a technical guarantee for the long-term safe operation of the salt cavern energy storage power station.

The invention discloses a dual-network-port EtherCAT master station redundancy communication method and device, and an EtherCAT master station runs in a bare computer environment without an operating system, and is configured with a master station redundancy management layer, a first physical Ethernet port and a second physical Ethernet port; the method comprises the following steps: in each EtherCAT communication period, controlling a first physical Ethernet port and a second physical Ethernet port to synchronously send the same EtherCAT data frame through a master station redundancy management layer; receiving and acquiring work counter WKC values in response frames returned by the first physical Ethernet port and the second physical Ethernet port respectively; the communication link state of the EtherCAT master station is judged according to the comparison result of the sum of the WKC values returned by the first physical Ethernet port and the second physical Ethernet port and the total number of the slave stations; and when the communication link is abnormal, the communication path of the EtherCAT master station is regenerated based on the pre-stored slave station connection information.

The invention relates to the field of material mechanical property testing, and provides an in-service material property detection system based on an indentation method, comprising: an excitation module for applying a composite dynamic excitation load to a pressure head when the spherical pressure head is pressed into a to-be-detected material; the signal acquisition module is used for synchronously acquiring a dynamic load signal and a dynamic displacement signal in the pressing-in process of the pressing head; the data processing module is used for calculating an instantaneous phase difference according to the dynamic load signal and the dynamic displacement signal, constructing an evolution curve of the instantaneous phase difference along with the change of the pressure depth, and extracting a first characteristic depth, a second characteristic depth and a phase hardening index from the evolution curve; and the performance resolving module is used for calculating the elastic modulus and the yield strength of the to-be-tested material according to the first characteristic depth, the second characteristic depth, the phase hardening index and the mapping relation model. According to the method, the plastic hardening constitutive relation of the material does not need to be preset, rapid and micro-damage detection of the tensile property of the material can be achieved through single-time micron-order pressing-in, and the method is particularly suitable for field evaluation of in-service equipment.

The embodiment of the application relates to the technical field of medical treatment, and discloses a focused ultrasound system and a tumor ablation device based on PET / CT functional image closed-loop control; the system comprises: a focused ultrasound unit used for emitting focused ultrasound energy to a target area to perform an ablation operation; a PET / CT imaging unit used for acquiring a PET / CT image containing anatomical structure information and metabolic function information; a control unit controls the focused ultrasound unit to perform a first round of ablation; after the first round of ablation, the PET / CT imaging unit is controlled to scan, and residual metabolic activity areas are extracted and detection confidence is calculated based on the scanning result; if the detection confidence is higher than a preset effectiveness threshold and it is determined that a treatment endpoint condition has not been reached, a supplementary ablation plan containing a supplementary target point set is generated; driving instructions are generated based on the supplementary ablation plan, and the focused ultrasound unit is controlled to perform the supplementary ablation plan, so that active residual foci can be found during an operation, and the determination of one-time complete ablation is improved.

The invention provides an AI intelligent calculation platform reasoning acceleration method and system, and relates to the technical field of artificial intelligence and deep learning. Retrieving cache records of historical memory access data in the to-be-retrieved data set, wherein the historical memory access data correspond to the query normal form to which the query request belongs; constructing a query logic tree according to the retrieval query request to obtain a cache hit probability of each piece of vector data in the to-be-retrieved data set; according to the query complexity and the cache hit probability, determining the partitioning granularity and the memory access step length of the vector data in the data set to be retrieved; establishing a mapping relationship between the query logic tree and the storage address to obtain target address space distribution; constructing a tensor layout mapping table matched with the query logic tree according to the block granularity; based on the tensor layout mapping table and the memory access step length, reasoning acceleration is realized, and the reasoning throughput and the real-time response efficiency of a large language model on an intelligent calculation platform are improved.

The invention belongs to the technical field of software automation testing, and provides an application state snapshot detection and bifurcation execution method and device, equipment and a medium, and the method comprises the steps: distributing instance identifiers and session identifiers for virtual instances of a running target Android application, and constructing a mapping table of the instance identifiers and the session identifiers; according to the mapping table, an incremental snapshot and fingerprint technology is adopted to collect snapshot content of the virtual instance according to the triggering opportunity, the snapshot content is stored in an associated mode according to the mapping table, and a snapshot generation process is determined through a session access index table; performing state recovery and operating environment isolation on the bifurcation instance according to the photo generation process; and creating a plurality of sub-process instances for each bifurcated instance, and debugging the bifurcated instances according to the debugging strategy and the modification configuration to obtain a debugging result. The method has the beneficial effects that the abnormal points can be carried and reproduced, and the detection accuracy and efficiency of faults, especially even faults, are improved.

