34 results about "Content awareness" patented technology

Embodiments of the present disclosure include systems and methods for providing query service of secured contents. A data collection service collects data and security context associated with the data from a data source and stores the data with the security attributes in a datastore, where the security attributes are derived from the security context and used to determine access to the data so that access to the data is consistent with the security context. Upon receiving a query and a user context of a requester making the query of the datastore, a set of query results is obtained. Based on the user context and security attributes, it is determined whether the requestor has a proper right to access the query results. If the requestor has a proper right to access the query results, access to the query results is granted.

The invention discloses a content awareness deep learning network-based remote sensing image super-resolution reconstruction method. A comprehensive measurement index and a calculation method for content complexity of images are proposed; based on this, the sample images are classified by the content complexity; deep GAN models with low, medium and high complexity are built and trained; and according to the content complexity of the to-be-classified super-resolution input images, corresponding networks are selected for performing reconstruction. In order to improve learning performance of a GAN, an optimized loss function definition is given. The method overcomes the over-fitting and under-fitting contradiction ubiquitous in machine learning-based super-resolution reconstruction, and effectively improves the super-resolution reconstruction precision of the remote sensing images.

The invention relates to an image scaling method based on content awareness, and relates to graphic image conversion in an image plane. An energy function is obtained by adopting a hybrid feature model fusing a salient map, an edge line map and a gradient map, and according to the energy function, a line clipping operation is carried out to complete image scaling. The image scaling method comprises the steps of: inputting a color image to carry out preprocessing; simultaneously carrying out extraction of the salient map and a salient target image of an original color image, extraction of the edge map of a gray scale image, which is fused with line information, and extraction of the gradient map of the gray scale image; fusing three feature maps by utilizing an HFPM (hybrid feature model) algorithm so as to obtain the energy function; and clipping the original image by using a line clipping algorithm. The method disclosed by the invention overcomes the defects that an existing line clipping method adopts a gradient map definition energy function of an image and still causes distortion and loss of partial image information in the image scaling process.

The invention discloses an adversarial sample generation method based on content awareness GAN, which changes a training process on the basis of WGAN_GP, directly generates an adversarial sample witha target by inputting random noise, adds a content feature extraction part, restrains the quality of the generated sample under the condition of not influencing an attack effect, and improves the accuracy of adversarial sample generation. Content characteristics of adversarial samples can be kept unchanged as much as possible. The system comprises a generator G, a discriminator D, a target model f, a disturbance evaluation part and a feature extraction network, wherein the generator is responsible for generating a sample from random noise, the generator is trained according to a loss functionof the discriminator D, the target model f, the disturbance evaluation part and the feature extraction network, and the generator directly generates an unlimited adversarial sample from the noise. Onthe basis of the generative adversarial network, the semantic information of the concerned sample and a mode of directly generating the adversarial sample instead of a superimposed disturbance mode, direct generation of the adversarial sample of the specified target is realized by using unsupervised GAN training, the sample generation speed is increased, and the quality of the generated sample isimproved; the change of the adversarial sample in the content feature region is reduced while the high attack success rate is maintained.

The invention discloses a non-reference video quality evaluation method based on deep spatio-temporal information, and the method comprises the following steps: S1, extracting content perception features, extracting semantic layer features of a top layer through a Resnet-50 pre-trained deep neural network, carrying out the aggregation, and carrying out the mean aggregation and standard deviation aggregation of a feature map; S2, carrying out modeling on the time memory effect, wherein in the aspect of feature integration, modeling is carried out on the long-term dependency relationship by adopting a GRU network, and in the aspect of quality aggregation, a time pooling model of subjective inspiration is proposed and embedded into the network. An existing NR-VQA method cannot well model a long-term dependency relationship in a VQA task; in order to solve the problem, GRU which is a recurrent neural network model with gate control and can integrate features and learn the long-term dependency relationship is used, and the GRU is used for integrating a content awareness function and predicting a frame-by-frame quality score.

