69 results about "Label quality" patented technology

The invention relates to a tag noise correction based crowd-sourced tagging data quality improvement method. The method comprises the following steps: running a tag integration algorithm in an initial crowd-sourced tagging data set to form a data set after tag integration, and estimating tagger quality and integrated tag quality information of samples in the process; performing multi-round K-fold cross validation by utilizing the data set after tag integration, and constructing a high-quality data set; determining a tag noise set in combination with the tagger quality and the tag quality of the samples by utilizing a prediction probability of a class tag of each sample in the multi-round K-fold cross validation process; and training a classification model by utilizing the high-quality data set generated in the multi-round K-fold cross validation process, and performing prediction and replacement on the class tag of each sample in the tag noise data set by using the model. With the tag noise correction method, the quantity of potential noise tag samples in the data set after original tag integration is reduced, so that the data quality is improved.

A system and method for label placement is disclosed that achieves the twin goals of practical efficiency and high labeling quality by employing cartographic heuristics. A caller defines map and label properties. Then labels are pulled within a map boundary. Labels are next ordered by priority in descending importance. The order of testing labels is determined. Attempts are made to move overlapping labels. This is an iterative process; therefore there must be criteria that halt the procedure. Upon reaching an acceptable solution, the label properties are adjusted to reflect the new label placements.

The invention discloses a label creation and management system. The system adopts a three-layer structure composed of a model label layer, a service label layer and a combined label layer in turn. Thesystem comprises a management module composed of a model management module, a label management module, a label quality control module, an integrated management module and a combined label module; anda label scheduling module, a label generation module and a label inquiry module. As the technical proposal is adopted, the invention obtains the beneficial effect of saving the label creation time, improving the label creation efficiency, increasing the label combination mode, increasing the intersection and union difference set function, expanding the number of the creation combination label andthe like under the condition that the number of the basic labels is unchanged by adopting the three-layer structure and a plurality of modules.

The invention provides a semi-supervised learning method based on target segmentation field self-learning. The method comprises the following steps: training an initial segmentation network by using marked data in a training data set; generating a pseudo label from unmarked data in the training data set through the trained initial segmentation network; performing shape quality evaluation and semantic quality evaluation on the generated pseudo label; fusing the shape quality and the semantic quality to obtain pseudo label quality; estimating the distribution of the real labels and the pseudo labels, and optimizing the distribution of the pseudo labels; adding data with relatively high pseudo label quality into the training data set to expand the training data set; optimizing the trained initial segmentation network by using the expanded training data set; and iteratively repeating the above steps until the performance of the segmentation network is saturated. The invention further provides a corresponding system, a terminal and a medium. The problem of low segmentation precision in the target segmentation field under the condition of a small number of sample annotations is solved, and good performance is realized.

A paper-like RFID tag is fabricated by bonding bumps of an RFID chip to an antenna to connect the chip to the antenna, the antenna being formed by a metal foil bonded onto a base film or a base tape, providing a protecting material around the RFID chip so that an upper surface becomes higher than that of the RFID chip, to constitute an RFID thread, and inlaying the RFID thread in between first and second paper layers to afford a paper-like structure. The thickness of the paper-like RFID tag fabricated as above is 0.1 mm or less. Further, the paper-like RFID tag is of high quality and has a surface smoothness almost equal to that of the ordinary type of paper, permitting a long-distance communication.

The invention relates to a label quality monitoring method and device, computer equipment and a storage medium. The computer equipment marks data according to the standard; adopting a preset quantification algorithm to quantify the labeling task quality of the labeling person; monitoring the labeling task quality of the labeling person according to the quantification result; Therefore, the qualityof the labeling task of the labeling person is automatically quantized through the computer equipment, so that the auditing of the labeling task is automatic, the quality of each labeling task can beintuitively observed according to the quantization result, the work of the labeling person can be conveniently examined, and the efficiency of monitoring the labeling quality is greatly improved.

