72results about How to "Improve labeling efficiency" patented technology

The invention provides a calculation amount drawing-oriented large model training data labeling method, device and equipment, which are used for labeling key information of a construction engineering cost calculation amount drawing. The method comprises the following steps: displaying a to-be-labeled calculation amount drawing, determining a to-be-identified area, and identifying to obtain text content and position information of a text block; displaying a label list according to the keyword field; after a user selects the tag, activating the interaction state of the to-be-identified area, so that the text block is presented as an operable hot area; after the user selects the hot area, generating a key value pair by using the selected label and the hot area text; and collecting the key value pairs of the to-be-identified areas to generate answer content, and filling the answer content into an answer part of a preset question-answer template to obtain a labeling result. By adopting the method, the field and text matching can be quickly completed, the structured annotation data is generated, the annotation efficiency and consistency are improved, and the method can be conveniently and directly used for large model training.

The invention provides a graph model-based instance relationship judgment method and device, equipment and a product. The method comprises the following steps: acquiring an annotation data set of target instances in a target scene, wherein the annotation data set is obtained by performing feature annotation on each target instance in the target scene according to a pre-training model; according to the annotation data set of each target instance, constructing a graph structure representing an association relationship between the instances; based on the graph neural network model, performing relation analysis on the graph structure to obtain an initial instance relation judgment result; and according to the spatial constraint rule, calibration processing is performed on the initial instance relationship judgment result to obtain a target instance relationship judgment result, so that the detection precision and the labeling efficiency of the target instance in the complex shielding scene are improved, and the instance relationship judgment efficiency is improved.

This invention discloses a method and system for labeling prompt words in a large-scale communication domain model based on knowledge graphs. Belonging to the fields of communication technology and artificial intelligence, it addresses the following technical problems: low efficiency in labeling specialized terms in communication domain prompt word annotation, low accuracy and efficiency of automatic recognition, limited prompt word dimensions, and lagging domain knowledge. The method includes: sequence labeling and entity extraction using a trained combined model to output entities; generating relation triples as candidate relations using a trained generator, and scoring and ranking the selected relation triples using a trained discriminator; semantic parsing of user input to extract key entities and intents; multi-hop reasoning from a dynamic knowledge graph to retrieve relevant entities and relations; eliminating illusory content using a trained GAN network; generating entity prompt words from three levels: basic, extended, and application dimensions; and introducing a triple quality verification mechanism.

The invention provides a data annotation system and method based on a large language model, a medium and a terminal, and the system comprises a data access layer which is used for collecting to-be-annotated data, and carrying out the preprocessing of the to-be-annotated data, so as to enable the to-be-annotated data to be converted into a unified data format; the intelligent annotation engine layer is used for retrieving context information of a vertical field for the received to-be-annotated data in the unified data format, splicing the retrieved context information and the to-be-annotated data and then inputting the spliced information and the to-be-annotated data into a plurality of pre-trained large language models; the large language model is used for generating a labeling result of the to-be-labeled data; and the quality control layer is used for monitoring the quality of the labeling results generated by the plurality of large language models and carrying out feedback analysis according to the monitored labeling quality problem. According to the method and the device, the automation of the whole process from the acquisition of the to-be-labeled data to the generation of the labeling result can be realized, the vertical domain knowledge is integrated, the labeling result is strictly controlled, and the process is continuously subjected to feedback optimization.

The application provides a metadata intelligent identification method and device based on semi-supervised learning, and the method comprises the following steps: generating a metadata keyword identifier according to a conditional random field, wherein the metadata keyword identifier is used for extracting metadata features corresponding to any identification; acquiring an initial labeled data set, training the metadata keyword identifier according to the initial labeled data set to generate a metadata classifier; acquiring an unlabeled data set, predicting the unlabeled data set according to the metadata classifier to generate a prediction result; generating an intermediate training data set according to the prediction result; cyclically self-training the metadata keyword identifier and the metadata classifier according to the intermediate training data set to obtain a cyclically self-trained metadata classifier; and identifying metadata intelligently according to the cyclically self-trained metadata classifier. Through the application, the problem of low metadata identification construction efficiency in the related art is solved.

