215 results about "Typing Classification" patented technology

The invention concerns a method and corresponding system for building a phonotactic model for domain independent speech recognition. The method may include recognizing phones from a user's input communication using a current phonotactic model, detecting morphemes (acoustic and / or non-acoustic) from the recognized phones, and outputting the detected morphemes for processing. The method also updates the phonotactic model with the detected morphemes and stores the new model in a database for use by the system during the next user interaction. The method may also include making task-type classification decisions based on the detected morphemes from the user's input communication.

A system for sensing the presence, position and type classification of an occupant in a passenger seat of a vehicle, as well as for sensing the presence of a rear-facing child seat therein, for use in controlling a related air bag activator control system to enable, disable or control inflation rate or amount of inflation of an air bag. The sensor system employs sensor fusion, a process of combining information provided by two or more sensors (24, 26), each of which "sees" the world in a unique sense. In a preferred embodiment, infrared sensor inputs (78) and ultrasonic sensor inputs (79) are combined in a microprocessor by a sensor fusion algorithm (80) to produce an occupancy state output signal (85) to the air bag controller.

A method and a device for detecting road users and obstacles on the basis of camera images, in order to determine their distance from the observer and to classify them. In a two-step classification, potential other parties involved in a collision are detected and identified. In so doing, in a first step, potential other parties involved in a collision are marked in the image data of a mono-image camera; their distance and relative velocity are subsequently determined so that endangering objects can be selectively subjected to a type classification in real time. By breaking down the detection activity into a plurality of steps, the real-time capability of the system is also rendered possible using conventional sensors already present in the vehicle.

A system and method are disclosed for providing semi-supervised learning for a spoken language understanding module using semantic role labeling. The method embodiment relates to a method of generating a spoken language understanding module. Steps in the method comprise selecting at least one predicate / argument pair as an intent from a set of the most frequent predicate / argument pairs for a domain, labeling training data using mapping rules associated with the selected at least one predicate / argument pair, training a call-type classification model using the labeled training data, re-labeling the training data using the call-type classification model and iteratively several of the above steps until training set labels converge.

Disclosed are a method and system, capable of performing adaptive intrusion detection proactively coping with a new type of attack unknown to the system and capable of training an intrusion type classification model by using a small volume of training data, the system including a data collector configured to collect host and network log information, an input data preprocessor configured to convert data acquired through the data collector into a feature vector, which is an input type of intelligence intrusion detection, and an intelligence intrusion detection analyzer configured to perform an intrusion detection and a model update by using the extracted feature vector, and an intrusion detection learning model configured to detect an intrusion and learn classification of the type of attack based on training data.

The invention provides a bearing fault classification diagnosis method based on sparse representation and an LDM (large margin distribution machine), overcomes the defects that signal decomposition is incomplete, a reconstructed signal cannot better keep features of an observed signal and the like in the conventional single-channel mechanical compound fault diagnosis method. According to the method, signal conversion from one dimension to high dimension is realized with a CEEMD (complete ensemble empirical mode decomposition) method, the decomposition completeness is guaranteed, and a mode mixing phenomenon is inhibited; meanwhile, a dimensionality reduction method based on sparse representation is introduced into a feature extracting and processing process of a blind source signal, data are subjected to sparse reconstruction through sparse representation, and data feature information is extracted from global data, so that the reconstructed signal can better keep the data features of the observed signal; further, the LDM classification method is introduced into a model fault type classification processing process of a to-be-detected bearing, and the accuracy and effectiveness of bearing fault diagnosis can be improved by aid of the generalization ability of the LDM classification method.

The invention relates to a convolution-neural-network-based text emotion classification method. The method comprises: a text linguistic data set is collected and data in a text are expressed into one sentence; pretreatment is carried out on the collected text linguistic data set and emotion text linguistic data are classified into training set linguistic data and testing set linguistic data; training is carried out on the text linguistic data set after pretreatment by using a word2vec tool to obtain a word vector model and a text vector is also obtained; the text vector of the training set linguistic data is inputted into a convolution neural network and training is carried out to obtain an emotion classification model; and the text vector of the testing set linguistic data is inputted into the convolution neural network, emotion type classification is carried out on the trained emotion classification model, and an accurate rate of emotion classification is calculated. Therefore, a problem that lots of artificial marks are needed during the previous classification process can be solved.

The invention relates to a judgement document-based structured processing method. According to the method, a natural language processing technology and an advanced machine learning technology are adopted to automatically realize case type classification on the basis of keyword extraction of brief texts, so that structured processing is carried out through constructing case hierarchical structuresand designing an extraction rule. Through constructing and extending a related lexicon, segmenting a judgment document module, designing and determining a cluster number K and an initial cluster center, and taking an increment of a word weight as second feature selection, the improvement of a kmeans cluster algorithm is realized, and class labels of cases are obtained; and through creating different hierarchical frameworks according to different case types and combining the designed extraction rule, structured processing of judgement documents is obtained. The method is capable of rapidly realizing the structured processing of judgement documents.

