112 results about "Task segmentation" patented technology

The invention relates to the technical field of low-power-consumption AI processors, and discloses a load balancing method of a low-power-consumption AI processor, a chip and a storage medium, the method comprises the following steps: performing cache resource division on a plurality of neural mimicry cores in the low-power-consumption AI processor to obtain a double-layer cache structure; performing characteristic analysis on the target calculation task, and constructing a calculation task directed acyclic graph and a dependency intensity matrix; performing task segmentation to obtain a subtask division scheme; creating a task-core mapping table; organizing the neural mimicry core into a plurality of concentric ring layers based on the task-core mapping table to obtain a multi-layer ring cooperative computing network; and calculating a plurality of performance indexes of each neural mimicry core based on the multilayer annular cooperative computing network, and performing dynamic load optimization to obtain a dynamic load optimization strategy. According to the invention, the load balancing of the low-power AI processor is realized, and the calculation efficiency and the energy utilization rate are improved.

The invention relates to the technical field of data processing, and discloses a multi-granularity extraction and enhancement method, device and equipment for multi-source heterogeneous data. The method comprises the following steps: performing multi-level feature extraction on a multi-modal data set and original feature parameters to obtain a multi-modal feature set; performing correlation analysis on the multi-modal feature set to obtain a feature correlation matrix and optimized feature representation; performing cross-modal knowledge migration processing on the feature incidence matrix and the optimized feature representation to obtain migration feature parameters; performing dynamic fusion processing on the migration feature parameters to obtain target fusion features; and performing multi-task segmentation processing according to the target fusion features to obtain a plurality of sub-task sequences, performing time sequence enhancement and multi-task fusion analysis on each sub-task sequence to obtain target extraction data, and performing visual display on the target extraction data. According to the method and the device, the efficiency and the accuracy of multi-granularity extraction and enhancement for multi-source heterogeneous data are improved.

The invention discloses an information processing method and system based on the Internet of Things, and relates to the technical field of Internet of Things services, and the method comprises the steps: dynamically obtaining the computing power, energy consumption and network state of each node through an environment perception module, and constructing a node capability vector and a link performance vector; the task analysis module is used for carrying out structured modeling on tasks and identifying candidate segmentation points and unloading targets; the cost modeling module establishes a joint cost function of time delay and energy consumption; the strategy generation module generates an optimal unloading strategy based on multi-agent reinforcement learning; and the scheduling communication module executes task unloading and result integration according to the strategy, and feeds back execution data to update the system state. According to the method, intelligent, low-delay and energy-consumption optimization processing of task segmentation unloading is realized, and the method is suitable for a dynamic heterogeneous Internet of Things computing environment.

The invention relates to an end-side cloud collaborative intelligent optimization and resource scheduling method and system for an industrial Internet of Things. The method comprises the steps that sparse processing and quantization are conducted on model parameters on end equipment through accelerator hardware, the storage and transmission requirements of a model are reduced, and meanwhile the accuracy of the model is kept; through the knowledge extraction technology, the end equipment can extract key information in the data, and high-quality input is provided for subsequent data processing and model training; by deploying the lightweight reasoning model and the routing model in the edge device, the uploaded data can be preliminarily classified, and only the data which is difficult to process is sent to the cloud, so that the overall delay is reduced to the greatest extent while the accuracy is kept; performance monitoring, bandwidth allocation and task segmentation are carried out on end equipment and edge equipment in real time through a cloud end, and global model optimization and resource scheduling are responsible for realizing optimal performance.

The invention relates to the technical field of information, and discloses a multi-machine collaborative operation control method and system based on a common inductance calculation control technology, and the method comprises the steps: obtaining task granularity distribution and a resource demand matrix, and carrying out the classification of the task granularity distribution and the resource demand matrix through a classification algorithm, and obtaining an initial task segmentation feature set; based on the optimal segmentation point set, a preliminary task allocation scheme is generated in combination with node calculation heterogeneity, and a task segmentation feature set is dynamically updated through a real-time sensing technology; a real-time task state set is generated through a group information sharing mechanism, a task allocation scheme is optimized through a dynamic programming algorithm, and it is ensured that tasks are reasonably allocated according to priorities and node resources; and the cooperation state matrix is updated through a real-time communication protocol, so that the accuracy and timeliness of a multi-node cooperation execution result are ensured. According to the invention, task allocation can be dynamically adjusted in a complex environment, resource utilization is optimized, the cooperation efficiency of a multi-robot system is improved, and the method can be widely applied to the fields of industrial production, warehouse logistics, disaster rescue and the like.