The application relates to the technical field of intelligent robots, solves the problems that high-end service robots in the prior art are excessively dependent on high-computing-power algorithms, leading to high cost and complex operation, and low-cost intelligent devices are single in function and lack of control logic cooperation, and cannot meet the home care needs of the elderly, and in particular relates to a home aging intelligent care auxiliary robot, which comprises a main control unit in the robot, an offline voice interaction module connected with the main control unit and used for analyzing natural language voice instructions in the state without external network connection, and an environment sensing module arranged on the robot and used for collecting state parameters of the environment where the robot is located. The application can complete the analysis of the natural language and the mapping of the control instructions in the state without the dependence on external internet connection, reduces the dependence of the system on complex cloud big model calculation and external network, greatly reduces the cost of the whole machine manufacturing and home deployment, and improves the safety of the robot in caring for the elderly at home.

The application discloses a multi-link intelligent switching method, which comprises the following steps: starting link state monitoring through a server, periodically starting an adaptive learning mechanism after being online on a wireless link; periodically sending a probe message through each wireless link, and a client replying a response message carrying a request serial number, a client unique identifier and corresponding link index information; introducing a multi-dimensional weighted scoring mechanism to calculate the comprehensive scores of each link and sort them based on the information reported by the client and comprehensive scoring weights; selecting the link with the highest score as a primary link; when the comprehensive score of a standby link is higher than that of the current primary link, or when the primary link is not responded within a timeout period, triggering link switching to the standby link with the highest score, sending a switching instruction through a link notification mechanism, and executing switching by the client and feeding back a confirmation; after learning, adjusting the comprehensive scoring weights according to the adaptive learning result, and updating the adjusted comprehensive scoring weights to the link scoring calculation.

This invention belongs to the field of check valve technology, specifically a float-type check valve, including a valve body, a mounting base, a float housing, and a hollow float. The float housing is rotatably assembled via a spherical seat on the mounting base. A ball bearing is provided between the spherical seat and the ball head of the float housing, allowing the float housing to automatically adjust its verticality when the valve body is tilted, maintaining its central axis perpendicular to the horizontal plane. The float housing has a through hole communicating with the inlet, a spherical cavity accommodating the float, and a through groove communicating with the outlet. A branch channel extending to the valve seat sealing surface is opened on the inner wall of the through hole, which can assist in pushing the float open during forward flow. A multi-stage staggered-hole throttling disc is provided on the inner side of the outlet end to reduce the reverse flow velocity. The float housing adopts a split plug-in structure. This invention solves the problems of poor sealing, difficulty in opening under low pressure, and susceptibility to water hammer in traditional float valves installed at tilted angles. It has advantages such as reliable sealing, sensitive response, and adaptability to complex installation environments.

The application provides a kind of real-time payment authorization method, device, equipment and medium for enterprise fund flow, it is related to payment authorization technical field, comprising: receiving the original payment request initiated by user through associated terminal, and the enhanced payment request is generated in combination with the business context information associated with original payment request;Through shadow account engine, according to the target budget unit of the enhanced payment request to which it belongs, perform budget pre-freezing operation;Generate compliance determination rules matched with enhanced payment request, to use compliance determination rules to determine the compliance of enhanced payment request;When budget pre-freezing operation is successful and compliance determination passes, grant payment authority for enhanced payment request, to release the payment instruction corresponding to enhanced payment request to external fund settlement account.The application realizes the prior forced binding of budget control and payment action, and significantly improves the compliance response speed and the efficiency of industry and finance reconciliation.

The application discloses a power distribution network distributed self-healing control method and device based on event-driven middleware, which comprises the following steps: deploying a hierarchical software stack in a power distribution terminal, including a hardware abstraction layer, a hardware service layer, a public middleware layer and an industry middleware layer, and performing event-driven communication between layers through a message bus; the hardware service layer collects electrical quantities at a predetermined sampling period and publishes a sampling ready event; a protection state machine middleware subscribes to the event to trigger reading of the sampling value; a criterion engine middleware extracts three types of characteristic parameters, i.e., a frequency domain, an energy domain and a waveform domain, from a current signal in parallel, compares the three types of characteristic parameters with preset inrush current threshold values respectively, and counts the number of characteristics meeting the determination condition; when the number of characteristics reaches a preset threshold value, the inrush current is determined and the protection is locked, otherwise, the fault current is determined and a self-healing control process is triggered. The application replaces polling with event-driven, realizes accurate identification of inrush current and fault current through multi-dimensional feature majority voting, and effectively reduces the false alarm rate.

