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

29957 results about "Network model" patented technology

The network model is a database model conceived as a flexible way of representing objects and their relationships. Its distinguishing feature is that the schema, viewed as a graph in which object types are nodes and relationship types are arcs, is not restricted to being a hierarchy or lattice.

System and method for automated placement or configuration of equipment for obtaining desired network performance objectives and for security, RF tags, and bandwidth provisioning

A method is presented for determining optimal or preferred configuration settings for wireless or wired network equipment in order to obtain a desirable level of network performance. A site-specific network model is used with adaptive processing to perform efficient design and on-going management of network performance. The invention iteratively determines overall network performance and cost, and further iterates equipment settings, locations and orientations. Real time control is between a site-specific Computer Aided Design (CAD) software application and the physical components of the network allows the invention to display, store, and iteratively adapt any network to constantly varying traffic and interference conditions. Alarms provide rapid adaptation of network parameters, and alerts and preprogrammed network shutdown actions may be taken autonomously. A wireless post-it note device and network allows massive data such as book contents or hard drive memory to be accessed within a room by a wide bandwidth reader device, and this can further be interconnected to the internet or Ethernet backbone in order to provide worldwide access and remote retrieval to wireless post-it devices.
Owner:EXTREME NETWORKS INC

Methods for testing biological network models

The present invention provides methods and systems for testing and confirming how well a network model represents a biological pathway in a biological system. The network model comprises a network of logical operators relating input cellular constituents (e.g., mRNA or protein abundances) to output classes of cellular constituents, which are affected by the pathway in the biological system. The methods of this invention provide, first, for choosing complete and efficient experiments for testing the network model which compare relative changes in the biological system in response to perturbations of the network. The methods also provide for determining an overall goodness of fit of the network model to biological system by: predicting from the network model how output classes behave in response to the chosen experiments, finding measures of relative change of cellular constituents actually observed in the chosen experiments, finding goodnesses of fit of each observed cellular constituent to an output class with which the cellular constituent has the strongest correlation, and determining an overall goodness of fit of the network model from the individual goodnesses of fit of each observed cellular constituent. Additionally, these methods provide for testing the significance of the overall goodness of fit according to a nonparametric statistical test using an empirically determined distribution of possible goodnesses of fit. This invention also provides for computer systems for carrying out the computational steps of these methods.
Owner:MICROSOFT TECH LICENSING LLC

Small target detection method based on feature fusion and depth learning

InactiveCN109344821AScalingRich information featuresCharacter and pattern recognitionNetwork modelFeature fusion
The invention discloses a small target detection method based on feature fusion and depth learning, which solves the problems of poor detection accuracy and real-time performance for small targets. The implementation scheme is as follows: extracting high-resolution feature map through deeper and better network model of ResNet 101; extracting Five successively reduced low resolution feature maps from the auxiliary convolution layer to expand the scale of feature maps. Obtaining The multi-scale feature map by the feature pyramid network. In the structure of feature pyramid network, adopting deconvolution to fuse the feature map information of high-level semantic layer and the feature map information of shallow layer; performing Target prediction using feature maps with different scales and fusion characteristics; adopting A non-maximum value to suppress the scores of multiple predicted borders and categories, so as to obtain the border position and category information of the final target. The invention has the advantages of ensuring high precision of small target detection under the requirement of ensuring real-time detection, can quickly and accurately detect small targets in images, and can be used for real-time detection of targets in aerial photographs of unmanned aerial vehicles.
Owner:XIDIAN UNIV

Image semantic division method based on depth full convolution network and condition random field

The invention provides an image semantic division method based on a depth full convolution network and a condition random field. The image semantic division method comprises the following steps: establishing a depth full convolution semantic division network model; carrying out structured prediction based on a pixel label of a full connection condition random field, and carrying out model training, parameter learning and image semantic division. According to the image semantic division method provided by the invention, expansion convolution and a spatial pyramid pooling module are introduced into the depth full convolution network, and a label predication pattern output by the depth full convolution network is further revised by utilizing the condition random field; the expansion convolution is used for enlarging a receptive field and ensures that the resolution ratio of a feature pattern is not changed; the spatial pyramid pooling module is used for extracting contextual features of different scale regions from a convolution local feature pattern, and a mutual relation between different objects and connection between the objects and features of regions with different scales are provided for the label predication; the full connection condition random field is used for further optimizing the pixel label according to feature similarity of pixel strength and positions, so that a semantic division pattern with a high resolution ratio, an accurate boundary and good space continuity is generated.
Owner:CHONGQING UNIV OF TECH

