Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
Hiro

119 results about "Comprehension approach" patented technology

The comprehension approach is methodologies of language learning that emphasise understanding of language rather than speaking. This is in contrast to the better-known communicative approach, under which learning is thought to emerge through language production, i.e. a focus on speech and writing.

Image understanding method based on depth residual error network and LSTM

The invention discloses an image understanding method based on a depth residual error network and an LSTM; the method comprises the following steps: firstly building a depth residual error network model so as to extract image abstract features, and storing the features as a feature matrix; using a dynamic attention mechanism in a LSTM model to dynamically form a proper feature vector according to the feature matrix; finally using the LSTM model to form a natural language (English) according to the feature vector. The method uses the advantages of the depth residual error network on image feature extraction and LSTM advantages on time sequence modeling; the depth residual error network and the LSTM model can form an encode-decode framework so as to convert the image content information into the natural language, thus extracting the deep information from the image.
Owner:SOUTH CHINA UNIV OF TECH

Indoor scene understanding method based on 2D-3D semantic data set

The invention discloses an indoor scene understanding method based on a 2D-3D semantic data set. The indoor scene understanding method comprises the main content of collection, arrangement, training and test division of data and comprises the following steps: firstly capturing an image, outputting a scanning region, an original color depth image (RGB-D) and a 3D texture grid, sampling the grid to generate point clouds, carrying out semantic annotation on the data, projecting each point label to a 3D grid and an image field, and representing certain regions with concentrated data to parts with similar buildings in aspects of appearances and architectural features, so as to define standard training and test division. According to the semantic data set disclosed by the invention, cross-modal study models and non-supervision methods can be developed and united by virtue of rules existing in large-scale indoor spaces; powerful prompts are provided for the detection of semanteme, layout, shields, shapes and modes; and the indoor scene understanding method is not limited by scale, diversity and number.
Owner:SHENZHEN WEITESHI TECH

Understanding method of non-parametric RGB-D scene based on probabilistic graphical model

The invention discloses an understanding method of a non-parametric RGB-D scene based on a probabilistic graphical model. The method comprises the steps of carrying out global feature matching between a marked image and an image marked in a training seat, and building a retrieval set of a similar image of an image to be marked; cutting the image to be marked and the image in the similar image retrieval set, so as to generate super-pixels, and carrying out characteristic extraction on the super-pixels extracted; calculating the proportions of all categories in the training set, building a dictionary of rare categories, and taking the training set and the retrieval set of the similar images as a label source of the image to be marked; carrying out characteristics matching on each super-pixel of the image to be marked and all super-pixels in an image label source; and building a probabilistic graphical model, converting the maximum posterior probability into a minimal energy function by using a Markov random field, and resolving the semantic annotation of each super-pixel of the image to be marked obtained by solving the problem through a graph cutting method. According to the method provided by the invention, the overall and local geometric information can be integrated, and the understanding performance of the RGB-D scene can be improved.
Owner:ZHEJIANG UNIV

Interactive-question semantic understanding method in intelligent customer services

The invention provides an interactive-question semantic understanding method in intelligent customer services. The method includes the following steps that the conversation content between current intelligent customer services and clients is subjected to co-text language environment expression, wherein co-text language environment expression comprises event expression and language environment expression; according to the co-text language environment expression, a conversation semantic event graph is constructed; according to multiple conversation corpora of the intelligent customer services and the clients, a business logic tree is constructed; according to a determined finite state automata, an order state machine is established; according to the semantic event graph, logic decision branches are selected from the business logic tree; according to the logic decision branches and the order state machine, a semantic processing template is returned to the intelligent customer services, and semantic expression generation is carried out. According to the interactive-question semantic understanding method in the intelligent customer services, interaction questions and answers based on aflow diagram are achieved, the accuracy of client interrogation understanding of the intelligent customer services is increased, the consistency between the intelligent customer services and client conversations is guaranteed, and the work efficiency of the intelligent customer services is improved.
Owner:KANGCHENG INVESTMENT CHINA

