252 results about "Structure recognition" patented technology

Described herein are systems and methods for performing multi-stage detection and classification of cancer regions from digitized images of biopsy slides. Novel methods for processing the digitized images to improve feature extraction and structure identification are disclosed, including but not limited to the use of quaternions, logarithmic mappings of color channels, and application of wavelets to logarithmic color channel mappings. The extracted features are utilized in improved machine learning algorithms that are further optimized to analyze multiple color channels in multiple dimensions. The improved machine learning algorithms include techniques for accelerating the training of the algorithms, making their application to biopsy detection and classification practical for the first time. The performance of the described systems and methods are further improved by the disclosure of a novel multistage machine learning scheme, in which additional classifiers are utilized to choose among the classes proposed by other classifiers in close cases.

The invention discloses an indoor unknown structure identification method based on laser scanning. The method comprises the following steps of designing a data structure of laser scanning data, and determining a global map coordinate system used in the scanning process; adopting a laser rangefinder to obtain laser scanning data, calculating the distance between scanning points with adjacent angles, and performing denoising pretreatment, generating a scanning contour of the indoor structure; saving the global coordinates of various points in the scanning contour into a database, and establishing a three-dimensional environmental point cloud by combining height sensor information; and according to the scanning contour and the environmental point cloud, identifying an indoor unknown environmental structure and barrier information. The method provided by the invention is complete in structure and real-time in data processing, and has strong adaptability and expandability.

The present invention discloses a residential house chart building structure recognition and functional area planning and designing method and system. The method comprises a first step of acquiring a residential unit vector diagram, wherein the residential vector diagram comprises walls and doors and windows, and walls and doors and windows form closed rooms; a second step of acquiring specific information about usage, walls and windows and doors of each closed room; and a third step of selecting a reference wall, traversing all walls, and recognizing building structures and planning and designing the functional areas in the closed rooms sequentially according to feature models of the residential building structure and the functional area structure, so as to form a residential vector diagram with all functional area names. According to the method, the functional areas beneficial to living are divided optimally by combining the feature models of the residential building structure and the functional area structure only according to the residential unit diagram or the residential unit vector diagram. The divided optimal functional areas are a reference for users or are used in household design later, and deliver a good effect.

A method and process provides structure recognition to a node-link diagram formed by a plurality of digital ink strokes drawn on a canvas of an electronic device by an input device. The method and process include grouping related strokes into multiple hypotheses reflecting structure interpretations of the strokes in a group. Confidence scores are computed for the multiple hypotheses based on local evidence regarding the strokes of the stroke groups and surrounding strokes. Constraints are applied among the hypotheses, and a collectively high scoring assignment of accept/reject values of the hypotheses are determined under the applied constraints. The hypotheses with collectively high scoring assignments are accepted as a representation of the node-link diagram, where structure information is provided to the strokes of the node-link diagram, making the node-link diagram electronically editable.

The invention discloses a spatial database synthesis method in land utilization information multi-scale expression; a spatial database in the land utilization information multi-scale expression is synthetically divided into two steps: semantics synthesis and graphics synthesis, wherein, the process of the semantics synthesis is synthesis constraint based on a thematic data base, under the control and guidance of knowledge inference, the geographic characteristics of land features in the database is judged, comprising the steps of identifying the spatial relation of land feature types, land feature grades, spatial position and other land features and determining the importance of the land features, therefore, the process is structural identification. The graphics synthesis is the synthesis of point features, line features and surface features under the guidance of cartography theory. The synthesis model based on the database is built, the standardization of database synthesis is realized, and the disadvantages that the traditional cartography synthesis only concerns graphics synthesis and neglects special semantics synthesis are overcome.

Specification is received regarding which structure identifying information, each corresponding to an anatomical structure, is specified as display target structure identifying information to be displayed with a display target image when displaying the display target image, in which the positions of each pixel correspond to positions within an original image. The position of the anatomical structure identified by the structure identifying information is specified, based on structure information, in which structure identifying information that identifies anatomical structures and structure position information that represents the positions of the anatomical structures within the original image are correlated, and the correspondence of positions between the original image and the display target image. The display target structure identifying information is displayed with the display target image such that a user can recognize that the anatomical structure identified by the display target structure identifying information is displayed at the specified position.

Systems and methods may include an application program interface that enables a user to: specify parameters associated with an evolutionary algorithm, where an execution of the evolutionary algorithm is in accordance with the specified parameters; define a chromosome data structure that includes a plurality of variables that are permitted to evolve in value in accordance with the execution of the evolutionary algorithm in order to generate one or more child chromosome data structures; identify one or more objective functions for evaluating chromosome data structures, including the generated one or more child chromosome data structures; and define an output format for providing one or more optimal chromosome data structures of the evaluated generated child chromosome data structures as designs to the identified objective functions.

