141results about How to "Improve determination accuracy" patented technology

The invention discloses a method and device for identifying a signal light. The method for identifying the signal light includes the following steps: a, extracting candidate luminance pixels and candidate color pixels from a traffic information image, wherein the color of the candidate color pixels is red or yellow or green; b, confirming judgment areas with the candidate luminance pixels as centers, counting the number of the candidate color pixels in the judgment areas and confirming the corresponding color of the candidate color pixels in the judgment areas, c, confirming the color of the signal lights which are represented by the candidate luminance pixels according to the number and the color of the candidate color pixels, d, aiming at the types of the different signal lights, confirming candidate areas according to the positions of the candidate luminance pixels and the color of the signal lights which are represented by the candidate luminance pixels, and judging whether the candidate areas are the signal lights through a shape detection method. By means of the method for identifying the signal light, the signal light candidate areas are confirmed with the combination of luminance factors and color factors, the number of finally confirmed candidate areas is reduced, the confirming accuracy of the candidate areas is improved, and thus the processing speed and accuracy of the follow-up shape detection process are improved.

Most automatic particle classification methods produce errors. The invention provides a method for improving the accuracy of particle classification while shortening the amount of manual review time required from the operator. The method uses class weights, which are statistically-derived correction factors that accounts for frequency of classification errors. A first class weight and a second class weight are assigned to the first class and the second class, respectively. The number of particles in each of the first and the second classes is multiplied by the first class weight and the second class weight, respectively, to generate a corrected number of particles in each of the classes. If particles are reclassified, the class weights are recalculated in response to the reclassification. The method is usable with a complete classification where all the particles in a sample are classified, or a selective classification of a subset of the particles in the sample.

The invention provides a method and a device for audio / video pushing, an apparatus and a storage medium. The method comprises: based on user information of a target user, obtaining audios / videos to which the target user prefers from an assigned resource pool, the user information including personal data, one or more historical watching records of the audios / videos; and pushing the audios / videos towhich the user prefers to the target user. The method and the device can prevent disturbance on users caused by pushing, and provide convenience for users to rapidly obtain audios / videos which meet preference of the users, and improve watching desire of the users to watch the audios / videos, so as to improve user satisfaction.

The invention discloses a processing method of a fuel valve hole of a marine diesel engine cylinder cover, comprising the following steps: a right-angle bent plate on a work bench is corrected, a marine diesel engine cylinder cover is fixed on the right-angle bent plate, and then the preset position of a fuel valve hole is drawn; the work bench alpha is vertically rotated, an upward hole is processed by three drills, and then heavy boring and finish boring are carried out; a model drilling body is put into the upward hole, the distance between the model drilling body and the upward hole is h, and a downward hole is processed by a common lengthening drill; the model drilling body is taken out to be put into a bush, parts of middle holes with the distance of d are axially milled by an expanding milling cutter, then a cutter is retreated to take the bush out, and a depth measuring tool is put in to measure residual depth; the depth measuring tool is retreated, then the bush and the expanding milling cutter are put in to mill residual depth so as to mill the middle hole; the cylinder cover on the right-angle bent plate is loosened, and is fixed after rotating; the above steps are repeated to process the residual two fuel valve holes. The above processing method has low processing cost, high quality of processing finished products, short processing time and low processing difficulty.

The invention relates to a parking space determination method and device, computer equipment and a storage medium. The method comprises the steps of collecting an initial image through a camera installed on a vehicle, and generating a to-be-processed spliced image according to the collected initial image; performing feature recognition on the to-be-processed spliced image to obtain image features;determining a preset type of useful information representing the to-be-judged parking space according to the recognized image features; matching the useful information representing the parking spaceto be judged with pre-stored memory parking space information; and taking the to-be-judged parking space represented by the useful information successfully matched with the memory parking space information as a target parking space. By adopting the method, the accuracy of parking space determination can be improved.

In an image data output processing apparatus of the present invention, a storage process section, when a reference document is duplex, stores, at the time of storage, a DocID indicative of the reference document, to correspond to an ID indicative of respective document images on front and back sides of the reference document, and in matching when a matching document is duplex, extracts an ID similar to that of the reference document and a corresponding DocID for each of the document images on the front and back sides of the matching document, to create candidate lists for the front and back sides of the matching document. If first candidates in the candidate lists correspond to an identical DocID, the images are similar to each other, and if first candidates do not correspond to the identical DocID, a determination result is corrected to a reference document corresponding to the identical DocID.

