109results about How to "Mapping implementation" patented technology

The invention discloses an asymmetrical voice conversion method based on deep neural network feature mapping, and belongs to the technical field of voice conversion. The asymmetrical voice conversion method based on deep neural network feature mapping disclosed by the invention is specific to asymmetrical data of source voice and target voice. The method comprises the following steps: firstly, performing probability modeling by using the pre-training function of a deep network, and extracting high-order statistics features in a voice signal to provide a standby preferred space of network coefficients; secondly, performing incremental learning by using a small quantity of asymmetrical data, and correcting network weight coefficients by using an optimized transmission error to realize mapping of feature parameters. According to the asymmetrical voice conversion method, a network coefficient structure is optimized and is taken as a parameter initial value of a deep forward prediction network, network structure parameters are further transmitted reversely and optimized in the incremental learning process of a small quantity of asymmetrical data, so that mapping of the personal feature parameters of a speaker is realized.

The invention designs a 5G wireless network virtualization system structure based on calculation and communication fusion. The system structure gets through a resource barrier of a heterogeneous wireless network, shares an access network resource, breaks through a traditional honeycomb cell structure, and supports separation of a control plane from a data plane. Communication, calculation and storage multi-dimensional resources are considered to perform joint optimization, a network taking a user as a center is established, and a development need of a 5G system is finally adapted. The system structure is divided into three parts, i.e., a wireless physical network, a wireless virtual network and a core network. The wireless physical network part comprises infrastructure, resource pools and a physical network controller; and network resources are divided and stored according to a wireless resource, a computing resource and a storing resource, are placed into the corresponding resource pools, and are managed and controlled in a unified manner by the physical network controller. The wireless virtual network part comprises a plurality of virtual subnets and a virtual network controller, and the virtual network controller is responsible for coordination with the physical network controller and allocates the virtual subnets according to user needs.

The present invention provides a memory data quick persistence method based on a storage-class memory. The method is based on a flat mixed memory architecture, and is implemented by collaborative design of a user layer and a kernel layer. The method comprises: abstracting an SCM with a certain capacity and data stored in the SCM into a persistent area; designing an application programming interface (API) at the user layer, and making a response to memory access of an application to the persistent area; extending a Buddy system of a kernel to implement heterogeneous mixed memory management; and designing a persistent area manager to implement the functions of persistent mapping in the persistent area, metadata management in the persistent area and the like. According to the memory data quick persistence method based on the storage-class memory, a data linearization process required for data persistence in a traditional storage architecture can be avoided, direct access and in-place update of persistent data can be implemented, hierarchy invoking of a software stack in the traditional architecture is simplified, the I/O bottleneck can be effectively alleviated, and the memory access performance of the persistent data can be improved.

The invention provides a method and device for acquiring a facial expression of a 3D face model and a storage device. The method comprises the steps of acquiring a 3D face model; acquiring a motion trajectory of a facial feature point by using an RGBD image sequence; and mapping the motion trajectory of the facial feature point to the 3D face model so as to enable a motion trajectory of a facial feature point of the 3D face model to be matched with the acquired motion trajectory of the facial feature point. The device comprises a processor, wherein the processor implements the method. The storage device stores program data which can be executed to implement the above method. The facial expression information acquired according to the invention is more comprehensive and accurate, so that the 3D face model can be enabled to reflect the facial expression of the human face more vividly.

The invention relates to a method and a device for mapping resources. The method comprises the following steps of: separately interlacing a virtual resource block on each carrier in a plurality of carriers, interlacing between the carriers by using a first and a second virtual resource blocks of each carrier in the carriers acquired by interlacing as an integrate body, and mapping the rearranged virtual resource blocks onto a physical resource block after the interlacing is finished. The method and the device of the invention can map resources in an LTE-A system under a big band width condition, increases the frequency diversity, decreases the influence on the LTE system by partially using a downlink distribution type resource mapping mode of the current LTE, and ensures the system compatibility.

The invention discloses a method and a device for mapping resources of a physical hybrid retransmission indicator channel. The method comprises the following steps: determining the position of PHICH resources of a first code stream based on a dynamic cyclic shift parameter of a DMRS in a PHICH resource mapping formula corresponding to the first code stream, wherein the dynamic cyclic shift parameter is the signaling value of the de-modulation reference signal (DMRS) in uplink scheduling/downlink control information (DCI); indirectly indicating the position of PHICH resources of a second code stream by using the position of PHICH resources of the first code stream; and carrying out resource mapping on the position of PHICH resources of the first code stream and the position of PHICH resources of the second code stream. The invention effectively realizes the mapping of PHICH resources under UL MIMO scenarios.