The invention relates to methods for determining a logical structure of a document. The system stores a collection of models, each of which describes one or more possible logical structures. At least one document hypothesis is generated for the whole document. For each document hypothesis, the system verifies the document hypothesis on each page, for example, by generating at least one block hypothesis for each block in the document based on the document hypothesis, selecting a best block hypothesis for each block, selecting the model that corresponds to a best document hypothesis the document hypothesis that has a best degree of correspondence with the selected best block hypotheses for the document, and forming a representation of the document based on the best document hypothesis described.

The invention provides a table recognition method fusing multiple text features and geometrical information. The method comprises: a data processing step of obtaining a picture of a table area, and respectively carrying out OCR and straight line recognition on the obtained picture to obtain key feature information; a graph convolutional neural network training step of performing graph convolutional neural network training according to the obtained key feature information, and constructing a table structure recognition model; and a table identification step of performing structure identification on the table in the picture format according to the constructed table structure identification model. The invention provides thea table recognition method fusing the multiple text features and the geometrical information. Improvements are made in the aspects of diversity of adopted data, a method for performing feature extraction on the data and the like, the accuracy of table recognition is effectively improved, a more accurate table structure reconstruction result is obtained, and compared with an existing table recognition mechanism based on traditional rules and a traditional deep learning method based on pictures, the table recognition method has the advantage that the effect is greatly improved.

The present invention relates to a structure recognition based Web table information extraction method. According to the method, table structure recognition is performed in two progressive ways. Firstly, a set of heuristic rules is used to determine structures of several common types of the tables, so as to define structure types of most Web tables; and tables not recognized by using the heuristic rules are processed by using a string matching method, and a matched cell is limited in a line or a column in which a ULC (upper-left-cell) is, so that content on which string matching needs to be performed is significantly reduced, thereby improving matching and recognition efficiency. Finally, in terms of two-dimensional tables, processing strategies of synthesizing cells in an information extraction process is proposed, which can reduce generated redundant data while ensuring that relationships between data in an extraction result are not damaged.

The invention discloses a sensor layout method suitable for structure health monitoring and a structure identification method. The sensor layout method comprises the following steps of through a finite element method, acquiring strain curves of a structure under various conditions; acquiring standard curves under the various conditions; carrying out linear superposition processing on the standard curves under the various conditions to acquire a characteristic curve; picking an inflection point, a concave point and a convex point on the characteristic curve as characteristic points; selecting new characteristic points till that all the units are selected as the characteristic points and acquiring a curve of an average error and a characteristic point number; and selecting a point with a maximum slope change on the acquired curve of the average error and the characteristic point number and determining an arrangement scheme of a sensor. In the invention, from two aspects of sensor optimization arrangement and area distribution sensing, a traditional problem that deflection of a civil engineering structure is difficult to be accurately measured is solved; and the limited sensors are used to realize effective monitoring and performance assessment of a large-scale civil engineering structure.

The invention discloses a three-dimensional reconstruction transformer substation implementation method based on a virtual simulation technology. According to the method, management is carried out based on millimeter-scale three-dimensional reconstruction of a multi-source modeling transformer substation and data association of a GIS three-dimensional model and equipment attributes; three-dimensional scanning is carried out on the whole transformer substation to obtain a whole structure; scanning details of GIS equipment to obtain a three-dimensional point cloud data structure and three-dimensional morphological characteristics of electrical equipment; image acquisition is made on the texture of the equipment; and finally, the structure and texture of the equipment are acquired in the difficult-to-scan area through panoramic images., which is used for structural recognition of a later-stage device.Simulation modeling is carried out by obtaining GIS equipment information, and information association is carried out on a system data interface and corresponding sensor data, so that visual display and operation can be carried out on the corresponding sensor information by selecting a model in the system, and equipment management work of places such as a transformer substation is greatly facilitated.

The invention relates to the artificial intelligence technology, and discloses a table structure identification method comprising the following steps: obtaining a to-be-identified image, and carryingout table area extraction and correction on the to-be-identified image based on semantic segmentation so as to obtain a table area graph; performing table frame line segmentation on the table area graph by using a pre-constructed segmentation model to obtain a table frame graph; performing image connected domain analysis on the table framework diagram to obtain table structure information; and performing text detection and identification on the table area graph to obtain a text detection result, and generating a spreadsheet document corresponding to the to-be-identified picture according to the text detection result and the table structure information. In addition, the invention also relates to a blockchain technology, and the to-be-identified image can be stored in a node of a blockchain.The invention further provides a table structure recognition device, electronic equipment and a computer readable storage medium. The accuracy of table structure identification can be improved.

