30results about How to "Resolve accuracy" patented technology

The disclosure provides a large model-based quality inspection method, device, equipment, storage medium and program, relating to the technical field of data processing, specifically relating to artificial intelligence, deep learning and large language model technology, which can be applied in the field of AI medical technology, wherein the method comprises: obtaining to-be-inspected data of a target platform and a target quality inspection prompt template; wherein the target quality inspection prompt template comprises quality inspection task instructions and quality inspection question classification content; splicing the to-be-inspected data and the target quality inspection prompt template to obtain a target quality inspection prompt; inputting the target quality inspection prompt into a target large model to automatically inspect the to-be-inspected data according to the target quality inspection prompt through the target large model, and obtaining first quality inspection data. The embodiment of the disclosure can reduce the artificial quality inspection cost of platform quality inspection and improve the efficiency and accuracy of platform quality inspection.

This invention relates to the field of intelligent water control terminal control and electromagnetic compatibility processing technology, and particularly to an anti-interference water control terminal control method and system with abnormal alarm. The method includes: reading raw signals in parallel from water shortage, anti-magnetic, and flow sampling interfaces; processing and mapping these signals into alarm events and writing them into a hardware alarm event set; performing board-level partitioning layout based on the information in the event set to generate a partitioning layout configuration set; configuring decoupling units, laying grounding wires and copper plating, and setting multiple grounding nodes accordingly; assigning power supplies to each detection interface to an analog power path and power supplies to the main control and communication modules to a digital power path, configuring filtering units and isolation nodes; laying grounding copper foil and vias for the crystal oscillator, and configuring metal shielding structures for key modules; reading alarm signals based on the obtained configuration set and encapsulating them into alarm messages, uploading them via Bluetooth or 4G channels, receiving management results, and generating a terminal alarm processing status. This invention improves the anti-interference capability and alarm reliability of the water control terminal through event-driven board-level layout and power supply link isolation.

Embodiments of the present application provide an identity authentication method and device, a storage medium and an electronic device. The method comprises: in response to a play instruction associated with a target action, obtaining K frames of face images to be processed, and then inputting the K frames of face images into a pre-trained recognition model respectively to obtain K recognition results, wherein each of the K recognition results comprises a probability of whether a frame of face image input is a real face and a probability of whether a target part of the frame of face image is in a target state, and the target part comprises different face parts performing different target actions; and then performing authentication on a face to be recognized according to the K recognition results to obtain a target authentication result, wherein the target authentication result comprises whether the face to be recognized is a real face and whether the face to be recognized performs the target action. Through the present application, the problem that identity authentication is too dependent on key point positioning accuracy in related art, resulting in low accuracy and efficiency of identity authentication, can be solved.

This invention relates to the field of medical device technology and provides a method, system, and medical device for measuring peak cough flow rate. The method includes real-time acquisition of inspiratory and expiratory flow rate signals and airway pressure signals; monitoring changes in the expiratory flow rate signal to identify burst cough events and lock a cough analysis window; calculating the difference in inspiratory and expiratory flow rates within the cough analysis window to obtain a baseline cough flow rate sequence; calculating the airway pressure signal within the cough analysis window using a system leakage factor to obtain a leakage flow rate compensation sequence; determining coefficients based on the ratio of tubing and patient airway resistance parameters and multiplying the expiratory flow rate signals to obtain a tubing resistance compensation sequence; superimposing the above sequences to obtain a corrected flow rate sequence; and extracting extreme value features and evaluating the effectiveness of the corrected flow rate sequence generated after a preset number of iterations to determine the peak cough flow rate value. This invention solves the problems of inaccurate and inefficient measurement results in existing peak cough flow rate measurements.

Embodiments of the present application disclose a serverless function tracking method, system, device and storage medium, wherein the method comprises: a proxy subsystem determines a hierarchical function association graph corresponding to each architecture level in the serverless architecture based on a function keyword; in the case of meeting a preset data sending condition, the proxy subsystem sends the hierarchical function association graph to a server subsystem; the server subsystem traces the application trace of the function between different architecture levels through the hierarchical function association graph, and obtains a function cross-layer traceability graph. The technical scheme of the embodiments of the present application solves the problem that in the prior art, it is difficult to uniformly track and accurately trace the cross-layer security events in the serverless architecture, can associate and track the function events across the levels in the serverless architecture, provides complete attack traceability analysis data for serverless function tracking, so as to meet the engineering requirements of security audit, compliance evidence and emergency response.

