156results about How to "Fully capture" patented technology

The invention discloses an active defense method based on a cloud platform. The active defense method comprises the following steps of: intercepting progress establishment by a client to acquire a program file route of a progress; calculating a characteristic value of a program file according to the program file route; judging whether the characteristic value of the program file exists in a local database; if the characteristic value of the program file does not exist in the local database, sending the characteristic value of the program file to a cloud end; judging whether the characteristic value of the program file exists in a database of the cloud end; if the characteristic value of the program file does not exist in the database of the cloud end, judging whether cloud end analysis is required to be performed on the program file; if the cloud end analysis is required to be performed on the program file, sending the program file to the cloud end; executing a command in the program file by the cloud end; judging whether the residual commands in the program file do not comprise commands related to a user environment; if the residual commands in the program file do not comprise commands related to the user environment, executing the residual commands; and detecting the security of the program file according to a behavior record of the program file, and returning a detection result back to the client.

A gasket system for sealed interfaces utilizes a carrier ring having inner and outer compression stops to account for and control flange rotation effects in order to achieve a uniform and constant gasket seating stress across the sealing region. The inner compression stop provides a fixed fulcrum for flange rotation. The inner and outer compression stops have different heights, dependent upon the characteristic flange rotation, that allow the orientation of the sealing material to be angularly matched to the flange side profile and that allow the sealing material to be fully captured within the sealing region upon full compression, minimizing creep relaxation. A visual compression indicator on the circumference of the carrier ring permits assembly of the interface without the use of specialized or calibrated tools.

The invention discloses a refining flux used for purifying and processing aluminum or aluminum alloy melt, and its preparation method. The invention is characterized in that the raw material formula of the refining flux is a powder mixture composed of the following components by mass: 25 to 35% of NaCl, 25 to 35% of KCl, 15 to 25% of Na3AlF6 and / or K3AlF6, 4 to 10% of Na2SO4 and / or K2SO4, 3 to 10% of CaCO3 and / or Na2CO3, 3 to 6% of REF3 and / or RECl3; wherein, a mass ratio of NaCl to KCl is 1:1; RE is Ce or La. The refining flux can be fully diffused after entering the melt without adsorption agglomeration and pore plugging; and has high efficient capability for producing dry slag without viscous aluminum melt; even the refining flux has a slight reaction with aluminum melt, the reaction products can be removed with slag; the refining flux provided in the invention has high efficiency of gas and slag removing.

In accordance with one aspect of the invention, a selector supports the selection of a first encoding scheme or the second encoding scheme based upon the detection or absence of the triggering characteristic in the interval of the input speech signal. The first encoding scheme has a pitch pre-processing procedure for processing the input speech signal to form a revised speech signal biased toward an ideal voiced and stationary characteristic. The pre-processing procedure allows the encoder to fully capture the benefits of a bandwidth-efficient, long-term predictive procedure for a greater amount of speech components of an input speech signal than would otherwise be possible. In accordance with another aspect of the invention, the second encoding scheme entails a long-term prediction mode for encoding the pitch on a sub-frame by sub-frame basis. The long-term prediction mode is tailored to where the generally periodic component of the speech is generally not stationary or less than completely periodic and requires greater frequency of updates from the adaptive codebook to achieve a desired perceptual quality of the reproduced speech under a long-term predictive procedure.

The anti-blooming structure of an image sensor is supplied with varying voltages during different integration periods such that charges generated in response to low level light are fully captured, whereas charges generated in response to a bright light spill over in a controlled manner. Accordingly, sensor's response may be generated to result in higher gains at low light levels and progressively lower gains at the higher light levels.

An image capturing system utilizes a plurality of image capturing apparatus to capture first image data for an object. A processor is utilized to combine the first images to produce first three-dimensional digital image data. The processor then utilizes a second algorithm to combine the intermediate digital image information to produce a complete three-dimensional digital image of the object. The three-dimensional image may be viewed on a display and viewed from any point on a three dimensional surface surrounding the object. The object can be non-stationary.