Multi-sensor-fusion-based unstructured environment understanding method

InactiveCN102564431ASolve the situation where road edge features cannot be extracted wellSolve the problem that cannot be effectively extractedInstruments for road network navigation2D-image generationFeature extractionRadar
The invention discloses a multi-sensor-fusion-based unstructured environment understanding method, which comprises the following steps of: firstly, registering and aligning the characteristic information of each vision sensor, and projecting the characteristic information to a current vehicle coordinate system; secondly, extracting fused road edges by adopting a confidence-weighting-based road edge characteristic extraction method; thirdly, performing inter-frame data comparison and judgment on the fused road edges to obtain more stable road edges; fourthly, extracting a passable area from three-dimensional radar data, and fusing the passable area with the stable road edges obtained from the vision sensors to obtain optimal road edge information; and finally performing inter-frame fusion on road edge results to reduce the change of inter-frame data and finally realize stable and reliable understanding in an unstructured environment. A confidence-weighting-based road edge fusion algorithm is adopted, so that the problem of incapability of effectively extracting road edge characteristics under the condition of a single sensor or a single frame of data is solved.
Owner:NANJING UNIV OF SCI & TECH

Scene understanding method based on multi-task learning

The invention provides a scene understanding method based on multi-task learning. The method includes: multi-task learning having unstable homoscedasticity, multi-task likelihood function, and scene understanding model. The method includes the following steps: first executing weighted linear sum on the loss of each individual task, learning the optimal task weight, inducting a multi-task loss function, defining a probability model, defining the probability as the gaussian function of the average values that are output by the model, eventually modeling pixel-level learning regression and classified output, comprising semantic division, incidence division and depth regression. According to the invention, the scene understanding model can learn multi-task weight, is more advantageous than models that independently train each task in that the method herein reduces computing amount, increases learning efficiency and prediction precision and can be real-timely operated.
Owner:SHENZHEN WEITESHI TECH

Text semantic understanding method, device and system

The invention provides a text semantic understanding method, device and system. The method comprises the steps that text pending to be subjected to semantic understanding is received; the text pending to be subjected to semantic understanding is subjected to semantic understanding according to a deep residual network model constructed in advance, and a semantic understanding result is obtained. By means of the method, the text semantic understanding effect can be improved.
Owner:IFLYTEK CO LTD

Multi-round interactive semantic understanding method and device

The application proposes a multi-round interactive semantic understanding method and device. The multi-round interactive semantic understanding method comprises: receiving speech data; segmenting thespeech data to obtain a speech data segment, and performing speech recognition on the speech data segment to obtain current segmented text data; and performing semantic understanding on the current segmented text data according to the historical data of user interaction to obtain a semantic understanding result. The method can improve the accuracy of the semantic understanding result, so as to feed more accurate interactive results back to a user and improve the user experience.
Owner:IFLYTEK CO LTD

Robot perception and understanding method based on man-machine collaboration

ActiveCN107150347AQuick understandingImprove online awarenessManipulatorMan machinePattern perception
The invention provides a robot perception and understanding method based on man-machine collaboration. The method allows an operator to help robot to perceive and understand the environment. The method comprises the steps of firstly, audio visual mode perception natural interaction, secondly, target describing and understanding and thirdly task teaching and learning. Intelligence of the human is used, the most natural and efficient interaction manner of the human is used for man-machine communication and assistance, and the robot perception and understanding flexibility, intelligence and adaptation are improved.
Owner:GUANGZHOU LONGEST SCI & TECH

Understanding method and device for queries, equipment and storage medium

The embodiment of the invention discloses an understanding method and device for queries, equipment and a storage medium. The method comprises the steps that word vectors of all words contained in annotated queries are determined; hidden layer parameters, convolutional layer parameters and pooled layer parameters in a query CNN model obtained through training according to all URL site names and clicked queries and non-clicked queries of all the URL site names in advance are used as hidden layer parameters, convolutional layer parameters and pooled layer parameters in an initial domain recognition model; and according to domain annotations of the annotated queries and the word vectors of all the words contained in the annotated queries, the initial domain recognition model is trained to determine fully-connected layer parameters in the initial domain recognition model, and a domain recognition model is obtained. Through the scheme, model ability and generalization ability under a smallquantity of samples can be improved, the model is trained optimally, and the understanding effect of the queries is improved.
Owner:BEIJING BAIDU NETCOM SCI & TECH CO LTD

Multi-task chapter-level event extraction method based on multi-headed self-attention mechanism

The invention provides a multi-task chapter-level event extraction method based on a multi-headed self-attention mechanism. The method comprises the following steps: converting single sentence-level event extraction into chapter-level event extraction of a packaged sentence set; carrying out the word embedding representation by utilizing a pre-trained language model BERT; taking all word embedding and position embedding in a single sentence as input, employing a convolutional neural network model for coding, and capturing the most valuable features in the sentence in combination with a segmented maximum pool strategy; utilizing a multi-head self-attention model to obtain chapter representation and attention weight fused with full-text semantic information; utilizing a classifier to obtain a predicted event type; taking event types as prior information, linking the event types to an input sequence for event element extraction, and extracting all related elements in the sequence by using a pre-training model in combination with a machine reading understanding method. The method can be used for text-level event extraction tasks, and the breakthrough of converting a sequence labeling problem into a machine reading understanding problem is achieved.
Owner:HARBIN INST OF TECH AT WEIHAI +1

