57 results about "Single label" patented technology

The invention provides an object retrieval method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a retrieval target which comprises at least one seed object; extracting semantic features of each seed object from the description text of each seed object; according to the semantic features of the seed objects, a candidate object set similar to the retrieval target semantics is recalled from a candidate object library; under the condition that the data volume of the candidate object library is smaller than a threshold value, sorting the recalled candidate object sets by utilizing the attribute tags of the seed objects; and determining a similar object recommendation list corresponding to the retrieval target from the candidate object set according to the sorting result. According to the method, the deep semantics of the unstructured text and the concrete index of the structured tag are integrated, so that the problems of single matching dimension and insufficient accuracy caused by only depending on single tag matching are avoided, and the accuracy and comprehensiveness of similar object retrieval are remarkably improved.

The invention discloses a TBM surrounding rock geology sensing method and system based on multi-source heterogeneous information, and the method carries out the feature fusion based on the tunneling parameter statistical features and rock slag image analysis features, carries out the model prediction based on the fusion features, achieves the deep fusion of the multi-source heterogeneous information, and improves the precision of the TBM surrounding rock geology sensing. The prediction accuracy of the model is improved, the model output is the surrounding rock score, so that the model training is not limited by the sparsity of a single label, the training precision of the model is improved, and in addition, the surrounding rock grade quantitative score and the surrounding rock state qualitative evaluation are combined, so that the surrounding rock score and the surrounding rock grade are in discontinuous correspondence. Continuous evaluation of the surrounding rock state is achieved, and the problem that a surrounding rock grading mode of a traditional method is discrete is solved.

The invention discloses a multi-label medical image classification method, belongs to the technical field of medical images, and solves the problem that an existing CNN-based multi-label medical image classification method is low in classification accuracy. According to the method, the image data in the training image set and the clinical risk data in the feature set are combined, the features of which the importance ranks in the front are selected from the single labels through the univariate decision tree model to form the top-N feature set, and then the top-N feature set is input into the MedFusionNet model, so that the multi-mode learning of the MedFusionNet model is expanded, and the accuracy of the multi-mode learning of the MedFusionNet model is improved. According to the method, input features can be represented more abundant, challenges of label dependency, data imbalance and modal integration are effectively solved, meanwhile, interpretability and predictability are enhanced, and classification accuracy is improved.

The invention provides a client follow-up information processing method and device and electronic equipment. The method comprises the steps that client follow-up information to be processed currently in a client management system is acquired; performing sentiment analysis prediction on the customer follow-up information through a preset sentiment analysis model, and outputting a sentiment classification tag corresponding to the customer follow-up information; determining a target customer index type corresponding to the customer follow-up information based on the emotion classification labels corresponding to the customer follow-up information; based on the sentiment classification label and the target customer index type, counting the customer follow-up information into existing billboard data; the billboard data comprises statistical data under a single label dimension and / or a combination of a plurality of label dimensions, and after an event meeting a decision triggering condition is monitored in the billboard data, decision reminding is carried out on salesmen. In the scheme of the invention, customer judgment standards are unified, automatic judgment, analysis and decision reminding can be realized, the transaction period can be shortened, the conversion rate can be improved, and the customer maintenance cost can be reduced.

The invention discloses a label portrait analysis method and device, and relates to the technical field of computers. A specific embodiment of the method comprises the steps of obtaining a target entity set and a historical entity set corresponding to a target tag; generating a link ratio difference set of the target label according to the target entity set and the historical entity set; feature data of each tag is constructed, a stability result of each tag is obtained according to the feature data of each tag, and the tags comprise a target tag and a portrait tag corresponding to the target tag; and for any portrait label, under the condition that both the stability result of the portrait label and the stability result of the target label indicate stability, generating a target portrait result of the target label according to the link ratio difference set. According to the embodiment, rapid portrait analysis can be carried out on a single label, the timeliness of label portrait analysis is improved, resource consumption is reduced, and large-scale expansion of the number of portrait types can be realized.

