149results about How to "Achieve learning" patented technology

The invention discloses a mobile terminal comprehensive learning system based on massive learning resources. The learning system comprises a resource library system, an intelligent knowledge diagnosis system, an online tutoring system, a student management system, a teacher management system, a parent management system, a learning PK system and an infinite extension system. According to the learning system, a learning resource library and a teacher team platform need to be built first, and after logging into the learning system, students acquire various learning resources such as synchronous videos, synchronous teaching materials, synchronous auxiliary teaching materials, synchronous question banks, knowledge points and micro videos. Questions which the students encounter in the learning process can be solved in two ways; in one way, the students communicate with teachers in a one-to-one mode through online tutoring, and in the other way, the students submit the questions to the question banks of the system and wait for other students or teachers to answer the questions. The students can carry out intelligent knowledge diagnosis on knowledge they have learnt to examine the knowledge mastering condition of themselves and can also interact with other students in communities.

The invention discloses a human body behavior recognition method and system based on a graph convolution network, and the method comprises the steps: extracting human body skeleton information from animage containing human body behaviors, obtaining a human body joint point position information sequence, and constructing a topological graph sequence with any length of a human body skeleton; performing feature extraction and topological structure adaptive evolution on the topological graph sequence through a topological learnable graph convolution-based space-time graph convolution network to obtain new node features fused with local space-time features and a topological graph sequence with a new topological structure; performing feature extraction through a graph convolution long-term andshort-term memory neural network; global spatio-temporal features are obtained through global pooling operation; and performing human body behavior recognition based on the global spatial-temporal features through a classifier. The features of a whole graph are directly learned, the weight matrix in graph convolution is expanded to the whole topological graph structure, the relation between any two nodes in the graph is learned, limitation of the topological structure is avoided, and the recognition accuracy is high.

The invention discloses a visual simultaneous localization and mapping (Visual-SLAM) method based on a depth convolution auto-encoder. The method comprises the steps of 1, performing data preprocessing on training data; 2, establishing a multi-task learning network; 3, taking three adjacent frames of binocular images in the image sequence as network input; 4, constructing a loss function; 5, training, verifying and testing the multi-task network; 6, the trained shared encoder network is used for loopback detection; 7, constructing a new Visual-SLAM system front end through the six steps, constructing a rear end of the Visual-SLAM system through pose graph optimization or factor graph optimization, and building a complete system, and 8, verifying the positioning accuracy and robustness. A depth convolution auto-encoder is used, a semi-supervised multi-task learning method is used to construct the front end of an SLAM system, depth estimation, camera pose estimation, optical flow estimation and semantic segmentation are included, and a feature map of a network is used to construct image representation to perform loop detection.

The present invention provides a vehicle position prediction method based on deep learning. The method comprises the steps of: S1, obtaining an original VLPR data set, performing cleaning and denoising of the original VLPR data set, obtaining a VLPR data set, and performing grouping of the VLPR data set according to a set time quantum; S2, from the VLPR data set, extracting vehicle pass record ofthe same vehicle, generating a vehicle track data set and screening out a track meeting a demand, and performing data conversion of the related feature information of the track; S3, establishing an algorithm model based on deep learning according to the obtained vehicle track features, and achieving analysis and learning of the track features; and S4, after performing feature learning of the vehicle track, employing a full-connection network layer to combine a Softmax classifier to output a next position vector, matching real geographic position information, outputting the real geographic position information, and achieving vehicle position prediction. The method provided by the invention analyzes operation features of a road network and employs the track features in the vehicle driving process so as to obtain high prediction precision.

The invention discloses a cell division detection method based on deep learning, relates to the technical field including cytobiology and the like, and solves the problem of inaccurate cell division detection results. The cell division detection method comprises the following steps: obtaining an image in a cell culture process, and forming a continuous image sequence according to the images obtained at certain time intervals; judging and comparing all cell division processes in the image sequence, and labelling a position where a process that a paternal cell is divided into two daughter cell membranes can be clearly observed in cell division, and forming a labelled dataset; constructing a full convolutional neural network model, determining the parameter of the full convolutional neural network model, loading the labelled dataset into the full convolutional neural network model, using a deep learning method to directly carry out feature learning on an image sequence, and obtaining the trained full convolutional neural network model; and deploying the trained full convolutional neural network model for automatically detecting the cell division of the cell culture data of a similar type. The cell division detection method is used for cell division detection.

