720results about How to "Improve efficiency and accuracy" patented technology

The invention discloses an advertisement playing system and an advertisement playing method based on image identification. The system comprises an advertisement display platform, an image acquisition system and a data analysis processing module. The implementation method comprises the following steps of: acquiring an attention degree of an advertisement audience to an advertisement in real time by the image acquisition system added into the advertisement display platform, automatically analyzing feedback of the advertisement audience to the advertisement according to human face expression identification, automatically analyzing and pre-estimating an age structure and genders of client groups according to human body feature analysis and identification, and supplying reference to subsequent playing of the advertisement. By the system and the method, advertisement audience data can be timely acquired and quickly transmitted to an advertisement supplier and an advertising product supplier, so that an advertisement effect and an advertisement investment value can be accurately analyzed, and relevant staff can timely adjust an advertisement strategy; and a relatively high push effect is achieved for popularization of products.

The invention discloses a feature item selection and weight calculation based text classification method, which comprises the following steps: for a corpus training set obtained through analyzing and arranging, based on the traditional TF-IDF (term frequency-inverse document frequency) feature item weight calculation, through considering the relevancy among feature items and different categories according to CHI chi-square statistics and considering the intra-category average distribution situation of the feature items according to the information entropy of the feature item, carrying out adjustment on the weights of the feature items; then, carrying out assessment and selection on the feature items according to the inter-category weight difference of the feature items; respectively expressing a text in the training set and a text to be classified into vector form by using the feature item in a feature subspace; and determining the category of the text to be classified through calculating the similarity between the text in the training set and the text to be classified. As in the processes of feature item selection and weight calculation, the inter-category and intra-category distribution situations of feature items are comprehensively considered based on TF-IDF, so that the accuracy of feature item selection is enhanced, and the dimensionality of feature items is effectively reduced, thereby improving the efficiency and accuracy of text classification.

The invention discloses an LSTM deep learning model-based hydropower unit fault diagnosis method and system. The method comprises the steps of obtaining sampling sequences of N different signal channels of a hydropower unit, and performing VMD on each time sequence to obtain K IMF components; constructing corresponding training sets and to-be-diagnosed sets; building an LSTM model for the trainingset of each IMF component, and performing feature extraction on each IMF component through two LSTM layers; connecting outputs of K LSTM layers of the same signal channel to a Dense layer; through aSoftmax layer, performing feature classification on outputs of multiple Dense layers; and training a deep learning neural network model through an RMSProp gradient descent algorithm, and diagnosing the to-be-diagnosed sets by the trained model. According to the method and the system, the relatively good signal-noise separation effect of the VMD is combined with the processing advantage of an LSTMto the time sequences, so that the hydropower unit fault diagnosis accuracy is effectively improved.

The invention discloses a realization method of atrial fibrillation based on a deep convolutional neural network. By the method, single-lead ECG data can be changed into a two-dimensional form through signal conversion to enable the same to be suitable for the deep convolutional neural network for processing two-dimensional data, so that automatic detection on atrial fibrillation is realized finally by automatically learning features and classifying the same by a machine. When the method is used for atrial fibrillation detection, P wave or R-R interval does not need to be detected, and artificial designing of features is not needed also, so that efficiency and accuracy in atrial fibrillation detection are improved greatly, wherein an atrial fibrillation detection method based on stationary wavelet transform combined with the deep convolutional neural network is 98.63% in accuracy, 98.79% in sensitivity and 97.87% in specificity, and an atrial fibrillation detection method based on short-time Fourier transform combined with the deep convolutional neural network is 98.29% in accuracy, 98.34% in sensitivity and 98.24% in specificity.

