225results about How to "Reduce computational overhead" patented technology

A method for tracking one or multiple objects from an input video sequence allows a user to select one or more regions that contain the object(s) of interest in the first and the last frame of their choice. An initialization component selects the current and the search frame and divides the selected region into equal sized macroblocks. An edge detection component computes the gradient of the current frame for each macroblock and a threshold component decides then which of the macroblocks contain sufficient information for tracking the desired object. A motion estimation component computes for each macroblock in the current frame its position in the search frame. The motion estimation component utilizes a search component that executes a novel search algorithm to find the best match. The mean absolute difference between two macroblocks is used as the matching criterion. The motion estimation component returns the estimated displacement vector for each block. An output component collects the motion vectors of all the predicted blocks and calculates the new position of the object in the next frame.

A system and method is provided for vectorizing misaligned references in compiled code for SIMD architectures that support only aligned loads and stores. In this framework, a loop is first simdized as if the memory unit imposes no alignment constraints. The compiler then inserts data reorganization operations to satisfy the actual alignment requirements of the hardware. Finally, the code generation algorithm generates SIMD codes based on the data reorganization graph, addressing realistic issues such as runtime alignments, unknown loop bounds, residual iteration counts, and multiple statements with arbitrary alignment combinations. Loop peeling is used to reduce the computational overhead associated with misaligned data. A loop prologue and epilogue are peeled from individual iterations in the simdized loop, and vector-splicing instructions are applied to the peeled iterations, while the steady-state loop body incurs no additional computational overhead.

Systems and methods for remote access of network-enabled devices that provide seamless, firewall-compliant connectivity between multiple users and multiple devices, that allow collaborative operations by multiple users of remote devices, that allow point to multipoint control of multiple devices and which allow rapid, secure transmission of data between remote users and devices. In general terms, the system includes at least one connection server, and at least two computers operatively coupled to the connection server via a public or global network. In an example where at least one client computer is operatively connected to at least one network-enabled device through a connection sever via the public or global network, the connection server is configured to route control instructions from the client to the network-enabled device, and route data from the network-enabled device to the client.

A vehicle navigation system helps guide a driver to a destination by enhancing visualization of landmarks and upcoming intersections. The navigation system stores resolution independent representations of the landmarks and intersections. The representations allow the navigation system to quickly and efficiently resize and render the landmarks and intersections without distortion and with reduced computational burden. When the vehicle approaches a landmark or intersection, the navigation system may retrieve the representation, mathematically scale the representation, adjust the perspective of the representation, and render a view of the upcoming landmarks and intersection to aid the driver with reaching the destination.

The invention relates to a data backup system based on fingerprints, belonging to the technical field of computer storage backup, which aims at reducing management, storage and network overhead of data backup and improving backup performance. The invention comprises a backup server, a backup agent, a storage server and a Web server which complete data backup and recovery through mutual network communication. The invention is characterized in that redundant data of backup files is recognized with the file segmentation technology based on anchors, thus the invention has the advantages that the modification stability is good and computation cost is low; data segmentations with fingerprints as the index are stored on a plurality of disk arrays of the storage server so as to eliminate backup of redundant data and save disk storage space; the data segmentations are not erased once stored and can be continuously appended on the disk so as to eliminate disk storage fragments; since the effective backup buffer strategy is adopted, the invention also has the advantages of reduced network overhead of backup, increased data backup speed and lowered backup influence to application servers.

The invention particularly discloses a dependence-based parallel task packet scheduling method used on an isomorphic multi-core cluster server. Through calculating the dependence between execution paths and under the premise of not increasing the maximum loads of the execution paths, execution paths with large dependence are combined, compared with the original algorithm, the method can reduce unnecessary task duplication expenditure, communication expenditure between processors and the idle time energy consumption of the processors are reduced, the load of the processors is more balanced, and the overall energy consumption is lowered. The dependence-based parallel task packet scheduling method has the greatest features that the number of the processors is effectively reduced, the resource utilization efficiency is improved, the load of the processors is more balanced, and accordingly, the overall energy consumption is reduced.

