114results about How to "Help identify" patented technology

The invention relates to the technical field of biomedicine, and relates to a biomedicine event trigger word identification method based on characteristic automatic learning. The biomedicine event trigger word identification method comprises the following steps of 1, data pre-processing; 2, construction of an event trigger word dictionary; 3, construction of candidate trigger word examples; 4, characteristic learning by means of a convolutional neural network model; 5, training by means of a neural network model; and 6, classification of event trigger words. The biomedicine event trigger word identification method is advantaged in that 1, complex preprocessing to data is simplified, and tedious steps for carrying out a characteristic design by people are saved; 2, domain knowledge is introduced, and a lot of external resources such as unlabeled linguistic data are effectively utilized; 3, characteristic automatic learning is carried out by means of a convolutional neural network, manual intervention is reduced, sentence level characteristics in a deeper level can be excavated and explored, through the fusion of local characteristics, implicit global characteristics are discovered, and the category of trigger words can be identified; and 4, a better experiment result is obtained in MLEE linguistic data, and the whole performance on event trigger word detection is improved.

The invention provides an intelligent traffic light auxiliary system based on a vehicle road coordination technology, and belongs to the technical field of intelligent traffic control. The intelligent traffic light auxiliary system comprises a traffic light control box, a vehicle-mounted wireless module, a vehicle-mounted traffic light display and auxiliary early-warning module and a vehicle-mounted lane recognition module, wherein, the traffic light control box is connected with the vehicle-mounted wireless module in a wireless manner and transmits traffic light signals and vehicle lane information; the vehicle-mounted wireless module is connected with the vehicle-mounted traffic light display and auxiliary early-warning module and transmits traffic light signals and vehicle lane information; the vehicle-mounted lane recognition module is connected with the vehicle-mounted traffic light display and auxiliary early-warning module and outputs information of a lane at which a vehicle is located; and the vehicle-mounted traffic light display and auxiliary early-warning module displays all steering traffic signals of the current intersection and emphatically displays traffic signals in the advancing direction of the vehicle, thus realizing functions of early warning for running a red light and prompting of change from a red light to a green light. By utilizing the intelligent traffic light auxiliary system, road traffic efficiency can be effectively improved and the auxiliary traffic signals can be provided to the vehicle.

The invention discloses a CNN-based emotion analysis method, apparatus and device, and a storage medium. In the scheme, a convolutional neural network CNN is used as a basic architecture, and a targettext to be analyzed is input into the CNN to obtain an analysis result. By using the CNN, parallel calculation can be well performed in the training process. A convolutional and gating mechanism layer in the CNN contains a GTRU unit with embedded aspect information, and emotion characteristics can be better controlled through a gating mechanism with the help of given aspect information through the GTRU unit. Furthermore, in a loss function of the CNN full connection layer, a regular item of at least one language rule is contained, and the regular item is formed by modeling through a linguistic resource and contributes to emotion polarity recognition. According to the scheme, the interpretability of a CNN model can be improved through regular items of language rules, and the performance ofthe CNN model is improved.

The invention discloses an intelligent control device for a water dispenser based on an RGB-D camera and a control method thereof. The device comprises: an RGB-D camera, a control and information storage / processing unit and a power source; the output end of the power source is connected with the input end of the power supply of the information storage / processing unit, the power supply is usedto directly or indirectly provide working power for the control and information storage / processing unit and the RGB-D camera; the control and information storage / processing unit is electrically connected to the RGB-D camera. The device obtains the depth image of a water taking container (cup) obtained by the RGB-D camera, extracts characteristic information such as the position and height of the mouth of the water container (cup) of the water dispenser, and realizes the function of intelligently controlling the start-up and closing of the water dispenser.

