54 results about "Computer learning" patented technology

In a first exemplary embodiment of the present invention, an automated, computerized method is provided for processing an image. According to a feature of the present invention, the method comprises the steps of providing an image file depicting an image, in a computer memory, assembling a feature vector for the image file, the feature vector containing information regarding a likelihood that a selected pair of regions of the image file are of a same intrinsic characteristic, for example, a same texture, providing a classifier derived from a computer learning technique, computing a classification score for the selected pair of regions of the image file, as a function of the feature vector and the classifier and classifying the regions as being of the same intrinsic characteristic, as a function of the classification score.

The red eye eliminating method based on human face detection in image treating technology includes three steps of: 1. human face detection with human face classifier obtained through computer learning algorithm training combined with skin color information to obtain the human face area from the picture; 2. red eye detection via dividing the detected area into typical red eye and wear red eye based on the red color distribution in the human face area; and 3. red eye correction via searching red eye area, color correction and edge smoothening to restore normal color of picture. The present invention adopts fast human face detecting algorithm to reduce red eye searching range and correction based on the color of pupils, and can detect precisely and correct red eye in digital picture in reasonable image processing speed even in condition of no attendance.

Various systems may benefit from computer learning. For example, robotics systems may benefit from learning actions, such as manipulation actions, from unconstrained videos. A method can include processing a set of video images to obtain a collection of semantic entities. The method can also include processing the semantic entities to obtain at least one visual sentence from the set of video images. The method can further include deriving an action plan for a robot from the at least one visual sentence. The method can additionally include implementing the action plan by the robot. The processing the set of video images, the processing semantic entities, and the deriving the action plan can be computer implemented.

In a first exemplary embodiment of the present invention, an automated, computerized method is provided for processing an image. According to a feature of the present invention, the method comprises the steps of providing an image file depicting an image, in a computer memory, assembling a feature vector for the image file, the feature vector containing information regarding a likelihood that a selected region of the image file is specular, providing a classifier derived from a computer learning technique, computing a classification score for the selected region of the image file, as a function of the feature vector and the classifier and classifying the region as being specular, as a function of the classification score.

A computer learning, assessment and recommendation system to generate new information about the quality and degrees of walkability of sidewalks and walking routes through the collection, assessment and manipulation of large amounts of visual and geographic data, a scoring system, a rule-based computational methods and selected data representation are provided. Micro and macro data, methods and systems for sidewalk assessment generate various characteristics and qualities of routes creating general, locational and navigational functions using attributes and categories identified as important to walkability or to the experience of walking. Computed scores and recommendations displayed in maps and other visual tools and apps can be used by multiple client and sector groups including navigation, real estate, fitness, tourism and urban and rural development planning. This abstract complies with rules requiring abstract submission but does not limit the scope, interpretation or full meaning of the claims.

The invention discloses a ship movement control experimental system and a work method thereof. The system comprises a main control computer, an Ethernet switch, an emergency remote control handle, a computer learning module, an on-shore wireless communication module, an onboard wireless communication module, an onboard computer, a Beidou navigation system, a ship compass and a reduced-scale ship model. The main control computer comprises a human-computer interaction interface which is achieved on the basis of LabView development, RT-LAB is applied to generate C codes of a Matlab / Simulink model of ship movement controllers to be verified by experiments, the C codes are compiled and uploaded to the onboard computer which runs a real-time operation system QNX, various ship movement controllers to be verified by experiments can be directly operated, researchers can conveniently and quickly carry out various complex experiment verifications on the ship movement controllers through the experimental system, experimental research workload is reduced, and the content and range of experiments are widened.

The invention provides an environment image multi-view-angle meaning cutting method and device, and the method comprises the following steps of: collecting multi-view-angle image data and carrying out pre-processing on the multi-view-angle image data, wherein the pre-processing includes establishing a three-dimensional coordinate system and dividing pixel points in an image into a plurality of super pixels; extracting two-dimensional characteristic vectors of the super pixels, wherein the two-dimensional characteristic vectors include an RGB value, a mean of lab color space composition, a derivation value, a bias or a peak state; extracting three-dimensional characteristic vectors of the super pixels, wherein the three-dimensional characteristic vectors include a three-dimensional point density; and substituting the two-dimensional characteristic vectors and the three-dimensional characteristic vectors into a smoothness value equation, and dividing the super pixels into a plurality of clusters. By adopting the environment image multi-view-angle meaning cutting method and device, an image identification effect in a computer system is achieved, and the problem of computer learning object classification is solved.