The invention relates to a virtual desktop transmission device and method. The virtual desktop transmission device comprises a client and a virtual desktop server, wherein a response server is also arranged between the client and the virtual desktop server; the response server is used for realizing the load balancing of requests and the buffer of virtual desktop frames. The method provided by the invention comprises the following steps: the client asynchronously sends requests to the response server; the response server sends the requests to the virtual desktop server through a load balancing algorithm; the virtual desktop server sends the virtual desktop frames to the response server; the response server stores the virtual desktop frames in a buffer; and the virtual desktop frames is sent to the client according to the bandwidth conditions of the client. By using the device and method provided by the invention, the transmission performance can be improved, the utilization of resources can be improved and the cost can be reduced.

The invention discloses an efficient condition privacy protection and security authentication method in an internet of vehicles. The efficient condition privacy protection and security authentication method comprises the following steps of: system initialization, generation of pseudonym identities and signature private keys of vehicles, signing and authentication of a message and tracing of real identities of the vehicles. The vehicles carry out cooperative communication with surrounding vehicles and roadside units deployed at both sides of a road by on-board units assembled on the vehicles, driving security of the vehicles can be effectively improved, and vehicle users can more conveniently and rapidly acquire related traffic services. The efficient condition privacy protection and security authentication method disclosed by the invention not only can meet security requirements in the internet of vehicles, but also optimizes the computing process of signature generation and verification in the communication. The efficient condition privacy protection and security authentication method is greatly improved on the aspect of efficiency of computing cost, communication cost and the like, and is more applicable to communication and application in the internet of vehicles.

The invention relates to a method and a device for forwarding a hyper text transport protocol (HTTP) request. The method comprises the steps of: judging whether a uniform resource locator (URL) of the HTTP request of a Web client is a URL requested by a Web form or a URL submitted by Web form data or not; forwarding the HTTP request when the URL of the HTTP request is the URL requested by the Web form and effective tokens are carried in parameters of the URL of the HTTP request; randomly generating a unique token if the token is not carried, splicing the URL of the HTTP request and the generated token into a new URL, discarding the HTTP request, and sending an HTTP response message requesting the redirection to the new URL to the Web client; and forwarding the HTTP request when the URL of the HTTP request is the URL submitted by Web form data, the HTTP request has the referrer value and the effective token can be extracted from the referrer. The method and the device have the advantages that the effective defense on cross site request forgery (CSRF) attack can be realized, and the computation overhead on a Web security gateway is greatly reduced.

The invention discloses an encryption and search method for revocable keyword search public keys in cloud storage and belongs to the technical field of network security. The method includes: setting system public parameters and dividing system time into z time segments, and creating a public-private key pair at a client; when data storage requests occurs, selecting a keyword collection of a data file, selecting optional symmetrical encryption algorithm to encrypt the data file, utilizing the public key and the current time segment to encrypt the keyword collection, and transmitting the encrypted keyword collection to the cloud server; when the next time segment arrives, generating keyword collection cipher again, and updating the cloud server on a server; when a keyword search request occurs, by the client, utilizing the private key and the current time segment to calculate trap door information of the keywords, transmitting the trap door information to the cloud server, and by the cloud server, obtaining the search results and sending back file cipher containing the search keywords to users. The encryption and search method is suitable for cloud storage with high safety requirements, safe, and efficient, and has the capability of revoking server search.

Systems and methods for distance and location tracking a wireless device are disclosed. More specifically, according to one aspect of the present disclosure, a Wireless Area Network-Location Based Services (WAN-LBS) algorithm that utilizes a hybrid Wireless Area Network (WAN) and Global Positioning System (GPS) circuit board and method for seamless indoor and outdoor tracking is disclosed. The WAN-LBS algorithm, in conjunction with the hybrid WAN/GPS circuit board, optimizes the degrees of precision and accuracy for distance measurements in locating a fixed or mobile IEEE 802® device. In one embodiment, the hybrid WAN/GPS circuit board according to the present disclosure integrates the data of a GPS receiver and several IEEE 802 standards based receivers. In another embodiment, the WAN-LBS algorithm according to the present disclosure utilizes received data, acquired by the hybrid circuit board, to calculate distances to the tracking devices, seamlessly in both indoor and outdoor environments.