The invention belongs to the technical field of information safety and discloses an android malicious software detecting platform oriented to the mobile internet. The Android malicious software detecting platform aims at overcoming the shortcoming that existing Android malicious software lacks dynamic detection, and the Android malicious software detecting platform achieves organic combination of static detection and dynamic detection of the Android malicious software. An application program is installed into an Android sandbox, the Android application program is automatically started and controlled, and various practical operations of a user to the application program are simulated. The operation situation of the application program is monitored to obtain API information of the application program and various dynamic behaviors so as to achieve the aim of detecting the malicious behaviors of the application program. According to the android malicious software detecting platform oriented to the mobile internet, the operations of the user to the application program in the Android sandbox is simulated, the practical behaviors of the application program are practically restored, and the efficiency and accuracy of the Android malicious software detection are improved.

The invention discloses a Bluetooth connecting method which comprises the steps that a mobile terminal sends requesting information to a Bluetooth device through a first infrared emitter, receives response information of the Bluetooth device through a first infrared outer receiver, and sends a connecting request to the Bluetooth device through the address of the Bluetooth device. Further disclosed is the mobile terminal. In practical implementation manner, due to the fact that the mobile terminal directly conducts exchange of measurement and control (MAC) address through the infrared emitter and the infrared receiver, the mobile terminal and the Bluetooth device are enabled to directly build Bluetooth connection according to the MAC address obtained through the exchange, scanning to devices around is not required, and therefore scanning time is saved, a condition that scanning is required to be repeated when names of the devices are the same, precision and efficiency of Bluetooth device connecting is improved, and meanwhile experience of users is improved.

The invention provides a multi-robot combined target searching method based on an animal-simulated space cognition technology, comprising the following steps of: (1) firstly, judging whether information is matched successfully or not; (2) establishing a kinematics model and a sensor model of each robot; (3) finishing calibration and matching of inner parameters of a camera; (4) utilizing a self-organization attractor network algorithm model to position the robots; (5) establishing fusion of an environment map and a multi-robot map; (6) navigating and searching a target; and (7) judging whether a task is finished or not; if so, finishing a searching task; otherwise, continually entering the step (2) and repeatedly carrying out a map construction and target searching process. According to the multi-robot combined target searching method, binocular vision can be used for automatically perceiving environmental information under the condition of no any artificial road sign; when the robots navigate and search a target and when the robots enter the same environment, an existing cognition map can be directly called to navigate, so that the working efficiency of the system is greatly improved.

The invention discloses a quick charging system for an unmanned aerial vehicle and a charging method thereof, wherein the quick charging system and the charging method belong to the field of unmanned aerial vehicle charging. The quick charging system for the unmanned aerial vehicle comprises a supporting module, a positioning module and a charging module. The supporting module is used for assisting the unmanned aerial vehicle body in landing and supporting the unmanned aerial vehicle body. The positioning module is used for performing return positioning on the unmanned aerial vehicle body so that the unmanned aerial vehicle body lands on the supporting module. The charging module is used for controlling a charging process. The charging method according to the quick charging system comprises an unmanned aerial vehicle body returning step, an image acquiring and positioning step, a quick landing step, a charging step, etc. The charging process comprises three periods, namely pre-charging, multi-step constant-current charging and pulse charging. According to the quick charging system, the supporting module is combined with other modules for forming a complete charging system, so that the unmanned aerial vehicle body can accurately and quickly land on the supporting module, thereby greatly improving charging efficiency and realizing high autonomy and convenient use.

The invention discloses an intention model training method and device, an intention identificationidentification method and device, equipment and a medium, and the method comprises the steps: inputting a training text word vector and an intention label into an intention training model, and obtaining an original intention identification model; inputting the test text word vector and the intention label into an original intention identification model, obtaining an output result corresponding to the original intention identification model, and determining the original intention identification model as a target intention identification model if the output result is greater than a preset accuracy rate; and inputting the target text word vector corresponding to the target text and the corresponding entity label into an entity training model to obtain an entity identification model. The target intention is obtained through the target intention identification model, and the target entity is obtained through the entity identification model, so that the effective speech corresponding to the target intention and the target entity is randomly selected from the speech template to communicate with the client, and the flexibility of the conversation between the robot and the client is improved.