The invention provides a tumor medical mode recognition method and system based on image detection, and relates to medical image processing and artificial intelligence auxiliary diagnosis, and the method comprises the steps: carrying out the partitioning of a digital pathological section image, evaluating the color distribution, light and shade contrast and microstructure geometric features of each block, and generating a local quality evaluation graph to recognize a quality damaged region; a local diagnosis judgment graph is generated based on tumor recognition model analysis, and the diagnosis confidence of the model on each region is visually reflected; performing correlation analysis on a low-confidence area and a quality damaged area in the two images, if the overlapping degree reaches the standard, generating quality correlation interpretation, and determining that uncertainty is derived from image quality; finally, an interpretability report containing the original image, the two evaluation images and interpretation is generated, a clear basis is provided for doctors, traditional compensation defects are avoided, recognition reliability and transparency are improved, precise clinical decision making is assisted, and practical value is achieved.

The application discloses a roof clip mounting and detecting system, which comprises a vehicle type information device, a vehicle type information display device and a tool table, wherein the vehicle type information device is in communication connection with the vehicle type information display device and the tool table; the vehicle type information device is used for obtaining pre-configured vehicle type information and transmitting the vehicle type information to the vehicle type information display device; the vehicle type information display device is used for obtaining vehicle type information of a target vehicle and transmitting the vehicle type information to a display device for display; and the tool table is used for placing at least one type of roof clip and a jig and generating regional light prompt in response to the vehicle type information of the target vehicle. Through the technical scheme, the roof clip can be mounted by the staff, and the mounting condition of the roof clip can be detected after the roof clip is mounted, so that the accuracy and work efficiency of the roof clip mounting and detecting work are improved.

The application provides a large image target positioning method and device, equipment and storage medium, which are applied to the technical field of image positioning. A plurality of local sub-images covering a to-be-detected object are acquired, then coarse positioning is performed on the local sub-images to quickly determine a region of interest, fine positioning is performed in the region of interest, sub-pixel level features are extracted, and local sub-pixel level coordinates are calculated, so that the positioning precision is greatly improved. Space transformation parameters between adjacent local sub-images are calculated, and the parameters are used to map local position information to a global coordinate system. The strategy of combining local fine positioning and global unified mapping avoids information loss and precision reduction caused by scaling in traditional large image processing, solves the problems of repeated detection and inaccurate splicing that may occur in slice processing, and has the advantages of improving the accuracy and efficiency of large image target positioning.

The invention discloses an image segmentation method, device and equipment, a medium and a product. The method comprises the steps of obtaining image segmentation prompt information in response to an image segmentation prompt information input operation for a first image; the image segmentation prompt information and the first image are input into a pre-trained image segmentation model to obtain a first image segmentation result, the first image segmentation result is displayed, the image segmentation prompt information comprises at least one of point prompt information, smearing prompt information and geometric prompt information, and the smearing prompt information comprises at least one of the point prompt information, the smearing prompt information and the geometric prompt information. The pre-trained image segmentation model is obtained by training a sample image, sample image segmentation prompt information and a sample image segmentation result. According to the technical scheme, the problem that the accuracy and efficiency of current image segmentation need to be improved is solved, image segmentation can be carried out by combining the model with the manual operation prompt information, the efficiency and accuracy of image segmentation are improved, and efficient cooperation of the model and manual operation is achieved.

The application provides a context-aware sensitive information identification method, system, device and medium, and relates to the technical field of data security. In the method, a preset sliding window is used to cut the to-be-identified data stream into a continuous text slice sequence and generate a global coordinate index; a corresponding prompt word template is dynamically assembled based on metadata characteristics; the text slice and the prompt word template are input into a preset context-aware model to output sensitive entities, local start coordinates, local end coordinates and confidence scores; after coordinate conversion, the sensitive entities corresponding to the same global position in the overlapping area of adjacent text slices are de-duplicated to generate a sensitive entity list; and the sensitive entity list is desensitized using a preset security protection strategy, and the desensitized sensitive entity list, global end coordinates and global coordinate index are used for stream reorganization to obtain an identified desensitized data stream, thereby improving the accuracy of sensitive information identification and data desensitization.