This invention discloses a vacuum furnace condition detection system based on pressure rise calculation, specifically relating to the field of industrial vacuum furnace sealing detection technology. It addresses the problem that existing methods, relying solely on single pressure change readings from bypass pipelines, cannot distinguish between the combined causes of valve internal leakage, pipeline venting, and external leakage, leading to uncertain fault location. This invention acquires the pressure time series of the bypass pipeline during the holding phase and the pressure decay time series after introducing a preset pressure gas. It constructs a decay phase trajectory and extracts the hysteresis loop envelope area as a decay dynamic feature. It also constructs a pressure rise phase trajectory and calculates the difference between its residence time percentage in each quadrant and that of the decay phase trajectory to obtain quadrant distribution asymmetry. This allows identification of whether the pressure change is dominated by valve internal leakage or pipeline venting. Furthermore, it assesses the internal leakage level for valve internal leakage and determines whether external leakage exists for pipeline venting, achieving the separation and quantification of the three causes: external leakage, internal leakage, and venting.

This invention relates to the field of computer security technology, specifically to a verifiable change access control method and system based on SET and dual atomic locks. The method includes generating a unique hash identifier for the request intent: intent_hash; constructing and issuing a certificate, which at least includes a secure execution type (SET), an expiration date, a replay protection field, a root of evidence, and a signature; binding the certificate to the intent_hash; performing expiration date checks through fast gating; and performing atomic occupancy operations based on the replay protection key. This verifiable change access control method and system based on SET and dual atomic locks achieves an end-to-end security closed loop, eliminating the possibility of execution bypass; providing industrial-grade consistency and replay protection guarantees through dual atomic operations and dual-channel consistency gating; constructing a self-verifying and tamper-proof chain of evidence, greatly improving auditability; forming a tiered defense-in-depth system, enhancing the overall system's robustness and resistance to censorship; and improving the system's determinism and observability.

The invention describes a device for evaluating sensor measurements (1.1), having: - a sensor (1), wherein for evaluating the sensor measurements (1.1) of the sensor (1) a model function suitable for least squares regression is provided, which can be defined by a parameter vector, wherein at least one parameter of the parameter vector forms a sensor output signal (3); and - a calculation and evaluation unit (2), having a neural network (2.1) for estimating the parameter vector on the basis of actually determined sensor measurements (1.1) and a least squares regression module (2.2), wherein the neural network (2.1) is trained with parameter vectors and associated sensor measurements, and the calculation and evaluation unit is set up to: ° determine at least one parameter estimate vector as an input variable for the least squares regression module (2.2) by means of the trained neural network (2.1) for sensor measurements (1.1) measured with the sensor (1), ° interrupt the least squares regression in the event that a convergence criterion is met when implementing the least squares regression, and ° output at least one parameter of the last determined parameter vector as a sensor output signal (3). An associated automated method for evaluating sensor measurements and the use of the device are also described.

The application discloses a continuous casting billet low-magnification subsurface crack digital rating method and relates to the technical field of continuous casting processes.The method comprises the following steps: acquiring the cross-sectional size parameters of a continuous casting billet, determining the subsurface crack evaluation area thickness and the risk critical distance based on the cross-sectional size parameters, acquiring the defect characterization data of the continuous casting billet, identifying the subsurface cracks in the subsurface crack evaluation area defined by the subsurface crack evaluation area thickness, and extracting the crack characteristic parameters, and performing combined grading based on the subsurface crack evaluation area thickness, the risk critical distance and the crack characteristic parameters to generate the subsurface crack rating result.The application takes the shortest distance d of the subsurface crack from the surface, the defect number n and the defect length l as the grading input, adopts the explicit threshold and the combination rule to output the 0-6 grade level G, and converts the subsurface crack evaluation from the experience judgment to the reproducible rule judgment.

The invention discloses a makeup extraction method and device, equipment, a medium and a product, and the method comprises the steps: firstly obtaining a target image, so that the target image comprises a first face and makeup used for modifying the first face; processing the target image by using a first model to obtain a first predicted makeup, so that the first predicted makeup can describe some makeup details in the target image; then, processing the target image and the first predicted makeup by using a second model to obtain a second predicted makeup, so that the second predicted makeup can describe makeup details in the target image more comprehensively and more accurately; and finally, according to the second predicted makeup, determining a makeup extraction result of the target image, so that the makeup extraction result can comprehensively and accurately describe the makeup in the target image, thereby facilitating improvement of a makeup determination effect.

The present disclosure relates to a cluster distributed virtual interrupt control method, device and system. The method comprises: in the case that a first distributed virtual interrupt controller (DVIC) node receives a virtual interrupt request, determining interrupt target routing information corresponding to the virtual interrupt; the first DVIC node constructs a cross-chip hardware-level direct injection message according to the interrupt target routing information; the first DVIC node sends the hardware-level direct injection message to a second DVIC node through an inter-chip interconnection, and the hardware-level direct injection message is directly written into an incoming message signal interrupt controller (IMSIC) interrupt file by the second DVIC node through hardware logic. The present disclosure can significantly reduce the cross-chip virtual interrupt delivery delay and improve system determinacy.