The invention relates to a self-adaptive learning method based on a graph search algorithm and a computer learning system, and the method comprises the following steps: 1) building a knowledge graph,and storing the knowledge graph in a graph database; 2) collecting and recording user learning data, judging whether a trigger condition of front knowledge point searching is met or not, if yes, executing the step 3), and if not, executing the step 5); 3) searching front weak knowledge points of the current knowledge points in the graph database by utilizing the graph search algorithm, to obtain apriority learning sequence of the front weak knowledge points; 4) pushing the corresponding learning contents based on the priority learning sequence, and 5) pushing the corresponding learning contents based on the current knowledge point. Compared with the prior art, the method has the advantages of being fast in search, improving the learning efficiency and the like.

In a first exemplary embodiment of the present invention, an automated, computerized method is provided for processing an image. According to a feature of the present invention, the method comprises the steps of providing an image file depicting an image, in a computer memory, assembling a feature vector for the image file, the feature vector containing information regarding a likelihood that a selected pair of regions of the image file are of a same intrinsic characteristic, providing a classifier derived from a computer learning technique, computing a classification score for the selected pair of regions of the image file, as a function of the feature vector and the classifier and classifying the regions as being of the same intrinsic characteristic, as a function of the classification score.

The invention discloses a media information tensor supervised learning-based stock price fluctuation prediction system and method. Tensors are built according to media information, and a supervised learning method is used for predicting stock price fluctuations. The stock price fluctuation prediction system and method are advantageous in that (1) the system and method are based on three dimension attribute information including market transaction information, stock investor emotion information and media news information; a quantitative information source and a qualitative information source are included, official new information and social media information are included, and therefore relations between Internet medias and stock markets can be analyzed comprehensively; (2) the tensors are used for representing media information space, and relations between information of different dimensions can be recorded; (3) expansion from vector-based computer learning algorithms to tensor-based computer learning algorithms is realized via the tensor supervised learning algorithm.

A computer learning system includes a database containing one or more sample essay questions and model answers prepared in advance by qualified instructors. The system assists a user in preparing for an essay examination by presenting the user with a sample essay question and prompting the user to identify an issue raised by the essay question until the user correctly identifies an issue raised by the essay question. The user is then enabled to provide a partial response related to the identified issue and to conduct a self-evaluation as to whether the partial response is correct by comparing the partial response to a partial model answer. Once the user has completed a question, the user's complete response, which combines the user's partial responses and organizes them by issue, is displayed and the user can compare the complete response to a complete model answer, which includes the partial model answers, organized by issue.

The invention belongs to the technical field of computers, in particular to a special placing device for computer learning books. The technical problem to be solved by the present invention is to provide a special placing device for computer learning books that can save space, be convenient to place, not easy to lose, and can improve learning efficiency. In order to solve the above-mentioned technical problems, the present invention provides such a special placement device for computer learning books, which includes a bottom plate, wheels, box body, hinge, box cover, pulley, placement plate, electric reel, first spring, second A motor, the first connecting plate, the first cylinder, etc.; the bottom of the bottom plate is provided with wheels, the top of the bottom plate is connected with a box by bolts, the left and right sides of the top of the box are connected with hinges by screws, and the top of the box is equipped with Two lids. In the present invention, the computer books are placed on the placing board in the box body, which is convenient for storage and not easy to be damaged and lost.

The invention provides a remote sensing identification method and device for grassland salinization and a computer readable storage medium, and belongs to the field of a remote sensing technology andenvironment survey. Preprocessing is carried out by existing salinization data, spectral features are obtained and used as feature input quantities, then computer learning is carried out, and an identification model is established; data of a to-be-detected region is identified by the model to obtain a salinization region; the identification device for salinization can rapidly and accurately identify the salinization region; and the computer readable storage medium can conveniently execute the identification method to obtain related data and then obtain data of the salinization region.

The invention relates to a riding posture recognition method which is suitable for execution in a computing device. The method includes the steps: acquiring posture parameters and generating posture attribute information according to the posture parameters; judging that the posture attribute information is as same as preset action information to generate control command information; transmitting the control command information to an indicating unit and executing the control command information by the indicating unit. According to the riding posture recognition method and an application device,based on computer learning and matching technique components, different postures of riders are recognized to generate corresponding steering indication information and realize bicycle steering promptduring riding, complicated connecting structures are omitted, and the device is simple in structure and can be effectively mounted and used.

In a first exemplary embodiment of the present invention, an automated, computerized method is provided for determining illumination flux in an image. According to a feature of the present invention, the method comprises the steps of performing a computer learning technique to determine spatio-spectral information for the images, and utilizing the spatio-spectral information to identify illumination flux.