The system comprises of a security label engaged to a package. The label comprises of a first layer with a first code and a second layer with a second code. The first layer is disposed over the second layer in a way that at least a portion of the second code on the second layer is covered by the first layer. The second code on the second layer is revealed upon irreversibly disengaging the first layer from the package. The first code is associated with a first information and the second code is associated with a second information. The second code on the second layer is associated with the first code on the first layer. These codes are authentication-wise interlinked and as one single label which can be pre-applied on product-label and still extended authentication challenge like refill-evidence or tamper-evidence in addition to basic counterfeit-evidence.

The invention belongs to the technical field of tag systems, and provides a user-defined tag generation method and system based on a metadata configuration tag rule, and the technical scheme is as follows: constructing a database by using obtained data information of each enterprise; reading table field information based on the database, and setting a label rule based on the read table field information; the method comprises the following steps: receiving batch label generation tasks, judging an associated label rule according to the generation tasks, determining the size of a required thread pool and a thread allocation scheme based on the label rule, hardware resources of a system and the complexity of the tasks, executing the label generation tasks by each thread, and generating corresponding labels; and assembling based on the generated single label result and a set combinatorial logic relationship to obtain a composite label. The custom labels can be timely and effectively generated in batches, and the real-time data analysis requirement of an enterprise is met.

A label feed system (15) for a labeling machine (1) is described, the label feed system (15) comprising: a sensor device comprising an optical sensor (17) having a fixed field of view (FV); a feed roller (11) for longitudinally feeding a web (4) of labeling material along a feed path (P) and through said field of view (FV), the web (4) comprising a plurality of longitudinally connected labels (3), each label (3) having a plurality of longitudinal portions and an aesthetic feature (AF) arranged at a determined longitudinal portion; and a reference member (16) fixed with respect to the feed path (P) and the field of view (FV) and arranged within the field of view (FV); the sensor device is configured to operate in a production mode according to which, for each label (3), the optical sensor (17) sequentially acquires an image (I) of the field of view (FV) at said determined longitudinal portion to detect a linear distance between the respective aesthetic feature (AF) and the reference member (16); the sensor device is further configured to operate in a recovery mode according to which the optical sensor (17) acquires a plurality of images (I1, I2, I3) of the field of view (FV) for searching for an aesthetic feature (AF), where each image (I1; i2; i3) of the plurality of images comprises at least one longitudinal portion which is different from the other images (I2, I3; i1, I3; i1, I2) to inspect all different longitudinal portions cumulatively by a single label (3) or by a plurality of labels (3a, 3b, 3c) during the recovery mode.

This invention relates to a labeling device (1) for affixing labels to a container, comprising: at least one unfolding device (7a) for a label strip (7); at least one cutting blade (5) for cutting the label strip into individual labels; and a vacuum transfer cylinder (2) for transferring the labels to the container, wherein the vacuum transfer cylinder (2) has a plurality of movable vacuum suction rods (16a, 16b, 16c, 17a, 17b, 17c, 18a, 18b, 18c, 19a, 19b, 19c, 20a, 20b, 20c, 21a, 21b, 21c) for suction, conveying, and transferring the labels. The label is described, wherein at least one vacuum suction rod is assigned to each label to be transported; at least one rod controller is used to control the movement of the vacuum suction rod, wherein the rod controller includes at least one control cam that interacts with at least one control roller (16aa) of the vacuum suction rod (16a) to control the movement of the corresponding vacuum suction rod (16a), and wherein the at least one vacuum suction rod (16a) assigned to the label to be transported is controlled by the control cam (11a, 12a) of the rod controller, or wherein all vacuum suction rods (16a, 16b, 16c) assigned to the label to be transported are controlled by the control cam (11a, 12a) of the rod controller.

1. Name of the product under this design: Thermal Label Printer (Y41BY‑A). 2. Purpose of the product of this design: equipment used for printing. 3. The key point of the design of this product lies in its shape. 4. The picture or photo that best illustrates the design points: Stereoscopic drawing 1.