The invention discloses a method and a special test paper for estrogen detection. The test paper for estrogen detection provided by the invention is a test paper obtained through step 1 or step 2 described below: 1, a test paper comprises a sample absorption pad, a reaction film, and a water absorption pad which are connected in sequence, wherein the reaction film comprises a detection zone which is coated with a conjugate of estrogen and carrier proteins and a quality control zone which is coated with quality control antigens; and 2, a test paper comprises a sample absorption pad, a maker pad, a reaction film, and a water absorption pad which are connected in sequence, wherein the tag pad is coated with fluorescein-labeled estrogen antibodies and fluorescein-labeled quality control antibodies, and the reaction film comprises a detection zone which is coated with a conjugate of estrogen and carrier proteins and a quality control zone which is coated with quality control antigens. The test paper for estrogen detection of the present invention has the advantages of high sensitivity, accurate quantification, strong specificity, simplicity and convenience, and short detection time.

The embodiment of the invention provides a labeling quality detecting method and device. The method comprises the following steps: acquiring a label image of a label to be detected; carrying out image pre-processing to the label image, wherein the image pre-processing at least comprises grayness processing; acquiring gray features of a feature point, in a set position, of the label image processed through image pre-processing; matching the gray features with the gray features of a corresponding feature point of a standard label image, and determining whether the label to be detected is qualified or not according to matching results. According to the scheme, manual visual detection is replaced by the acquisition of the gray features and the matching of the gray features, defects of strong subjectivity, high misjudgment ratio, low detection efficiency and the like of manual visual detection can be overcome, the detection precision to the labeling quality is improved, and the detection efficiency is improved.

The invention provides a semi-supervised medical image segmentation method and a semi-supervised medical image segmentation device based on double-model interactive learning, which are assisted by a stability judgment strategy and are used for solving the false label quality problem of semi-supervised medical image segmentation. According to label data features, supervision constraints of cross entropy and DICE are introduced when label data knowledge is effectively learned. According to features of the false labels, adjoint variables are introduced in the method, and the adjoint variables are mainly used for relieving the influence of wrong false labels on the model learning process. And performing noise enhancement on the samples, and providing consistency loss based on a sample prediction result and a noise sample prediction result in a formal training stage. According to the learning of double models on label-free data, the method proposes a false label screening mechanism based on stability judgment, and realizes double-model interactive learning; the medical image segmentation method is simple and convenient to implement and flexible in means, and remarkable segmentation effect improvement is achieved on training data of the medical image.

The invention relates to a labeling quality detection method used for carrying out qualified detection on an automatically-pasted label. The labeling quality detection method comprises following steps of carrying out image acquisition on the automatically-pasted label; carrying out initial processing on the acquired image; carrying out presence detection on the processed image, judging whether the label exists or not, entering the next step when the label exists, and determining the label to be unqualified when the label does not exist; carrying out character detection on the processed image, judging whether the character in the image is consistent with the specified character of the label, indicating the label to be qualified when the character in the image is consistent with the specified character of the label, and indicating the label to be unqualified when the character in the image is not consistent with the specified character of the label. Compared with the prior art, the labeling quality detection method is advantaged by being convenient for detection, accurate in examination, low in cost and convenient to implement and the like.

The invention discloses a printing quality visual detection method, and the method includes the following content: a label quality detection system which can be installed on a printing machine, a rewinding inspection machine and a slitting machine, a detection data server and a machine table. The label quality detection system detects a label of a to-be-detected product, and defect information of the product location, image and product label is uploaded to the detection data server; the detection data server is responsible for receiving, storing and transmitting the defect information of the product label; the machine table is used for carrying a product with a defective label, and can download the defect information corresponding to the product label from the detection data server; workers can choose the defective label needed to be eliminated on the machine table and eventually complete the label uncovering and labeling processes. The beneficial effects of the printing quality visual detection method are that the whole testing process does not stop, and the overall efficiency is 2-4 times of that of a single point mode.

A system and method for label placement is disclosed that achieves the twin goals of practical efficiency and high labeling quality by employing cartographic heuristics. A caller defines map and label properties. Then labels are pulled within a map boundary. Labels are next ordered by priority in descending importance. The order of testing labels is determined. Attempts are made to move overlapping labels. This is an iterative process; therefore there must be criteria that halt the procedure. Upon reaching an acceptable solution, the label properties are adjusted to reflect the new label placements.

The invention discloses a radio frequency identification label quality detection method and device applied to the technical field of communication. The method comprises the steps of detecting the distance between a radio frequency identification label and an antenna in a preset antenna power value, starting to read the signal intensity feedback value of the radio frequency identification label when the distance is equal to a preset distance, confirming that the quality of the radio frequency identification label is unqualified if the signal intensity feedback value can not be read, continuously reading the signal intensity feedback values for preset times if the signal intensity feedback value is read and obtained, judging whether the average value of the read signal intensity feedback values is in accordance with a preset value or not, and confirming that the quality of the radio frequency identification label is unqualified if so. According to the radio frequency identification label quality detection method and device, the operation of detecting the quality of the radio frequency identification label is simplified, the detection efficiency is improved, and the detection cost is saved.