The present invention has multiple aspects relating to a bone-tendon-bone graft and components thereof. Embodiments of the present invention comprise an intermediate bone block that is used to adjustably secure soft tissue (e.g., tendon) in a patient. Preferred intermediate bone blocks comprise at least one machined segment of cortical bone, cancellous bone or both. Embodiments of intermediate bone blocks of the present invention are used singly or in combination with one or more bone blocks to form a bone block assembly for securing soft tissue. The present invention further relates to an assembled bone-tendon-bone graft suitable for implantation in humans comprising the intermediate bone block and a length of soft tissue. In a preferred embodiment, a bone-tendon-bone graft comprises a length of soft tissue (e.g., tendon) extending from a first assembled bone block to a second bone block and then doubles back to said first assembled bone block. Depending upon the embodiment, the second bone block fixedly or slideably attaches to the length of soft tissue and facilitates it doubling back to the first assembled bone block.

The invention discloses an interactive attention emotion analysis method based on dependency syntax. According to the method, the overall mutual influence of a target set and an original sentence is considered by carrying out dependency syntactic analysis on the sentence and designing interaction attention; The problems that the given target contains limited information, the relation between the target and the context cannot be fully established based on a simple splicing mode, and the emotion polarity of the text comments cannot be accurately analyzed are solved. According to the invention, the interaction attention based on the dependency syntax is designed; the fine-grained emotion analysis accuracy can be further improved; The emotional tendency of the commodity comments in the e-commerce industry is accurately mined, more detailed and directional commodity evaluation information is provided for the user before the user purchases the commodity, and the merchant can more fully understand the comment tendency of the consumer to a certain commodity, so that the service quality is improved, and the method has very high practical value and practical significance.

The invention discloses a session recommendation method based on a space-time sequence diagram convolutional network. The method comprises the following steps: S1, modeling all session sequences intoa directed session graph; S2, constructing a global graph by taking common commodities in the session as links; S3, embedding an ARMA filter into a gated graph neural network, extracting a topologicalgraph signal which changes over time from the graph model, and obtaining a feature vector of each node involved in the session graph; S4, obtaining global preference information from historical sessions of the user by adopting an attention mechanism; S5, obtaining local preference information of the user from the last session clicked by the user, and obtaining final preference information of theuser in combination with the global preference information; S6, predicting the probability of possible occurrence of the next clicked commodity in each session, and giving a Top-K recommended commodity. According to the method, rich context relationships of clicked commodities can be captured from the global graph, global and local preferences of the user are accurately learned, the time attenuation effect of historical preferences of the user on current preferences is effectively evaluated, and accurate commodity prediction is provided.

The invention discloses a human skeleton action recognition method, and relates to a method for identifying graphics. The method is a human skeleton action recognition method combining space-time attention and a graph convolution network. Diversity and complementarity of different feature information are fully mined; the weight value of each joint point of the space structure and the importance ofeach frame of the video sequence are adaptively adjusted by using an attention mechanism; the motion recognition of the human skeleton is carried out by using the graph convolution network, and the defects that in the prior art, space-time feature information cannot be better captured, and errors are likely to occur in the recognition of the difficult motion of the human body are overcome.

The invention discloses a remote sensing image classification method, a storage medium and computing equipment, and the method comprises the steps: building a remote sensing image set, carrying out the standardization of the remote sensing image set, and obtaining a training sample set and a test sample set; setting a multi-scale feature extraction module, and generating feature maps of two scalesby setting different hole convolution in two parallel convolution modules; setting a self-adaptive feature fusion module, wherein the self-adaptive feature fusion module can adaptively select usefulinformation in the two generated features of different scales and perform fusion; building a whole neural network model; performing iterative training on the whole neural network model by using the training sample set; and randomly selecting a sample from the test samples as a position category sample, and classifying unknown samples needing to be predicted by using the trained neural network. According to the method, redundant information is reduced, multi-scale features are selected more flexibly, the stability of the network is improved, and then the classification capacity of the network model is improved.