The invention discloses a computer teaching system, which comprises a main control unit, wherein the main control unit is connected with a teacher terminal, a signal conversion module, a student terminal and a loudspeaker unit, one side of the teacher terminal is connected with a teacher computer, a projection module, a display module and a pickup module, and one side of the student terminal is connected with a student computer, a calling module, a monitoring module and a check-in module. According to the computer teaching system, the teacher can give lessons by means of the teacher computer, project teaching content by means of the projection module, and can utilize the pick-up module to explain, so as to ensure that each student can hear the knowledge explained by the teacher clearly; and the student can learn the knowledge by means of the student computer, can ask questions to the teacher and answer teacher's questions by using the calling module, and the classroom discipline of the student can be monitored by means of the monitoring module; and the computer teaching system has comprehensive functions, is stable in operation, and can effectively improve the teaching efficiency of the teacher.

The invention relates to a computer learning machine with a detachable keyboard. The computer learning machine comprises a host and the detachable keyboard. The host is provided with an accommodation cavity. The detachable keyboard is accommodated in the accommodation cavity, and comprises a first panel, a second panel and a control circuit main board, wherein the first and second panels are arranged on a shell, and are opposite to each other; the control circuit main board is arranged in the shell; the first panel comprises a full keyboard and a first touchpad; the second panel comprises direction keys and a second touchpad; the control circuit main board comprises a microprocessor, a universal serial bus (USB) data interface connected with the host and a wireless communication module; when the USB data interface is detected to be in signal connection with the host, all the operation of the second panel fails under the action of the microprocessor; and when the signal connection between the USB data interface and the host is detected to be interrupted, all the operation of the first panel fails under the action of the microprocessor, and the wireless communication module is started to be connected with the host. The keyboard can be simply and conveniently taken down and arranged, and can be used in a double-sided way.

The invention relates to a self-adaptive learning task intelligent generating device and a computer learning system. The generating device comprises a first storage module, a second storage module, aninput module and an analysis module, wherein the first storage module is used for obtaining and storing basic information of a user; a second storage module is used for obtaining and storing historical learning behavior data of the user; the input module is used for obtaining learning target data in real time; the analysis module is used for analyzing the basic information of the user, the historical learning behavior data and the learning target data to generate learning task data. Compared with the prior art, the self-adaptive learning task intelligent generating device and the computer learning system have the advantages that the learning efficiency is effectively improved, and the like.

An architecture and processes enable computer learning and developing an understanding of arbitrary natural language text through collaboration with humans in the context of joint problem solving. The architecture ingests the text and then syntactically and semantically processes the text to infer an initial understanding of the text. The initial understanding is captured in a story model of semantic and frame structures. The story model is then tested through computer generated questions that are posed to humans through interactive dialog sessions. The knowledge gleaned from the humans is used to update the story model as well as the computing system's current world model of understanding. The process is repeated for multiple stories over time, enabling the computing system to grow in knowledge and thereby understand stories of increasingly higher reading comprehension levels.

The disclosure discloses estimation of the implicit likelihoods of generative adversarial networks, and relates to the field of computer learning. The thriving of deep models and generative models provides approaches to model high dimensional distributions. Generative adversarial networks (GANs) can approximate data distributions and generate data samples from the learned data manifolds as well. Presented herein are embodiments to estimate the implicit likelihoods of GAN models. In one or more embodiments, a stable inverse function of the generator is learned with the help of a variance network of the generator. The local variance of the sample distribution may be approximated by the normalized distance in the latent space. Simulation studies and likelihood testing on data sets validate embodiments, which outperformed several baseline methods in these tasks. An embodiment was also applied to anomaly detection. Experiments show that the embodiments herein can achieve state-of-the-art anomaly detection performance.

Systems and methods are disclosed to label data collected by one instrument based on data collected by another instrument, train a computer learning model with the labeled data, and then use the trained computer learning model to estimate a condition from new unlabeled data. For example, weather-related pavement conditions may be estimated from camera images according to such systems and methods. Systems and methods are also disclosed to estimate road weather safety or hazard conditions using two different types of pavement conditions, such as road state and friction or grip coefficient, estimated from unlabeled camera images.

There are provided a learning processing method and device capable of improving the learning speed and computer learning accuracy when computer learning is performed by using a plurality of learning data. A similar learning data generation unit (4) selects similar learning data SSDq having a high similarity with the data to be processed, from n learning data SDq. A computer learning device (5) performs computer learning by using the similar learning data SSDq.