The invention relates to a building engineering computation and evaluation system and method. The system comprises an associating module, a parameter reading module, an analysis module and a computation and evaluation module, wherein the associating module is used for identifying a building information model object from a building information document and associating a corresponding building system project in a master database with the building information model object; the parameter reading module is used for acquiring the structural parameters of the building information model object according to the structural item corresponding to the building system project and generating a parameter list file of the acquired structural parameters; the analysis module is used for analyzing the parameter list file and determining the definitions of the structural parameters according to the building system project; and the computation and evaluation module is used for estimating the engineering quantity according to the definitions of the structural parameters, evaluating the price of required building resources by querying a unit price catalogue database, and further generating an engineering quantity and evaluation list by combining the engineering quantity and the evaluated result. According to the invention, the efficiency and accuracy of engineering computation and evaluation are greatly increased.

The invention discloses an omnidirectional mobile platform vehicle with automatic line patrolling and obstacle avoiding functions for a factor area indoor space. The omnidirectional mobile platform vehicle comprises a main frame, a vehicle-borne computer, a storage battery group, a motor driver, and a Mecanum wheel component. The automatic line patrolling is realized through the combined working of a magnetic navigation sensor and a digital camera, the magnetic navigation sensor detects a vehicle offset through the position of an induction sensor relative to the a magnetic strip, the digital camera collects a navigation magnetic strip sideline in real time, the offset of the sideline to a theoretical calibration line is calculated, the numerical value returned by the sensors are integrated to adjust a vehicle body direction in real time, and thus the platform vehicle can move forward along a preset magnetic strip. According to the omnidirectional mobile platform vehicle, with an omnidirectional mobile vehicle as a basis, the multiple sensor integrated function modules of automatic line patrolling and automatic obstacle avoiding are developed, the vehicle can carry out transverse movement and oblique movement and rotate for any angle on a spot in a mode of zero gyration radius, point-action position fine adjustment is achieved, the automatic line patrolling is realized, the material delivery time can be reduced, and the working efficiency is improved.

The invention discloses a method and a system for testing a mobile phone mainboard to improve the accuracy and efficiency of the test of the mobile phone mainboard. The system comprises a mobile phone mainboard to be tested, function components corresponding to items to be tested of the mobile phone mainboard to be tested and a computer used for testing. When the system is used, the function components corresponding to the items to be tested of the mobile phone mainboard are tested by using the computer used for testing, and according to the testing result, whether the testing items corresponding to the functional components are qualified is determined. By using the technical scheme of the invention, the computer used for testing is used for controlling and realizing the test of the mobile phone mainboard, thereby overcoming the problem of low testing accuracy and efficiency caused by testing functions of the mobile phone mainboard by man power in the prior art, and improving the accuracy and efficiency of the test of the mobile phone mainboard.

The invention provides a system and a method for identifying association relations among enterprises. The system for identifying the association relations among the enterprises comprises an enterprise credit reporting system, an intermediate table module, a association query system and a plurality of commercial bank clients; the intermediate table module is connected with the enterprise credit reporting system, the association query system is connected with the intermediate table module, and the commercial bank clients are connected with the association query system; the intermediate table module comprises an association relation computation module and an association relation intermediate result storage module; the enterprises in investment association relations, the enterprises in association relations among part-time senior executives, the enterprises in personal investment distribution association, the enterprises in personal guarantee association, the enterprises in guarantee association relations, the enterprises in group parent-subsidiary association relations, the enterprises in family association, the enterprises in address association and the enterprises in telephone association are computed by the intermediate table module, and information of the enterprises is stored in the intermediate table module. The system and the method have the advantage that the association relations among the enterprises can be quickly displayed by the aid of the system and the method.