The invention discloses an electroluminescence device, which structurally and sequentially includes a substrate / a conducting layer / a hole injection layer / a hole transmission layer / an electronic blocking layer / a luminous layer / a hole blocking layer / an electronic transmission layer / an electronic injection layer / a cathode layer, wherein the luminous layer adopts a composite layer structure; the composite layer structure sequentially includes a first blue light luminous layer / a first quantum well / a red light-green light phosphorescence luminous layer / a second quantum well / a second blue light luminous layer; the first blue light luminous layer is prepared on the surface of the electronic blocking layer; and the hole blocking layer is prepared on the surface of the second blue light luminous layer. According to the electroluminescence device provided by the invention, the first and the second blue light luminous layers are placed on the two sides of the red light-green light phosphorescence luminous layer respectively, so that the illumination of the red light and the green light can be completely excited by the emitted blue light, the energy utilization ratio is improved, and further the luminous efficiency can be enhanced.

The invention relates to an aloe extractive as methanal trapping agent and a preparation method thereof. The prior art has the problems of large volatility, irritant rare delicacy, secondary harm on human body and unfitness for mass use in the preparation technology of the methanal trapping agent. The invention takes natural aloe as raw material and modifies amylose in aloe gel hydrolysate on amino derivatization and then the amylose and amination collagen protein are compounded for obtaining the methanal trapping agent. Outcome of the invention contains collagen protein modified outcome, aloe protein, the amylose and the like. The amino in molecules is mainly used to eliminate the methanal, has small volatility and almost no rare delicacy and does not cause secondary hurt on human body in the use process. The invention can not only eliminate free formaldehyde in the air and reduce methanal release in man-made boards but also reduce free formaldehyde content in leather and fur. The invention has good application foreground.

The invention discloses a white light electrogenerated light-emitting apparatus, which comprises the structure including a substrate, a conductive layer, a hole injecting layer, a hole transmitting layer, an electron stop layer, a light emitting layer, a hole stop layer, an electron transmitting layer, an electron injecting layer and a cathode layer in sequence; the light emitting layer is a composite layer structure, and the composite layer structure comprises the following layers of a red phosphor light emitting layer, a first interval layer, a blue light emitting layer, a second interval layer and a green phosphor light emitting layer in sequence; and the red phosphor light emitting layer is prepared on the surface of the electron stop layer, and the hole stop layer is prepared on the surface of the green phosphor light emitting layer. According to the invention, light emitting layers of blue light emitting materials, red and green light phosphor materials are applied in the white light electrogenerated light-emitting apparatus, red and green light phosphor light emitting layers are respectively arranged on both sides of the blue light emitting layer, therefore the triplet state energy of blue light can be fully captured, and the light emitting efficiency is further increased.

The invention discloses an asymmetric multi-modal fusion saliency detection method based on an attention mechanism. The method comprises: inputting an RGB image and a depth image of an original three-dimensional image into a convolutional neural network for training to obtain a corresponding saliency detection image; calculating a loss function between a set formed by the saliency detection imagegenerated by the model and a set formed by corresponding real human eye gaze images to obtain an optimal weight vector and an offset term of the convolutional neural network classification training model; and inputting the three-dimensional image in the selected data set into the trained convolutional neural network model to obtain a saliency detection image. According to the saliency detection method, RGB and depth map features are fully extracted by adopting an asymmetric coding structure, rich image information of RGB is effectively utilized after an internal sensing module is added, channel and spatial attention mechanisms are added, the expression of a saliency region and saliency features is enhanced, and the detection accuracy of visual saliency detection is improved.

The invention discloses a swing wing type ocean current energy power generating device. The swing wing type ocean current energy power generating device comprises a supporting frame, a motion mechanism is connected to the supporting frame and connected with an energy extracting mechanism, the energy extracting mechanism comprises a crank and rocker mechanism which transmits energy extracted from the motion mechanism to an energy conversion mechanism, and the energy conversion mechanism comprises a power generator; and the motion mechanism is connected with an attack angle adjusting mechanism, and the attack angle adjusting mechanism adjusts and controls an attack angle degree of the motion mechanism so as to enable the motion mechanism to achieve preset pitching motion. The hydrofoil attack angle adjusting mechanism is arranged, so that the problem in control over the hydrofoil attack angle degree is solved; and as the hydrofoil pitching motion is accurately controlled by monitoring the position of a swing rod, the swing wing type ocean current energy power generating device can adapt to different water flow speeds, kinetic energy of the water flow is sufficiently captured, and the power generation efficiency and power are improved. A transmission system composed of a four-link mechanism and a synchronization belt in a transmission manner is used, so that the structure is simple, mechanical losses are small, and the conversion efficiency is high.