Method for understanding questions in question type automatic question-answer systems on basis of rule

The invention discloses a method for understanding questions in question type automatic question-answer systems on the basis of a rule. The method comprises the following steps of: regular expression-based question preprocessing: deleting useless texts such as polite formulas and the like, adding omitted punctuations, completing sentences, and carrying out basic natural language processing such as word segmentation, part of speech tagging and component syntax analysis on the completed sentences; designing a set of type-argument template-based question understanding intermediate forms, and applying a trigger word and a question template conversion rule summarized through manually pre-tagged data; and further verifying the generation result by utilizing a knowledge base so as to obtain a reasonable and complete question logic expression. According to the method, key points of questions can be extracted aiming at intermediate logic forms in question text design, so that proper input is reasonably and effectively provided for the question solution and then the question answering correctness is improved.
Owner:NANJING UNIV

Natural language question understanding method and device and electronic equipment

The invention provides a natural language question understanding method and device and electronic equipment, and relates to the technical field of natural language processing. The natural language question understanding method comprises the following steps of: obtaining natural language question information input by a user side, wherein the natural language question information is question information related to data query; analyzing the natural language question information to obtain a minimum analysis unit; generating a query instruction corresponding to the natural language question information on the basis of the minimum analysis unit and a preset instruction set; and carrying out retrieval in a preset knowledge base according to the query instruction so as to obtain a data result corresponding to the natural language question information, wherein the preset knowledge base is generated according to database data provided by a user, input information data of the user and / or third-party data. The method is capable of correctly recognizing natural language question information and matching data results with high correctness, and can be applied to professional scenes in the field of data analysis.
Owner:杭州汇数智通科技有限公司

Man-machine conversation understanding method and system for specific field and relevant equipment

The invention relates to the field of artificial intelligence, in particular to a man-machine conversation understanding method and system for a specific field and relevant equipment and aims to improve accuracy of conversation understanding. The conversation understanding method of the man-machine conversation system comprises the steps as follows: receiving a word currently input by a user and mapping the word to a vector space; expressing a historical word vector, semantic annotation information and intent category information as vectors by using a semantic representation layer; acquiring asemantic label of the current word by using the semantic labeling layer; acquiring the intent category of the current word by using an intent recognition layer. Extra part-of-speech information is introduced during model training, the part-of-speech prediction layer is used for predicting the part-of-speech of the next input word, semantic information shared by semantic annotations, intent recognition and part-of-speech prediction is fully used and promoted mutually by joint processing of the three tasks; the system and the method have clear logic, high efficiency and high accuracy, and the technical problem that the existing man-machine conversation system cannot effectively perform real-time conversation understanding is properly solved.
Owner:INST OF AUTOMATION CHINESE ACAD OF SCI

Method, device, equipment and storage medium for semantic understanding of dynamic human posture

ActiveCN108197589APrecise deliveryReduce the difficulty of human gesture recognitionCharacter and pattern recognitionSemanticsPosture recognition
The embodiment of the invention discloses a method, device, equipment and storage medium for semantic understanding of dynamic human posture. The method for semantic understanding of dynamic human posture comprises the following steps: the target person is identified from the acquired video streaming image frame; a non-mandatory image frame sequence is determined between adjacent command image frames as a semantic image frame sequence according to a command image frame in the video streaming, wherein the said command image frame is an image frame of the target person with command postures; thecorresponding semantics are identified according to the postures in the semantic image frame sequence. The embodiment of the invention is capable to adapt human posture recognition under a richer scene to improve the accuracy of recognition.
Owner:北京如布科技有限公司

Dialogue command understanding method based on knowledge graph

The invention provides a dialogue command understanding method based on a knowledge graph. The conversation command understanding method based on the knowledge graph includes: based on a preset domainknowledge set, correspondingly constructing a domain knowledge spectrum library; converting an input dialogue command into a candidate list with a plurality of candidate matching sequences; converting each candidate matching sequence in the candidate list into a knowledge graph vector through a preset coding mode; forming a dialogue command vector related to the input dialogue command according to the knowledge graph vector, training the deep learning neural network model based on the dialogue command vector, and executing dialogue command semantic understanding operation through the traineddeep learning neural network model.
Owner:BEIJING UNISOUND INFORMATION TECH