The invention provides an intelligent flower irrigation system comprising a power supply module, a main control module, a planting environment data acquisition module, a plant category identificationmodule and an automatic irrigation module. The plant category identification module is used for acquiring a state image of a planted flower and feeding back the data to the main control module, comparing and judging the variety of the planted flower through the main control module and automatically setting growth habits and / or irrigation methods required for planting the flower. The planting environment data acquisition module is used for acquiring the planting growing environment and / or the irrigation methods and feeding back the acquired data to the main control module for a comparison withset physical environment requirement data, to judge whether irrigation is required, and the irrigation methods of the automatic irrigation module are controlled through the main control module to achieve automatic control. The intelligent flower irrigation system can be used for irrigation of large-scale flowers, fruits and vegetables, as well as for household irrigation. The intelligent flower irrigation system not only reduces the waste of water resources, but also protects the ecological environment, and also makes it easier and more intelligent to cultivate and grow flowers.

The invention relates to a light-transmitting key, a light-transmitting keycap and a light-transmitting keycap manufacturing method. The light-transmitting keycap is provided with a shading layer and a semi-transparent layer. The shading layer is provided with a hollowed-out pattern. The semi-transparent layer at least covers part of the hollowed-out pattern. The semi-transparent layer is provided with the light penetration rate ranging from 20% to 80% under the irradiation of the visible light and is provided with the color similar to the shading layer. In this way, when the backlight of the interior of a keyboard is switched on, sends out light rays and penetrates through the hollowed-out pattern, the brightness of the hollowed-out pattern is far greater than that of the shading layer, so that the hollowed-out pattern can be observed and identified easily; when the backlight is switched off, only the ambient light irradiates, the difference between the hollowed-out pattern, covered by the half-transparent layer, on the shading layer and other portions in visual sense is limited, so that the hollowed-out pattern is shaded and not obvious. The effect that the hollowed-out portion in the shading layer is shaded without the backlight is achieved, and the operation mode of the light-emitting keyboard can be identified effectively.

The invention provides a method of identifying cell types based on single-cell RNA sequencing data. According to the method, a low-rank representation model of a high-dimensional matrix is effectivelycombined with a graph regularization theory; an optimization model is constructed by considering the global structure and local structure characteristics of data, the model is solved by adopting an alternating direction multiplier method (ADMM) to obtain a reliable inter-cell similarity matrix, and then the similarity matrix is clustered by adopting a spectral clustering method, so that single cells are clustered, and the cell types are identified. According to the method, the clustering effect of the single-cell RNA sequencing data can be remarkably improved.

The disclosure relates to an image processing method and apparatus and an electronic device. The method may include the following steps: obtaining an initial gray value of each pixel point on an image to be processed; and according to gray value difference degrees among pixel points in a neighborhood corresponding to each pixel point, processing the initial gray value of each pixel point into an adjusted gray value, and obtaining adjusted images, wherein when the gray value difference degree corresponding to any pixel point is increasingly lower, the adjusted gray value of any pixel point increasingly approaches to an intermediate gray value of the corresponding adjusted image. Through the technical scheme of the disclosure, images can be preprocessed effectively to enable edges and texture details of objects to be shot to be improved, which is helpful in identifying objects and realizing other subsequent functions.

The invention relates to a segmentation method of a Chinese food image. A texture image of a Chinese food image is collected and follow-up processing is carried out to realize image segmentation; and during the segmentation process, collection of several kinds of image features is not required. Moreover, with the method, the accuracy of segmentation of a Chinese food image can be improved and thus a Chinese food image can be identified well.

The invention discloses a management method, system and device for nodes in a cluster. The method comprises the steps that the grouping condition of each node in a cluster is determined according to apreset cluster grouping deployment strategy when a deployment instruction representing a cluster grouping mode is received; a configuration file corresponding to each node in the cluster is configured according to the grouping condition of each node in the cluster; after the configuration files corresponding to all the nodes in the cluster are configured, all the nodes in the cluster are restarted, so that the cluster grouping mode takes effect; and in the cluster grouping mode, a target node is used for monitoring the node states of the rest nodes through message transmission with the rest nodes in the same group. Visibly, under a large-scale cluster, the nodes in the cluster can be deployed in a cluster grouping mode, and only the nodes in the same group need to be monitored by one another in the cluster, so that stable monitoring of states among the nodes is facilitated, and misjudgment is not easily caused; and moreover, the adoption of the cluster grouping mode is helpful for identifying a fault node in the cluster so as to avoid the influence of the fault node on the service.