The invention discloses an underground deep-buried meander point dam sand body historical reconstruction method. A preliminary-stage riverway flow line and terminal-stage riverway flow line enveloping method is adopted, and combined with a sand body thickness diagram, a point dam is recognized; through extraction of the point dam lateral deposition body scale, the tendency, the inclination and lateral deposition body spacing parameter in a research region, a point dam building structure recognition key parameter database is built, through a skeleton section, fine dissection is conducted on the meander riverway sand body of a stratigraphic unit, riverway flow lines are obtained, then the riverway flow lines are sequentially recovered and arranged according to the development sequence, and the riverway sedimentary evolution history is represented. The underground deep-buried meander point dam sand body historical reconstruction method has the advantages that the sand body building structure dissection result for riverway historical process rebuilding is made to be more reasonable and closer to the actual underground situation.

The invention relates to a PDF table structure recognition method based on a graph attention mechanism, and belongs to the technical field of document analysis in a data mining technology. The methodcomprises the following steps of 1, preprocessing, wherein all cells in a table and position coordinates of the cells are obtained; 2, graph construction: establishing an undirected graph for the obtained cells; and 3, relationship prediction: classifying the edges on the constructed undirected graph, and predicting the adjacency relationship between the cells by using a neural network model. Compared with the prior art, the method for identifying the complex table structure in the PDF is proposed for the first time, the best effect is achieved on two table structure identification data sets,and particularly, the effect is obviously improved on complex table structure identification.

The invention discloses a prediction method for disease dynamic states on the basis of a network structure. The prediction method comprises the following steps: firstly, analyzing a great quantity of spatio-temporal data collected by mobile equipment carried by a volunteer in an epidemic situation region to evaluate influence of the network structure on the disease dynamic states; adopting three features including a clustering coefficient, degree distribution and degree correlation to describe the network structure; then, on the basis of an evaluation result, designing a structure identification model to identify a dynamically changed network structure; and finally, according to a relationship model of disease dynamic property and the network structure, predicting the dynamic change of a disease. According to the method, disease prediction accuracy can be effectively improved, and the method has an important significance for infectious disease control and prevention.

The invention provides an intelligent area recognition method based on a wireless mesh network, and relates to an area recognition method. At present, an intelligent area generally takes an intelligent distribution transformer terminal as an information data center, and the updating timeliness is poor. According to the technical scheme, the method comprises the steps that when slave equipment is powered on, the slave equipment actively registers from a network server to acquire an IPv6 address, and the network server synchronizes the IPv6 address and an equipment ID of the slave equipment to amain station; main equipment sends a query command to the slave equipment according to received slave equipment archive information; the main equipment actively reports the detailed archive information to the main station to form a list and sends the white list to the distribution transformer terminal main equipment of the corresponding area; and the distribution transformer terminal starts equipment and line topology structure recognition of the area according to the received white list and synchronously sends information returned after recognition to the main station, and an equipment and line topology graph of each area is formed on the main station. According to the technical scheme, the updating timeliness and convenience of wireless slave equipment archives and topology structures of the areas are improved, and install and play of the wireless slave equipment is achieved.

Some embodiments of the invention provide a method for pre-tabulating sub-networks. This method (1) generates a sub-network that performs a function, (2) generates a parameter based on this function, and (3) stores the sub-network in a storage structure based on the generated parameter. In some embodiments, the generated sub-network has several circuit elements. Also, in some embodiments, the generated sub-network performs a set of two or more functions. Some embodiments store each generated sub-network in an encoded manner. Some embodiments provide a method for producing a circuit description of a design. This method (1) selects a candidate sub-network from the design, (2) identifies an output function performed by the sub-network, (3) based on the identified output function, identifies a replacement sub-network from a storage structure that stores replacement sub-networks, and (4) replaces the selected candidate sub-network with the identified replacement sub-network in certain conditions. In some embodiments, this method is performed to map a design to a particular technology library. Some embodiments provide a data storage structure that stores a plurality of sub-networks based on parameters derived from the output functions of the sub-networks.

Some embodiments of the invention provide a method for pre-tabulating sub-networks. This method (1) generates a sub-network that performs a function, (2) generates a parameter based on this function, and (3) stores the sub-network in a storage structure based on the generated parameter. In some embodiments, the generated sub-network has several circuit elements. Also, in some embodiments, the generated sub-network performs a set of two or more functions. Some embodiments store each generated sub-network in an encoded manner. Some embodiments provide a method for producing a circuit description of a design. This method (1) selects a candidate sub-network from the design, (2) identifies an output function performed by the sub-network, (3) based on the identified output function, identifies a replacement sub-network from a storage structure that stores replacement sub-networks, and (4) replaces the selected candidate sub-network with the identified replacement sub-network in certain conditions. In some embodiments, this method is performed to map a design to a particular technology library. Some embodiments provide a data storage structure that stores a plurality of sub-networks based on parameters derived from the output functions of the sub-networks.