The invention discloses an intelligent book arrangement robot and an arrangement method thereof. The intelligent book arrangement robot is characterized by comprising a central control unit, a RFID reader/writer, a RFID electronic label, an image identifying module, a book arrangement module, a wireless communication module, a moving module, a detecting module, a human-computer interaction module and a power supply module. The central control unit is connected with the RFID reader/writer, the image identifying module, the book arrangement module, the wireless communication module, the moving module, the detecting module, the human-computer interaction module and the power supply module. The power supply module is connected with the central control unit, the RFID reader/writer, the image identifying module, the book arrangement module, the wireless communication module, the moving module, the detecting module and the human-computer interaction module. The intelligent book arrangement robot and the arrangement method thereof realize full-automatic and intelligent arrangement of books returned by readers, improve the book arrangement accuracy and efficiency, and simplify the work of book administrators.

The invention discloses a road condition state determination method and device, and belongs to the traffic field of road traffic. The method comprises the steps of acquiring driving information of multiple first vehicles of a target road segment, wherein each first vehicle is a vehicle of which the vehicle position is arranged at the target road segment within preset time, and the driving information comprises driving track and lamp waiting information; determining road condition characteristic of the target road segment according to the driving information of multiple first vehicles, whereinthe road condition characteristic of the target road segment comprises numbers of the first vehicles at different driving information zones within at least one time frame in the preset time; and calling a road condition identification model, inputting the road condition characteristic of target road segment into the road condition identification model, and outputting the road condition of the target road segment. By the method, the accuracy and the efficiency of road condition determination can be improved.

An optical touch panel including a light guide plate, at least one light-emitting element and a plurality of optical sensing elements and a touchscreen using the same are provided. The light guide plate has a plurality of lateral surfaces, a top surface, a bottom surface and a light extraction structure. The light-emitting element provides a light beam entering the light guide plate. The optical sensing elements are disposed under the bottom surface of the light guide plate. Each of the optical sensing elements has a sensing surface not parallel to the bottom surface of the light guide plate. The optical sensing elements are disposed at an illuminated region of the light beam provided by the light-emitting element. By the light extraction structure, a second portion of the light beam entered the light guide plate is scattered to the optical sensing elements from the bottom surface.

The invention relates to a method for extracting the text information in a new webpage, belonging to the field of webpage information analyzing and processing technique. The existing technique ordinarily adopts a packer to extract the interesting data from the web pages and the obtaining of information mode recognition knowledge of the packer is a time-taking and laborious, higher intelligence-demanding operation. The invention uses a stack data structure to convert the hiberarchy information of webpage data into vectors, constructs and analyzes Html tree, then making compression on the data of various layers of the Html tree, making data filtration, thinning, recognition and recombination to extract the needed data information. The invention is applied to extracting the template-generated news information in the news web pages from a fixed website.

The embodiment of the invention provides a method and a server for processing on-line transaction fault events of a mainframe. The server comprises a node information acquiring unit, a hierarchical model storage unit, a mapping table storage unit, a fault information receiving unit, a fault node state updating unit, a relevant node state updating unit and a node state display unit, wherein the node information acquiring unit is used for obtaining node information and node relation information; the hierarchical model storage unit is used for storing business service hierarchical models; the mapping table storage unit is used for storing a node state mapping map and a node incidence relation mapping table; the fault information receiving unit is used for receiving fault event information sent by an external event monitoring system; the fault node state updating unit is used for updating corresponding node state in the node state mapping table; the relevant node state updating unit adopts the node state of a father node generated by calculation to update the node state corresponding to the identification of the father node in the node state mapping table; and the node state display unit is used for displaying the node of which the state is updated in an imaging mode. The monitoring facing to business service is used for quickly determining and displaying an IT unit fault event.

The invention is applicable to the technical field of optical cable detection, and provides an optical cable circuit fault positioning and detection method and terminal equipment. The method includesthe following steps: establishing an optical cable circuit fault reflected light information sample library according to the reflected light information; establishing an optical cable circuit fault positioning and detection model according to a neural network; determining the theoretical result of the fault position of the optical cable circuit corresponding to the reflected light information in the sample library; judging whether a first difference value of the fault position theoretical result and the optical cable circuit fault position true result corresponding to the reflected light information exceeds a preset first threshold value or not; determining the fault theoretical result if the judgment does not exceed the preset first threshold value; judging whether the second difference value of the fault theoretical result and the fault true result of the optical cable circuit exceeds a preset second threshold value or not; detecting the fault condition of the optical cable circuit to be tested if the judgment does not exceed the preset second threshold value. By adopting the scheme, the accuracy of the fault identification of the optical cable circuit is improved, and the intelligent identification of the fault of the optical cable circuit is realized.