The invention provides a multi-task parallel processing method for user problems under an AI platform, and belongs to the technical field of digital data processing of the AI platform. Task resource requirements are accurately calculated through a video memory pre-estimation function and a memory pre-estimation function, a resource consumption mode of concurrent execution is analyzed through a multi-task resource prediction model; the load capacity of the system is evaluated based on a video memory utilization rate gain index, an optimal task segmentation strategy is determined through a data set splitting degree calculation function, multi-stage video memory sub-pools are constructed to realize differentiated resource management, and a Nash equilibrium point of resource allocation is solved by adopting a data set video memory allocation game model. Collaborative optimization allocation of resources is achieved through the video memory and memory coupling allocation equation set, the task state is monitored in real time in the multi-task concurrent execution process, the resource allocation weight is dynamically adjusted, and the technical problem that the system resource utilization rate is low during AI platform multi-task parallel execution is solved.

The invention discloses a network scanning task intelligent segmentation and load balancing method and system. The method comprises the following steps: identifying abnormal fragments by executing integrity verification on a scanning task, analyzing a task dependency relationship to determine a divisible position, and generating a segmentation scheme in combination with granularity constraint; node resource states are collected to calculate load indexes, and scheduling weights are hierarchically configured according to load levels to establish parallel scheduling channels; constructing a node cooperation group based on load difference, formulating an inter-group circulation rule to form a cooperation scheduling space, fragmenting tasks according to node capacity, and generating a parallel execution plan; monitoring the execution progress to identify overstocked nodes, diagnosing bottleneck types and planning a dredging path to execute task migration; the resource time sequence data is collected to identify the low-load time period, and the task is distributed to the optimal time slot to be executed, so that the task segmentation reasonability and the node load balancing degree are improved, and the distributed scanning execution efficiency is improved.

The invention belongs to the field of live-action three-dimensional modeling, and discloses a live-action three-dimensional modeling method and device based on cloud native task batch processing scheduling, and the method comprises the steps: calculating aerial triangulation encryption data of a three-dimensional modeling task, and carrying out the block processing of the aerial triangulation encryption data through a task segmentation module; the task scheduling module adjusts operation and task issuing of sub-tasks in real time according to available computing resources of the cloud architecture system, and optimizes the utilization rate of distributed computing resources in real time; the task state monitoring module monitors the execution state of the sub-task in real time; the subtask processing module performs corresponding generation on the image data in each block, and uploads a generation result to a storage resource pool after operation is completed; and after all the subtasks are operated, calling a model merging module to perform overall merging. According to the method, the live-action three-dimensional modeling task and the cloud native task batch processing architecture are combined, operator-level scheduling of the live-action three-dimensional modeling function is achieved through the cloud native system architecture and the resource container pool, and cloud native system resources are efficiently utilized.

The invention discloses a pet size intelligent measurement system and method based on a multi-task segmentation network. According to the system, the three-dimensional visual perception capability of a scene is constructed by using a binocular stereoscopic vision technology, and accurate depth information is obtained; meanwhile, an advanced multi-task segmentation network is adopted to process the pet image, efficient target detection, category recognition and pixel-level instance segmentation are achieved, and the pet contour (especially the hair edge) is accurately extracted; and finally, depth data provided by three-dimensional vision and the accurate contour output by the multi-task segmentation network are fused, and the real three-dimensional size of the pet is calculated. According to the method, the problems of low precision and poor contact stress and dynamic edge processing of a traditional method are solved, and non-contact and high-precision pet size intelligent measurement is realized.