The invention discloses a universal remote control method, which is characterized by comprising the following steps of: identifying infrared remote control signals of each piece of equipment to acquire primary data, compressing infrared codes in the primary data by using a compression method, storing the compressed infrared codes in a primary data sheet, and when the infrared codes are used, performing decompression reduction to control given infrared remote control equipment to work. By the compression method used by the universal remote control method, a data volume can be reduced to be within 1K by compression, and the highest compression ratio of 16:1 can be realized, thereby reducing a storage space required by the learning of a remote controller code to further make possible all remote controller code learning of a remote controller.

The invention discloses a transesophageal echocardiography visual simulation system, comprising an intelligent phantom, a heart ultrasonic image stimulation and acquisition device, a computer connected with the heart ultrasonic image stimulation and acquisition device and a transesophageal ultrasonic probe attitude data acquisition device, wherein, the heart ultrasonic image stimulation and acquisition device is used for stimulating and acquiring heart ultrasonic images, and transmitting the acquired heart ultrasonic images to the computer; when the heat ultrasonic image stimulation and acquisition device acquires the heart ultrasonic images, the transesophageal echocardiography probe attitude data acquisition device and a sensor are used for synchronously acquiring probe attitude data of the heat ultrasonic image stimulation and acquisition device, and transmitting the acquired probe attitude data to the computer; and the computer is used for analyzing, processing and visually displaying the received heart ultrasonic images and the probe attitude data.

The invention relates to a radar and multi-network fusion based pedestrian pose recognition method and system. The method comprises: pre-processing an echo signal of a radar signal to obtain an outputsignal; suppressing a stationary target in the output signal; searching a pedestrian located distance unit in the suppressed output signal; performing time-frequency analysis on the echo signal corresponding to the pedestrian located distance unit, and obtaining a an echo signal time-frequency map; using a plurality of convolutional neural networks to recognize the echo signal time-frequency map,and obtaining a recognition result of each convolutional neural network; and fusing the recognition results of all the convolutional neural networks, and obtaining the fused pose recognition result.The pedestrian pose recognition method provides a pedestrian pose recognition result in real time by analyzing the radar echo, can ensure high pose recognition accuracy through multi-neural network fusion, is free from interference of factors such as illumination conditions, weather, smoke and the like, and is able to work all day and all weathers.

The invention discloses an algorithm of walking along an edge of a dust collection robot. The dust collection robot comprises a frame, wherein two driving wheels, two driving wheel motors connected with the driving wheels, and a supporting wheel are arranged on the frame. The dust collection robot further comprises a fore sight obstacle detection sensor installed in the front of the dust collection robot, a side sight obstacle detection sensor installed on the left side, a step detection sensor installed at the bottom of the frame, a rechargeable battery which provides a power source, a dust collection motor which is used for carrying out vacuum dust collection, and a processor which is used for carrying out centralized control. The driving wheel motors, the dust collection motor, the fore sight obstacle detection sensor, the side sight obstacle detection sensor and the step detection sensor are connected with the processor. The processor has the algorithm of walking along the edge, and the algorithm of walking along the edge is used for designing different strategies according to different positions of walls or obstacles, and guiding the dust collection robot to walk along the walls or the obstacles.

The invention discloses an air control voice command recognition method based on deep learning. The method comprises the following steps: acquiring a voice signal to be recognized, and converting thevoice signal into 16-bit 16-kHz PCM audio data; building a deep network model; training the deep network model by using training data to obtain a voice recognition engine; performing voice segmentation on the audio data; and inputting effective audio clips obtained by voice segmentation into the voice recognition engine, and outputting a character recognition result. According to the deep networkmodel, a convolution module is used as a feature extractor; extracted feature data is processed through a reshape layer and a full-connection layer; sequence learning is carried out through a gating circulation unit; finally classification learning and decision making are carried out through the full-connection layer, so that a prediction result is obtained. According to the method, an artificialintelligence deep learning engine is adopted as a core, so that the method has the advantages of extremely high professional applicability and accent generalization ability, and lower data quantity dependence, and is obviously superior to a general voice recognition system in air control voice recognition.