The invention provides a music recommendation method and device and a mobile terminal. The music recommendation method comprises obtaining application operating information generated when a user operates applications; generating into music recommendation information according to the application operating information; recommending music for a user according to the music recommendation information. According to the technical scheme of the music recommendation method, the application operating information generated when the user operates the applications is rich in valuable contents which can show user focus points or likes, so that the music recommendation information can be generated according to the application operating information generated when the user operates the applications, individual recommendation can be performed aiming at different users, music recommendation efficiency and accuracy are improved, time wasting is reduced, a user potential demand for discovering unknown music is satisfied, and the user experience is improved.

A rack and conveyor shuttle system for placing and retrieving articles of various sizes including a rack system having a shelving framework of at least two tubes spaced apart from one another at a distance, wherein the distance between the tubes is sufficient to support the article in a single stack and a motorized, conveyor shuttle system having at least two finger-like conveyors spaced apart from one another at a distance, wherein the distance between the finger-like conveyors is sufficient to move in between the two tubes so as to engage the article wherein the conveyor shuttle system is adapted to interlace in between the rack system for both placing and retrieving the article.

The invention provides a failure type identifying method and system of a power distribution network, wherein the failure type identifying method of the power distribution network comprises the following steps: obtaining a real-time waveform of a signal in a working process from a transmission line of the power distribution network; matching the real-time waveform with a waveform in a waveform library, finding out a waveform the matching degree with the real-time waveform is greater than or equal to a first preset threshold to obtain a matched waveform; respectively extracting edge feature points of the real-time waveform and the matched waveform, calculating a Hausdorff distance between the edge feature points; screening the matched waveform with minimum Hausdorff distance from the real-time distance to obtain a failure waveform; and determining a failure type of the real-time waveform according to the type of the failure waveform. According to the failure type identifying method and system of the power distribution network, a complicated mathematical model is built so that the failure type in the power distribution network can be accurately judged, thus the failure identifying efficiency and accuracy of the power distribution network are improved.

An image capturing apparatus including a plurality of image sensing modules and at least a light source is provided for capturing an image of an eye. Each of the image sensing modules includes an image sensor and a lens. The light source emits an illumination light, and the illumination light irradiates the eye. The eye reflects the illumination light into an image light. The image light includes a plurality of sub-image beams, and the sub-image beams are transmitted to the image sensors of the image sensing modules through the lenses of the image sensing modules, respectively. A capturing method is also provided.

The invention relates to a context-based forward input error correcting method and device. The method comprises the following steps: acquiring a confirmed text, wherein the confirmed text is the text which is confirmed before the current input is started; searching in a text corpus based on the correcting form of the confirmed text to acquire a search result which contains the correcting form of the confirmed text; performing error correcting on the confirmed text based on the search result. The invention provides an intelligent context-based forward input error correcting method and an intelligent context-based forward input error correcting method; the user operation is greatly simplified; the input efficiency and the accuracy are improved; the user experience is improved.

The invention aims to provide a method and equipment of screening a video. According to the method, a target video segment such as an important or abnormal video segment can be automatically screenedout through a deep-learning video model on the basis of image feature vectors, the existing problem that picking of target video segments from massive videos is difficult can be solved, and efficiencyand accuracy of video screening are greatly improved. In addition, through identifying whether video segments contain the target video segment such as the important or abnormal video segment, the target video segment is extracted, the extracted target video segment is used to select image key-frames, key characters/words are extracted on the basis of image content understanding of the generated key-frames, text information such as news is generated through semantic templates and a classification dictionary, a method of extracting the target video segment from the video, and using the extracted target segment to automatically generate the text information such as the news can be realized, and the problem that efficiency of a process of manually viewing the video to form text information corresponding to the video is low is solved.