The invention discloses a video data labeling method and device, electronic equipment and a storage medium. The method comprises the following steps: determining target video data and target information in the motion process of the intelligent equipment, wherein the target video data is video data collected by the intelligent equipment; determining a target image based on the target video data, and generating a target labeling result of the target image; determining a plurality of reference images based on the target video data, and generating respective target labeling results of the plurality of reference images based on the target labeling results of the target images; based on the target information, the target annotation result of the target image and the target annotation results of the multiple reference images, determining the confidence degree of the target annotation results of the multiple reference images, and based on the confidence degree of the target annotation results of the multiple reference images, determining the annotation result of the target object in the target video data. According to the scheme, the efficiency and accuracy of labeling the video data collected by the intelligent device with the body can be improved.

The application relates to a video labeling method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a playing time corresponding to a to-be-labeled image frame in a playing video; the to-be-labeled image frame is located in a video layer; labeling the to-be-labeled image frame in a labeling layer to obtain labeling information corresponding to the playing time; the labeling information is information of a target labeling frame corresponding to the to-be-labeled image frame, and the target labeling frame is located in the labeling layer; saving the playing video and the labeling information corresponding to the target labeling frame of each playing time in the labeling layer to obtain a labeled target video. The method can improve the efficiency of video labeling.

The invention discloses a multi-modal medical data encryption method and device based on a large model, equipment and a storage medium, relates to the field of artificial intelligence, is applied to a cloud service layer carrying a target labeling large model, and comprises the steps of obtaining original multi-modal medical data and corresponding target identity data sent by edge end equipment corresponding to the cloud service layer; performing feature extraction on the target identity data to obtain a corresponding biological feature vector, generating a target dynamic key based on the biological feature vector by using a target dynamic key generation algorithm, and encrypting the original multi-modal medical data by using the target dynamic key to obtain corresponding encrypted medical data; and obtaining a target labeling result corresponding to the encrypted medical data by using the target labeling large model, and binding the biological feature vector with the target labeling result by using the target dynamic key so as to encrypt and store the target labeling result in a target database. According to the method and the device, the data annotation efficiency is improved while the privacy disclosure risk is reduced.

The application discloses a virtual object marking method and device, a terminal and a storage medium, and relates to the field of human-computer interaction. The method comprises the following steps: receiving a touch operation on a virtual object, displaying a marking mode prompt based on the pressing degree of the touch operation, the marking mode prompt being used for indicating the marking mode currently in, wherein different marking modes correspond to different pressing degrees, and the marking modes are different in the marking mode of the virtual object, and in the case of being in a target marking mode, in response to an end signal of the touch operation, the virtual object is marked by a target marking mode indicated by the target marking mode. The embodiment of the application improves the efficiency of marking the virtual object in human-computer interaction.

This invention provides a method, apparatus, device, and medium for extracting water body information based on remote sensing data, comprising: acquiring multispectral image data of a target area; constructing a water body mask image corresponding to the target area based on the multispectral image data; initializing a threshold population corresponding to the water body mask image, iteratively updating the threshold population, and determining a target segmentation threshold corresponding to the water body mask image based on the iteratively updated threshold population; generating a binary image of the water body corresponding to the target area based on the target segmentation threshold; and constructing sample data based on the binary image of the water body, so as to extract water body objects contained in the target area based on the sample data using a water body information extraction model. This invention can automatically generate high-precision sample images, thereby improving the tedious and time-consuming problems caused by manual sample annotation, improving sample annotation efficiency, and significantly improving the accuracy of water body object extraction.