The invention provides a residual oil distribution prediction method and device based on a deep convolutional neural network, and belongs to the technical field of oil development. The prediction method comprises the steps of simulating and establishing a training data set according to geological parameters and development parameters of an oil reservoir of a type to which a target oil reservoir belongs; training by using the training data set to obtain a deep full convolutional coding and decoding neural network prediction model; and predicting the remaining oil distribution of the target oil reservoir at any moment by adopting the deep full convolutional coding and decoding neural network prediction model. According to the prediction method, through geological parameters and development parameters of a target oil reservoir, factors influencing remaining oil distribution and influences of time sequences are considered, different numerical simulation schemes are set for the oil reservoir of the type, and a data set is established through oil reservoir numerical simulation; and then a deep full convolutional coding and decoding neural network is used as a framework for training, so that rapid and accurate prediction of remaining oil distribution of the oil reservoir at any moment is realized, and guidance is provided for efficient development of the oil reservoir.

The invention discloses an invulnerability mapping method of a multicast virtual network. The invulnerability mapping method of the multicast virtual network comprises the steps that leaf node sets capable of being covered with each bottom layer data center are obtained according to the zone requirements of each bottom layer data center where each leaf node in the multicast virtual network has access, a set scheme that all the leaf nodes are covered with the sets in the minimum number is obtained by the adoption of the minimum set covering algorithm, mapping of the leaf nodes is conducted through the corresponding bottom layer data centers, and the leaf nodes covered with the same bottom layer data center share node protection resources; the bottom layer data centers where root nodes have access serve as standby bottom layer data centers, primary link mapping is conducted on each standby bottom layer data center, the leaf nodes covered with the same bottom layer data center share a protecting route, and after link mapping is completed, a link mapping scheme with the smallest total resource consumption is selected from the link mapping schemes corresponding to the standby bottom layer data centers to serve as the optimal link mapping scheme. According to the invulnerability mapping method of the multicast virtual network, invulnerability mapping with low cost is achieved.

The present invention provides a reference signal mapping method and device. The method comprises the steps of determining the RE positions where the channel state indication reference signals (CSI-RS) are mapped according to N-port reference signal patterns, and carrying out the resource mapping on the CSI-RS according to the determined RE positions, wherein N is equal to 12 or 16, and the RE positions wherein the CSI-RS are mapped in the N-port reference signal patterns are determined according to the RE positions wherein the CSI-RS are mapped in one or more of the 2-port reference signal patterns, the 4-port reference signal patterns and the 8-port reference signal patterns. According to the present invention, the 12-port or 16-port CSI-RS mapping can be realized, and further, the12-port or 16-port CSI-RS transmission is realized.

The embodiment of the invention discloses a multi-core online patching method and a multi-core online patching device. By the method and the device, all patch data can be mapped into a patch area of a shared memory. The method of the embodiment of the invention comprises the following steps of: separating a shared global variable from a private global variable in a patch; and mapping the shared global variable into a shared data segment of the patch area in a direct memory address mapping way, and mapping the private global variable into a private data segment of the patch area in a user-specified variable address mapping way. The embodiment of the invention can be used for a multi-core digital signal processor (DSP) system of carrier-class software.

The invention discloses a mapping method, which is used in a next generation network field. The method comprises the steps as follows: the resource the service data flow is distributed by a TRC-FE according to the resource information of a MPLS LSP, a PE-FE updating flow label mapping table and other information are announced by the TRC-FE through an Rc interface; the update of a FTN table is finished by the PE-FE, and the service data flow is mapped to a named LSP tunnel. With the Rc interface and the TRC-FE announcing the PE-FE the information of mapping the data flow to the named LSP tunnel, the invention realizes the mapping between the service data flow and the MPLS LSP tunnel. So the problem that in the prior standard the service data flow of NGN network which based on MPLS-TE and the mapping of LSP tunnel resource is incapable of being realized is solved.

The invention relates to a bidirectional mapping queue researching information system based on an index storage model. According to the system, a no-structure data storage mode is utilized; an index model and an index database are established for indexes in data acquisition; a basic acquisition template which can be modified by a user is supplied; corresponding analysis data are supplied accordingto the user requirement; a data sequence with time sequence can be supplied to the same index; and the system can directly perform modeling on user data for analysis without database disposition by an administrator. The bidirectional mapping queue researching information system has advantages of simplifying the operation, saving manpower and physical resources, and maximally ensuring a fact thata requirement is satisfied by acquisition.

The invention relates to the technical field of communication and discloses a transmission method and a device for an E-PDCCH. The E-PDCCH comprises at least one E-PDCCH unit. The method comprises that the at least one E-PDCCH unit is mapped to preset positions centered by at least one resource block, wherein, the sizes of centered resources of each resource block are larger than or equal to those of resources occupied by two E-PDCCH unit, and resources occupied by the centering of each E-PDCCH unit in the resource block comprise part of resources of each resource block centered by the resource block; and the E-PDCCH is transmitted to time-frequency resources centered and mapped by the resource block. According to the transmission method and the device for the E-PDCCH, good enhanced PDCCH performances can be obtained under the situation that the time-frequency resources are fully used.