The invention relates to a method and a system for identifying particles in insulating oil driven by deep learning, which can realize automatic identification and classification of particles in insulating oil by combining a micro-flow digital imaging technology and a semantic segmentation method of deep learning, constructing a unified training data set and combining a multi-task segmentation model. Images of the particles in the transformer oil are obtained through the micro-flow digital imaging module, and pixel-level classification and particle recognition are carried out on the images through the multi-task segmentation model. The segmentation branches in the model are responsible for extraction of particle boundaries, and the classification branches are used for analyzing particle morphology and outputting particle categories. The model adopts multi-scale feature fusion, an attention mechanism and a deep residual network structure, and the problems of particle morphology diversity, size change, image boundary blur and the like are solved. Through the high-precision particle recognition algorithm and system, the accuracy and efficiency of detection of the particles in the insulating oil are effectively improved, manual intervention is reduced, and the safety and long-term stability of the transformer are improved.

The invention relates to the technical field, in particular to a multi-task AI video analysis and detection method and system of a shared network mechanism, and the method comprises the steps: freezing CLIP-ResNet50 as a unified feature extractor, utilizing the structural advantages of a YOLOv8 detection head, and combining an innovative MoE module (supporting temperature adjustment, intelligent routing, self-attention fusion and an expert enhancement mechanism), thereby achieving the multi-task AI video analysis and detection of the shared network mechanism. The problems of task conflict, model redundancy, inflexible reasoning and the like are effectively solved. The scheme is suitable for real application scenes such as multi-task edge reasoning and video stream concurrent detection and has high practical value and technical advancement, secondly, multi-label merging is not adopted, an independent detection head is established for each task, and inter-task interference is avoided; meanwhile, a detection head is dynamically selected through the MoE module, task switching or sharing in the reasoning stage is achieved, and therefore the method is suitable for scenes with high task segmentation and obvious feature conflicts.

The invention relates to a task execution method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a to-be-processed target task of a server cluster; according to a preset segmentation rule, segmenting the target task to obtain a to-be-executed target sub-task of each server in the server cluster; predicting an execution score of each server when the server executes the corresponding target sub-task to obtain a prediction score set corresponding to the server cluster; and when the predicted score set satisfies a preset condition, adjusting a segmentation rule according to the predicted score set, so as to re-execute segmentation of the target task based on the adjusted segmentation rule. Under the condition that the performance of the server is abnormal, the task segmentation rule executed by the server cluster is adjusted in time, and compared with current static load configuration, the response speed of the service is remarkably improved; besides, when the performance of the server is reduced, the task is transferred to the server of the same type, so that the task execution speed is prevented from being reduced, and the task execution efficiency is further improved.

The invention discloses a general pathological image segmentation method based on depth style features. The method comprises the following steps: 1, preprocessing a pathological image through a preprocessing module; 2, inputting the preprocessed pathological image into a shared encoder module, and extracting multi-scale semantic features of the preprocessed pathological image to obtain a deep feature map; 3, inputting the deep feature map into a depth style feature extraction module, and outputting style features; 4, the task label and the learnable task embedding matrix are input into a task feature coding module, and dynamic task features are generated; 5, a task prompt generation module fuses the style features and the dynamic task features to generate a task prompt; 6, the task prompt is input into a dynamic convolution module, and dynamic convolution parameters matched with the current task are generated; and 7, inputting the shallow layer features and the dynamic convolution parameters output by the decoder into a multi-task dynamic head module, and predicting a segmentation mask. According to the invention, the multi-task segmentation precision and efficiency of the pathological image are improved.

The invention belongs to the technical field of medical information processing, and relates to a laparoscopic surgery full-granularity recognition system and method based on feature extraction and task segmentation. Firstly, a video stream composed of multiple frames of images of an endoscope is input into the system, and then the spatial feature of each frame of image is obtained through a multi-frame spatial feature extraction module; the spatial features of all the frames are then input into a multi-frame time feature fusion module to obtain time features fused with different time lengths, and then the time features are used for outputting operation stages, operation steps and operation ternary body labels; meanwhile, only the spatial features of the current frame are extracted from the spatial features of all the frames, and an operation scene segmentation label is output through a single-frame feature restoration module. According to the laparoscopic surgery full-granularity recognition system and method based on feature extraction and task segmentation, computing resources can be greatly reduced, and compared with an existing single-granularity surgery scene technology, the laparoscopic surgery full-granularity recognition system and method based on feature extraction and task segmentation can output surgery scene information of multiple granularities at a time by means of the mutual coordination effect of multiple tasks.