The utility model discloses a packaging box label attaching device, which relates to the technical field of label attaching and comprises a base and a control device, the base is provided with a conveying part electrically connected with the control device, and the conveying part is used for conveying a packaging box to be attached with a label. The packaging box labeling device further comprises a limiting mechanism used for limiting the packaging box in the conveying process, an adjusting mechanism arranged on the arch plate and a labeling mechanism used for attaching labels to the packaging box. According to the device, the limiting mechanism is arranged, so that the device can limit a packaging box in the transportation process, the situation that a label cannot be attached to an area, to be attached with the label, of the packaging box subsequently due to the fact that the packaging box deviates in the transportation process is avoided, and meanwhile, an electric push rod is arranged, so that the packaging box is convenient to transport. By arranging the limiting mechanisms on the two sides, the distance between the limiting mechanisms on the two sides can be adjusted through the device, packaging boxes of different sizes can be limited through the device, and therefore the application range of the device is widened to a certain degree.

The utility model belongs to the technical field of circuit board processing, specifically disclose circuit board labelling positioning device, including bottom plate, conveyer, set up on the bottom plate, be used for transmission circuit board, jacking mechanism, set up on the bottom plate, jacking mechanism includes driver and sets up the material assembly of driver output end, and material assembly can pass through driver drive circuit board vertical movement, and separate conveyer, labelling mechanism, set up on the bottom plate, labelling mechanism includes feeder and mechanical arm, and mechanical arm is used for fetching the label of feeder and pastes on circuit board. The utility model in through setting jacking mechanism, can make circuit board separate conveyer, and realize accurate positioning, again through labelling mechanism carries out the labelling operation, has promoted the labelling efficiency of circuit board, has promoted circuit board productivity.

The application discloses an active domain adaptation semantic segmentation method, system, device and storage medium, selects and labels taking superpixels as units, which is different from image-level and pixel-level labeling methods, and the superpixel-level labeling greatly improves labeling efficiency by only assigning a semantic category to each superpixel; in addition, different from the existing scheme based on the uncertainty selection labeling strategy, the application focuses on difficult example samples in the domain adaptation scene, and proposes a selection strategy based on domain information quantity to label superpixels most valuable for domain adaptation learning; by adopting the superpixel-level labeling method and the selection strategy based on the domain information quantity, the application greatly reduces the labeling cost while improving the labeling quality, and guarantees the performance of the domain adaptation semantic segmentation.

The application discloses a map annotation method and device, electronic equipment and storage medium, the method comprises the following steps: generating the annotation control point of each lane according to the road information of map data; drawing a plurality of annotation straight line segments of each lane connected in a loop according to the annotation control point of each lane; for each group of two connected annotation straight line segments, drawing a circular arc tangent to the two annotation straight line segments to obtain a plurality of annotation circular arc segments of each lane by passing through the points on the two annotation straight line segments at a distance of half of the length of the shorter annotation straight line segment from the connecting point; deleting the line segment between the connecting point and the corresponding annotation circular arc segment to obtain the lane annotation line of each lane; and using the lane annotation line to annotate each lane in the map data. The application can provide a universally applicable map lane annotation method, support annotation of different map formats, and facilitate users to realize map annotation without relying on existing map annotation software.

This application provides a method and related apparatus for processing landmarks in a map. The method includes: a first object inputting and confirming a first landmark code; a terminal obtaining import information of a first landmark set exported by a second object corresponding to the first landmark code; the first object confirming the import information; and the terminal importing the first landmark set into the application map page based on the import information. Based on this, the first object does not need to manually mark each landmark in the application map to gradually explore the application map. Instead, it obtains the import information of the first landmark set exported by the second object by inputting and confirming the first landmark code corresponding to the first landmark set exported by the second object. After confirming the import information, the first landmark set exported by the second object is successfully imported into the application map page. This method is relatively simple to operate, consumes less time, improves the efficiency of landmark marking in the map, and thus improves the map exploration efficiency.