The present invention discloses a method and device for answer sentence selection by fusing hierarchical capsules with multi-view information. First, question texts containing multiple answers on the Internet are collected and preprocessed; then, a hierarchical capsule network is used to perform information aggregation and word aggregation on the representation of the encoded output through its progressive aggregation feature, thereby ensuring the accurate transmission of effective information and being able to use the same feature for parameter transmission; then, in order to solve the problem of instability in the fine-tuning step and the problem that the final output of the pre-trained model can only predict a single label, the weight of the entire BERT encoding intermediate layer output is calculated using a multi-view information fusion method to weight the learning information. The present invention can improve the problems in previous answer sentence selection methods, such as ignoring the influence of the multiple layers of intermediate output within the model on the answer sentence selection and being able to only predict a single result, thereby improving the accuracy and efficiency of answer sentence selection.

A hot melting labeling machine capable of preventing label warping comprises a pinch roll, a tension roll, a cutting rotary roll, a container conveying disc, a glue spreading roll, a fixed cutting mechanism, a telescopic cutting mechanism, a first sensor, a second sensor and a third sensor, the telescopic cutting mechanism is arranged on the cutting rotary roll, and the fixed cutting mechanism is arranged on the side of the cutting rotary roll. When the cutting rotating roller rotates until the inner cutter and the outer cutter are aligned, the inner cutter acts to cut off the label belt, then a label to be pasted is attached to the surface of the cutting rotating roller under the action of negative pressure, the label is pasted on a container after being coated with glue of the glue coating roller, and in the process, the label belt is cut into single labels, and then the single labels are pasted on the container. And the end part of the label cannot be wrinkled or twisted, so that the label adhered to the container cannot be warped.

The present invention relates to a test tube labeling machine. It includes a casing structure, and multiple groups of tube storage cavities arranged side by side are provided in the upper part of the inner cavity of the casing structure; it also includes a tube dropping device, a tube discharging funnel structure, a test tube positioning device and a tube blocking device; a driving pinch roller device is arranged beside the discharging port of the tube discharging funnel structure, and it also includes a driven pinch roller structure. A sliding material guiding plate is arranged below the driven pinch roller structure, and a material taking box is arranged below the discharging end of the sliding material guiding plate; it also includes a roller moving device; a tube pushing device is installed on the roller moving device; a label discharging assembly is arranged below the sliding material guiding plate; it also includes a label supply device, and the label supply device and the driven pinch roller structure are respectively located on both sides of the driving pinch roller device. The present invention can realize the automatic dropping, conveying, labeling and discharging operations of test tubes, with high automation degree and high working efficiency. In addition, the present invention can provide specific specimen test tubes and single label papers, with strong applicability.

The invention provides a single label type paper transfer device based on vision, and the device comprises a vision positioning system which is used for positioning the position of single label type paper; the transfer device comprises a robot, a servo motor and a suction cup, the suction cup is detachably connected with the robot, the servo motor is connected with the suction cup and used for driving the suction cup to rotate, the robot comprises a driving system and a mechanical arm, the suction cup comprises an upper aluminum block, a lower aluminum block and four ball spline linear bearings, and an air cavity is formed in the lower aluminum block; the upper end face of the lower aluminum block is used for being connected with negative pressure, and a first air hole is formed in the lower end face of the lower aluminum block. And the transfer device control system is used for controlling the driving system to drive the mechanical arm to reach a designated position according to the deviation value and controlling the electromagnetic reversing valve to connect or disconnect the vacuum generator and the suction cup. Accurate transfer of single label type paper in the sucking, transferring and stacking processes can be achieved.