The invention discloses a video intelligent identification system for abnormal and illegal behaviors on a customs supervision operation site, and belongs to the technical field of intelligent identification and detection. The system comprises an input module, an edge cloud cooperation module and an output module. The input module comprises a video access module and a video preprocessing module; the edge cloud cooperation module comprises an edge end calculation module, a cloud calculation module and a communication optimization module; the edge end calculation module comprises a task complexity evaluation module, a dynamic task scheduling module, a task segmentation module and a dual-channel detection module; the cloud computing module comprises a cloud detection module and a fusion analysis module; and the output module comprises a report generation unit and an abnormity notification unit. According to the invention, the customs monitoring video is intelligently identified and analyzed based on edge cloud collaboration, the dynamic scheduling of tasks, the optimization of computing resources and the improvement of data transmission efficiency are realized, and the real-time performance and accuracy of identification and detection are obviously improved.

The invention discloses an intelligent driving task unloading method based on mobile edge computing. The intelligent driving task unloading method comprises the following steps of S1, environment perception and task identification, wherein the surrounding environment of a vehicle is perceived in real time through a vehicle-mounted sensor; and the intelligent driving task needing to be processed is identified according to the sensed environment information. According to the method, unloading task evaluation is divided into simple tasks and intensive tasks, cooperative unloading work of the vehicle-mounted equipment and the edge server is carried out, efficient unloading and execution of the intelligent driving tasks are achieved, the performance and the response speed of the intelligent driving system are improved, and through task segmentation, priority determination and dynamic adjustment of the execution sequence, the efficiency of the intelligent driving system is improved. The overall execution efficiency is optimized, further efficient and real-time processing of the intelligent driving task is achieved, the optimal dynamic adjustment strategy of the edge server is additionally selected according to the information monitored in real time when the load capacity is large, smooth and rapid execution of task unloading is ensured, and the unloading efficiency and stability are further optimized.

The invention discloses an intelligent super-division coding system for a low-bit-rate video, which comprises multiple stages of assembly lines and a merging module, each stage of the multiple stages of assembly lines is responsible for different GPUs / threads, and the multiple stages of assembly lines are integrated into a pipeline parallel task. The multi-stage assembly line comprises a video input and analysis module, a task segmentation and dynamic scheduling module, an AI super-division module, an HDR / SDR self-adaptive conversion module, a video coding module and a quality control and closed-loop optimization module which are sequentially processed. And time consistency and efficient utilization are ensured through dynamic scheduling, overlapping slicing and inter-frame constraint on multiple GPU hosts, so that the system can still improve subjective and objective image quality on the premise of remarkably reducing the code rate.

The invention discloses a method and device for detecting particulate matter in transformer oil, a medium and a product, and relates to the field of power transformer detection.The method comprises the steps that an image of the particulate matter in the transformer oil is obtained, and a multi-task segmentation data set is constructed; according to the multi-task segmentation data set, building a multi-task segmentation model by adopting a double-branch framework comprising segmentation branches and classification branches; the segmentation branch comprises a global attention up-sampling module, a mixing pool module, a feature pyramid attention module and a self-attention module; the classification branch comprises a full attention module; the multi-task segmentation model takes a particulate matter image as input and takes a detection result of the particulate matter image as output; obtaining a to-be-detected particulate matter image; and determining a detection result of the to-be-detected particulate matter image by adopting the trained multi-task segmentation model. According to the invention, accurate and efficient automatic identification and classification of particulate matters can be realized.

The invention discloses an automatic control method and system for injection molding equipment, and relates to the related technical field of production process control, and the method comprises the following steps: reading basic configuration information and work order injection molding tasks of the injection molding equipment; performing process control segmentation, and determining a task segmentation grid; traversing task segmentation grids, and carrying out grid independent control analysis and cooperative control adjustment; establishing communication connection between the programmable controller and the injection molding equipment, and setting a control program; a control response instruction is generated, equipment injection molding control is carried out, monitoring is synchronously carried out, and real-time control data is returned; and measuring control deviation data, performing deviation quality influence evaluation, and performing feedback operation regulation and control. The technical problems that control parameters cannot be intelligently and accurately determined through existing injection molding equipment control, consequently, injection molding task parameter control schemes are not matched, the positioning and calibration node quality problem is poor in accuracy and pertinence, and the injection molding quality cannot reach the expected effect are solved, and the technical effect of improving the injection molding quality is achieved.