The invention provides a software definition service reorganization method based on content awareness and weighted graph in fog computing. The method runs in hierarchical fog node architecture, a fog node is formed by three layers, namely a service reconfiguration layer, a content-driven control layer and a user-defined reconfiguration layer, and meanwhile, the fog node is divided into three corresponding storage spaces, namely a service resource library, a content-based strategy library and a demand library which are respectively used for storing a service module, a content-operation mapping table and a preprocessing demand of a fog user. Compared with a service provision mechanism in the conventional fog node, the novel software definition service reorganization mechanism in the fog node, which is provided by the invention, can reduce the time cost and the computational cost to a great extent, so that the overall energy consumption of a fog system is reduced.

Embodiments of the invention provide a method and system for playing audios and videos at a multiplied speed based on content awareness, a storage medium and equipment. According to a technical scheme provided by the embodiments of the invention, the method comprises the following steps: reading a target audio / video; inputting the target audio / video into an AI model with a specified function, and outputting an identification result; according to the identification result, determining a speed multiplication instruction from a plurality of set speed multiplication instructions to be selected; and playing the target audio / video based on the speed multiplication instruction. According to the invention, the playing speed is dynamically adjusted based on the psychological perception of a user in auditory sense and visual sense, thereby improving user experience, and deeply mining the low power consumption potential of playing associated equipment such as hardware acceleration equipment.

The invention discloses a dynamic rule chained recursive triggering method based on message content awareness and a system. According to main technical characteristics, on the basis of establishing adynamic rule chained recursive triggering system, for the problem that an existing TAP device cannot carry out dynamic association analysis on message streams, triggers and an association relationshipbetween the triggers and triggering rules are added, through adoption of steps such as trigger configuration, static rule and trigger definition configuration, message matching, new dynamic rule generation, rule aging deletion and rule table updating, dynamic information such as such as IP addresses, ports and user identities in the externally input message streams are dynamically extracted, andthe new dynamic rules are recursively triggered, so the message streams can be precisely output to a rear end analysis system, input bandwidths of the rear end analysis can be clearly saved, and the performance loss of the rear end analysis system is avoided. The method and the system have the advantages that the realization method is simple, the rule triggering is rapid and efficient, and the message traffic of the rear end analysis system can be clearly reduced.

The invention discloses a vehicular ad hoc network video transmission method based on an overlay structure. The method mainly solves the problem that in the prior art, a node with needs for video cannot find a proper video source node is solved. The method comprises the implementation steps that (1) a cluster structure based on content awareness is built according to the position of the node, video stored in the node and the needs of the node for the video; (2) a distribution list of the video in a cluster is built at a cluster head, and is used for constructing the overlay structure; (3) the node in the cluster searches for the optimal source node according to the overlay structure, and request information is sent to the optimal source node; (4) the cluster head sends video request information to an RSU firstly, the RSU returns video information to the cluster head, and the communication between the cluster head and the RSU is achieved. According to the method, the node with the needs can find the optical video source, transmission delay is reduced, video playing quality is improved, and the method can be used for video transmission in a vehicular ad hoc network.

The invention discloses a content-awareness-based resource scheduling method. The method comprises: a transmitting terminal carries out statistics on IDs of requested contents by all receiving terminals and correlated channel state information; the transmitting terminal maintains corresponding request queues by using the content IDs as identifiers and calculates the user number of current service requests that can be met by employing different coding modulation ways under different request contents; the transmitting terminal calculates the potential service user number for scheduling a certain content currently by using a certain coding modulation way is calculated according to a predicted future content heat degree, a user caching state within a coverage range, and the connecting number and adds the potential service user number into a scheduling weight; according to the scheduling weight, content sorting is carried out and the contents are sent in a descending order according to a preset coding modulation way based on the scheduling weight; and a receiving terminal receives and decodes air interface data and a user that does not send a request and has a high channel can cache an air interface high-attention content selectively. According to the invention, the highest expected return under unit consumption resources during resource scheduling can be realized. Because the user having the high channel can carry out caching, repeated transmission of the hot content is avoided, the content transmission efficiency is improved, and the good user experience is provided.