The invention discloses a task prompt perception industrial defect detection model fine tuning method, and belongs to the field of model fine tuning, and the method comprises the steps: generating a task prompt feature for a detection model based on a segmentation target of a to-be-segmented image, and determining the image feature of the to-be-segmented image; performing feature fusion on the task prompt feature and the image feature to obtain a fusion feature of the to-be-segmented image; and performing segmentation reasoning based on the fusion feature to obtain a segmentation result of the to-be-segmented image under the segmentation target. According to the method, a task prompt perception mechanism is introduced in the fine tuning process, so that the model can further understand task semantics during segmentation reasoning, the accuracy and stability of a detection result are improved, negative effects caused by single labels or insufficient task understanding are reduced, and the detection efficiency is improved. And the generalization ability of the model under multiple defect categories and multiple scenes is enhanced.

The invention discloses a user interest modeling method and system based on user explicit interest labels and high-intention behavior weighting, and the method is characterized in that a user can independently select an interest label, and can add or cancel the interest label at any time; the system carries out matching recall on the platform information based on the interest label set to generate a candidate information set, and sorts the candidate information set to generate a recommendation list; the system monitors behaviors of the user in the recommendation list, low-intention behaviors and high-intention behaviors are distinguished, the high-intention behaviors trigger interest label weight updating, meanwhile, attenuation processing is executed on interest labels without behaviors for a long time, and a weight capping value is set; when the user actively cancels a single tag or all tags, the system stops tag recommendation or switches to first-level classification bottom recommendation; the updated interest label set and the corresponding weight participate in the next round of recommendation, dynamic closed-loop recommendation is achieved, and it is ensured that the recommendation result continuously fits the user interest and behavior data.

The present invention discloses a method for generating a backdoor attack defense model based on object detection. First, benchmark data is selected, and a clear image X is selected from a single label class Y to be detected as the benchmark data; the SinGAN model is used to perform multi-scale learning on the benchmark data, the parameters of the SinGAN model are randomly initialized, and the data of the 0th layer with relatively large differences is selected as the enhanced clean data set D c ; a training backdoor set D with a quantity of N and a size of (w, h) is given t , and D t is placed at any position in the clean data set D c . The synthesized sample set is denoted as the poisoned training set D p ; the trained object detection tool Yolov5 model is used to generate a backdoor defense model M p for the poisoned training set D (x,y,w,h) . Thus, the final model has a high defense effect against backdoor attacks on deep neural networks, improving the robustness and security of deep neural networks.

To provide a labeling device.SOLUTION: A conveyor of containers to be labelled along a path of travel, and one labelling unit comprising a respective transfer drum arranged along the path of travel and rotatable about its own axis of rotation with respect to the supporting base, and configured to receive from a supply station labels to be applied to the containers from the supply unit, and to transfer the labels received at the supply station to the respective containers at a transfer station angularly spaced from the supply station about the axis of rotation of the transfer drum; A sector is defined on the lateral surface of the transferring drum, extending along a portion of the circumferential extension, and containing a single label to be transferred to a corresponding container, a portion of the sector being rotatably movable about a rocking axis substantially parallel to and spaced apart from the rotation axis of the transferring drum, about a rocking axis arranged close to the axis of the transferring drum, about a rocking axis arranged spaced apart from the rotation axis of the transferring drum.SELECTED DRAWING: Figure 1

The invention belongs to the technical field of in-image target recognition, and relates to a training set labeling and automatic generation method for in-image military target recognition. According to the specific technical scheme, the method comprises the steps of obtaining original pictures and unifying the size, dividing the original pictures into multiple groups of working directories, processing all the pictures in a two-focus manual labeling mode for each target category and forming a labeling file, labeling each picture in each group of working directories for the military target of the same target category, the method comprises the following steps of: performing first-stage merging processing on a single label file under each file directory, then performing sampling review, performing second-stage merging processing on a plurality of label files of the same picture under different groups of working directories, and automatically generating a training set by using the formed label files and original pictures according to a directory organization form required by a deep learning framework. The method is beneficial for realizing the automation of training set construction, and the manual labeling process adopts a simple and easy-to-implement method to quickly execute bifocus labeling, so that the manual efficiency can be effectively improved.