The invention discloses a calcium-based thermochemical particle bed health status assessment method and a related device. The method comprises the following steps: acquiring a surface image of a calcium-based thermochemical particle bed; performing data enhancement on the surface image of the calcium-based thermochemical particle bed layer; inputting the surface image after data enhancement into a multi-task segmentation model to obtain a pixel-level segmentation mask; calculating the expansion rate, the agglomeration index, the crushing degree, the collapse depth and the collapse area ratio of the particle bed layer according to the pixel-level segmentation mask; the performance of the calcium-based thermochemical particle bed layer is evaluated according to the expansion rate, the agglomeration index, the crushing degree, the collapse depth and the collapse area ratio of the calcium-based thermochemical particle bed layer. The method and the related device can accurately detect the health of the calcium-based thermochemical particle bed layer.

The invention provides an ultrasonic image segmentation method, device and equipment. The method comprises the following steps: acquiring an ultrasonic image to be segmented; inputting the ultrasonic image into a multi-scale expansion convolution module in a multi-task segmentation model to obtain a first feature map; the multi-scale expansion convolution module is used for extracting cross-scale feature information; inputting the first feature map into an efficient channel attention module in a multi-task segmentation model to obtain an attention weight of each channel in the first feature map; obtaining a second feature map according to the first feature map and the attention weight of each channel; and inputting the second feature map into a decoder module in the multi-task segmentation model to obtain a segmentation result of the ultrasonic image. According to the method provided by the embodiment of the invention, accurate segmentation of the ultrasonic image of the intermuscular brachial plexus retardation scene is realized.

The application discloses a medical data processing method and product based on multi-stage transfer learning and multi-modal data collaborative fusion, relates to the technical field of medical data processing and artificial intelligence, and adopts MedicalNet medical special pre-training weights to initialize a classification model backbone network; a three-stage progressive fine-tuning framework is constructed on the basis, field adaptive coarse classification fine-tuning and target task fine classification are performed, and clinical structured data is introduced in the third stage; after high-dimensional image features are reduced in dimension and low-dimensional clinical features are increased in dimension through an adaptive multi-branch multi-level MLP architecture, mid-term deep fusion is performed, the application can effectively mine the complex relationships such as complementation, correlation and cooperation of multi-modal heterogeneous data, improve the lung disease classification precision and model generalization capability, and significantly inhibit the small sample overfitting phenomenon.

An AI model-based task allocation method in a communication system, and a related product. The method comprises: a terminal / core network element sends first information to an access network device, wherein the first information indicates a first model identification (ID), and the first model ID indicates an AI model deployed on the terminal / core network element; the terminal / core network element receives second information from the access network device, wherein the second information indicates a second model ID, the second model ID is a model ID in the first model ID, and the second model ID is associated with a first task; and the terminal / core network element executes the first task on the basis of an AI model corresponding to the second model ID. In the example, computing power alignment between devices can be implemented, a condition is provided for task segmentation, implementation of distributed training is enabled, and an overall inference delay is reduced. The present application can be applied to the technical field of AI, the technical field of communications, etc.

The application discloses a kind of blood vessels and lesion multi-task segmentation method based on ultra-wide-angle fundus image, belong to image processing field.The steps include as follows:S1.UWF fundus image dataset is obtained, and is assigned as training set and test set;S2.build multi-task semi-supervised learning network based on weight control mechanism, including encoder, decoder, cross-level non-local graph module and loss weight control mechanism, and training set image enters encoder and carries out feature extraction;S3.feature that encoder output is decoded and feature reconstruction by decoder part;S4.combined loss function and total loss are optimized to network by training, obtain the blood vessels and lesion multi-task segmentation model based on ultra-wide-angle fundus image;S5.data in test set are input into model, and segmentation result is obtained.The application can improve the ability of model to extract image detail features, realize fine and accurate multi-task segmentation.