The invention relates to the field of network information retrieval and mining, in particular to a system and method for identifying webpage types. The method comprises the following steps that a heuristic rule is predefined, and a heuristic rule list is generated; a predetermined feature is extracted from a training webpage to form a standard feature vector which is optimized twice to form a simplified feature set, a classifier and a feature extractor are established, and a classification model is generated through the classifier; based on an URL and a source code of a webpage to be identified, rule matching is carried out on the heuristic rule list; if matching succeeds, the webpage type of the webpage to be identified is output; if matching fails, the classifier is used for carrying out webpage type classification on the webpage to be identified. The system and method for identifying webpage types are flexible and convenient to use, high in identifying speed and high in identifying accuracy, big change is of no need when a cross language webpage is identified, identifying efficiency is high, and high actual use value is achieved.

The invention relates to the field of automobile classification technologies and discloses a method and system for classifying automobile types based on a neural network. The method comprises the steps that a plurality of training samples are collected, and the training samples are classified based on the convolutional neural network, so a classifier containing label results is obtained; when the automobile types are classified, a video image to be detected is read in, a motion object in the image is detected, and sub-block processing is performed on the image according to the motion object; afterwards, classification processing is performed on each block of image through the classifier, so a detection result is obtained. Accordingly, a neural network system can be constructed easily and conveniently to serve as the classifier, the system is trained by using different automobile samples, the system is made to automatically study complex class conditional density of the samples, and therefore the problems caused by class conditional density functions of artificial hypothesis are avoided. Compared with an existing automobile type classification method, the method for classifying the automobile types based on the convolutional neural network has the advantages that the accuracy of classification is improved, and the classification speed is increased.

The invention discloses a high-resolution remote-sensing multifunctional urban land spatial information generation method. According to the method, a complex system rank theory is introduced; an urban land spatial information classification system with three rank scales, i.e. a multifunctional target landscape type, a functional area type and a land cover type, which self-adapts to urban planningmanagement and environmental renovation, is proposed; and on the basis of realizing the treatment of fine correction and alignment on remote-sensing images of a Landsat TM, a Google Earth and auxiliary maps, the Landsat TM is applied to carrying out urban landscape type classification, a three-level rank classification type-merged combination and an information mining knowledge base are constructed, the classified information is merged to form a first-level classification result of urban land, the digital functional areas are classified into a second level and the land cover is classified into a third level under the constrained control of higher-level classification information. The method has the characteristics of low cost, high accuracy of classification and strong targeted application, and thus, the requirements on target applications, such as ecological urban design, urban environmental management and the like, are better met.

A method for selecting a digital image having controlled sharpness characteristics from a set of candidate digital images of a common scene, each digital image having different sharpness characteristics. An image segmentation process is used to segment each of the candidate digital images into a subject region and a background region. For each candidate digital image the subject and background regions are analyzed to determine an associated subject and background sharpness levels. An output digital image is selected by comparing the determined subject and background sharpness levels to respective aim subject and background sharpness levels. In some embodiments, the aim subject and background sharpness levels are defined in accordance with a scene type classification.

The invention discloses a heartbeat type classification method and device, and electrocardiogram equipment and belongs to the field of medical apparatuses. The method comprises steps that heartbeat waveform characteristic parameters are extracted from ECG signals, heartbeat and template heartbeat matching degree characteristic parameters are further extracted, analysis is carried out through a decision tree classifier, and the heartbeat type analysis result can be acquired through measurement. An extracted characteristic vector can be better present difference of the ECG signals, and disadvantages of utilizing only a waveform analytical method or a template analysis method are avoided. The invention further comprises the electrocardiogram equipment comprising the method. The method is advantaged in that the characteristic vector extraction algorithm is simple and easy for realization, the operation speed is fast, classification accuracy is high, and the method is suitable for a wearable electrocardiogram real-time monitoring automatic analysis diagnosis system.

The invention provides an optical fiber vibration identification system based on a phi-OTDR technology and an optical fiber vibration identification method thereof. The monitoring distance of the system is greatly enhanced through a dual-path detection structure; the adaptability of the system for environmental noise changes is enhanced by the method of characteristic threshold dynamic updating, and a vibration event is accurately positioned; background noise in signals can be greatly reduced through spectral subtraction noise reduction under the condition of maintaining the signal characteristic and energy of the vibration signals, and the signal-to-noise ratio of the signals and the sensitivity of system detection can be enhanced; and multi-characteristic parameter mode identification is performed on the vibration signals from the time domain and the wavelet domain so that the influence of other complex time-dependent interference noise can be effectively avoided, the correct rate of vibration event detection and vibration type classification can be enhanced, the false alarm rate of the system can be reduced, the detection performance of a vibration detection system based on the OTDR technology in the actual complex noise environment can be enhanced, and the national major project application requirements in the aspects of boundary safety and long-distance pipeline safety can be met.