The invention discloses a method and apparatus for determining an EVM (error vector magnitude), and relates to a communication technology. The method comprises: performing timing estimation, synchronization, phase correction and frequency offset correction on received data, after the corrected received data is obtained, further performing intersymbol interference elimination on the corrected received data so as to remove intersymbol interference introduced by synchronization errors in a TD-SCDMA (time division synchronized code division multiple access) system, and determining the EVM according to data after the intersymbol interference is eliminated, so the determination accuracy of the EVM is improved.

Systems and methodology for determining not only precise and repeatable results, but accurate values of the refractive index of solids, fluids and liquids.

The invention discloses an interaction target determination method and device for intelligent equipment. The method comprises the steps of: acquiring an environment image in a monitoring range of theintelligent equipment, and carrying out target identification on the environment image; using targets identified out from the environment image as candidate targets, and acquiring state information ofthe candidate targets; for each candidate target, according to the corresponding state information, judging whether the candidate target has an interaction intention of interacting with the intelligent equipment; and selecting an interaction target of the intelligent equipment from the candidate targets with the interaction intention. In the embodiment, by screening out the candidate targets withthe interaction intention from all the candidate targets according to the state information of the candidate targets and further selecting the interaction target from the candidate targets with the interaction intention for the intelligent equipment, an effect that the selected interaction target is the most possible to be a target with the interaction intention with the intelligent equipment isachieved, a case of using a target without the interaction intention as the interaction target is avoided, and false startup of the intelligent equipment is reduced.

The invention discloses a method and apparatus for determining target image location information. The method includes the following steps that: for each matching area, the score value of the matching area is determined according to each pixel in the matching area as well as the horizontal gradient value, vertical gradient value and gradient amplitude of each template pixel in a pre-stored template image; a target area is determined, pixels in the target area which are located in a first neighborhood are determined according to the gradient directions of the template pixels; a preset algorithm is adopted to determine the perspective matrix of the target area; and each edge line of the template image is corrected according to the perspective matrix, and target image location information in the target area is determined. According to the method and apparatus for determining the target image location information provided by the embodiments of the present invention, after the target area is determined, the perspective matrix of the target area is determined; the edge lines are corrected according to the perspective matrix; the target image location information in the target area is determined; and therefore, the accuracy of the determination of the target image location information under perspective deformation can be improved.

A pathological diagnosis assisting apparatus according to the present invention provides the diagnosis assisting information for assisting diagnosis of tissue fibrosis from the image of the stained sample, and includes an image reading unit configured to read the image of the stained sample and an image processing unit configured to process the image. The image processing unit includes an image classification unit configured to classify the collagenous fiber and the elastic fiber, a tissue area extraction unit configured to extract a tissue area, an occupancy rate calculation unit configured to calculate the occupancy rates of the collagenous fiber and the elastic fiber in the tissue area, and a diagnosis assisting information providing unit configured to provide the diagnosis assisting information based on the calculated occupancy rates of the collagenous fiber and the elastic fiber.

The invention relates to a video title generation method, a video title generation device, a computer readable storage medium and computer equipment. The method comprises the steps: extracting a plurality of video frames of a video, and enabling the video frames to serve as a plurality of target video frames; obtaining image contents of the plurality of target video frames, and obtaining text information of the video under each target video frame according to the image contents of the plurality of target video frames; determining alternative titles of the video according to the text information of the video under each target video frame; and obtaining the comprehensive word frequency of the alternative titles in the text information, and screening the alternative title with the maximum comprehensive word frequency from the alternative titles as the title of the video. According to the scheme provided by the invention, the purpose of automatically generating the title of the video according to the plurality of target video frames of the video can be achieved. The global information of the video is comprehensively considered. The word frequency information of the alternative titles is combined, so that the determination of the video titles is more accurate. The determination accuracy of the video titles is improved.