A liner-label combination is provided for a tamper evident label. The combination includes a label substrate and a liner substrate. The label substrate includes label substrate die cuts that define a top portion, a middle portion, and a bottom portion for the label. The liner substrate includes liner substrate die cuts that correspond to the middle portion of the label. The top portion, middle portion, and bottom portion adapted to be separated from the liner substrate as a single label with a backside of the middle portion including the corresponding liner substrate defined by the liner substrate die cuts. The middle portion of the label adapted to cover an opening of an item and when removed from the opening, the middle portion separates from the top portion and the bottom portion along the label substrate die cuts providing evidence that the opening to the item was exposed.

The application discloses a data labeling method and system, a model training method, a medium and an electronic device, and relates to the technical field of data processing. The data labeling method comprises the following steps: in response to a labeling personnel completing labeling of a piece of to-be-labeled data, determining a labeling duration of completing the labeling; periodically aggregating labeling durations of each labeling personnel; for each labeling personnel, determining a statistical feature of labeling durations of multiple pieces of to-be-labeled data completed by the labeling personnel; according to the statistical feature, quantifying the labeling durations of the multiple pieces of to-be-labeled data into corresponding reliability coefficients respectively; and outputting a structured data set, wherein the structured data set comprises labels of each piece of to-be-labeled data and corresponding reliability coefficients. Thus, by quantifying the labeling duration into the reliability coefficient, the labeled data is expanded from a single label to a structured data containing reliability information, which can provide differentiated sample weight basis for model training, and improve training efficiency and model precision.

This invention relates to the field of labeling machine technology and discloses a fully automatic labeling machine, comprising: a computer labeling machine, wherein a label suction module is provided at the end of a four-axis labeling robot in the computer labeling machine; the label suction module includes a base plate, a label suction seat, and a pushing module, the base plate is fixedly connected to the end of the four-axis labeling robot and moves synchronously with the four-axis labeling robot; multiple sets of label suction seats are arranged in a matrix array below the base plate, and each label suction seat is equipped with a vacuum adsorption unit at its bottom for independently adsorbing a single label; the pushing module includes two pushing arms and a kinetic energy supply module, and a transverse groove is opened through the center of the label suction seat, and the two pushing arms are slidably assembled in the transverse groove; this invention, by setting a pushing module including two pushing arms and a kinetic energy supply module, and performing a two-stage movement of "first pressing vertically down on the center, and then rolling horizontally in both directions", makes the center of the label paper first contact the surface to be labeled, and then continuously rolls from the center to both sides.

The invention discloses a map-image semantic change detection method and system based on multitask-single label learning. According to the method, efficient learning of semantic change detection of a previous land cover map and a newest optical image is realized through semantic change detection and distribution transformation. The method comprises the following steps: firstly, extracting a change region between map data and an optical image and a semantic category of image pixels by using a multi-task model, and then converting a semantic category probability into a binary change probability through a distribution transformation function; through the design, on one hand, the latest and best semantic change detection model can be integrated; and on the other hand, learning of the semantic change detection model can be efficiently driven only by using a binary change tag or a semantic segmentation tag, and finally, high-performance and low-cost semantic change detection is realized. According to the method, modeling of map-image pair change area positioning and pixel category identification can be realized by virtue of an existing optical remote sensing image semantic change detection method based on mature research.

Embodiments of the present disclosure disclose a method, device, system and electronic medium for recognizing annotated incomplete images. A specific embodiment of the method includes: annotating an annotated incomplete image with a single label, inputting the annotated incomplete image into a feature extraction backbone network to extract a plurality of features at different levels; constructing a semantic mask module to generate a semantic mask according to a class activation map of the image and extract features of a single object from the features at different levels; constructing a mask optimization module to filter a certain amount of significant semantic masks for each class of object from all images; and constructing a bootstrap learning framework based on an expectation maximization algorithm to gradually recover and supplement constraint information in the network training stage by comparing semantic masks of objects of different labels, so that the network can more accurately recognize annotated incomplete images. The embodiment enhances the ability to distinguish, locate and recognize different objects in complex images, and improves accuracy through bootstrap.