The invention provides a data deduplication method and device based on content awareness, the device comprises an IO processing device running in an upper computer and a data deduplication device running in a storage controller, and the data deduplication device comprises an analysis unit, a feature value comparison unit and a data deduplication execution unit. The method includes: Dividing the to-be-stored data into texts according to contents, Audio, image, and non-timing numerical data and timing numerical data, calculating the characteristic values by using different algorithms according to types; combining the obtained characteristic value with the initial IO write request to form a new IO write request; and according to the analyzed IO write request data type, enabling the characteristic value comparison unit to read the characteristic value from the corresponding characteristic value sub-table, calculating the Hamming distance between the characteristic value of the to-be-storeddata and the read characteristic value, and sending a judgment result to the data deduplication execution unit to perform data deduplication operation. According to the invention, a feature value calculation algorithm based on content awareness and a data repetition judgment standard are adopted, so that the duplicate removal success rate and the duplicate removal efficiency are improved.

The invention relates to the field of a computer network, and particularly relates to a high-quality QoS guarantee method for content centric networking. Content identification tables can be added on routing nodes of the content centric networking to implement content awareness, priority differential transmission of different content feature data packets is implemented by active values, and throughput of the networking is improved, so that high-quality QoS guarantee is provided for contents required by users. The high-quality QoS guarantee method comprises the following contents of: according to content data information of the content centric networking, obtaining traffic classification identification tables in the networking; according to popularity information of content features in the nodes of the content centric networking, obtaining active value categories of different content features; according to active value information of different content features, establishing different priority transmission queues; according to the active value information of different content features and a priority judgment threshold, obtaining probabilities for different content data packets to obtain high priorities; and according to the probabilities for different content data packets to obtain the high priorities, completing the scheduling operation on different priority transmission queues.

The invention provides a crowd density estimation method and system based on a content perception module, and relates to the technical field of crowd density estimation, and the method comprises the steps: carrying out the preprocessing of to-be-predicted image data; extracting deep semantic information and multi-layer fusion context information according to the processed image data; dynamically generating content perception parameters according to the deep semantic information based on a content perception module; initializing the weight and offset of a crowd density estimation module convolution layer according to the content perception parameters; and inputting the multilayer fusion context information into the initialized crowd density estimation module to obtain a crowd density image. When the crowd density is predicted each time, the deep semantic information of the picture data is obtained, and then the parameters of the convolution layer of the crowd density estimation module are dynamically generated, so that crowd density estimation is carried out more flexibly and specifically, and a better prediction result can be obtained in a complex and changeable scene.

The invention discloses a content awareness-based mobile terminal adaptive video multi-speed playing method. The method comprises the following steps of: downloading video data online; decoding and playing the video data; analyzing the video data in real time and identifying a specific scene in the video data; and determining the playing speed according to the identification result. According to the content awareness-based mobile terminal adaptive video multi-speed playing method, an artificial intelligence deep learning algorithm is applied to a player of an intelligent mobile terminal and is used for analyzing a specific scene. And when the required scene does not exist, a high-speed playing mode is adopted for quickly skipping, and when a specific scene is found, the playing speed is adaptively reduced to be normal. On one hand, the time for positioning the specific content during video playback is shortened, on the other hand, battery consumption during multi-speed playing of the player is reduced, and the service time of the mobile equipment is prolonged.