This invention provides a method and system for intelligent layout of directional boreholes for gas control, relating to the field of gas borehole control technology. The method includes: acquiring coal seam occurrence information and gas source information; collecting roadway video and then performing feature frame recognition and stitching to construct a three-dimensional scene; performing initial fitting of gas boreholes based on gas control demand information, coal seam occurrence information, and gas source information; fitting and correcting the borehole positions of the initial borehole fitting results based on roadway features in the scene; obtaining the adjusted borehole fitting results for numerical simulation; generating numerical simulation results for multi-angle gas control evaluation; generating corrected borehole layout information for gas borehole control; solving the technical problem that existing technologies rely more on manual experience for borehole placement, resulting in inaccurate placement; achieving rational and precise placement of gas boreholes, thereby improving the accuracy of gas borehole placement.

Embodiments of the present application disclose a method and device for synchronously detecting output images, electronic equipment and a medium. The method comprises: in response to a frame synchronization signal and a unit pixel group synchronization signal of an image decoding output device, receiving at least two target images output by the image decoding output device; wherein the unit pixel group synchronization signal comprises a horizontal synchronization signal or a vertical synchronization signal; dividing each unit pixel group of the at least two target images into a preset number of cache pixel groups for caching, and recording the time stamp of each cache pixel group; and according to the time stamp, detecting whether the corresponding cache pixel groups in the at least two target images are synchronized. The above scheme can cache the images before displaying by setting a cache space, and determine whether the at least two output images are synchronized according to the cached time stamp, thereby solving the problems of low efficiency and low accuracy of the current scheme which requires manual observation and only uses synchronization signals for judgment.

This invention relates to the field of medical information processing technology, and discloses a method, apparatus, model, and medical device for surgical third-party security verification. The method includes: acquiring multimodal information to be verified; verifying the information using a surgical third-party security verification model; the model includes a dedicated encoder, an uncertainty gating unit, a feature fusion unit, and a decision unit. The dedicated encoder maps the multimodal information to be verified to a high-dimensional latent space, obtaining the corresponding latent representation and uncertainty variance, which serve as the feature information of the information to be verified; the uncertainty gating unit outputs gating values ​​corresponding to the latent representation and uncertainty variance of the multimodal information to be verified; the feature fusion unit fuses the feature information of the multimodal information to be verified based on the gating values; and the decision unit judges the verification result based on the fused feature information. This invention can improve the accuracy and reliability of surgical third-party security verification.

The utility model discloses a kind of crab automatic dormancy sequencing feeding assembly line, belong to crab sorting technical field, in turn including automatic dormancy device, sequencing feeding device and grading sorting device. Crab is carried out low-temperature treatment by low-temperature dormancy cavity and dormancy conveyor belt to automatic dormancy device, so that it enters quiet state, reduces stress reaction. Sequencing feeding device includes the feeding conveyor belt of the link of dormancy conveyor belt and is equipped with the guide structure of inclined guide surface, and single-column sequencing passage is formed between guide structure and feeding conveyor belt, realizes the ordered output of crab. Grading sorting device is equipped with image acquisition module, control system and actuating mechanism, generates sorting instruction according to shape feature data automatically, accurately push crab to corresponding sorting passage. The assembly line realizes the automation of live crab sorting whole process, improves processing efficiency and classification accuracy, solves the problem of low efficiency and poor precision of traditional manual sorting.