The utility model discloses a cosmetic bottle bottom labeling device which comprises a conveying belt, a feeding rubber sleeve is arranged on the outer side of the conveying belt, a plurality of sets of evenly-distributed containing grooves are formed in the outer side of the feeding rubber sleeve, cosmetic bottles are arranged on the inner sides of the containing grooves, an L-shaped plate is fixedly connected to the right end of the rear side wall of a mounting plate, and a discharging air cylinder is fixedly connected to the middle of the rear wall of the L-shaped plate. The output end of the discharging air cylinder is fixedly connected with a synchronous plate, and the front wall of the synchronous plate is fixedly connected with a plurality of evenly-distributed push rods. According to the cosmetic bottle bottom labeling device, when a cosmetic bottle needs to be labeled, a conveying belt works to enable a top feeding rubber sleeve to move and drive the cosmetic bottle body to move at the same time, in the cosmetic bottle body moving process, a first feeding air cylinder works, and a second feeding air cylinder and an electric sliding table work; the first feeding air cylinder works to enable the material taking head to move backwards to make contact with a label at the left end of the front side of the label plate for material taking, then the second feeding air cylinder works to push the label to the bottom of the cosmetic bottle, and automatic labeling operation is completed.

This application relates to the field of autonomous driving technology, specifically providing a data generation method, perception method, device, and storage medium for traffic lights, aiming to solve the problem of how to quickly and accurately acquire traffic light image annotation data. To this end, the method provided in this application includes tracking and detecting individual lights in each frame of a video clip to obtain the shape and state of the light head; clustering all individual lights in the target image to obtain at least one cluster, and using the cluster as a traffic light intersection; acquiring the lit individual lights within the traffic light intersection, acquiring the traffic direction of the virtual light corresponding to the shape of the lit individual light's light head, and determining the state of the traffic direction based on the state of the lit individual light; generating virtual lights for the traffic light intersection based on the traffic direction and its state; and annotating the target image based on the virtual lights to form image annotation data. Based on the above method, traffic light image annotation data can be acquired quickly and accurately.

This invention discloses a method and system for instrument reading recognition based on image segmentation, comprising: acquiring an instrument image to be recognized; extracting mask information of the pointer region, the center region of the dial, and a predetermined triangular region from the instrument image using a trained semantic segmentation model; wherein the predetermined triangular region is the triangular region formed by the first and last scale values ​​of the dial and the center region of the dial; separating the pointer region, the center region of the dial, and the predetermined triangular region into three independent channels based on the mask information of each region, and performing rotation correction on the instrument position in the instrument image; performing triangle fitting on the predetermined triangular region, selecting key points in the center region of the dial, the pointer region, and the fitted predetermined triangular region, and calculating the instrument reading based on the key points. The segmentation method of this invention has higher annotation efficiency and fewer, larger, labeled target regions, resulting in an overall improvement in algorithm performance.

This invention provides an automatic fiber extraction method, a training method for a fiber segmentation model, and an electronic device. The automatic fiber extraction method includes: segmenting a patient's medical image using a fiber segmentation model to obtain a segmentation result; wherein the fiber segmentation model is a pre-trained deep learning model for extracting slender structures; and performing connected component analysis on the segmentation result to extract the fiber region. This invention pre-trains a fiber segmentation model, which is particularly suitable for extracting slender structures in three-dimensional medical images, improving the accuracy and efficiency of fiber segmentation. After obtaining the fiber mask through segmentation, connected component analysis is performed to further accurately screen out the fiber region, eliminating interference from artificial filling materials, blood vessels, and other noise, thus more accurately determining the fiber region.

The present application relates to the technical field of computer vision and image processing, and discloses a remote sensing weakly supervised fine-grained target detection and recognition method and device, based on a general text prompt corresponding to a small amount of labeled remote sensing data samples and a large amount of unlabeled remote sensing data samples, by extracting the geometric features of the samples and the cross-modal features representing the visual text differences, combining the fine-grained class labels of the labeled remote sensing data samples, prior knowledge prototypes of various fine-grained classes are constructed; then, the prior knowledge prototypes of various fine-grained classes are used to construct fine-grained soft labels of the unlabeled remote sensing data samples, which are used as supervision signals to realize the training of the model, solve the cost consumption problem of fine-grained labeling, improve the label labeling efficiency of the unlabeled remote sensing data samples, and further improve the model training efficiency and target recognition efficiency.