A recommendation method fusing content awareness and feature similarity relates to the cross technical field of recommendation systems and distributed computing, and comprises the following steps: firstly, decomposing playing frequency data by using a weighted matrix decomposition model: decomposing the playing frequency data into a product of a transpose matrix WT of a user preference matrix and an item (namely music) attribute matrix H; introducing content information into weighted matrix decomposition in consideration of the characteristics of a project (music): adopting a mode of rewriting a project attribute matrix H, specifically, rewriting the project attribute matrix H by adopting a content characteristic zi to obtain a rewritten project attribute matrix Hi; obtaining a prediction matrix A (WT * Hi) by adopting the product of a transpose matrix WT of the user preference matrix and the rewritten project attribute matrix Hi; performing cosine similarity calculation on the timbre and pitch of the item (music) to obtain a prediction matrix B; and fusing the obtained prediction matrix A and the prediction matrix B by using a fusion factor to obtain a recommended value F (mu, i), of which the specific expression is F (mu, i) = theta A (mu, i) + (1-theta) B (mu, i). According to the invention, the cold start problem in the recommendation process can be effectively solved.

The invention provides a remote sensing target detection method based on content awareness. The method comprises the following steps: obtaining coordinate information of a prediction frame according to information output by a regression module, and obtaining coordinate information of a real frame from an original remote sensing image; positioning on the maximum feature map according to the coordinate information of the prediction frame and the coordinate information of the real frame to obtain feature information v* and v contained in the prediction frame and the real frame respectively; determining a common minimum horizontal enclosing rectangle of the prediction frame and the real frame, and taking out an area where the minimum horizontal enclosing rectangle is located from the maximum feature map, and recording the area as a feature map x; respectively setting contents which do not belong to the real frame and the prediction frame on the feature map x to be 0 to obtain a feature map f1 corresponding to the real frame and a feature map f1* corresponding to the prediction frame; calculating the similarity between the feature map f1 and the feature map f1*, taking the similarity as a content consistency loss Lfeed, and constraining the regression of the anchor frame based on the content consistency loss Lfeed.

The invention provides a blockchain and security multi-party technology combined content awareness method, which belongs to the field of financial data, can be widely applied to financial service scenes such as financial security data sharing and the like, and comprises the following steps: acquiring a deployment request of a smart contract, and assembling a deployment message to all blockchain nodes to broadcast; each block chain node analyzes the deployment message and deploys the smart contract; issuing a multi-party security computing transaction request, assembling a transaction message, and broadcasting the transaction message to all block chain nodes; each block chain node analyzes the transaction message and performs joint calculation on encrypted data in the transaction message; and the smart contract on each block chain node verifies a multi-party security computing transaction result through a block chain consensus mechanism and evaluates a similarity result. On the basis of a block chain and multi-party security computing framework, a semantic perceptual hash algorithm is creatively embedded into the block chain and multi-party security computing framework, and the problems of digital file content tracking and operation are effectively solved.

The invention discloses a content awareness-based power transmission line aerial image data cleaning method, which comprises the following steps of: firstly, constructing a self-classification model, and cleaning an image which does not contain an insulator part in an aerial image; training a model for extracting image content by using the residual data set; the method comprises the following steps: acquiring local and global features of an image, establishing a quality perception rule according to different influence degrees of local and global contents on image quality, mapping an image quality score function, and finally screening the image according to image quality scores. According to the invention, mass aerial image data can be effectively cleaned, so that clearer and higher-quality images can be obtained.

The invention discloses a resource scheduling method based on content awareness, which includes the steps of: the transmitting end counts the content ID and related channel state information requested by each receiving end; The number of users who can satisfy the current service request with different coding and modulation methods under the request content; the transmitting end calculates the number of users currently scheduling a certain content with a certain coding and modulation method according to the predicted future popularity of the content, the user cache status and the number of connections within the coverage area The number of potential service users in the future, and the number of potential service users is added to the scheduling weight; the content is sorted according to the scheduling weight, and the content is sent from high to low according to the scheduling weight according to the preset coding and modulation method; the receiving end receives and decodes For air interface data, for users with good channels who have not sent requests, high-interest content on the air interface can be selectively cached. The present invention realizes the maximum expected return under resource consumption per unit when scheduling resources. Since users with good channels can cache, repeated transmission of hot content is avoided, the transmission efficiency of content is improved, and a better experience is brought to users.