The invention relates to a dynamic text cluster method based on network flow and laminated knowledge base, belonging to information processing and network safety technical field. The inventive method comprises that first assuming the clustered file type is provided with vector character, using TFIDF method to extract and normalize character of single clustered text, using the method that defining meaning distance in the knowledge base to calculate the distance of text and type, adjusting and refreshing keyword and weight of the type of new added file, when present file can not be combined with known types, needs to build new type. And the algorism comprises dynamic character vector extraction, type classification, distance calculation, type combination and new type construction. The invention is characterized in that the cluster process is based on the meaning information provided by the laminated knowledge base but not keyword, the invention can dynamically remove noise data, the similarity is calculated network flow algorism, to confirm optimized match, and the invention can meet real-time refresh cluster of Web text, in particular as non-detect type, without pre-appointing type group.

The sharpness of a digital image is adjusted according to defined aim subject and background sharpness levels. An image segmentation process is used to segment an input digital image into a subject region and a background region. The subject and background regions are analyzed to determine corresponding subject and background sharpness levels. An enhanced digital image is formed wherein the sharpness of the subject region is adjusted responsive to the subject sharpness level and the aim subject sharpness level, and the sharpness of the background region is adjusted responsive to the background sharpness level and the aim background sharpness level. In some embodiments, the input digital image is analyzed to determined a scene type classification and the aim subject and background sharpness levels are defined in accordance with the determined scene type classification.

The invention discloses a language ability testing method and system. The language ability testing method includes the following steps: by setting a test question database, classifying the test questions according to the degree of difficulty and setting corresponding grades; according to the test question type Choose moderately difficult test questions as test benchmark questions to test the testers; adjust the difficulty level of the test questions according to the answering behavior and answering results; Analyze and determine the tester's final language proficiency level, and test the tester through the test benchmark questions to distinguish the different levels of the tester's language ability, such as strong, medium, and weak, and conduct personalized tests according to the relevant levels and adjust the test questions according to language ability. It is easier to know your language proficiency level, realize targeted learning, and improve language learning efficiency.

The invention relates to a traffic signal coordination control optimization method for an emergency rescue vehicle driving path. The method comprises emergency rescue prioritization, path type classification, signal priority control, signal coordination control, and a bi-level programming model; relevant coordination control research on paths is conducted; first, emergency rescue characteristics and objects are analyzed, research on upper level dynamic phase difference and lower layer path green wave coordination in different path forms is conducted, and relevant case analysis and verificationare performed. The method disclosed in the invention is characterized by real time performance, traffic signal coordination priority control over an emergency rescue path can be exerted based on timewhen an emergency rescue team reaches a crossroad and real time vehicle load information of general vehicles, emergency rescue response time can be shortened, and casualties and property loss can beeffectively reduced.

The invention provides a method work-out the bottom water cone advancing height, cone advancing radius and water cone volume according to the bottom water cone predicted model, researches the water submerging degree and residue oil latent capacity with the water cone shape, and the oil wells are classified according to the water containing rising rule, researches the change rule of the water capacity in oil; determines a bottom water treatment rule and solves flowing nine problems: 1. oil well type classification; 2. reservoir submerged degree estimation; 3. residue oil latent capacity analysis; 4. the selecting rule of bottom water treating method; 5. the selecting condition of bottom water treating well; 6. the decided rule of the isolating board in manual isolating board method; 7. creates the sifting standard and the assembly rule of blocking and sealing agent for isolating board; 8. creates the design method of the blocking agent quantity; 9. the effect estimation and the economic efficiency prediction of the bottom water treatment.

The invention discloses a method for processing multiple foreground tasks on a screen. The method comprises the following steps that: a plurality of application windows are operated on the same display screen through a multitasking mechanism; user events are received; the received user events are subjected to type classification; and a task management module is used to allocate the received user events with marks to different tasks to be processed, and the corresponding results are returned to the corresponding application windows. The method and device for processing multiple foreground tasks on the screen provided by the invention are used for realizing that the user events among a plurality of foreground tasks on the same screen are reasonably allocated and the task windows are displayed at the same time, and thus the plurality of foreground tasks can be operated on the same screen at the same time.