The application discloses a sample data determination method, a data processing method, a device, equipment and a medium, wherein the method comprises: acquiring an unlabeled data set; taking part of the unlabeled data from the unlabeled data set and labeling the part of the unlabeled data to obtain a first data set with labels corresponding to the part of the unlabeled data; training a type recognition model by using the first data set, the type recognition model being used for recognizing the unlabeled data to obtain a recognition result; acquiring corresponding remaining unlabeled data from the unlabeled data set; and determining a labeled sample data set based on the first data set, the type recognition model and the remaining unlabeled data. The application can save labor cost and time cost in the data labeling process, thereby improving the efficiency of data labeling.

The invention discloses a DNA structure for saccharomyces cerevisiae single-cell barcode labeling and chromatin state conjoint analysis and application thereof, and belongs to the technical field of molecular biology and single-cell multi-omics analysis. From the 5'end to the 3 'end, a fluorescent protein gene, a Tn5ME structure, a reverse transcription primer binding region, a CYC1 terminator fragment and a linearized shuttle plasmid skeleton are subjected to homologous recombination, and a pCM carrier is obtained; after NdeI / BamHI enzyme digestion, oligonucleotides with random sequences, which are subjected to annealing, amplification, Klenow filling-in and enzyme digestion, are connected, and transformation extraction is carried out, so that the DNA structure for saccharomyces cerevisiae single-cell barcode labeling and chromatin state combined analysis is obtained. On the basis of the constructed DNA structure, ATAC-seq and a bar code reverse transcription method are combined, important data related to the transcription state and chromatin state of a single cell can be obtained at the same time through bar code information, single cell pedigree tracking of the saccharomyces cerevisiae is achieved, and research on single cell epigenomics of the saccharomyces cerevisiae and other microorganisms is promoted.

This application discloses an intelligent assisted annotation method and apparatus for remote sensing images, relating to the field of image annotation technology. The method includes: constructing a total cost function based on a preprocessed remote sensing image; treating each pixel of the remote sensing image as a node in a connected weighted graph, and determining the cost value of weighted edges between adjacent pixels based on the total cost function; wherein the total cost function is configured such that the cost value of weighted edges between pixels located on the edges of features in the remote sensing image is less than a preset cost value; and drawing the shortest path between adjacent marker points based on marker points sequentially input by the operator along the edges of features in the remote sensing image, forming a preliminary outline. This method can improve the annotation efficiency of annotators, save manpower and resources, and effectively solve the problem of decreased annotation accuracy due to annotator fatigue.

The present disclosure relates to an image sample labeling method and device, a storage medium and an electronic device, and relates to the technical field of computers. The method displays a target image pair, and in response to a selection operation on a target image region of any image in the target image pair, synchronously determines a to-be-labeled image block corresponding to the target image region in the first image and the second image, then labels the to-be-labeled image block, and obtains an image labeling result of the target image pair. The method can not only improve the labeling efficiency of image samples, but also make the sample labeling result more accurate by comparing the first image and the second image for labeling.

The invention provides an automatic material labeling method and device, equipment and a storage medium. The method comprises the following steps: identifying a material type of a material uploaded by a user; preprocessing the material uploaded by the user according to the material type to obtain a to-be-processed material; performing label extraction on the to-be-processed material to obtain a plurality of candidate labels meeting output specifications; vector-based recall sorting and frequency calculation are carried out on the candidate tags, and fusion sorting is carried out on the candidate tags according to a recall sorting result and a frequency calculation result; and outputting a preset number of labels according to the fusion sorting result. According to the method, by identifying the material types and adopting the corresponding preprocessing and label extraction strategies, unified and accurate label generation can be performed on various types of marketing materials such as documents, pictures and videos at the same time, and the efficiency of labeling the materials is effectively improved.