The invention relates to a video processing method and device, electronic equipment and a storage medium. The video processing method comprises the steps: determining a plurality of target video frames in a plurality of video frames of a to-be-processed video; respectively carrying out feature extraction processing of a plurality of channels on the plurality of target video frames to obtain firstfeature maps of the plurality of channels of the target video frames; performing space-time extraction processing on the first feature map of the target video frame to obtain space-time feature information of the target video frame; and recognizing the to-be-processed video according to the space-time feature information to obtain a recognition result. The video processing method obtains the spatial-temporal feature information through sequential replacement processing and feature extraction processing, and does not need to use modes such as optical flow or three-dimensional convolution with large operation expenditure to obtain the spatial-temporal feature information, so that the processing efficiency is improved, and good spatial-temporal feature information expression capacity can be obtained under the condition that the number of parameters of the neural network is not increased.

The invention discloses an anisotropic rotor holographic dynamic balancing method based on equivalent initial phase vector. According to the method, the holographic spectrum principle is taken as the basis, vibration data in two directions are taken into account and are subjected to data fusion, so the anisotropic characteristic can be completely formulated, and the balance fault diagnosis is facilitated; furthermore, the rule of angular momentum conservation is utilized, a holographic spectrum ellipse is converted into an equivalent rotating frequency circle, the equivalent initial phase vector is taken as the characteristic for unbalanced vibration, and the requirement on fine balance is met. Through the equivalent initial phase vector, the defects of traditional initial phase vector on describing the amplitude of unbalance and phase position caused by anisotropy can be eliminated, the unbalance characteristic can be accurately characterized, and the spatial dynamic property of a rotor can be truly indicated. Therefore, the anisotropic rotor holographic dynamic balancing method based on the equivalent initial phase vector, provided by the invention, facilitates improving the accuracy and efficiency of on-site dynamic balance, so that the operating precision and service life of a rotary machine can be effectively improved and prolonged.

This driving-support-image generation device is provided with a character recognition unit and an image processing unit. The character recognition unit recognizes text in an image that represents the posting content of a posted object posted on or near a road. On the basis of the text recognized by the character recognition unit and driving-support information associated with said text, the image processing unit generates a driving-support image by changing the image of the posted object into an image corresponding to the aforementioned driving-support information.

The embodiments of the invention disclose a method and a device for extracting a banknote image, equipment and a storage medium. The method includes the following steps: acquiring a to-be-extracted banknote image after tilt correction, and integerizing the to-be-extracted banknote image through a first integerizing method; taking a target image containing a to-be-extracted image from the integerized to-be-extracted banknote image; determining the boundary of the to-be-extracted image on the target image through a row projection method and a column projection method based on a sliding window, and carrying out image extraction to get the to-be-extracted image; and integerizing the to-be-extracted image through a second integerizing method. The problem that the original banknote image extraction method is slow and the extracted image is of poor quality is avoided. The method improves the quality of image extraction, and is conducive to subsequent banknote image identification.

The invention provides a cloth defect detection model and a cloth defect detection method based on improved YOLOv4-tin. According to the detection model, a dense connection convolution block CSPDenseBlock is added to a residual block in a YOLOv4-tiny backbone network, and an spp module is added to the backbone network finally, so that a new feature extraction network is formed; the new feature extraction network outputs two feature maps with different scales; and the two feature maps with different scales are processed by respective corresponding convolution blocks and then enter respective YOLO layers to predict a target. According to the model and method, the backbone network is adjusted by adding the dense connection convolution block into the original residual block, so that the neuron receptive field of the detection model is expanded, and the extraction of shallow information is facilitated; a deep backbone network framework is constructed by using dense convolution blocks, so that a plurality of targets which are difficult to distinguish can be identified, and the accuracy of model detection is improved; the whole model can be used for optimizing a cloth defect detection task in a complex scene, and the detection precision is better than that of a traditional model.