The embodiments of the invention provide a human face shape classification method and a system, wherein the method comprises: obtaining the three-dimensional point cloud data of a target user's head part and two-dimensional image data; according to the three-dimensional point cloud data, generating a three-dimensional model for the target user's head part; according to the corresponding relationship between the three-dimensional point cloud data and the two-dimensional image data, mapping the two-dimensional image data into the three-dimensional model; based on the texture information and the color information of the two-dimensional image data, detecting the human face area; according to the mapped three-dimensional model and the human face area, extracting the human face characteristics points; and according to the human face characteristics points, performing classification to the target user's human face shape. The embodiments of the invention increase the accuracy to extract the human face characteristics points, overcome the defects of poor Robustness of purely extracting human face characteristics points from a two-dimensional image and increase the accuracy for human face type classification.

The invention discloses a safety on-demand supply method. The method comprises the following steps of: determining configuration parameters of a safety function module according to a safety level which is set for service request by a user, an application scene in which a user terminal is positioned, and / or service type; and configuring the safety function module according to the configuration parameters, and safely protecting service data of the user. The invention also discloses a service type acquisition method for safely protecting a specific user and / or a service. Service identifiers of data are acquired by using a service type classification function of a quality of service (QoS) function module, so the specific user and / or the service are safely protected. The invention also discloses a safety on-demand supply system and the service type acquisition method. By the invention, according to safety requirements of different users on different services, different service safety guarantee can be provided. The safety on-demand supply system meets the safety requirements of various users and various services, personalized safety guarantee is provided for the users, and a user experience effect is improved.

The invention discloses a photovoltaic power station generation power prediction method comprising the steps that six meteorological characteristics are daily extracted by using the historical meteorological data of a photovoltaic power station so that a meteorological characteristic library is established; the daily characteristic data in the meteorological characteristic library are clustered through a KFCM algorithm so as to realize weather type classification, and class marking is performed on the daily power data and the meteorological data; an SVR sub-model is established for each classof power data and meteorological data according to the class mark; the weather type of the target day is identified by using the SVM through the target day weather characteristics provided by numerical weather prediction and the corresponding SVR sub-model is selected; an ARIMA model is established by using the real-time monitoring data of the target day, and real-time prediction of the irradiation intensity and the temperature can be realized by using the rolling prediction model; and the prediction values of the irradiation intensity and the temperature are inputted to the selected SVR sub-model so that the photovoltaic power station power prediction result can be obtained. The photovoltaic power station generation power prediction accuracy can be enhanced.

A speech compression system capable of encoding a speech signal into a bitstream for subsequent decoding to generate synthesized speech is disclosed. The speech compression system optimizes the bandwidth consumed by the bitstream by balancing the desired average bit rate with the perceptual quality of the reconstructed speech. The speech compression system comprises a full-rate codec, a half-rate codec, a quarter-rate codec and an eighth-rate codec. The codecs are selectively activated based on a rate selection. In addition, the full and half-rate codec are selectively activated based on a type classification. Each codec is selectively activated to encode and decode the speech signals at different bit rates emphasizing different aspects of the speech signal to enhance overall quality of the synthesized speech. The overall quality of the system is strongly related to the excitation. In order to enhance the excitation, the system contains a fixed codebook comprising several subcodebooks. The invention reveals a way to apply a pitch enhancement efficiently and differently for different subcodebooks without using additional bits. The technique is particularly applicable to selectable mode vocoder (SMV) systems.

A medical imaging device (300) comprising a magnetic resonance imaging system (302). The medical image device further comprises a memory (334) containing machine executable instructions (370, 372, 374, 376, 378, 380, 382, 384, 386) for execution by a processor (328). Execution of the instructions causes the processor to receive (100, 204) a pulse sequence protocol (340). Execution of the instructions further causes the processor to determine (102, 206) a pulse sequence type classification (342) descriptive of the pulse sequence protocol. Execution of the instructions further cause the processor to determine (104, 208) a magnetic resonance contrast classification (344), wherein the choice of the magnetic resonance contrast classification depends upon the pulse sequence type classification. Execution of the instructions further causes the processor to determine (106, 210) a pulse sequence protocol classification (346). The pulse sequence protocol classification is determined by the pulse sequence type classification and the magnetic resonance contrast classification.

The invention discloses a method and a system for detecting vehicle storage battery service life. The method includes: the OBD interface voltage change curves of various car types are collected in advance, type classification is performed on the voltage change curves according to the service life of storage batteries, the real-time voltage change curve of the OBD interface of a current car type is monitored and is matched with the collected voltage change curve of the same car type so as to predict residual service life of the storage battery of the current car type, if the residual service life of the storage battery is smaller than a preset value, and a warning for replacing or maintaining the storage battery is given out. By the method, a user does not need to worry that a car fails due to the fact the storage battery has no signs before being used up, pre-warning of the storage battery service life can be achieved, the user can replace or maintain the storage battery in advance, damage to car electric elements is lowered, and system safety is increased.