The invention discloses a language model intelligent training task scheduling method based on reinforcement learning, and the method comprises the following steps: constructing a task pool, and generating a task attribute set; performing rhythm coding and memory priority processing to generate a training state vector, inputting the training state vector to the improved PPO model, and outputting scheduling action distribution; executing task segmentation, and inserting fragment boundary mark information to generate a scheduling action set; calculating a step length difference value, a training duration difference value and a rhythm difference value between tasks, generating a conflict weight, and executing scheduling conflict constraint and buffer control; scheduling behaviors and training feedback are collected, training efficiency, parameter stability, task switching smoothness and task distribution diversity scores are calculated and fused to generate a smooth reward vector, and a task pool is updated and refreshed. According to the method, self-adaption, high efficiency and rhythm coordination of task scheduling are realized, and the self-adaption capability and execution efficiency of training task scheduling can be improved.

The application discloses a multi-granularity and multi-level computing power scheduling method in an end-side computing power network, which comprises the following steps: deploying an online deep reinforcement learning agent and a system state monitoring module on a base station side; detecting and collecting channel states and computing resources and sending them to the deep reinforcement learning agent; outputting a task segmentation strategy of each user and a scheduling decision of each subtask; a task-type device (TD) segments tasks according to the instruction and schedules the tasks to target computing nodes according to the instruction for parallel processing; the task-type device (TD) receives the computing results of all the subtasks scheduled out in parallel; and the scheduling method segments the tasks of each user in multiple granularities according to current system state information, including channel gain, idle resources provided by a resource-type device (RD) and resources allocated to each user by an edge computing server, and schedules the segmented subtasks to suitable computing nodes in parallel, so that the execution time delay of an application program is further shortened by taking advantage of parallel execution.

The invention discloses an intelligent marine environmental pollution early warning method based on a biomarker, and belongs to the technical field of marine ecotoxicology, and the method comprises the steps: building an image quality control standard, obtaining a marine mussel histopathology image, building a data set according to the standard, carrying out the data marking and quality control, and carrying out the preprocessing. A model is constructed based on a YOLOv11 architecture, a multi-task segmentation function is realized by taking an image segmentation module as a core, and model training, verification and evaluation are carried out; different feature pixel points in the histopathology image are extracted through the trained model, then the shortest distance between the different feature pixel points is calculated through the Euclidean distance, and a result image containing distance marks is generated after averaging; and the marine environmental pollution early warning grade is divided by calculating the response index H. According to the method, the tissue pathology image is accurately analyzed by means of deep learning, identification and segmentation of different tissue structures of the marine mussels are realized, and an efficient technical means is provided for marine environment monitoring.

The invention provides a dynamic fragmentation method for a front-end rendering task. The execution pressure degree of the rendering task can be quantified in real time by monitoring the style calculation time consumption and the layout calculation time consumption in the rendering process of the front end of the browser. Based on the execution pressure degree, the system dynamically determines the segmentation granularity of the current front-end rendering task, so that the task is divided into a plurality of sub-tasks, and the sub-tasks are scheduled and executed according to priorities. According to the method, task segmentation can adaptively respond to actual rendering pressure, and the problem of unreasonable resource allocation caused by solidification of a static fragmentation strategy is avoided. Through accurate task division and scheduling, the system can effectively reduce blockage of a main thread by a long task, the rendering efficiency and the response speed of user interaction are improved, and meanwhile, the rendering stability and fluency in a complex scene are ensured.

A computing task segmentation method and a related apparatus are provided, and may be applied to an extended reality (XR) service or another low-latency service. The method includes: obtaining an amount of to-be-transmitted data and a computation amount of a terminal device that correspond to a first computing task, where the first computing task is obtained by segmenting a computing task of a neural network model based on a segmentation point; determining that the segmentation point is a target segmentation point based on the amount of to-be-transmitted data, the computation amount, and a status of a channel between the terminal device and a radio access network device; and sending indication information to the terminal device, where the indication information indicates the target segmentation point.

The present application belongs to the field of multi-agent reinforcement learning, and specifically relates to a collaborative reasoning task scheduling method based on multi-agent reinforcement learning, which initializes a multi-agent environment, sets up multiple user devices, each of which is an agent, deploys an independent Actor neural network on each user device, sets up multiple cloud servers and deploys a central Critic neural network on a central control server, and provides multiple communication channel options for user devices. The advantage is that this solution is used to solve the combinatorial optimization problem of task segmentation points, communication channels, and server selection.