Network models of complex systems

InactiveUS20050171746A1Simulator controlData visualisationComplex dynamic systemsNetwork model
This invention describes computer based virtual models of complex systems, together with integrated systems and methods providing a development and execution framework for visual modeling and dynamic simulation of said models. The virtual models can be used for analysis, monitoring, or control of the operation of the complex systems modeled, as well as for information retrieval. More particularly, the virtual models in the present implementation relate to biological complex systems. In the current implementation the virtual models comprise building blocks representing physical, chemical, or biological processes, the pools of entities that participate in those processes, a hierarchy of compartments representing time-intervals or the spatial and/or functional structure of the complex system in which said entities are located and said processes take place, and the description of the composition of those entities. The building blocks encapsulate in different layers the information, data, and mathematical models that characterize and define each virtual model, and a plurality of methods is associated with their components. The models are built by linking instances of the building blocks in a predefined way, which, when integrated by the methods provided in this invention, result in multidimensional networks of pathways. A number of functions and graphical interfaces can be selected for said instances of building blocks, to extract in various forms the information contained in said models. Those functions include: a) on-the-fly creation of displays of interactive multidimensional networks of pathways, according to user selections; b) dynamic quantitative simulations of selected networks; and c) complex predefined queries based on the relative position of pools of entities in the pathways, the role that the pools play in different processes, the location in selected compartments, and/or the structural components of the entities of those pools. The system integrates inferential control with quantitative and scaled simulation methods, and provides a variety of alternatives to deal with complex dynamic systems and with incomplete and constantly evolving information and data.
Owner:INTERTECH VENTURES

Method and system for integrated reservoir and surface facility networks simulations

Integrated surface-subsurface modeling has been shown to have a critical impact on field development and optimization. Integrated models are often necessary to analyze properly the pressure interaction between a reservoir and a constrained surface facility network, or to predict the behavior of several fields, which may have different fluid compositions, sharing a common surface facility. The latter is gaining a tremendous significance in recent deepwater field development. These applications require an integrated solution with the following capabilities: * to balance a surface network model with a reservoir simulation model in a robust and efficient manner. * To couple multiple reservoir models, production and injection networks, synchronising their advancement through time. * To allow the reservoir and surface network models to use their own independent fluid descriptions (black oil or compositional descriptions with differing sets of pseudo-components). * To apply global production and injection constraints to the coupled system (including the transfer of re-injection fluids between reservoirs). In this paper we describe a general-purpose multi-platform reservoir and network coupling controller having all the above features. The controller communicates with a selection of reservoir simulators and surface network simulators via an open message-passing interface. It manages the balancing of the reservoirs and surface networks, and synchronizes their advancement through time. The controller also applies the global production and injection constraints, and converts the hydrocarbon fluid streams between the different sets of pseudo-components used in the simulation models. The controller's coupling and synchronization algorithms are described, and example applications are provided. The flexibility of the controller's open interface makes it possible to plug in further modules (to perform optimization, for example) and additional simulators.
Owner:SCHLUMBERGER TECH CORP

Unsupervised domain-adaptive brain tumor semantic segmentation method based on deep adversarial learning

The invention provides an unsupervised domain-adaptive brain tumor semantic segmentation method based on deep adversarial learning. The method comprises the steps of deep coding-decoding full-convolution network segmentation system model setup, domain discriminator network model setup, segmentation system pre-training and parameter optimization, adversarial training and target domain feature extractor parameter optimization and target domain MRI brain tumor automatic semantic segmentation. According to the method, high-level semantic features and low-level detailed features are utilized to jointly predict pixel tags by the adoption of a deep coding-decoding full-convolution network modeling segmentation system, a domain discriminator network is adopted to guide a segmentation model to learn domain-invariable features and a strong generalization segmentation function through adversarial learning, a data distribution difference between a source domain and a target domain is minimized indirectly, and a learned segmentation system has the same segmentation precision in the target domain as in the source domain. Therefore, the cross-domain generalization performance of the MRI brain tumor full-automatic semantic segmentation method is improved, and unsupervised cross-domain adaptive MRI brain tumor precise segmentation is realized.
Owner:CHONGQING UNIV OF TECH