A gradient sensitivity based method is described for reducing Peak to Average Power Ratio in multicarrier communication systems. The method involves definition of a differentiable penalty function which has non-zero values whenever there is a peak violation. The gradient of the function with respect to the symbol magnitudes and phases is evaluated at the given symbol vector. Symbol perturbations are computed in the opposite direction of the gradient in such a way that the peaks are reduced and the amount of errors introduced in the symbols is not large enough to affect the decoding process. In order to reduce computation, a variant of the method is proposed where a peak reduction kernel is precomputed and stored. The kernel is rotated to the location of the peak and added with proper sign to reduce PAR.

The invention discloses a method and a device for controlling hyper text transport protocol (HTTP) traffic. The method includes the following steps: according to HTTP traffic sent to a Web server by a Web client, deciding that the Web client is an automatic tool or artificial browsing; only allowing the HTTP traffic generated by the Web client which is decided to be artificial browsing to enter the Web server. The method and the device are capable of limiting abnormal HTTP traffic, simultaneously allow HTTP traffic of normal Web business access to normally pass and accordingly ensure Web access service quality of normal users to the maximum degree.

The invention provides an agricultural product traceability system based on a block chain. The system comprises an agricultural product big data platform, an agricultural product traceability query platform, and a 5G network; the agricultural product big data platform comprises an agricultural product growth information module, a storage information module, a processing information module and a sales information module; the agricultural product traceability query platform comprises a data restoration module, an anti-counterfeiting query module and a traceability information query module; and each module carries out data uploading and traceability query through a 5G network. According to the invention, further optimization is carried out on the basis of a traceability system based on a block chain; an agricultural product tracing system combining the advantages of a private chain and a public chain is designed. On the basis of preventing the agricultural product information from being attacked and tampered, the certified product label is further prevented from being stolen and copied, the data storage efficiency is improved, the calculation expenditure is reduced, and consumers or supervision departments can quickly position production links and responsible persons of problematic agricultural products.

The invention provides a private protocol message format inference method based on an extended prefix tree. The method comprises the following steps of message preprocessing, protocol keyword extraction, message structure and semantic inference and protocol format combination. The method aims to solve the problems that an existing message format inference method based on network traffic is high intime complexity and low in accuracy. The N-gram segmentation method is used, protocol keywords in a private protocol are accurately judged on the basis of pointwise mutual information, a protocol keyword sequence corresponding to a message is described through the extended prefix tree, segmented multi-sequence comparison is implemented, the calculation expenditure of sequence comparison is reduced, and the accuracy of comparison results is improved. In addition, according to the characteristic that a network protocol is high in flexibility, the redundant message format is reduced through effective format combination, and the practicability of inference results is enhanced.

A road-crossing driving control method for intelligent vehicles belongs to the technical field of unmanned driving. Firstly, a monocular camera is mounted on a rear-view mirror in an intelligent vehicle and used for collecting video images so as to perform the lane line detection, the stop line detection, the stop line distance measurement, the pedestrian detection and the traffic light identification. Afterwards, the lane virtual center line is calculated according to the lane line detection result, and the intelligent vehicle is controlled to run along the center line by means of a PD control algorithm. The distance from the stop line, the pedestrian detection result and the traffic light identification result are synthesized to make the driving decision so as to control the vehicle to move forwards or stop. By means of only one camera, the intelligent vehicle can stably and safely pass many road crossings; when a pedestrian is detected or a red traffic light is identified, the system controls the intelligent vehicle to stop within 20 cm away from the stop line; and when a green traffic light is identified and no pedestrian is detected, the system controls the intelligent vehicle to normally move or make a turn. The execution cycle is controlled within 50 ms, and thus the requirement that the driving control period is 100 ms is met.

The invention discloses a recommending method and equipment for sequencing-oriented collaborative filtering. The method comprises the steps of: acquiring preference data of a user to sequenced item pairs; establishing a probability potential preference analysis model based on a user preference according to the preference data and data related to the user and items; and predicting item sequencing based on the user preference by utilizing the probability potential preference analysis model so as to make recommend to the user. By means of the recommending method and equipment, the user preference sequencing of other items can be directly predicted by using the predicted user preference to the item pairs, thus, sequencing prediction can be instantly made without the searching in the whole preference database. Therefore, not only can calculation expense be reduced, but also the accuracy and the convenience in making recommends to the user are improved.