A means and methods are provided to give a gift and simultaneously and unobtrusively record images and / or sound to capture the emotion and excitement of the gift recipient upon receipt of the gift and discovery of the gift-container contents. A gift container is combined with miniaturized video and / or audio recording technologies for use in conjunction with microelectronics and provides a means for recorded data transfer to external devices. Appropriate methods associated with related rentals, sales, and services to the consumer using the gift container are provided.

The invention provides a short message monitoring method comprising the following steps: counting the number of short messages sent by users; during a time granularity, recording user information of the users after the number of the short messages sent by the users reaches the threshold monitored according to the flow within regular duration and recording all short messages sent by the users within regular duration before the ending time of the time granularity is after the time granularity ends; calculating all recorded short messages and marking the users as suspicious users when the calculated value reaches a user level monitoring threshold corresponding to the user level attribute of the users; and displaying the user information of the suspicious users. The invention also provides a short message monitoring system. The invention introduces the user level monitoring based on the flow monitoring and achieves the aim for monitoring high-end users and low-end users at the same time.

The invention discloses an attack perception method, device and equipment based on honeypot induction and a medium. The method comprises the following steps: filtering a captured data package to obtain a first target data package, and outputting the first target data package to a user mode process; determining a target communication induction strategy corresponding to the first target data packetin a user mode process, and communicating with a source Internet protocol address corresponding to the first target data packet based on the target communication induction strategy to capture at leastone second target data packet sent by the source Internet protocol address; and determining the first target data packet and / or the second target data packet corresponding to the first target data packet as a to-be-analyzed data packet, analyzing the to-be-analyzed data packet based on a preset analysis rule set, and analyzing an attack behavior according to an analysis result. The system has thedual advantages of an intrusion detection system and a honeypot system, attack information can be captured more comprehensively, and the safety of the network is protected.

The invention provides a water treatment device and method for cleaning the same, which can reduce the ratio of cleaning water volume to the treated water volume (filtered water volume), thereby can improve water treatment efficiency and cause the structure compacter and simplier. A water treatment device comprising: a jet nozzle (7) for mixed raw water flow; a filtration tank (5) for containing a filtration material layer (4) which comprises two layers which are an upper layer (2) and a lower layer (3) and are arranged below the jet nozzle for mixed raw water flow, the upper layer consisting of filtration materials having a smaller specific gravity and a greater particle diameter than the filtration materials of the lower layer; a water collecting / distributing pipe (13) for extracting filtered water and supplying reverse washing water to the filtration material layer in reverse washing operation; a filtration material vibrating means for vibrating each particle of the filtration materials, the filtration material vibrating means being provided with a vibration element (9) connected to the power source and also with vibration expansion elements (10) embedded in the upper layer (2) in such a manner that the vibration of the vibration element (9) is transmitted to the vibration expansion elements (10); and a drainage trough (12) provided above the filtration material layer.

The invention discloses an osteochondral stent as well as a preparation method and application thereof. The osteochondral stent comprises a cartilage layer, an adhesion layer and a bone layer, wherein two sides of the adhesion layer are respectively connected with the cartilage layer and the bone layer; one or more of the cartilage layer, the adhesion layer and the bone layer are of a porous structure. The osteochondral stent can be used for repairing osteochondral defects, can fully capture cells, has a strong osteochondral injury repairing function, and is simple and convenient to operate.

The invention discloses a microblog friend recommendation method based on a heterogeneous information network. The microblog friend recommendation method comprises the following steps: S1, constructing heterogeneous information network data of a microblog user; S2, extracting meta-paths in the heterogeneous information network, and converting the heterogeneous information network into a homogeneous information network; S3, obtaining vector representation of the user by using a graph convolution network; and S4, calculating the similarity between different users and carrying out similarity fusion. By using the heterogeneous information network and the graph convolution network, rich user behavior information is captured, and high-performance user vector representation is obtained, so that the recommendation accuracy is effectively improved.

An image capturing system utilizes a plurality of image capturing apparatus to capture first image data for an object. A processor is utilized to combine the first images to produce first three-dimensional digital image data. The processor then utilizes a second algorithm to combine the intermediate digital image information to produce a complete three-dimensional digital image of the object. The three-dimensional image may be viewed on a display and viewed from any point on a three dimensional surface surrounding the object. The object can be non-stationary.