The application discloses an adaptive DVFS method and device for optimizing energy efficiency of a deep neural network. The method comprises the following steps: extracting global features of the deep neural network (DNNs) by a global feature extractor and inputting the global features into a clustering hyperparameter prediction model, and predicting clustering hyperparameters of the current DNNs by the clustering hyperparameter prediction model; performing power consumption behavior similarity clustering according to the predicted clustering hyperparameters and DNNs fine-grained deep features extracted by the deep feature extractor, so as to divide the DNNs into multiple power consumption blocks and form power consumption views, and then inputting global features of each power consumption block into a decision model to obtain a target decision frequency of each power consumption block; setting a DVFS plug-in point before each power consumption block of the DNNs, and setting a target frequency in each power consumption block of the DNNs in advance according to the target decision frequency. Through pre-analysis of the DNNs task, the target frequency in each power consumption block of the DNNs is set in advance according to the decision information, so that the energy efficiency optimization of the neural network can be realized while the performance is maintained.

The application relates to a window width and window position automatic generation method and device, computer equipment and a storage medium, wherein the method comprises the following steps: inputting an acquired sample medical image, an initial minimum value of image grayscale and an initial maximum value of image grayscale into a neural network model for prediction, and constructing a loss function of the neural network model based on a prediction result; based on the loss function, the neural network model is iteratively optimized to obtain a target neural network model; a target medical image is input into the target neural network model for prediction to generate a window width and window position. Through the application, the problem that the window width and window position generation efficiency is low and the accuracy is not high due to human factors in the related art is solved, the loss function is constructed by using a prediction result, the training of the target neural network model is completed, and finally the window width and window position are automatically generated based on the target neural network model, so that the influence of human factors is avoided, and the generation efficiency and accuracy are improved.

The application discloses a battery pack state of charge prediction method, a battery pack and a computer readable storage medium. The method comprises the following steps: obtaining a static temperature value of the battery pack after the battery pack is discharged to a load device for a preset time length; determining a voltage change rate of the battery pack within the preset time length; and predicting the state of charge of the battery pack according to the static temperature value and the voltage change rate. The method can greatly shorten the static time, solve the problem of low accuracy and efficiency caused by the long static time of the battery pack in the related art, and effectively improve the efficiency and accuracy of the prediction of the state of charge of the battery pack.

This invention provides a specific primer set for identifying the sex of Eucommia ulmoides and its identification method, relating to the field of plant molecular identification technology. The specific primer set includes primers INV53, INV54, INV55, and INV56. The upstream primer of primer INV53 is shown in SEQ ID NO.1, and the downstream primer is shown in SEQ ID NO.2; the upstream primer of primer INV54 is shown in SEQ ID NO.3, and the downstream primer is shown in SEQ ID NO.4; the upstream primer of primer INV55 is shown in SEQ ID NO.5, and the downstream primer is shown in SEQ ID NO.6; the upstream primer of primer INV56 is shown in SEQ ID NO.7, and the downstream primer is shown in SEQ ID NO.8. The specific primer set for identifying the sex of Eucommia ulmoides provided by this invention targets the Chr10a inverted region, achieving 100% identification accuracy. It features high specificity and accuracy, is adaptable to samples with different genetic backgrounds and cultivation environments, and solves the problems of low accuracy and poor stability in existing methods.

The present application relates to the technical field of image processing, more particularly, the present application relates to an AOI defect detection method and system based on industrial vision, comprising: acquiring a gray scale image of an electronic screen to be measured; determining the physical period of pixel arrangement of the electronic screen; constructing a local period difference degree, which includes calculating the gradient difference value between any pixel point in the gray scale image and its adjacent pixel point along the physical period. The present application quantifies the degree of dirtiness of the image by calculating the local period difference degree and constructing a period confusion factor combined with statistical characteristics. On this basis, the ADMM acceleration coefficient is dynamically generated: a conservative step size is used for low-quality images to prevent algorithm divergence, and an aggressive step size is used for high-quality images to improve efficiency. This method effectively solves the problem that the traditional fixed step size strategy cannot balance detection speed and accuracy, significantly improving the robustness of the AOI system.

This application provides an image recognition method and system for digital twins of aquatic organisms, applied in the field of image recognition technology. By dynamically determining the priority biological features for recognition of aquatic organisms, it effectively solves the problems of decreased recognition accuracy and data link breakage caused by changes in body surface features during the growth process. It has the advantages of effectively solving the problem of decreased individual recognition accuracy caused by dynamic changes in body surface features of aquatic organisms during the growth process, and ensuring the continuity and accuracy of health data in the digital twin model.