The invention discloses an arrhythmia classification method based on multilead information fusion. The method comprises the steps: 1. carrying out downsampling, filling-in and pretreatment of fixed length on original multilead electrocardiogram signals; 2. constructing a classification model, which is composed of a plurality of single branch networks and is responsible for specially processing corresponding mono-lead electrocardiogram signals, and meanwhile, optimizing parameters of the classification model by using a multi-loss cooperative optimization strategy; 3. employing ten-fold cross verification training on a public data set, and evaluating the above-mentioned model; and 4. achieving an arrhythmia classification task by using the trained model. According to the method, informationlearning of the multilead electrocardiogram signals is maximized through fusing integrity and diversity of the multilead electrocardiogram signals by the multi-loss cooperative optimization strategy,and high-accuracy automatic detection on arrhythmia can be achieved, so that assistance is provided for clinical diagnosis, and early warning is provided for a patient employing wearable equipment.

The present invention discloses a primary school Chinese electronic learning system identified and read by two-dimension codes. The primary school Chinese electronic learning system identified and read by the two-dimension codes comprises a two-dimension code generating module, a matching teaching-aid book, a network server and an intelligent terminal, wherein the two-dimension code generating module generates the two-dimension code images corresponding to the learning content in each page of the teaching-aid book separately according to the learning content of the teaching-aid book and then prints the two-dimension code images in the teaching-aid book correspondingly, the network server stores the learning content corresponding to the two-dimension code images, the intelligent terminal is used to identify the two-dimension code images in the matching teaching-aid book and calls the corresponding learning content from the network server, and the reading of the learning content and the learning of various formats are carried out on the terminal. The primary school Chinese electronic learning system of the present invention combines the books and an electronic technology, realizes the effective binding of the APP and the paper presswork (the school teaching material and teaching aid), is suitable for the lower grade students to use, and enables the price to be reduced greatly without needing to purchase the hardware individually.

A method for classify cell defects base on unbiased embedded zero-sample learning includes such steps as 1) obtaining an image block containing defects with suitable size from a cell picture collectedby a high-definition camera through a sliding window, allocating common defect categories as labeled source data, and allocating rare defect categories as unlabeled target data; 2) end-to-end training of the quasi-fully supervised learning network model QFSL using annotated source class data and unannotated target class data; (3) In the testing phase, the input image block is embedded by the visual embedding subnetwork, and then the image block is embedded by the visual embedding subnetwork, and the image block is embedded by the visual embedding subnetwork. Semantic linking sub-network completes the mapping from visual embedding to semantic embedding, then gets the scores of visual embedding and semantic embedding by inner product calculation, and finally sends them to Softmax classifierto generate the prediction probability of all categories, and takes the category with the highest probability as the classification result.

The embodiment of the invention provides a network intrusion detection method and system. The method comprises the steps of detecting to-be-detected network data based on a trained network intrusion detection model, analyzing and judging an obtained detection result, and if it is judged that the to-be-detected network data is known network data, performing fine-grained classification on the knownnetwork data according to the detection result; if it is judged that the to-be-detected network data is the network unknown category intrusion data, clustering the network unknown category intrusion data to obtain clustering centroids of all categories in the network unknown category intrusion data; and updating the classifier of the trained network intrusion detection model according to the clustering centroid of each category in the network unknown category intrusion data so as to perform intrusion detection on the network unknown category intrusion data in the subsequent to-be-detected network data according to the updated network intrusion detection model. According to the embodiment of the invention, recognition and learning of unknown invasion of the network are realized.

The invention relates to the field of air conditioners, and provides a learning method for remote control codes of an air conditioner based on a cloud platform. Learning of remote control codes of an infrared remote controller of the air conditioner is realized. The method comprises the following steps in summary: code matching: receiving infrared codes generated by an air conditioner remote controller, and transmitting the infrared codes to a cloud platform; initial analysis: retrieving in a database, obtaining an infrared code rule and a data format, and verifying the obtained infrared code rule and data format; gradual analysis: carrying out code value change, setting parameters, uploading to a cradle head for analysis, and obtaining the infrared code rule and data format; and downloading: transmitting the code rule and the data format to an embedded module and a user operation terminal. The learning method provided by the invention is suitable for learning of the air conditioner remote controller.