Compressed low-resolution image restoration method based on combined deep network

The present invention provides a compressed low-resolution image restoration method based on a combined deep network, belonging to the digital image / video signal processing field. The compressed low-resolution image restoration method based on the combined deep network starts from the aspect of the coprocessing of the compression artifact and downsampling factors to complete the restoration of a degraded image with the random combination of the compression artifact and the low resolution; the network provided by the invention comprises 28 convolution layers to establish a leptosomatic network structure, according to the idea of transfer learning, a model trained in advance employs a fine tuning mode to complete the training convergence of a greatly deep network so as to solve the problems of vanishing gradients and gradient explosion; the compressed low-resolution image restoration method completes the setting of the network model parameters through feature visualization, and the relation of the end-to-end learning degeneration feature and the ideal features omits the preprocessing and postprocessing; and finally, three important fusions are completed, namely the fusion of the feature figures with the same size, the fusion of residual images and the fusion of the high-frequency information and the high-frequency initial estimation figure, and the compressed low-resolution image restoration method can solve the super-resolution restoration problem of the low-resolution image with the compression artifact.
Owner:BEIJING UNIV OF TECH

Traffic prediction method based on attention temporal graph convolutional network

The invention belongs to the field of intelligent transportation, and discloses a traffic prediction method based on an attention temporal graph convolutional network. The method includes the following steps that: firstly, an urban road network is modeled as a graph structure, nodes of the graph represent road sections, edges are connection relationships between the road sections, and the time series of each road section is described as attribute characteristics of the nodes; secondly, the temporal and spatial characteristics of the traffic flow are captured by using an attention temporal graph convolutional network model, the temporal variation trend of the traffic flow on urban roads is learned by using gated cycle units to capture the time dependence, and the global temporal variation trend of the traffic flow is learned by using an attention mechanism; and then, the traffic flow state at different times on each road section is obtained by using a fully connected layer; and finally,different evaluation indexes are used to estimate the difference between the real value and the predicted value of the traffic flow on the urban roads and further estimate the prediction ability of the model. Experiments prove that the method provided by the invention can effectively realize tasks of predicting the traffic flow on the urban roads.
Owner:CENT SOUTH UNIV

Unmanned aerial vehicle patrol detection image power small component identification method and system based on Faster R-CNN

The invention discloses an unmanned aerial vehicle patrol detection image power small component identification method and system based on Faster R-CNN. The method comprises the following steps: carrying out pre-training on a ZFnet model, and extracting a feature graph of an unmanned aerial vehicle patrol detection image; training an RPN region proposed network model obtained through initialization to obtain a region extraction network, generating a candidate region frame on the feature graph of the image by utilizing the region extraction network, and carrying out feature extraction on the candidate region frame to extract position features and in-depth features of a target; carrying out training on a Faster R-CNN detection network obtained after initialization by utilizing the position features and in-depth features of the target and the feature graph to obtain a power small component detection model; and carrying out actual power small component identification detection by utilizing the power small component detection model. The beneficial effects are that Faster R-CNN is utilized to realize identification and positioning of a plurality of types of power small components, so that identification speed of about 80 ms per picture and 92.7% accuracy can be achieved.
Owner:STATE GRID INTELLIGENCE TECH CO LTD

Cascaded neural network-based face key point detection method

The invention relates to a cascaded neural network-based face key point detection method. The method includes the following steps that: a) a training-used face image set is established, and a key point position requiring detection is marked; b) a first-layer depth neural network is constructed and is used to train a face region estimation model; c) a second-layer depth neural network is constructed and is used to perform face key point preliminary detection; d) local region division is continued to be performed on an inner face region; e) a third-layer depth neural network is constructed for each local region respectively; f) the rotation angle of each local region is estimated; g) correction is performed according to the estimated rotation angles; h) a fourth-layer depth neural network is constructed for the correction data set of each local region; and i) any face image is given, and the above four-layer depth neural network model is adopted to perform key point detection, such that final face key point detection results can be obtained. With the cascaded neural network-based face key point detection method of the invention adopted, face key point detection can be improved, and especially the accuracy and real-time property of dense face key point detection.
Owner:BEIJING KUANGSHI TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products