A method for identifying a seismic even makes it possible to determine the direction to a source of a seismic disturbance both on the surface and in a three-dimensional space and increases the probability of correct identification of a seismic event and the accuracy of the determination of the direction to the source. A compact seismic detector for implementing the method reduces the cost of scanning and servicing a perimeter protection system constructed on the basis of such detectors. In the method, two or three pairs of seismic sensors which are divided by a distance and are oriented at an angle with respect to one another, for example mutually perpendicularly along the axes X, Y or along the axes X, Y, Z, respectively, are combined to form one seismic detector. The resultant seismic data are processed to determine the direction to the source of the seismic disturbance.

The embodiment of the invention discloses a medical image noise data identification method based on artificial intelligence and a related device. For a medical image with a label, if the medical imagewith the label needs to identify whether noise data exists in the medical image, the medical image can be taken as to-be-identified label data and be identified through a first identification model.The first identification model has a good anti-noise capability. According to the identification result obtained by the first identification model, the reflected confidence coefficient can effectivelyexpress the possibility that the to-be-identified label data belongs to the noise data, so that the medical images with wrong label annotations are effectively screened out from the medical images with labels, and the determination accuracy of the noise data is improved.

The invention belongs to the technical field of heat exchange equipment, and particularly relates to a control method of an air conditioning system. The invention aims to solve the problems that the uniform distribution degree of cold and warm airflow in a room cannot be guaranteed when the indoor temperature is adjusted by the existing air conditioning equipment, and the temperature adjusting effect of each indoor position is not uniform. According to the control method of the air conditioning system, the temperature equalization degree in the room can be determined based on the difference between the maximum temperature value and the minimum temperature value in the current environment temperatures of the multiple areas in the room, so that the rotating speed of an indoor fan is regulated and controlled according to the current temperature equalization degree of the room, and the indoor air supply capacity of the air conditioning system is selectively adjusted, and therefore, the indoor temperature environment can be uniformly and quickly adjusted when the air conditioning system operates, and a good temperature adjusting effect can be achieved in any position in the room, and thus, the overall temperature adjusting effect of the air conditioning system is greatly improved.

The invention discloses an operation recovery method. A client end records a local operation execution state; a server records a local operation execution state; when the client end and / or the server restarts due to failure, the client end and the server interchange the execution states of the records; the client end and the server confirm a breakpoint according to the execution state of the local record and the execution state of the counterpart record; and the client end and the server recover operation from the breakpoint. The invention also discloses the client end, the server and the system for operation recovery. The invention can realize the interactive failure recovery between the client end and the server in a C / S construction.

The invention relates to a field-level data blood relationship determination method and device based on a knowledge graph. The method comprises the steps of obtaining table names and table structure information of two data tables in a first data table pair, wherein the first data table pair comprises any two data tables in a database; calculating a first table name similarity between table names of two data tables in the first data table pair, and calculating a similarity between fields in the table structure information to obtain a first field similarity matrix; according to the first table name similarity, the first field similarity matrix and a pre-acquired target weight, determining whether a blood relationship exists between two data tables in the first data table pair or not; acquiring a plurality of first data table pairs with a blood relationship as target data table pairs; and generating a field-level data relationship graph according to the blood relationship between the knowledge graph and the data table in the target data table pair. The method and device can be used to reduce the difficulty in determining the blood relationship of the data table, and improves the determination accuracy of the blood relationship.

The embodiment of the invention discloses a retransmission message determination method and device. The method comprises the steps of recording a message disorder phenomenon to form message disorder information at a first moment, wherein the message disorder information at least comprises a defective message serial number and disorder forming time of a to-be-completed defective message; matching a received first message serial number of a first message with the defective message serial number at a second moment, wherein the second moment is later than the first moment; if the first message serial number matches with the defective message serial number successfully, determining the time difference of the receiving time of the first message and the disorder forming time; comparing the time difference with retransmission timeout, thereby forming a comparison result; and determining whether the first message is a retransmission message or not based on the comparison result.