The invention puts forward a finite-element reconstruction method for space weighting of auto-fluorescence imaging. The method uses a diffuse optical tomography (DOT) technique to precisely reconstruct optical parameters of main biological tissue. The method limits a possible real light-source region, and divides a whole reconstruction region into a light-source possible region and a light-source impossible region so as to improve the numerical stability and effectiveness of reconstruction problems and reduce the ill-conditioned property of BLT reconstruction problems. In order to avoid famous 'Inverse Crime' problems, the method adopts a Monte Carlo (MC) random method to simulate the process of transmitting light in the biological tissue. Then the invention puts forward a finite-element method based on a space-weighting unit and adopts a tolerant algorithm for linear constrained optimization problems to reconstruct light-source information.

The invention discloses a secret key changing and signature updating method for cloud data audit. The method comprises the following steps of a before-updating stage, an updating stage and an after-updating stage. In the before-updating stage, a data owner generates an initial secret key and a signature through calculation, a cloud server obtains initial data integrity authentication parameters through calculation according to the signature of the data owner and parameters received from the third audit party, and the third audit party is used for auditing the initial data integrity authentication parameters. In the updating stage, a data owner selects a random number generation secret key, calculates a secret key needed by a shared updated signature and sends the secret key to the cloud server, and the cloud server calculates a new signature according to data and the received parameters. In the after-updating stage, the data owner generates a signature through calculation according to the new secret key generated in the updating stage, the cloud server calculates new date integrity authentication parameters, and the third audit party is used for audit. The scheme guarantees safety and reduces cost of communication between the data owner and the server end and calculation cost of the data owner.

A distributed absorbing surveillance apparatus comprising a plurality of surveillance branches at which data of a video image frame is characterised, indexed and stored locally in real-time upon capturing means browsing and searching can be conducted locally at a relatively low computational overhead upon receipt of searching instructions from a central query processor. Such a distributed surveillance apparatus also facilitates enhanced target searching speed and efficiency.

The invention provides a compressed spectrum sensing method based on autocorrelation matrix reconstitution and mainly solves the problem that an existing sensing algorithm is high in sampling speed and large computation overhead. The compressed spectrum sensing method comprises that a secondary user obtains an observation sequence of a frequency spectrum environment through compressed sensing and obtains an autocorrelation matrix estimated value of Nyquist sampling by utilizing autocorrelation vector quantity of a autocorrelation matrix reconstitution Nyquist sample sequence of the observation sequence; a multi-signal classification MUSIC algorithm is adopted to obtain an estimated value of occupied channel number according to a characteristic value of the autocorrelation matrix estimated value; a characteristic spectrum is constructed according to the characteristic value and the estimated value of the occupied channel number, spectral amplitude values corresponding to characteristics of channels are added to obtain a sum, and mark numbers of the occupied channels are judged. By means of the compressed spectrum sensing method, the sampling speed of a secondary user receiving machine can be reduced, algorithm complexity at the reconstitution end is low, and spectral amplitude occupying situation in a cognitive radio system can be judged quickly.

The invention discloses an Internet of Things device identity authentication method based on a block chain, and belongs to the field of Internet of Things. According to the method, an emerging block chain technology is adopted to construct an identity authentication alliance block chain suitable for Internet of Things equipment; in the registration stage, an Internet of Things device identifier and a public key pair (ID, PK) are stored in a block chain in a block chain transaction mode, and in the authentication stage, the public key PK in the registration process is used for authenticating the Internet of Things device identifier ID. According to the method, the authentication process can be reduced, the authentication efficiency can be improved, the cost requirement on the Internet of Things equipment is low, the method does not depend on a certain specific third-party mechanism, and tamper-proofing and traceability of the identifier and the public key pair (ID, PK) are ensured.

The invention concerns a method of communication in a communication network, using a communication device comprising means for exchanging data with at least another communication device, comprising the following steps:

• • receiving, at said device:
    • i. information representing interconnections between users of a social networking application;
    • ii. information concerning at least one contact between said device and at least another device, occurring in said communication network;
  • determining, upon occurrence of a new contact between said device and at least another device, based:
    • on said information representing interconnections between users of a social networking application and
    • on predetermined rules related to data forwarding;
  • to transmit data or not to transmit data from said device using said new contact.