The invention discloses an integralheart three-dimensional mapping system for complex arrhythmias. The system comprises an epicardial three-dimensional mapping module, an endocardial three-dimensional magnetic localization mapping module, an endocardial epicardial mapping module and a three dimensional mapping common basic module, wherein, the three dimensional mapping common basic module provides the construction of the three dimensional anatomical model of heart and the fusion between electrophysiological information and the anatomical model; the endocardial three-dimensional magnetic localization mapping module is embedded with magnetic field generation and localization hardware to provide the magnetic field localization for the whole system. The epicardial mapping module is embedded with epicardial segmentation software; the endocardial epicardial mapping module is embedded with endocardial epicardial combined mapping algorithms to achieve the integral heart mapping. The integralheart three-dimensional mapping system for complex arrhythmias can organically fuse the three-dimensional electro-anatomical information mapped collectively by the endocardium and the corresponding epicardium to achieve the accurate mapping in the situations of both the fixed heart shape and position and the dynamic conditions of chest opening, and therefore the requirement of clinical diagnosis and treatment and the mechanism research of the arrhythmias can be satisfied.

The application relates to a container persistent-storage mounting method, and belongs to the technical field of computers. The method includes: receiving a container creation request, wherein the container creation request includes a container identifier of a to-be-created container; reading a pre-stored configuration file according to the container creation request, wherein the configuration file includes a volume directory address and a mapping address of mapping a volume into the to-be-created container; establishing a binding relationship among the container identifier, the volume directory address and the mapping address; and creating the container with the container identifier, wherein data provided by the container synchronously change with data, which are indicated by the volume directory address, through the binding relationship. The persistence storage problem of partial data in container instances can be solved, and an effect that the data in the container can be saved intoa local file system is realized.

The invention discloses a method for realizing the mapping of service data flow and label switching path tunnel, which is used in a next generation network field. The method comprises the step as follows: the resource of the service data flow is distributed by a TRC-FE according to the resource information of a MPLS LSP, and a PE-FE updating flow label mapping table and other information are announced by the TRC-FE through an Rc interface; the update of a FTN table is finished by the PE-FE, and the service data flow is mapped to a named LSP tunnel. With the Rc interface and the TRC-FE announcing the PE-FE the information of mapping the data flow to the named LSP tunnel, the invention realizes the mapping between the service data flow and the MPLS LSP tunnel. So the problem that in the prior standard the service data flow of NGN network which based on MPLS-TE and the mapping of LSP tunnel resource is incapable of being realized is solved.

The invention discloses a mass knowledge resource storage method based on a knowledge graph. The mass knowledge resource storage method comprises: establishing a knowledge map meta-model by a tripletof an entity-relationship-entity manner;; Establishing a knowledge resource ontology model by taking the knowledge graph model layer, the knowledge resource ontology layer and the knowledge resource instance layer as hierarchical structures; And constructing a knowledge resource ontology description model by using the knowledge resource ontology entity and the knowledge resource ontology model through an instance mapping relationship to form corresponding knowledge graph storage. According to the technical scheme, the storage space can be saved, and the knowledge resource processing efficiencycan be improved.

A method for classifying MPLS service stream includes defining identification rule of MPLS service stream, mapping MPLS countermark of MPLS service stream to be CID by utilizing defined identification rule and carrying out classification on MPLS service stream according to CID.

The invention relates to a power transmission line wire icing thickness detection system and method, and the method comprises the steps: collecting and uploading an image and video data of a field power transmission line wire through an image collection terminal, and carrying out the wire icing thickness detection through an icing detection early-warning system according to the following steps: 1)preprocessing the image; 2) extracting an edge in the image by adopting a Canny edge detection algorithm; 3) adopting a Hough straight line detection algorithm to obtain wire edge sub-line segments;4) determining the edge to which each sub-line segment belongs by adopting an improved K-means algorithm; 5) fitting to obtain a complete and continuous wire edge; 6) mapping the wire pixel width intoan actual width based on a monocular vision distance mapping model; and 7) comparing the current actual thickness value of a wire with the thickness value of a wire which is not iced, calculating theicing thickness value of the wire, and if the icing thickness value exceeds the icing thickness threshold, carrying out early warning. The system and the method are beneficial to accurately and conveniently detecting the icing condition of the wire of the power transmission line.

The embodiment of the invention discloses an establishment method and system of a network semantic model of a target system. The method comprises the following steps: establishing the network semanticmodel of a similar network according to the network topology of the similar network of the target system; mapping the network semantic model of the similar system as the initial network semantic model according to a mapping rule between the target system and the similar system; and modifying and verifying the initial network semantic model according to the predicable information in the target system to obtain a verified network semantic model, and taking the verified network semantic model as the network semantic model corresponding to the network topology of the target system. Through the technical scheme disclosed by the embodiment of the invention, the network semantic model which cannot be obtained by the network topology and is applied by the system application can be portably and efficiently established.