The invention relates to a natural language processing technology in artificial intelligence, in particular to a text recognition method and device, computer equipment and a storage medium, and can be applied to scenes such as electronic commerce, news information, microblog forum and vehicle-mounted recommendation. The method comprises the following steps: acquiring a grammatical relationship sequence of a to-be-recognized text; if the grammatical relationship sequence of the to-be-recognized text comprises at least two grammatical relationship sequences in preset grammatical relationship sequences, determining a feature word position tag in the to-be-recognized text according to one grammatical relationship sequence in the at least two grammatical relationship sequences; determining feature words in the to-be-recognized text according to the feature word position tags in the to-be-recognized text; and determining a text recognition result of the to-be-recognized text according to the feature words in the to-be-recognized text. By adopting the method, the determination accuracy of the feature words in the to-be-recognized text is improved, so that the text recognition accuracy is improved, and the effectiveness of a data analysis result of big data is ensured.

The embodiment of the invention discloses a problem matching method and device, equipment and a medium, and relates to the field of cloud computing and data processing. According to the specific implementation scheme, the method comprises: using a problem classification sample to train a basic network layer multiplexed by a problem classification model and a problem sorting model and a classification output network layer in the problem classification model; and training the basic network layer and a sorting output network layer in a question sorting model by utilizing a question sorting sampleto obtain a trained question sorting model for question matching. According to the problem matching method and device, the equipment and the medium provided by the embodiment of the invention, the robustness of the problem sorting model is improved, and then the accuracy of problem matching is improved.

The embodiment of the invention discloses an interest point processing method and device, equipment and a medium, and relates to the field of electronic maps. The method comprises the following steps:determining missing interest point information of a retrieved fruitless object and a retrieval position associated with the missing interest point information; determining retrieval times of varioustypes of missing interest point information associated in an area to which the retrieval position belongs; and determining newly added interest point information of the belonging area from the missinginterest point information according to the retrieval times of the associated missing interest point information in the belonging area. The embodiment of the invention provides an interest point processing method and device, equipment and a medium. The newly added interest point information is determined based on searched fruitless interest point information.

In a method for wheel suspension alignment, the following steps are carried out: providing a wheel suspension alignment system having four measuring heads situated in a known position with respect to one another, of which each has a monocular picture recording device, the relative position of the measuring heads with respect to one another being known; recording in each case one wheel or a measuring target mounted on it in an initial position of the vehicle, using each of the four measuring heads; shifting the vehicle from the initial position to at least one further position; recording in each case one wheel, or a measuring target mounted on it, of the vehicle standing in the further position, using each of the four measuring heads; recording a reference target having a known pattern mounted on the vehicle, using one of the four measuring heads in at least one of the initial position and the further position of the vehicle, and determining from this an absolute scale for the measuring heads; carrying out local 3D reconstructions for determining the translation vectors, the rotation vectors and the wheel rotational angles between the at least two positions, as well as the wheel rotational centers and the wheel rotational axes of the wheels; and determining the wheel suspension alignment parameters of the vehicle, especially the camber, the single toe and / or the total toe, the wheelbase and the tread width.

Embodiments of the invention disclose a rhythm prediction method and device, equipment and a medium, and relate to the field of data processing, in particular to a voice synthesis technology. The method comprises the following steps of segmenting a Chinese and English mixed text to be predicted to obtain a Chinese text and an English text; determining character vectors of characters in the Chinesetext and word vectors of words in the English text; and determining a rhythm prediction result of the Chinese and English mixed text according to the determined character vectors and word vectors. According to the rhythm prediction method and device, the equipment and the medium, the rhythm prediction accuracy of the Chinese and English mixed text is improved.

The invention discloses a convolutional neural network model training method and device, an electronic device and a computer readable storage medium. The method comprises: establishing a convolutional neural network, a convolutional layer of the convolutional neural network comprising a plurality of parallel convolutional cores, and each convolutional core corresponding to one training channel; inputting the training sample set into a convolutional neural network, and independently training each training channel until respective convergence conditions are met to obtain a convolutional neural network model comprising a plurality of training channels; the plurality of training channels being respectively used for predicting a plurality of feature data associated with the target area. According to the embodiment of the invention, the training sample set is trained through the plurality of parallel training channels, the convolutional neural network model obtained through training comprises a plurality of training channels, and the plurality of training channels are respectively used for predicting a plurality of feature data associated with the target area, so that more features associated with the target area can be obtained, and the target area determination accuracy can be improved.