The present invention relates to division method for video motion object, which comprises that first, take time sequence division to the image, separate the initial zone and background contained motion object; take the following space division and the classification and combination for zones only on initial zone, decrease largely compute consume, increase division speed; according to zone space, time sequence and similarity rate of neighboring region, add space constraint, time sequence constraint, and neighboring region constraint to MRF mode; by computing maximum posterior probability of MRF, classify the zone; finally, divide accurately motion object, and overcome the shortcoming that motion estimation is easy to be affected by irregular motion and light lamination.

A method for mapping a network topology request to a physical network is described. For each of a plurality of primary nodes included in the network topology request a plurality of nodes in the physical network is determined that meet a primary resource requirement associated with the primary node, and from the nodes determined one or more node pairs connected in the physical network by a path are selected that meets a backup connection requirement for a connection between a primary node and its backup node. Then, the paths in the physical network between node pairs associated with interconnected primary nodes are determined.

The invention discloses a high-efficiency polymerization signcryption solution in an Internet-of-things environment, wherein the high-efficiency polymerization signcryption solution mainly settles theproblems of relatively small application range and relatively low practicability caused by relatively high calculation expenditure in an existing polymerization signcryption solution. The high-efficiency polymerization signcryption solution comprises the following realizing steps of 1, performing system initialization by a secret key generating center KGG and a certificate issuing mechanism CA; 2, registering a perception entity Ni in a CLC system and a user U in a PKI system to KGC and CA; 3, broadcasting a common random number and a public key certificate by the user U in the PKI system, and performing verification and receiving by the Ni; 4, selecting a message generating signcryption cryptograph by the perception entity Ni in the CLC system, transmitting the message generating signcryption cryptograph to an aggregator Nn for forming aggregated signcryption, and transmitting the aggregated signcryption to the user U by the aggregator Nn; and 5, performing polymerization de-signcryption and restoring corresponding plaintext information. The high-efficiency polymerization signcryption solution has isomerism and higher practicability. The high-efficiency polymerization signcryption solution can be used for smart home, intelligent traffic, intelligent medical service and intelligent industrial detection.

A federated learning global model training method based on differential privacy and quantification comprises the following steps that a central server issues a pre-trained federated learning global model, a local model gradient of each local user is generated, and noise adding, threshold quantification and compression quantification is sequentially performed on the local model gradient of each local user, and the compressed and quantified local model gradient is uploaded to the central server, the central server performs weighted aggregation on the uploaded local model gradient by the central server, updates a global model, issues the global model to each local user, and training is ended when the privacy budget value of each local user is exhausted or the federal learning global model is converged. According to the method, on the premise that the precision of the federal learning global model is not lost, the privacy of local users is protected, the communication overhead in the transmission process is reduced, and the training efficiency of the federal learning global model is improved.

The invention discloses a segmentation-based multi-scale feature pyramid text detection method. The method comprises the steps of obtaining data; constructing a pyramid feature extraction model, and extracting features from the acquired data; sampling input data to obtain input images with different scales, respectively inputting the input images into a pyramid feature extraction model, extractingtext features, fusing and processing the text features of the input images with different scales through a multi-scale detection network to obtain a feature map, and predicting the feature map; and processing a prediction result to obtain a contour boundary line of the text region. The method has high robustness, can be directly applied to text detection of any shape in a natural scene, and achieves high accuracy, recall rate and F value.

A method and apparatus for providing point-in-time consistent backup images while reducing a computing overhead associated with copy-on-write operations for snapshot images is described. In one embodiment, the method comprises processing at least one data block that is inconsistent with a backup image; and updating the backup image with the at least one inconsistent data block from a snapshot image, wherein the snapshot image is created after a backup process, wherein the updated backup image is point in time consistent.

The invention discloses an edge computing-based end-edge collaborative federated learning optimization method, and the method comprises the following steps: constructing a federated learning optimization system architecture which comprises a cloud data center, an edge server and a mobile terminal; the mobile terminal and the edge server cooperatively perform local updating of the model; after local updating is completed, the edge server sends the updated model parameters to the cloud data center for global aggregation, and a new model is obtained; and the cloud data center performs model precision testing on the new model. According to the invention, the federated learning optimization system architecture is established, and the edge server and the mobile terminal are added for cooperativelocal updating, so that the computing overhead of the mobile terminal is greatly reduced; and after the collaborative updating is finished, the operation results of the plurality of mobile terminalsare aggregated on the edge server, so that the required global aggregation frequency is reduced, the communication delay is greatly reduced, and the model training speed of federated learning is increased.