The invention belongs to the field of biological medicine and molecular biology, relates to establishment and application of a disease-related factor scoring parameter analysis model, in particular toa progression appendicitis nomogram prediction system, which comprises model software, and is characterized in that the model software consists of a back-end database, a model algorithm and a front-end graphical user interface; the model algorithm is realized through computer programming of a scoring analysis model, a predictor is analyzed through logistic regression, and a Nomogram is adopted toestablish the model; the predictor comprises a fibrin degradation product, C-reactive protein and Na < + >. The established Nomogram diagnosis nomogram based on multivariate logistic regression analysis can help determine the risk index of the appendicitis in the progression stage according to the score in clinical practice; furthermore, the method can be used for analyzing scoring parameters ofserum markers of the progressive appendicitis and the early appendicitis, and is beneficial to identification of the progressive appendicitis and the early appendicitis.

The invention relates to a stay point identification method and system based on high-frequency equidistant space-time trajectory data. The method includes the following steps: (1) collecting and preprocessing high-frequency equidistant space-time trajectory data; (2) traversing position mark points one by one, and calculating the density value within a buffer region range of certain radius relative to each position mark point; (3) drawing a density-time curve, and finding peak points; (4) traversing all the peak points one by one, finding a stay center point, and determining the total number of times of stay and the stay point positions of a travel individual; and (5) calculating the instantaneous speed of each position mark point, drawing a speed-time curve, and comparing the speed-time curve with the density-time curve to determine the arrival time, the departure time and the length of stay. The method and the system have high identification accuracy, will have great influence on thesubsequent analysis of travel mode, travel purpose, behavior prediction and so on when applied, and will provide an important theoretical basis for urban planning management.

The invention discloses a garbage multi-target detection method based on improved YOLOv3. In the convolutional neural network, the neuron receptive field of the model is expanded by using 7 * 7 convolution with the step length of 2, so that extraction of shallow information is facilitated; a deeper backbone network framework is constructed by using dense convolution blocks, so that a plurality oftargets difficult to distinguish can be identified, and meanwhile, the phenomenon of overfitting easily occurring when the model is trained to a deep network can be reduced due to the parameter transfer characteristic of dense convolution; on the basis of the whole framework, in cooperation with adjustment of training parameters, the whole model can be used for optimizing a multi-target detectiontask in a complex scene, the detection precision is better than that of a traditional model, and a thought and a method for solving problems are provided for intelligent treatment of household garbageand multi-target detection.

The invention relates to an alarming system used for monitoring a plant inclination state. The alarming system comprises an inclination detection device, a monitoring platform, and a shooting device. The inclination detection device comprises an MEMS gyroscope, a main controller, a power supply module, and a wireless communication module. The MEMS gyroscope is connected with the input end of the main controller, and the wireless communication module is connected with the output end of the main controller. The wireless communication module is used for establishing communication connection with the shooting device or the monitoring platform. By evaluating an inclination damage state, the shooting device is activated automatically to capture on-site pictures, and the wireless communication module is used to transmit early warning information to the monitoring platform, and then monitoring blind areas and manual patrol are reduced, and manpower and material resources are saved, and then the plant inclination state is timely monitored, and alarms are given timely. Under severe weather conditions, the alarming system is used for monitoring of abnormal states and fault elimination, and then identification of sources of alarming abnormal data is facilitated, and plant monitoring management level is improved.

The invention provides a reagent composition for screening and / or diagnosing clonal diseases by a one-step method and application of the reagent composition. The reagent composition comprises eight groups of antibodies and is a flow cytometry detection composition capable of being used for one-step screening and / or diagnosis of clonal diseases, a five-tube parallel scheme is used during application, the first group of antibodies and the second group of antibodies are respectively used for different tube samples, the third group of antibodies and the sixth group of antibodies are used for the same tube sample, the fourth group of antibodies and the seventh group of antibodies are used for the same tube sample, and the fifth group of antibodies and the eighth group of antibodies are used for the same tube sample. The reagent composition disclosed by the invention can be applied to flow cytometry one-step screening / diagnosis of various clonal diseases and other abnormalities, comprising acute leukemia, mature lymphocyte tumor, chronic myeloid tumor, night paroxysmal sleep hemoglobinuria, metastatic cancer, precancerous lesions, rare normal human polymorphisms, aplastic anemia, other immune cell subset abnormalities and the like.