The invention relates to a dual-adhesive-label high-speed label machine, which is used to stick label on the drug bottle. Wherein, it comprises a rotation drum, an adsorption drum, a paste wheel, a drug bottle transfer band and a bottle dial disk; the transfer band and bottle dial disk are matched with the circumference of rotation drum; the adsorption drum is relative to the rotation drum; one side of adsorption drum is mounted with label box; a label adsorption claw is arranged between the label box and the adsorption drum; the rotation drum is mounted with label compress hook; the descending part of rotation drum is arranged with paste wheel brusher; there are two adsorption drums arranged on the circumference of rotation drum; the rotation drum has two label compress hooks; each adsorption drum has own label box and label adsorption claw; the invention adds adsorption drums based on former label machine, while said two adsorption drums rotate alternatively to supply label; therefore, it can reduce the label adsorption frequency, to confirm the label quality and improve the speed.

The invention discloses a round bottle label quality detecting device. The round bottle label quality detecting device comprises a conveying belt, a label discharging mechanism, a bottle rolling plate, a bottle rolling belt and an imaging component. The label discharging mechanism discharges a label, the front end of the label makes contact with a bottle body and is pasted to the bottle body, thebottle rolling plate and the bottle rolling belt are arranged on the two sides of the conveying belt, the bottle body is clamped, the bottle rolling belt rotates, the bottle body is driven to rotate,and the label is pasted to the bottle body. The bottle rolling plate is provided with a label detection window, the upper edge of the label detection window is higher than the upper edge of the label,and the lower edge of the label detection window is lower than the lower edge of the label; and the length is larger than the perimeter of the bottle body. The bottle rolling plate is provided with the long-strip-shaped detection window corresponding to the label in height, and the length of the detection window is larger than the perimeter of the round bottle. The bottle rolling belt is utilizedfor rotating the round bottle, the imaging component can collect the integral state of the label, the detection dead angle is avoided, the detection accuracy rate is effectively increased, and the phenomenon of mistaken removal is avoided.

The invention discloses a supervised learning method for noise tags, a data classification processing method and device. The method comprises the following steps: determining a first loss function fora classification model, and constructing a second loss function according to the loss function; and replacing the first loss function with the second loss function, and training the classification model according to the second loss function. According to the invention, the technical problem that noise tags exist due to inaccurate sample tag labeling is solved. According to the invention, the tolerance degree of the label quality is improved, and the labeling cost is reduced.

The invention relates to the technical field of labeling machines for complicated-shape bottles and particularly discloses a semiautomatic labeling machine for the complicated-shape bottles. The semiautomatic labeling machine comprises a machine board as well as a conveying mechanism, a transferring mechanism and a hold-down mechanism which are arranged on the machine board, wherein the transferring mechanism comprises a first sliding plate, a first driving motor, a rotary table, a second driving motor, a jacking plate, a third driving motor and a first cam; the first sliding plate is connected to the machine board in a sliding manner; the rotary table is rotationally connected to the first sliding plate; the jacking plate is connected to the first sliding plate in a sliding manner; duringuse, the complicated-shape bottles are placed between the rotary table and the jacking plate; the third driving motor drives the jacking plate to jack the complicated-shape bottles on the rotary table; the first driving motor drives the complicated-shape bottles to move to the hold-down mechanism; the second driving motor drives the complicated-shape bottles to rotate; the hold-down mechanism compresses and attaches label paper conveyed by the conveying mechanism to the complicated-shape bottles; compared with a mode of holding down the complicated-shape bottles on the rotary table by using an air cylinder, the semiautomatic labeling machine has the advantages that the stability of labeling quality is improved, and noise generated by using an air compressor configured for the air cylinderis avoided.

A reagent bar used for quickly detecting staphylococcus enterotoxin consists of reagent, reagent bar and micro-scale pipet. It is featured as using cell clasmatosis solution as reagent and carrying 20 micron L of specific single clone antibody of anti-staphylococcus enterotoxin, collaurum labeled single clone antibody of anti-staphylococcus enterotoxin, collaurum labeled quality control antigen and antibody of relevant antigen on reagent bar.