The application discloses a folding cloth label sticking device which comprises a belt conveying mechanism provided with at least two belts and a label sticking mechanism; the two belts are arranged at intervals to form a gap; the label sticking mechanism comprises a connecting plate, a label pressing assembly, a unwinding roller fixed on the connecting plate and a label stripping plate, the bottom surface of the label stripping plate is a label stripping surface, the label stripping surface is arranged at intervals from the belt and is arranged opposite to the gap; the label pressing assembly comprises an upper label pressing assembly and a lower label pressing assembly; the upper label pressing assembly comprises a fixed connecting piece fixed on the connecting plate and an upper pressing wheel rotatably connected to the first connecting piece; the upper pressing wheel is located at the rear side of the label stripping plate and is arranged opposite to the gap; the lower label pressing assembly comprises a fixed second connecting piece and a lower pressing wheel rotatably connected to the second connecting piece; the lower pressing wheel is located below the upper pressing wheel; the distance between the upper pressing wheel and the belt and the distance between the upper pressing wheel and the lower pressing wheel are matched with the thickness of the folding cloth to be labeled. Compared with the prior art, the device can automatically stick labels on the folding cloth, reduces the labor cost and improves the label sticking efficiency.

The invention discloses a cross-modal knowledge collaboration blast furnace condition evaluation method, and belongs to the technical field of blast furnace production. The method comprises the following steps: a data acquisition module for acquiring sensor data, process parameters and historical data in blast furnace production; the data preprocessing module is used for cleaning, standardizing and extracting features of the acquired data; the cross-modal knowledge cooperation module comprises knowledge extraction, knowledge fusion and knowledge graph construction, and is used for extracting feature and semantic information from multi-modal data, realizing association and complementation among modals and forming a comprehensive knowledge system; the model training module is used for training a LightGBM model by utilizing the marked historical data and learning a mapping relation between the data and the furnace condition; and the evaluation and feedback module is used for inputting multi-modal data in real time, outputting a furnace condition evaluation result and dynamically adjusting a production control strategy. According to the method, through cross-modal knowledge cooperation, high-precision, high-adaptability and real-time evaluation of the furnace condition of the blast furnace is achieved, the production efficiency can be remarkably improved, energy consumption and accident risks can be reduced, and blast furnace production intellectualization is promoted.

The application discloses a kind of multi-level man-machine collaborative data labeling method and system based on task ambiguity evaluation, first constructs and trains ambiguity prediction model, utilizes lightweight ambiguity prediction model, predicts the cognitive ambiguity score and logic ambiguity score of each data instance;Four-stage labeling resource pool containing low-performance large language model, high-performance large language model, ordinary crowd sourcing labeler and expert labeler is constructed;According to the predicted two-dimensional ambiguity score, the data instance is automatically routed to the optimal labeling resource by dynamic scheduling strategy, for the task with high cognitive ambiguity, the upgrading mechanism combining crowd sourcing voting and expert arbitration is used, and the final labeled data set is output.The application can intelligently match labeling cost and ability according to the intrinsic properties of the task, significantly reduce the labeling cost and improve the labeling efficiency under the premise of ensuring the data labeling quality, solve the problem of high cost of traditional crowd sourcing and unstable quality of single model labeling.

An automatic postage label pasting device includes: a frame; a label pasting machine installed on the frame for conveying postage label tape, the postage label tape including a postage label and release paper; a mail conveyor belt installed on the frame for conveying mail along the conveying direction; and an inclined guide disposed above the conveying path of the mail conveyor belt, a gap being formed between the lower surface of the end of the inclined guide and the upper surface of the mail conveyor belt; the inclined guide has a folding portion for guiding the postage label tape to fold back; wherein the height of the gap is configured to be greater than the thickness of the mail to be pasted; and the inclination angle of the inclined guide is configured as follows: This utility model has a novel structure. Through the design of the inclined guide, the pushing force of the mail itself during the conveying process is used to completely peel the postage label from the release paper and flatly paste it onto the surface of the mail, ensuring the stability and consistency of the pasting quality.