The present invention relates to a bone-tendon-bone graft and components. Embodiments of the present invention comprise an intermediate bone block that is used to adjustably secure soft tissue (e.g., tendon) in a patient. Embodiments of intermediate bone blocks of the present invention are used singly or in combination with one or more bone blocks to form a bone block assembly for securing soft tissue. The present invention further relates to an assembled bone-tendon-bone graft for implantation in humans comprising the intermediate bone block and a length of soft tissue. A bone-tendon-bone graft comprises a length of soft tissue extending from a first assembled bone block to a second bone block and then doubles back to said first assembled bone block. Depending upon the embodiment, the second bone block fixedly or slideably attaches to the length of soft tissue and facilitates it doubling back to the first assembled bone block.

The invention relates to a tracheal tree hierarchical extraction method combining a multi-information fusion network and regional growth. The method comprises the following steps: S1, acquiring a CT image of a lung, and preprocessing the CT image; S2, grading the preprocessed CT image set of the lung, and dividing the CT image set into two training sets, namely an overall tracheal tree and a tiny tracheal branch; S3, respectively sampling the overall tracheal tree training set and the tiny tracheal branch training set to obtain an overall tracheal tree training subset and a tiny tracheal training subset; S4, constructing a multi-information fusion segmentation model, and training according to the overall tracheal tree training subset; S5, constructing a voxel classification network model, and training according to the tiny trachea training subset; S6, sequentially inputting to-be-segmented image data into the trained multi-information fusion segmentation model and the trained voxel classification network model to obtain a preliminary tracheal tree; and S7, processing the preliminary tracheal tree by a geometric reconstruction method based on a center line to obtain a final tracheal tree. According to the invention, the classification accuracy is effectively improved.

The invention provides a cellular network traffic prediction method based on a three-dimensional convolutional neural network. The cellular network traffic prediction method comprises the following steps: modeling network traffic data into a three-dimensional tensor input form to obtain a three-dimensional network traffic data model; obtaining training set data and test set data according to the three-dimensional network flow data; constructing a basic three-dimensional convolutional neural network; performing three-dimensional convolutional neural network training on the short-time dependence data to obtain short-time features, and performing three-dimensional convolutional neural network training on the long-time dependence data to obtain long-time features; carrying out fusion training on the short-time features and the long-time features to obtain a feature matrix, and taking the feature matrix as the output of the basic three-dimensional convolutional neural network to form a training model; and predicting the network traffic data to be predicted by using the training model to obtain a network traffic prediction result. According to the prediction method provided by the invention, the short-term correlation and the long-term tendency of the network traffic data are considered at the same time, and the characteristic correlation of the network traffic data in time sequence is captured.

The invention discloses a fault recording device, a decentralized fault recording system and a method. The scheme is a recording device which is freely arranged according to the input channel of analog quantity and switch quantity, dual-network redundant connection is conducted between devices, network synchronization can be carried out through measurable time delay, inter-device network triggering is performed, each device records and sends its own sample data, and also records sample data sent by other devices, and access from any device can view the technical realization scheme of the decentralization of all data. As the invention is different from the traditional fault recording realization scheme, the invention has the following characteristics: configuring by channel number saves screen cabinets, double network redundancy improves the reliability of the connection, the delay measurable network timing saves the connection lines and interfaces, the network synchronous triggering realizes the panoramic capture of disturbing data, the redundant records of each device improve the reliability of the data, and the access to all data at any point solves the problem that the recordingdevices of the protection and control room cannot access each other directly.

The invention relates to the field of natural language processing, in particular to a user abnormal comment detection method based on hierarchical multi-channel attention, which comprises the following steps: collecting user comment data in real time, inputting the collected user comment data into a trained user comment data detection model, and judging whether a comment is abnormal or not; according to the method, tedious feature engineering design in detecting the abnormal comments of the user based on a machine learning method is avoided, the features of the user comment text are analyzed,and the problem that global semantics cannot be captured or cannot be fully captured by a single deep neural network model and a single-layer attention mechanism is solved.