Video understanding method based on deep learning

The invention provides a video understanding method based on deep learning. The method comprises the steps that 1 a model based on an LSTM network is acquired through training; a C3D algorithm is usedto acquire image features; a PCA algorithm is used to reduce dimensions; the dimension of a feature vector is reduced from 4096 to 128; time-domain aliasing and normalization are carried out to acquire a normalized feature vector; an MSR-VTT database is used to train in the LSTM network to acquire the LSTM network model; and 2 through the LSTM network-based model, the statement information of a video image sequence to be detected is acquired; a C3D algorithm is used to acquire the feature vector of the video image sequence to be detected; a PCA algorithm is used for dimension reduction, and time domain aliasing and normalization are carried out to acquire a normalized feature vector; and through the LSTM network-based model, a statement output by the video image sequence to be detected isacquired. According to the invention, the accuracy of the existing model can be improved, and an original model can be further optimized based on new data.
Owner:TIANJIN UNIV

DLLE model-based data dimension reduction and characteristic understanding method

The invention discloses a DLLE (Linear Local Embedding of Difference) model-based data dimension reduction and characteristic understanding method, and belongs to the field of computer vision. The method comprises the steps of firstly, obtaining an image sequence through a visual sensor, then analyzing an input motion image sequence, extracting a foreground human body contour region through a background subtraction method, performing binarization, researching a periodic characteristic of a motion, performing key frame extraction on each motion sequence, and extracting a complete motion periodic sequence; performing manifold dimension reduction through a DLLE algorithm to obtain a low-dimensional eigenvector, and storing the low-dimensional eigenvector in a motion database; and performing identification through a nearest neighbor classifier by comparing a mean Hausdorff distance between a test sequence and a motion sequence in a training sample library. According to the method, the application of a differential function and category information-based neighborhood preserving embedding algorithm to human body motion identification is proposed; a DLLE model can not only keep a manifold local geometric structure during dimension reduction but also fully utilize category information of original high-dimensional data; and the extension from unsupervised extension to supervised extension is realized.
Owner:BEIJING UNIV OF TECH

Global semantic understanding method and device, computer device and storage medium

The embodiment of the invention discloses a global semantic understanding method and device, a computer device and a storage medium. The method comprises the following steps of acquiring a plurality of target texts in a display interface of a user terminal; detecting keywords in the target text according to the preset keyword list; querying the semantic comprehension results corresponding to the key words from the preset network resources; generating a synopsis of the key word and a network link according to the semantic understanding result, wherein the network link is one or more network page links obtained after the network query is carried out according to the key words. By acquiring the target text in the display interface of the user terminal, and then filtering out the key words inthe target text, querying the semantic understanding result corresponding to the key word in the preset network resources, and generating the semantic understanding result into the key word brief introduction and the network link to display to the user, the user can understand the key word in time and accurately, the query operation steps of the user are reduced, and the use is convenient and reliable.
Owner:西安欧思奇软件有限公司

Viewpoint type problem reading understanding method based on multilayer bidirectional LSTM and verification model

The invention discloses a viewpoint type problem reading understanding method based on a multilayer bidirectional LSTM and a verification model. According to the method, the advantages of an extraction type model and a judgment type model are fused. Based on reading understanding and feature learning, inference modeling is added, to obtain rational feature representation of a problem and an article, and the relationship between the candidate answers and the correct answers is used as a classification question in subsequent learning. A corresponding Loss function is designed, a verification model is designed for the situation that a part of questions cannot obtain correct answers in a given article, and compared with an existing method, the method is higher in accuracy when the viewpoint type problem reading understanding data set is processed.
Owner:海南中智信信息技术有限公司

Chinese spoken language semantic comprehension method and system

PendingCN110516253AReduce demandTraining time will not skyrocketSpecial data processing applicationsSpoken languageUser input
The embodiment of the invention provides a Chinese spoken language semantic comprehension method. The method comprises the steps of obtaining a generalized label-free text sequence training set, and performing forward prediction and reverse prediction on the training set in sequence to train a word-level and a word-level bidirectional language model; receiving spoken language voice audios input bya user, and carrying out sequence word segmentation to obtain character sequences and word sequences; decoding the character sequence and the word sequence by using the character-level bidirectionallanguage model and the word-level bidirectional language model respectively to obtain character-level implicit strata vectors and word-level implicit strata vectors; performing vector alignment on theimplicit strata vectors of the character sequence and the word sequence to obtain an implicit strata vector of spoken language voice audio input by the semantic comprehension model; and inputting thehidden layer vector of the spoken language voice audio into a semantic comprehension model, and determining the semantics of the spoken language voice audio. The embodiment of the invention further provides a Chinese spoken language semantic comprehension system. The embodiment of the invention has good generalization ability, combines word and character sequences, and improves the performance ofChinese semantic comprehension.
Owner:AISPEECH CO LTD