The application discloses a control method and device of a battery thermal management system, a processor and a vehicle. The method comprises the following steps: obtaining working condition data of a battery of the vehicle, wherein the working condition data is used to represent the working condition of the battery during the operation of the vehicle; extracting feature data of the battery from the working condition data, wherein the feature data is used to represent the performance condition of the corresponding battery; determining a target performance state of the battery based on the feature data; determining a control strategy of the battery based on the target performance state; and controlling the thermal management system of the battery to perform heating operation or cooling operation on the battery in response to the control instruction corresponding to the control strategy. The application solves the technical problem of low accuracy of heating or cooling of the battery.

An elastography method (200) and an ultrasound imaging system are disclosed. The method (200) includes: transmitting a first push pulse toward a target tissue to generate a first shear wave (S210); transmitting a first ultrasound wave that tracks the first shear wave and receiving a first ultrasound echo to obtain first ultrasound echo data (S220); generating a first elastic image based on the first ultrasound echo data (S230); determining a target region in the first elastic image (S240); transmitting a second push pulse based on the target region to generate a second shear wave propagating within a target sub-tissue corresponding to the target region (S250); transmitting a second ultrasound wave that tracks the second shear wave and receiving the second ultrasound echo to obtain second ultrasound echo data, wherein the penetrating power of the second ultrasound wave is greater than that of the first ultrasound wave, and / or the penetrating power of the second shear wave is greater than that of the first shear wave (S260); generating a second elastic image based on the second ultrasound echo data (S270); and outputting the second elastic image (S280). This method (200) can solve the problem of insufficient penetration power during elastography.

A cold rolling process control system inlet and outlet saddle steel coil tracking method and system, the method comprises the following steps: step 1, establish a coordinate system, calibrate the position interval of each saddle in the coordinate system; step 2, real-time acquisition of each saddle occupancy signal, and steel coil trolley moving position; step 3, based on the occupancy signal of the steel coil and each saddle, the moving position of the steel coil trolley and the position interval of each saddle, the source saddle and the target saddle of the steel coil movement are obtained, and the steel coil movement trajectory tracking is carried out. The present application establishes a tracking coordinate system, collects the saddle occupancy signal of the basic automation system (PLC) and the moving position of the steel coil trolley, monitors the steel coil movement trajectory throughout, determines the source saddle and the target saddle of the steel coil movement according to the saddle position matching algorithm, and updates the client tracking picture accordingly, solves the problems of low tracking efficiency and low accuracy of the traditional method, and is a kind of efficient and accurate automatic tracking method.

The application provides a multi-dimensional identity recognition method and device, electronic equipment and program product, and relates to the technical field of artificial intelligence. The method comprises the following steps: collecting video stream data containing a to-be-recognized object, extracting multi-dimensional feature information of the to-be-recognized object based on a first target image in the video stream data, recognizing the identity of the to-be-recognized object according to face information, if the recognition fails, performing target tracking on the to-be-recognized object based on a second target image before the first target image in the video stream data to determine whether a third target image containing the identity recognition result of the to-be-recognized object exists in the second target image, and if not, recognizing the identity of the to-be-recognized object according to body key point information. Through face recognition, target tracking and identity recognition based on body key points, the high accuracy of face recognition is retained, and when face recognition fails, identity recognition can be accurately performed based on body key points, so that the balance between the high accuracy and the easy failure of face recognition is achieved.

This application provides a method, training method, and device for determining the health status of a device model. The method includes: obtaining a health status confirmation instruction; wherein the health status confirmation instruction indicates a corresponding device type; determining at least one corresponding sub-health status type based on the device type; obtaining sub-source data corresponding to the device type and sub-health status type, and obtaining a sub-health status model corresponding to the device type and sub-health status type; wherein the sub-health status model is trained based on labeled historical sub-source data; there is a one-to-one correspondence between the sub-source data and the sub-health status model; and determining the health status corresponding to the health status confirmation instruction based on the sub-source data and the sub-health status model corresponding to the sub-source data. This method aims to improve the accuracy and relevance of determining the health status of a device.