The invention discloses a malicious application detection system for an android mobile phone, and belongs to the technical field of mobile phone detection. The malicious application detection system for the android mobile phone comprises a positive and negative sample collection module, a static feature extraction module, a dynamic feature extraction module, a neural network module and a monitoring result output module. According to the system, software code-based static analysis is combined with software behavior-based dynamic monitoring, so that a detection method is no longer static analysis purely based on a signature technology and judges whether software is malicious software or not by analyzing codes of the software and a behavior after running of the software; the detection methodis more accurate; and malicious software samples and secure software samples are collected to obtain static eigenvectors and a dynamic feature matrix, and an MLP neural network and an RNN are trainedthrough a large amount of samples to perform automatic learning and detection, so that the efficiency and the accuracy are greatly improved.

The invention relates to a method for identifying the flow type of soft grain two-phase turbulence based on artificial intelligence, comprising the following steps: 1), collecting the pressure signals of the soft grain two-phase turbulence; 2), extracting and analyzing the pressure signal characteristics: carrying out multi-level division on a frequency band by using a small wave packet method, further decomposing a high-frequency part without dispersed multiple resolution ratios, adaptively selecting corresponding frequency bands according to the characteristics of the analyzed signals so as to match the corresponding frequency bands with signal frequency spectrums; and 3), training and identifying flow type samples composed of defined characteristic parameters by using a probabilistic neural network, wherein learning samples are characteristic vectors of information entropy of a small wave packet after normalization according to a relationship between a probabilistic neural network structure and the learning samples, determining the structure of the probabilistic neural network for flow type identification as well as carrying out relative set on the network and learning training, and identifying the samples with different flow types by using the probabilistic neural network. The method has high veracity, good applicability and high reliability, and can effectively meet on-line identification requirements.

The invention discloses an electricity metering equipment fault analysis method and device. The method comprises steps that source data related to faults of electricity metering equipment are acquired; data cleaning of the source data is carried out; characteristic item selection of the source data after data cleaning is carried out; selected characteristic items are made to be sample data, and the sample data comprises training data, test data and prediction data; the training data and the test data are loaded into a machine learning model established based on the Xgboost algorithm, and the model is trained and tested; the prediction data is loaded into the trained model, and the electricity metering equipment fault analysis result is acquired . The method is advantaged in that accuracy of electricity metering equipment fault analysis can be improved, manpower resource cost is reduced, and efficiency is improved.

The invention discloses a method and system that inserting advertisement in network TV living broadcast program, the method includes following steps: A1. Generate advertisement program flow; A2. Receive switching request which switches living broadcast program to advertisement program; obtain data of said advertisement program, and living broadcast program medium flow played is changed to said advertisement program flow, at the same time information of medium flow played by living broadcast program is saved; A3. Receive the switching request which switches current program to original living broadcast program. The method and system in the invention can fulfill network TV advertisement requirement.

The invention discloses a positioning method of fault points in power distribution network mixed circuits. The positioning method comprises steps of adopting the method of monitoring external faults and measuring propagation speed of fault traveling waves in a cable circuit in an on-line manner; supposing one end in any one circuit zone breaks down and calculating time difference that the initial fault traveling waves achieve two measuring ends of the mixed circuit; setting the time difference as the node time difference, establishing corresponding relations between ends in circuit zone and the node time difference and storing the corresponding relations into a data base; when the mixed circuit breaks down, measuring the actual measurement time difference that the initial fault travelling waves reach two measuring ends of the mixed circuit; comparing the actual measurement time difference with the node time difference in the data base and determining fault circuit zones; and calculating the position of the actual fault point according to the factor that the difference value between the actual measurement time difference and the corresponding node time difference of the end in the fault circuit zone multiplied by a half of the propagation speed value of the fault travelling waves in the fault circuit zone equals the distance between the fault point and the fault circuit zone.