The invention relates to a fire index prediction method and device, a readable storage medium and electronic equipment. The method comprises the following steps: acquiring data of various fire disaster causing factors at the current moment; Generating a first feature map corresponding to the plurality of fire disaster causing factors at the current moment according to the data of the plurality offire disaster causing factors at the current moment; Performing Fourier transform on the data of the plurality of fire disaster-causing factors at the current moment; Generating a second feature map corresponding to a plurality of fire disaster-causing factors at the current moment according to the data obtained after conversion at the current moment; According to the first feature map and the second feature map, generating a target feature map corresponding to various fire disaster-causing factors at the current moment; And inputting the target feature map into a prediction model corresponding to the fire index to obtain a predicted value of the fire index. Therefore, recognition of the image features by the prediction model corresponding to the fire index is facilitated, so that the fireindex prediction accuracy is improved, and the probability of fire occurrence and the loss caused by the fire are reduced to the maximum extent.

The invention discloses a fusion method of infrared polarization images based on YUV and multi-feature separation. Firstly, a Stokes vector is used to express the polarization state of light, and a polarization degree image and a polarization angle image are calculated. Then the unique parts P 'and R' of the polarization degree and polarization angle images are obtained. Then the image P 'and R' are fused by using the multi-feature separation method of dark primary color theory, and the fused result and the total intensity image I are fused in YUV space to obtain the final fusion result of infrared polarization image. This method is based on the human visual characteristic YUV color space, effectively improves the visualization effect of the polarization image fusion result, enhances the detail information of the image such as the edge, improves the image contrast, fuses the complementary information among the polarization quantities, makes the fusion image scene richer, and is helpfulto the recognition of the camouflage target.

The invention discloses a preparation and conversion method of soybean protoplasts, and relates to the technical field of plant gene engineering. The method comprises the following steps of (1) performing induction to obtain soybean embryogenic calluses; (2) adding an enzyme solution, performing enzymolysis on the embryogenic calluses, performing filtration and centrifuging, adding a W5-23% solution, and performing centrifuging so as to obtain purified protoplasts; and (3) putting the purified protoplasts on ice, performing treatment, performing re-suspending with MMG suspension liquid, then sequentially adding plasmids and a PEG4000 solution, performing uniform mixing, performing standing, adding a W5 solution, ending a conversion reaction, performing re-suspending on protoplasts obtained through centrifuging with a WI solution, performing shifting into a culture dish, and performing light-away room-temperature putting. According to the preparation and conversion method disclosed by the invention, the compounding ratio of the enzyme solution and PEG instantaneous conversion reaction conditions are optimized, the yield and the conversion rate of the soybean protoplasts are significantly increased, identification of soybean gene functions is facilitated, and a reliable basis is further provided for exploration of the preparation and conversion conditions of the protoplasts of other plants and other positions.

The disclosure relates to a method and device for prompting an incoming call, and belongs to the technical field of communication. The method includes: after receiving an incoming call, if it is detected that the incoming call number does not match any contact number in the local address book, obtaining the tag information and answering status information of the incoming call number from the server, and the tag information is used for The identity of the caller is marked, and the answering status information is used to represent the answering status of the call initiated by the calling number; the marking information and answering status information are displayed on the caller ID page. Since the answering status information can reflect the status of calls initiated by the incoming number being answered, it can not only assist users in identifying harassing calls and advertising calls more accurately, but also help users maliciously mark normal numbers or In the case of a wrongly marked number, it is possible to effectively prevent the user from missing an important call due to an error in the marked information.