A multifunctional desk allowing English learning can effectively solve the problem the function of the conventional learning desks is single. The multifunctional desk can achieve English word learning on the basis of an original desk, and can achieve switching of a plurality of modes in a learning process, can spell words, call allow a user to understand the words all round, and can deepen the impression of word spelling; and the multifunctional desk can allow the user to copy words, can regulate the writing format of the user, can correct wrong fonts and the style of writing, can perform other recreational and sports activities, can enrich the mental activities of the users, and can be used as a common desk.

The invention discloses a virtual detection teaching platform with industrial ultrathin material production detection equipment and a use method thereof. The platform comprises a virtual simulation laboratory three-dimensional environment module, a virtual simulation laboratory entity detection device, a virtual simulation laboratory three-dimensional simulation detection device, a virtual detection data module and a virtual detection software program. The use method comprises getting access to content in the catalogue of a virtual detection simulation laboratory; getting access to content in the catalogue of an entity detection device display area; getting access to content in the catalogue of a virtual detection display area; getting access to content in the catalogue of a virtual software display area; and providing the guidance of eight experiments by using a website experimental handout. The platform combines an actual industrial production line and detection equipment, virtual dynamic simulation production line and detection equipment and the application software of the detection equipment so as to provide students with empathy during the operation and improve the students' abilities to linking theory with practice.

Disclosed is a voice conversion method of united frequency-spectrum modeling based on a restricted boltzman machine. The method comprises the implementation steps: extracting voice spectrum envelope characteristics, extracting voice high-layer spectrum characteristics, conducting dynamic time warping, training a GMM, dividing acoustic subspaces of the united spectrum envelope characteristics, training a Gaussian-Bernouslli RBM or a Gaussian-Gaussian RBM, converting frequency spectrums and synthesizing conversion voices. According to the voice conversion method of the united frequency-spectrum modeling based on the restricted boltzman machine, the precision of the frequency-spectrum modeling is improved, and the tone quality and the naturalness of the conversion voices are improved.

The invention discloses a human motion tracking method based on a deep nuclear information image feature. The human motion tracking method based on the deep nuclear information image feature mainly solves the problems that in human motion tracking of the prior art, features of a video image are not accurately expressed, so that a tracking result is caused to be not accurate. The method comprising the steps: obtaining an articulation point three-dimensional coordinate matrix Y of the video image from a data bank; extracting the deep nuclear information image feature X of the processed video image; serving the deep nuclear information image feature X as an input, serving the three-dimensional coordinate matrix Y, in the video image, of a human body as an output, and learning a regression function by using of gaussian process; learning an obtained regression function by using of the gaussian process, serving a new deep nuclear information image feature X of the video image as an input, and estimating data of three-dimensional poses of a moving body. Compared with an existing human body tracking method, the human motion tracking method based on the deep nuclear information image feature has the advantages of being high in training speed, accurate in express of image features, and capable of being used in motion catching, human-computer interaction, video surveillance, recognition of human body goals and restoration of the three-dimensional poses.

The invention provides a learning scene processing method and device, and the method comprises the steps: displaying a learning scene interface which comprises a plurality of virtual objects, and enabling the position relationship and size of the virtual objects in the learning scene interface to simulate the layout of a real scene; receiving a selection operation on the learning scene interface,determining a virtual object selected by the selection operation in the learning scene interface, and determining the virtual object as a target object; and obtaining learning resources of the targetobject, and displaying a learning interface corresponding to the learning resources. Due to the fact that the learning scene simulates the layout of a real scene, the combination of learning and the scene is achieved, the association between learning and the scene is enhanced, learning gamification and memory consolidation can be achieved at the same time, learning interest is stimulated, and thelearning effect is improved. Especially for learning of English words of children, real scenes, gamification learning and teaching through lively activities, the learning effect of the English words can be improved for the children.