A new technology is provided for a portable bar code verification reader and is designed to create a video image of a bar code label in accordance to strict industry bar code verification standards. The barcode verification reader will effectively capture a bar code label and then send the resulting video image to a computer. The resulting bar code label quality may be displayed on the portable verification reader via one or more LED's, an embedded display monitor, or the like.

The invention provides a label quality detection method and device and a computer readable storage medium. The label is defaulted to be in a defective product state; the method comprises the followingsteps of: acquiring quality detection information of a label; if the label information is collected, the collected quality detection information is compared with a standard information template, whenthe collected quality detection information is consistent with the standard information template, the default state is changed into a good product state and then is output as a detection result, andwhen the collected quality detection information is inconsistent with the standard information template, the default defective product state is output as the detection result; and if the label information is not collected, outputting a detection result based on the default state of the defective product. Through the implementation of the scheme, the quality of the label is detected in an intelligent mode, the workload of label quality detection is effectively reduced, and the detection efficiency and the detection accuracy are improved; and when the label information is not effectively collected, the detection result is output based on the default state, so that the yield of labels used on line is improved.

The invention discloses a dyeing silicone oil label, comprising base papers orderly stacked, a first back adhesive layer, separation paper and face paper with information, wherein between the separation paper and the face paper, a second back adhesive layer is arranged on one side, and a dyeing silicone oil layer and a third back adhesive layer which are stacked are arranged on the other side; the outer side close to the third back adhesive layer is provided with a white ink layer. The invention also discloses a formula of the dyeing silicone oil and a method for preparing the dyeing silicone oil. In the invention, the dyeing silicone oil replaces the conventional colorless silicone oil and is coated on the part of the label separation paper which needs to be uncovered. When the improved label is illuminated, the bottom part of the label is clear in color, and whether the label is coated with the silicone oil can be differed only depending on naked eyes, so that the label quality can be conveniently tested. The invention can save manpower, reduce wrong frequency and improve working efficiency.

The invention relates to the technical field of active learning, in particular to an active learning method combined with labeling quality control, which comprises the following steps: S1, training a model; S2, carrying out pseudo labeling on all unlabeled samples; S3, analyzing calculation and observation distance; S4, selecting to-be-labeled and to-be-checked samples based on the observation distance, selecting the to-be-labeled and to-be-checked samples based on the observation distance, labeling the to-be-labeled and to-be-checked samples, adding the labeled samples into a labeled sample set, performing additional labeling on the to-be-checked samples once, determining the label based on the existing historical label, updating the labeled sample set, and returning to the step S1 after the step is completed until the model performance can be accepted or the label budget limit is reached; according to the active learning method combined with labeling quality control, the labeling quality can be controlled while active learning is carried out.

The invention provides a label quality control method and device. The method comprises the following steps: 1, establishing m libraries for M known sequences, wherein each library contains one or morelabels, each known sequence and each label form a known label-sequence combination, the M libraries include N labels, any two or more libraries at most contain one same label, M and N are natural numbers, and M is less than or equal to N; sequencing the M libraries in the same lane to obtain sequencing original data; and judging whether pollution exists in the N labels or not by analyzing whetherlabel-sequence combinations which do not conform to known label-sequence combinations exist in the sequencing original data or not. According to the method, the number of constructed libraries is small, the number of quality control labels is large, the cost is low, and label low-frequency cross contamination can be found.

The invention discloses a steel coil sequence accurate positioning method based on a multi-carrier identification technology, and relates to the field of steel coil transfer in the metallurgical industry. According to field actual requirements and support of various multi-carrier identification technologies, the data flow and logic relation of the steel coil information are scientifically plannedand utilized, the shunting design is adopted, information accumulation is avoided, and the steel coil information tracking efficiency is improved. The method comprises the specific steps of steel coilinformation automatic identification, RFID label accurate positioning and sequence judgment, image data intelligent matching and steel coil data system verification. According to the method, code spraying errors are effectively reduced; the code spraying quality and the influence of the confusable characters on the correct identification of the sprayed codes are improved; the influence of factorssuch as the pasting position of the RFID tag, the tag quality and the surrounding environment on the tag identification sequence in the identification range of the identification and reading equipment is greatly reduced, accurate positioning of the steel coil data is realized, the integrity and the accuracy of the material information are comprehensively improved, and the reliability of the material information data is improved.