The invention relates to an unmanned aerial vehicle night plant protection operation method based on multi-view vision, and belongs to the technical field of aviation systems. The method comprises thefollowing steps of arranging a plurality of binocular vision sub-platforms around a farmland to be subjected to plant protection to form a multi-view vision platform; carrying out parameter calibration on the binocular vision sub-platforms, and then calibrating the position relation between the binocular vision sub-platforms; calculating the initial position of the plant protection unmanned aerial vehicle platform relative to the farmland through a multi-view vision system; planning a flight path of the plant protection unmanned aerial vehicle by using a ground control platform; starting plant protection work, in the operation process, enabling the multi-view vision platform to capture infrared feature identification points on the plant protection unmanned aerial vehicle platform, calculate the position of the plant protection unmanned aerial vehicle relative to a farmland in real time, and feed back position information to an autopilot and a ground control platform of the plant protection unmanned aerial vehicle, which is used for correcting the route of the unmanned aerial vehicle in the plant protection operation process in real time. A pesticide spraying platform is carried onthe unmanned aerial vehicle platform and is an executing mechanism for plant protection operation of the unmanned aerial vehicle.

The invention provides a contact processing method and equipment. The method comprises the following steps: displaying a contact recording interface; receiving a preset operation of a user on the displayed call history interface specific to the preset call history record; and storing telephone numbers corresponding to the preset call history as temporary contacts. Through adoption of the contact processing method and equipment, the user can be assisted in identifying the telephone numbers which are stored in the contact, and can conveniently contact the telephone numbers again.

The invention relates to a liver cancer tumour marker AFP, AFP variant and GP73 antigen synchronous immunodetection kit, a method and application, and belongs to the technical field of immunodetection. The kit comprises magnetic particles, AFP and GP73 antibody coupled mixed acridinium ester, GP73, an AFP calibration product, first stimulation solution, second stimulation solution, washing solution and all kinds of diluting solution; and the magnetic particles are respectively marked by AFP, LCA-BSA and GP73. By adoption of the kit and the detection method, the sensitivity, the specificity andthe accuracy are high; the detection time is short; identification on benign liver diseases and malignant liver diseases and early diagnosis on liver cancers can be easily carried out; and the kit and the detection method have an important purpose for increasing the early diagnosis rate of patients with liver cancers and prolonging the survival rate, and have wide clinical application value.

The invention discloses human TSHR extracellular fusion protein and a preparation method thereof. The N end of the human TSHR extracellular fusion protein is connected with exogenous secretion signal peptide, and the C end thereof is connected with a purification marker sequence; when the fusion protein is expressed and purified, the fusion protein the N end of which is connected with the exogenous secretion signal peptide is resected, and the specific amino acid sequence is shown in SEQ ID No.1. The invention obtains high-expression soluble recombinant protein with the antigenic secretion form of TSH receptor in a good in-vitro expression system by constructing recombinant pichia pastoris containing pPICZ alpha A-hTSHRecd expression vector.

The invention provides a fuzzy character processing method applied to an electronic device. The method comprises steps: a picture containing fuzzy characters is displayed; when an instruction for processing the picture is received, a character block is drawn for each character in the picture; the stroke line of the character in the character block is analyzed, and according to the stroke line, a character with the highest similarity with the character is recognized as the distinct character from a pre-stored character set; and the fuzzy character is replaced to be the distinct character. The invention also provides a fuzzy character processing system and an electronic device. The distinct character can be used for directly replacing the fuzzy character in the picture, and the purpose of helping the user to recognize the fuzzy character in the picture is achieved.

The invention provides an infrared polarization image fusion method based on multi-feature and feature difference driving, which comprises the following steps: using a Stokes vector to represent the polarization of light, calculating a polarization degree image P and a polarization angle image R; linearly weighting the polarization angle images R and U to obtain an image R '; Between that computerimages R ', I, and P, the respective unique part of the common part is remove, denoted as R1, I1, and P1; the images P1, I1 and R1 are mapped to R channel, G channel and B channel in RGB space to obtain RGB image. The RGB image is converted to YUV image and the luminance component Y is extracted. The image I1 and P1 are fused by a method based on multi-feature separation to obtain F; F is substituted for Y to obtain the substituted YUV image, and the substituted YUV image is inversely transformed to obtain RGB image, that is, the result of polarization image fusion. Multiple polarization images of infrared polarization images are fused, which makes the fused image scene richer and helps to identify the camouflage targets. The invention is applied to the field of computer vision.