1. Name of the product under this design: Thermal Label Printer (Y41BY‑B). 2. Purpose of the product of this design: equipment used for printing. 3. The key point of the design of this product lies in its shape. 4. The picture or photo that best illustrates the design points: Stereoscopic drawing 1.

The utility model discloses a waterproof self-adhesive label, and belongs to the technical field of labels, the waterproof self-adhesive label comprises a main body module and a label module, the main body module comprises a substrate, the top and bottom of the substrate are each provided with two first waterproof layers, the label module is arranged on the substrate, and the label module comprises a plurality of label assemblies arranged on the substrate; the label assemblies are arranged on the base plate, the splitting assemblies are arranged in the middle of the base plate, the top of each label assembly is provided with a second waterproof layer, through the arrangement of the label assemblies, a single label can be split to be carried and used, through the arrangement of the splitting assemblies, two labels can be split to be carried and used, and through the arrangement of the second waterproof layers, the two labels can be split to be carried and used. Through the arrangement of the waterproof layer, the label assembly can have the waterproof capability, and through the arrangement of the first waterproof layer, the substrate can have the waterproof capability.

A robot includes a controller programmed to: when a single tag is detected in an image captured by an imaging sensor, apply a first tag detection algorithm to the image to obtain pose data of the single tag; when two or more tags are detected in the image, apply a second tag detection algorithm to the image to obtain pose data of the two or more tags; obtain pose data of the single tag in a map frame or pose data of the two or more tags in the map frame; determine pose data of the robot in the map frame based on a comparison of the pose data of the tag in the image and the pose data of the tag in the map frame; and operate one or more motors to autonomously navigate the robot based on the pose data of the robot in the map frame.

The invention relates to high-speed package and single RFID label coding and detecting equipment, and relates to the technical field of RFID labels, the high-speed package and single RFID label coding and detecting equipment comprises a cabinet, a touch screen, an alarm lamp and a UV ink-jet printer, the touch screen is arranged on the upper end face of the cabinet, the alarm lamp is arranged on one side of the touch screen, the UV ink-jet printer is installed on the side end face of the cabinet through an installation frame, a UV lamp set is installed on one side of the UV ink-jet printer, and the UV lamp set is installed on the other side of the cabinet. A package working platform is arranged on the side, away from the UV ink-jet printer, of the UV lamp set, a first package antenna is arranged below the package working platform, and a second package antenna is arranged above the package working platform. According to the device, through combined use of the single-piece label and the coiled label, integrated arrangement of the single-piece label and the coiled label is achieved, the printing, reading and writing detection functions are achieved, when the single-piece label is used, the conveying belt and vacuum adsorption are used, the detection and kicking function can be achieved, and when the coiled label is used, the magnetic panel is used for covering and isolating the conveying belt, and the forward and reverse rotation and back-off function can be achieved.

The invention relates to the technical field of data statistics, in particular to a data statistics method and related equipment, and the method comprises the steps: obtaining to-be-counted data; analyzing the to-be-counted data, and generating a corresponding data label; judging whether the data label is a single label or not; if the data label is a single label, generating a statistical scheme based on the data label; if the data tags are not single tags, comprehensive tags are generated based on different data tags; generating a statistical scheme based on the comprehensive label; and generating a statistical result based on the statistical scheme. According to the method and the device, the accuracy of data statistics can be improved.

Generally disclosed herein is an approach for modifying use of segment routing multiprotocol label switching (SR-MPLS) allowing an arbitrary MPLS control plane and traditional MPLS data plane to utilize a single MPLS label to represent two or more edges in a path. MPLS labels may be divided into smaller sub-labels, which together uniquely represent a pair of edges along a route. In one example, a single MPLS label may be divided into two sub-labels, the first sub-label representing a first edge, and the second sub-label representing a second edge. In this regard, longer source routes may be supported in a packet header in network designs that implement strict source routing.