The invention relates to a verifiable multi-party kmeans federated learning method witha privacy protection function, and belongs to the technical field of data mining. The data is horizontally distributed on multiple users, and each user encrypts and uploads the respective data to the cloud server; the cloud server randomly selects an initial centroid, and calculates the square of the Euclidean distance between the data and the initial centroid by using a secure multiplication protocol and a secure distance calculation protocol; the cloud server performs distance comparison by using a security bit decomposition protocol and a security comparison protocol, and divides the data; each user updates the clustering center by using a secret sharing protocol, encrypts the clustering center and uploads the clustering center to the cloud server; and the cloud server calculates the distance between the new clustering center and the original clustering center, if the distance is smaller than a threshold value, clustering operation is ended, and otherwise, the clustering center is updated for next iteration.

The invention relates to the field of laboratory safety management, and discloses a laboratory safety admittance platform system, comprising a student registration module, a virtual simulation learning module, an examination module, an admission authentication module and a credit feedback module connected with a background management module through wired or wireless mode. After the trainees register in the laboratory safety access platform system and log in, they learn the knowledge of safety management system, safety operation and emergency plan through virtual simulation, and obtain the laboratory access authority after passing the examination. The invention adopts 3D simulation technology to increase interaction and interest in safety knowledge learning process, through the VR simulation technology to achieve the emergency plan of experiential learning, at the same time, through the background management module to receive feedback from the credit feedback module to the students in the examination and experimental operation process of disciplinary violations, an admission blacklist is established to achieve the dynamic management of admission rights.

The invention discloses a disaster monitoring and early warning platform construction method based on an SOA architecture. By using existing geological disaster monitoring and early warning, early warning is performed through manual inspection and monitoring and monitoring equipment threshold alarm; but due to the diversity and the uncertainty of factors inducing geological disasters, complex geographical environment factors are usually interpenetrated and blended and data management and evaluation become extremely difficult. By using the method of the invention, the above defect is solved. The low coupling characteristic of the SOA architecture is fully used to integrate an AI early warning service with the monitoring system, and simultaneously the advantage of a machine learning algorithm which is used to process complex data and accurate classification is used so that the data of each monitoring object is monitored in real time based on the risk assessment of a risk source and monitoring object evaluation. The grades of a series of geological disasters with different intensities which may occur in a danger area for a certain period of time are predicted, and for the characteristics of different risk areas, the various countermeasures of risk reduction are proposed to provide an auxiliary decision for the geological disaster monitoring and early warning.

The invention discloses a high spectral image denoising method combining Bayes layered learning with space-spectrum combined priori; the method comprises the following steps: carrying out three dimensional slide block segmentation for a high spectral image according to a space spectrum correlation and non-local self-similarity thereof, and using relative distance priori based on fused features tonon-locally select a plurality of block data pieces to serve as synergy block data, wherein the selected block data is most similar to to-be-observed block data; using layered priori to build a Bayeslow-rank decomposition model, and realizing synergy block data learning and expression. The mode uses low-rank decomposition to carve statistics characteristics of synergy block data, and combines Dirichlet process mixing Gaussian distribution to express noise statistics characteristics; the method uses a variation Bayes method to solve the model, thus effectively reducing image noises. An existing method cannot simultaneously inhibit a plurality of noises in the high spectral image; the high spectral image denoising method combining Bayes layered learning with space-spectrum combined priori can solve said problems, is accurate in result, and can provide strong analysis basis for follow up applications.

An embodiment of the invention discloses a radio frequency control learning method and device, equipment and a medium. The method includes: according to set acquiring sequence, acquiring signals received by a set-frequency receiver and transmitted by to-be-learned remote control equipment; determining the signal strength indicator value of the signals according to the received signals; if the signal strength indicator value is not smaller than a set signal strength threshold, determining the set frequency as the target frequency; learning the remote control signals of the to-be-learned remotecontrol equipment according to the target frequency; associating the learned remote control signals and the target frequency with the equipment identifier of to-be-remotely-controlled equipment controlled by the to-be-learned remote control equipment. The method has the advantages that the learning of the remote control equipment of at least two different frequency is achieved, the control of theto-be-remotely-controlled equipment of different frequency is achieved, and the problem that exiting remote control equipment can only perform remote control on the basis of single fixed frequency issolved.