The invention discloses a hot-melt glue labelling machine. The hot-melt glue labelling machine comprises a machine frame, a dispensing mechanism, a label cutting mechanism, a labelling mechanism, a control box, a conveying belt and a label conveying mechanism. The label conveying mechanism comprises a label wheel and a vacuum drum. A negative pressure space with one face being provided with an opening is arranged in the vacuum drum, and the opening is connected with a draught fan. Air suction holes communicating with the negative pressure space are formed in the side face of the vacuum drum. Labelled paper can be adsorbed onto the side face of the vacuum drum after extending out of the label wheel, and the vacuum drum rotates to drive the labelled paper to pass through the dispensing mechanism, the label cutting mechanism and the labelling mechanism in sequence. The dispensing mechanism coats glue onto the labelled paper, the label cutting mechanism can cut off the labelled paper and the labelling mechanism can paste the labelled paper onto bottles. According to the hot-melt glue labelling machine, the labelled paper can be adsorbed through the air suction holes in the side face of the vacuum drum, and the effect of fixed conveyance of labels can be achieved; and meanwhile, the labels can be conveyed flatly due to arrangement of the multiple air suction holes, wrinkling of the labels is prevented, the labels can be pasted onto the bottles more flatly and firmly, and thus labelling quality is improved.

The invention discloses a method and a device of evaluating label quality of pitch information and a computer-readable storage medium, and belongs to the technical field of information processing. Themethod includes: obtaining lyric information and the pitch information of a target song, wherein the lyric information includes a playing starting time point and continuous-playing duration which areof each sentence of lyrics in multiple sentences of lyrics of the target song and in playing duration of the target song, and the target song is a song of which the label quality of the pitch information is to be evaluated; and then evaluating the label quality of the pitch information of the target song on the basis of the playing starting time point and the continuous-playing duration of each sentence of lyrics in the multiple sentences of lyrics. Therefore, in the embodiment of the invention, the pitch information of labels can be inspected through the lyric information of the song, and thus an evaluation result is obtained; and compared with methods of manual inspection in the related art, the method of the invention saves manpower and time, also eliminates, at the same time, label errors caused by human factors, and improves accuracy of an evaluation result.

The invention relates to an automatic optical identification system suitable for radio frequency identification label quality checking, and belongs to the technical field of optical automatic identification. The automatic optical identification system is characterized in that a planar light source is fixed on a cabinet on one side of a product roll at the front end of a reader-writer of a winding part of composite equipment; a camera is fixed on the cabinet on the other side of the product roll of the planar light source; the output of the camera is connected with an AOI controller in a control box; the reader-writer detects a triggering signal sent by each label, and the triggering signal is connected with the vision controller Trigger1 end in the AOI controller. According to the automatic optical identification system, the reader-writer is used for detecting that each label sends the triggering signal, and each RFID label is shot at ultra-high shooting speed; after high-speed image processing calculation of the AOI controller is carried out, whether the various sizes are greater than an upper set limit and a lower set limit is judged. The automatic optical identification system can be used for detecting each RFID label and providing a deflection value of the relevant size of each label; the detection speed can be matched with the highest speed of the cabinet, so that abnormal products caused by human eye precision and efficiency are reduced, and the qualification rate and the yield of products are improved.

The invention discloses a high-precision automatic labeling machine, which comprises an electric box bottom frame. The top of the electric box bottom frame is fixedly provided with a labeling machinemounting board. The labeling machine mounting board is equipped with a label belt clamp, a guide wheel set, a power traction unit, a label covering unit and a paper taking roller; the label belt clampis mounted on the top of the labeling machine mounting plate through a label belt clamp shaft, the guide wheel group and the paper taking roller are mounted on the lower part of the label belt clamp,in the middle space of the guide wheel set, the power traction unit is fixed on the side wall of the labeling machine mounting plate, the lower part of the guide wheel set is provided with a stripping unit, the label covering unit is fixedly disposed on the side part of the labeling machine mounting board by a label covering fixing bracket, and the lower part of the label covering unit is provided with a jig adjustment unit. The high-precision automatic labeling machine is wide in application scope, high in degree of automation, and simple in operation, the production process is effectively improved, the labeling efficiency is improved, labeling quality and precision are ensured, the manpower greatly is saved, the production cost is reduced, and the waste of resources is reduced.