Video understanding method and device

The invention proposes a video understanding method and device, and the method comprises the steps: obtaining video data; carrying out the scene depth estimation of the obtained video data; and carrying out the understanding of scene contents based on the scene depth estimation result. According to the embodiment of the invention, the method achieves the analysis and understanding of the scene contents according to the depth information of the scene through the depth estimation of the scene, and can achieve the efficient and accurate recognition and location of an object in the video contents.
Owner:BEIJING KUANGSHI TECH +1

Object and scene image understanding method based on text-object-scene relations

The invention relates to an object and scene image understanding method based on text-object-scene relations. The method comprises steps as follows: information of three domains are integrated through a conditional random field, and the three domains include a scene, an object and a text; features of different domains are correlated through three relations which include a scene and object relation, a scene and text relation and an object and text relation; the text is related with the object and the scene in two different manners. The relation of the text and the object is represented through matching probability of noun and the object, the relation of the text and the scene is represented through occurrence probability of the noun in description of different scenes; the matching probability of the text and the object is obtained through solution of a problem about bidirectional matching optimization constrained under the weak supervision condition; the method only requires simple annotated information, the accuracy is improved, network resources can be sufficiently utilized for image understanding, and huge manual annotation cost is saved.
Owner:SHANGHAI JIAO TONG UNIV

Image understanding system based on layered temporal memory algorithm and image understanding method thereof

The invention relates to an image understanding (IU) system based on a layered temporal memory algorithm and an IU method thereof. The IU system consists of a layered temporal memory network training module, a layered temporal memory network database and an IU module. Besides, the IU method comprises the following steps: constructing a training image set P; training layered temporal memory networks by utilizing the training image set P; storing the trained layered temporal memory networks into the layered temporal memory network database; and carrying out understanding on a target image by utilizing all the layered temporal memory networks stored in the layered temporal memory network database. According to the invention, a novel time mode set learning method and a novel database technology are applied to rapidly convert several image content attributes into natural semantic description, thereby realizing deep understanding on the image. Compared with a traditional IU method, the provided IU method has advantages of simpleness, high practicability and flexibility.
Owner:SHANGHAI INST OF OPTICS & FINE MECHANICS CHINESE ACAD OF SCI

Video event processing method based on and-or graph

The invention discloses a video event processing method based on an and-or graph. The method includes defining a video event needed to be processed to be a motion combination participated by a single person or multiple persons and having connection, and defining node type, node distribution mode and relation among nodes of the and-or graph in the video event; on the basis of the defined node type, node distribution mode and relation among the nodes of the and-or graph, expressing the video event; using an inference method combining bottom-up and top-down to infer the occurrence process of an event in the video event. By the video event processing method, a commentary text of video content can be automatically generated according to an inference result of the video event, so that on one hand, a video editor or a video commentator can be helped to reduce working intensity in the field of video rebroadcast, and on the other hand, a good video content understanding method is provided for video retrieval related application.
Owner:BEIHANG UNIV

End-to-end unsupervised scene passable area cognition and understanding method

ActiveCN108876805AAddressed difficulty with traversable area labelingImprove applicabilityImage enhancementImage analysisData setNetwork architecture
The invention discloses an end-to-end unsupervised scene road surface area determination method. A road position prior probability distribution diagram is constructed; furthermore, the road position prior probability distribution diagram is used as characteristic mapping of a detection network, and directly added into a convolutional layer; a convolutional network framework fused with position prior characteristics is constructed; then, in combination with a fully convolutional network and a U-NET, a deep network architecture-UC-FCN network is constructed; a constructed passable area positionprior probability distribution diagram is used as characteristic pattern mapping of the deep network architecture-UC-FCN network; a UC-FCN-L network is generated; a passable area is detected based ona vanishing point detection method; furthermore, the obtained detection result is used as the truth value of a training dataset, and used for training the UC-FCN-L network; a deep network model used for extracting a passable area is obtained; the problem that the passable area is labelled difficultly can be solved; the applicability is high; steady operation in multiple road environments can be realized; furthermore, the real-time performance is better; the method in the invention is high in detection accuracy rate, and good in adaptation, real-time performance and robustness; and in addition,the method is simple and effective.
Owner:CHANGAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products