Techniques are disclosed relating to automating authentication decisions for a multi-factor authentication scheme based on computer learning. In disclosed embodiments, a mobile device receives a first request corresponding to a factor in a first multi-factor authentication procedure. Based on user input approving or denying the first request, the mobile device sends a response to the first request and stores values of multiple parameters associated with the first request. The mobile device receives a second request corresponding to a factor in a second multi-factor authentication procedure where the second request is for authentication for a different account than the first request. The mobile device automatically generates an approval response to the second request based on performing a computer learning process on inputs that include values of multiple parameters for the second request and the stored values of the multiple parameters associated with the first request. The approval response is automatically generated and sent without receiving user input to automate the second request.

The invention relates to the field of computer learning and data detection. The invention particularly relates to a batch automatic detection method for background color compliance of electronic certificate photos. The method comprises the steps of dividing electronic certificate photos into a sample set and a detection set; training the sample set through a background color detection model; obtaining a stable deep neural network background image detection model; therefore, the electronic certificate photos in the detection set can be automatically detected through the background colors of theelectronic certificate photos; the background color detection efficiency of the electronic certificate photos is greatly improved, meanwhile, the number of manual detection is greatly reduced, and the problems that manual detection needs to be conducted on the background colors of existing electronic certificate photos one by one, the detection result of the images cannot be automatically given,the efficiency is low, and missing and misjudgment are likely to happen are solved.

The invention discloses a computer teaching system which comprises a master control unit. The master control unit is connected to a teacher terminal, a signal conversion module, a student terminal and a voice amplifying unit. One side of the teacher terminal is connected to a teacher computer, a projection module, a display module and a pickup module; one side of the student terminal is connected to a student computer, a calling module, a monitoring module and a sign-in module. The teacher can teach through the teacher computer and projects the taught content through the projection module; the teacher can explain by means of the pickup module, so that every student can clearly hear knowledge explained by the teacher; the student learns knowledge through the student computer, asks questions to the teacher and answers questions through the calling module and monitors the classroom discipline of the student through the monitoring module. The system is comprehensive in function and stable to operate and can effectively increase the teaching efficiency of the teacher.

A method and system of training a computer learning model to improve medical care. The method and system may include training a computer learning model to obtain a predictive model that can identify success rates, potential complications, anesthesia time, and predicted life expectancy. The method and system can include inputting physiological data to generate and / or remove one or more predictive models.

Past electronic records may be studied using a computer learning algorithm in order to make predictions of future use of the electronic accounts.

The invention discloses a large-space simple desk. The large-space simple desk comprises an upper desk top, a drawer, desk legs, a bottom plate and lifting devices. A learning tool storage space is formed above the upper desk top, a front desk top is arranged at the front end of the upper desk top and is connected with the upper desk top through hinge pins, and the front desk top is connected with the bottom plate through a fixing plate; the drawer is located below the upper desk top, a lower desk top is arranged on the drawer and located under the upper desk top, and the bottom plate is arranged under the drawer. Each lifting device is connected with the upper desk top, the lower desk top and the bottom plate, the desk legs are fixed to the bottom plate through fasteners, universal wheels are arranged at the bottoms of all the desk legs, and a brake mechanism is arranged on each universal wheel. As the lifting devices are arranged on the desk, the upper desk top and the lower desk top can be lifted, and the desk is adaptive to new learning tools such as a tablet personal computer and a learning machine and can ensure sufficient writing space. The universal wheels are arranged under the desk legs of the desk, and the desk is convenient to move with less effort.

The invention discloses an application cleaning method, device and equipment and a readable storage medium, and belongs to the technical field of mobile terminals. The method comprises the steps: carrying out preprocessing on application use data to obtain an input vector; selecting a preset parameter regression model, and training the parameter regression model through the input vector to obtainhabit data of a user in using the application; predicting the waiting time required for restarting after the application is suspended according to the habit data; and cleaning the application according to the waiting time. Through a computer learning algorithm, the application use habit of the user is learned, and the application is intelligently cleaned, so that the user experience is improved, and the endurance time of the mobile terminal is prolonged.

The invention relates to a thinking mathematics learning method based on capability grading and a computer learning system. The learning method comprises the steps of obtaining the thinking mathematics portrait information of a student user; taking the thinking mathematics portrait information subjected to discretization processing as the input of a classifier to obtain the grade information corresponding to the student user; based on the grade information and the knowledge point learning model of the current class hour, obtaining the screening information of the student user for each knowledge point in the current class hour; and pushing the learning contents based on the screening information. Compared with the prior art, the method has the advantages of being beneficial to learning efficiency and the like.