The application relates to the technical field of data encryption, and particularly discloses a searchable public key encryption method supporting semantic query simultaneously, which comprises the following steps: performing semantic expansion on a query keyword set Q input during searching to obtain a query expansion keyword set; converting a document keyword set W in a data set and the query expansion keyword set into a coefficient vector and a root vector respectively; generating a public key and a private key for the data set; performing encryption processing on the coefficient vector of the document keyword set W and the root vector of the query expansion keyword set respectively by using a predicate encryption method according to the generated public key and private key; and obtaining an encrypted index I of the document. W The encryption trapdoor of the query is T Q The application can guarantee data security and perform semantic expansion on user query information, and solves the accuracy problem in the field of ciphertext retrieval.

The present application relates to the technical field of vehicle-mounted human-computer interaction, and provides a vehicle-mounted multi-modal human-computer interaction method and system, which comprises the following steps: based on a specific change state, adjusting the collection parameters and processing mode of a visual input signal to improve the quality of the visual input signal, obtaining an adjusted visual input signal; at the same time, adjusting the preprocessing parameters and processing mode of a voice input signal to improve the quality of the voice input signal, obtaining an adjusted voice input signal; according to the influence degree of the environmental state judgment result on the visual input signal and the voice input signal, adjusting the weight of the visual input signal and the voice input signal in the judgment of the driver's intention, obtaining a modal weight; comprehensively processing the adjusted visual input signal, the adjusted voice input signal and the modal weight to judge the operation intention of the driver, and outputting an operation intention judgment result. The present application has the effect of improving the accuracy and safety of vehicle-mounted human-computer interaction.

This invention discloses a soluble microneedle patch with drug-enriched needle tip and its preparation method, belonging to the field of microneedle technology. The preparation method includes: (1) preparing a drug-loaded solution using a hydrophilic drug and a soluble polymer as solutes, injecting the drug-loaded solution into the needle body part of a microneedle mold, and performing low-temperature centrifugal drying; (2) preparing a polymer solution using a soluble polymer as a solute, injecting the polymer solution into the remaining part of the microneedle mold, centrifuging and drying, and then demolding to obtain the soluble microneedle patch with drug-enriched needle tip. Compared with the conventional natural drying molding method, the two-step centrifugal drying method of this invention can enrich the drug at the tip of the microneedle, thereby enabling rapid release of the drug from the microneedle, improving drug utilization, and reducing waste caused by the inability of the soluble microneedle body to fully enter the skin. At the same time, the microneedles prepared using a soluble matrix avoid organic solvent residue, have good biocompatibility, and good puncture effect.

Embodiments of the present application disclose a power construction document checking method and device, electronic equipment and a storage medium. The power construction document checking method comprises: obtaining a to-be-checked power construction project document, and determining at least one first to-be-checked document and at least one second to-be-checked document in the to-be-checked power construction project document; sequentially inputting each first to-be-checked document into a pre-trained document classification model to determine the document type of each first to-be-checked document through the document classification model; performing complete specification checking on each first to-be-checked document according to the document type of each first to-be-checked document; and performing document logic checking on each second to-be-checked document in the case that the complete specification checking result of each first to-be-checked document meets the complete specification checking condition. The technical solution of the embodiments of the present application can automatically check the power construction document, and improve the checking speed and accuracy of the power construction document.

This invention discloses an intelligent maintenance and fault diagnosis system for wind turbine pitch systems, specifically relating to the field of wind turbines. It includes a multi-sensor data acquisition module, a data preprocessing module, a real-time status monitoring module, a fault feature identification module, an intelligent diagnostic decision-making module, a predictive maintenance strategy module, and a human-machine interface module. Addressing the limitations of existing wind turbine pitch system maintenance methods, such as limited fault diagnosis capabilities, low maintenance efficiency, and difficulty in cost control, this invention employs a modular design and integrates signal processing, machine learning, and visualization technologies to construct an intelligent maintenance and fault diagnosis system. Through the coordinated operation of functional modules such as multi-sensor all-around data acquisition, high-precision fault feature identification, intelligent diagnostic decision-making, and predictive maintenance strategy formulation, it achieves early fault warning and accurate diagnosis of key components of the pitch system, significantly improving equipment operational reliability and maintenance efficiency.