The invention discloses a three-dimensional sensor network node positioning method and system. The method comprises the following steps: acquiring positions of a one-hop anchor node and a two-hop anchor node of a known node; according to a position relationship between the one-hop anchor node and the unknown node, a position relationship between the two-hop anchor node and the unknown node and therange of an RSSI value of the unknown node and one-hop anchor node, preliminarily determining a prediction area of the unknown node; in the prediction area, determining an initial position of the unknown node according to a magnitude relationship of the RSSI value between the unknown node and the plurality of anchor nodes; and after the position of the unknown node is updated, judging whether theposition of the unknown node meets the condition or not, then obtaining the confidence degree obtained by the last iteration, calculating the confidence degree of the iteration by combining a variational message passing algorithm, updating a position parameter and the position of the unknown node by utilizing the confidence coefficient until the number of iterations reaches a preset number of iterations. The method and the system provided by the invention can improve the positioning accuracy.

The invention discloses a roadside parking management system based on a longitudinal multi-view image and a sensor signal, and the system comprises at least two camera groups, a sensor, at least one installation rod, or an installation position and a main controller. At least two camera groups are used to respectively collect respective corresponding monitoring area images; the sensor is used to collect sensor signals of at least one parking space in a monitoring area; and the main controller is used to determining the parking event information of each parking space in the monitoring area according to the corresponding monitoring area images corrected by the camera groups and the sensor signal, collected by the sensor, of at least one parking space of the monitoring area. By arranging a plurality of camera groups and sensors at different heights in the longitudinal direction, the system can determine the parking event information of each parking space in the monitoring area by jointlyanalyzing the image of the monitoring area and the sensor signals of the various parking spaces collected by the sensor, thereby improving the determining accuracy of the parking event information, and improving the roadside parking management precision.

The invention belongs to the technical field of heat exchange equipment, and particularly relates to a control method of an air conditioning system. The method aims at solving the problem that a way of determining an indoor temperature regulating demand by existing air conditioning equipment is inaccurate and thus, the initial opening degree of an electronic expansion value determined is not matched with an actual operating requirement of air conditioning equipment. In order to achieve the purpose, according to the control method of the air conditioning system, the room can be divided into the multiple areas, a room load is calculated based on the initial environment temperatures of the multiple areas, the load size of each area in the room is refined, and the room load value corresponding to the whole room is accurately calculated; the difference between the current temperature condition and the expected temperature condition of the room can be accurately determined on the basis of the difference between the room load value and the preset load value, and the determination accuracy of the temperature regulation requirement in the room is improved, so that the actual operation requirement of the air conditioning system can be more accurately obtained, and the initial opening degree of the electronic expansion valve is better matched with the actual operation requirement of the air conditioning system.

The present invention includes a system for use in a vehicle for determining the terrain type in proximity to a vehicle. The system comprises a processor configured to receive primary output data from at least one vehicle-mounted sensor, and secondary output data from a secondary data source; and a data memory configured to store pre-determined primary data relating primary output data for the at least one vehicle-mounted sensor to a particular terrain type, and store at least one data set relating to one or more terrain types. The processor is configured to compare the primary output data with the pre-determined data to determine an indication of the particular terrain type corresponding to the primary output data, and compare the secondary output data with a data set corresponding to the indication of the particular terrain type.

The invention discloses a method and a device for determining a fetal section based on an ultrasonic dynamic image. The method comprises the following steps of: sequentially inputting each frame of fetal ultrasonic image in a plurality of frames of acquired fetal ultrasonic images into a pre-determined feature detection model for analysis; obtaining analysis results sequentially output by the feature detection model to serve as feature information of each frame of fetal ultrasound image; and determining a tangent plane corresponding to the fetal ultrasound image according to the category of the part feature of each frame of fetal ultrasound image and the category of the structural feature of the fetal ultrasound image. Visibly, the tangent plane of the fetus ultrasonic image is determinedby acquiring the part features and the structural features of the continuous multiple frames of fetus ultrasonic images and combining the part features and the structural features of the fetus ultrasonic images, so that the determination accuracy of the tangent plane of the fetus ultrasonic image can be improved; the standardization degree of the section of the fetal ultrasound image can be improved; and the determination efficiency of the tangent plane of the fetal ultrasound image can be improved.