18results about How to "Achieve interaction" patented technology

The application provides a blind guiding method and system based on real-time interaction and scene recognition, divides a route into path points and path lines, takes an entry point of a path line as a navigation focus, obtains scene information through a camera arranged on glasses, detects a current scene through software, interacts with local software through a button arranged on a blind stick when needed, and is silent for detection in most time, does not interfere with self detection and judgment of a user on a road, and has the advantages of simple structure, convenient use and low cost.

This application discloses a data-aware intelligent analysis method, system, storage medium, and computer device. The method includes: receiving a target problem and initializing state information, wherein the state information includes key sub-conclusions, a reasoning state logic diagram, and key data descriptions; a language model-based agent iteratively performs the following operations: obtaining input information for the current step; updating the state information of the current step based on the action and environmental feedback information of the previous step, and making decisions based on the state information of the current step to obtain the action and tools to be invoked for the current step; invoking the tools to execute the action of the current step, and obtaining environmental feedback information for the current step after the action is completed; checking whether the target problem has been analyzed; if not, proceeding to the next loop; if completed, generating analysis results. This method can improve the reasoning ability, response consistency, and interpretability of deep reasoning models.

The utility model provides three operating system detection data processing equipment, including the casing cover and the power module, network switch module, three main control system modules, three hard disk extension modules and three communication extension modules that are all arranged in casing cover, the casing cover has the backboard for power distribution and data interaction between each module, and the power module, network switch module, each main control system module, each hard disk extension module and each communication extension module all take this backboard as the basic connection carrier, the power module is connected network switch module, each main control system module, each hard disk extension module and each communication extension module through the backplate, network switch module is connected with each main control system module through the backplate, and each main control system module is connected with a hard disk extension module and a communication extension module respectively through the backplate, and the multi -source data parallel processing is realized through the construction three operating system framework, and the space burden and power consumption burden of the application scene such as railway detection vehicle are reduced.

The application provides an interactive method and device based on a virtual role, equipment, a medium and a product, applied to the field of artificial intelligence technology, and the method comprises the following steps: when it is detected that a client plays video data, a target video emotion vector corresponding to a current playing time point of the video data is acquired from a video emotion vector sequence generated offline; a user emotion vector collected in real time is acquired; a preset AI language model is called, interactive information is generated according to the target video emotion vector, the user emotion vector and a prompt word template for a video watching scene, and the interactive information is sent to the client so as to feed back the interactive information through a virtual role in the client. Through the embodiment of the application, the virtual role is combined with the offline generated video emotion and the real-time collected user emotion to interact with the user, a companion watching scheme is provided for the user, and the immersive watching experience of the user is improved.

This invention discloses a progressive multimodal remote sensing image super-resolution reconstruction method and apparatus. The method includes: inputting a low-resolution image; extracting primary modal features and secondary modal features from the low-resolution image; inputting the primary and secondary modal features into a context-aware adaptive multimodal alignment module, and processing the input modal features to obtain processed secondary modal features; fusing the processed secondary modal features at different levels through a progressively gated fusion mechanism to obtain the final fused features; inputting the final fused features into a preset reconstruction module to obtain the final super-resolution reconstructed image; and outputting the final super-resolution reconstructed image. This method employs a context-aware adaptive multimodal alignment module to achieve interaction and alignment of multimodal features; and uses a progressively gated fusion mechanism to achieve progressive reconstruction from local detail enhancement to high-level semantic guidance.

The application provides a kind of multi-modal sensitive label identification method and system for short video content security, anchors short video sensitive representation subject and disassembles to form multidimensional fine-grained representation unit, excavates the content association logic chain between multidimensional fine-grained representation unit, obtains relevant set;Perform cross-dimension feature interlacing and link fusion processing to construct sensitive association feature network;Sensitive association feature network is matched with preset sensitive label item set to generate initial anchoring association relationship;Then, the initial anchoring set is executed feature network reverse tracing and label anchoring counteracting calibration processing, and the anchoring association deviation is corrected;Based on the calibrated label feature node anchoring set, extract sensitive label and execute label interrelation aggregation and hierarchical division processing, output the sensitive label set corresponding to short video.The application comprehensively utilizes short video multi-modal information, and improves the accuracy of sensitive label identification.

A dynamic time-series data prediction method based on static knowledge graphs, belonging to the fields of artificial intelligence and industrial IoT, is proposed. This method constructs a graph model embedding representation of the static attributes of industrial equipment and uses it as a query vector in an attention mechanism. This vector is then cross-attentionally calculated with real-time operational data collected during equipment operation to achieve dynamic weight allocation for specific equipment. Through this mechanism, the system can adaptively adjust the importance of real-time control commands (such as parameter adjustments and switching operations) based on the static background information of different equipment (such as model and design parameters), generating context-aware control representations. These representations are then fused with time-series features such as sensor data and operational events and input into a sequence prediction model, thereby improving the prediction accuracy of future key performance indicators of the equipment. This method balances the synergistic effect of static knowledge and dynamic signals, effectively enhancing the model's sensitivity to individual differences among different equipment and reducing prediction bias.

The application discloses a three-dimensional feature representation-based intelligent text reading comprehension method and device, a storage medium and an electronic device, and belongs to the fields of natural language processing and artificial intelligence; the technical problem to be solved by the application is how to capture direct interaction features between three-sequence data and how to enhance sufficient interaction among an article, a question and options, so as to improve the prediction accuracy of an intelligent text reading comprehension system; the technical scheme adopted is as follows: ① a three-dimensional feature representation-based intelligent text reading comprehension method, comprising the following modules: a pre-training embedding representation module, a feature filtering module, a 3D CNN interaction feature module and a label prediction module; ② a three-dimensional feature representation-based intelligent text reading comprehension device, comprising: an intelligent text reading comprehension data set acquisition unit, an intelligent text reading comprehension model construction unit and an intelligent text reading comprehension model training unit.

The application provides a semantic iterative fusion depth prediction method and system based on monocular depth estimation, and an automatic driving environment perception field, which comprises the following steps: taking original bottom visual features and global text semantic embedding vectors as multi-modal prompts, injecting each layer of a pre-trained encoder through a cross attention mechanism to obtain multi-modal output; performing multi-scale feature aggregation on the multi-modal output to obtain compact single-scale aggregated features; inputting the global text semantic embedding vectors and the compact single-scale aggregated features into a semantic guidance attention module to obtain semantic alignment visual features; inputting the semantic alignment visual features into an iterative semantic fusion module for iterative geometric calibration to obtain iterative optimized features; and inputting the iterative optimized features into a semantic fusion decoder to obtain monocular depth prediction results of automatic driving. The method can guarantee the real-time performance of automatic driving environment perception and improve the depth prediction accuracy.

The utility model relates to a multifunctional EPD visual acuity detector, including EPD display screen, aluminium frame and back cover, EPD display screen is connected in the front end of aluminium frame, the back cover is fixed in the back of aluminium frame, and the space inside surrounded between back cover and aluminium frame is equipped with control mainboard, EPD display screen is electrically connected with control mainboard, and EPD display screen can switch and show different visual acuity detection image, the upper portion of aluminium frame is equipped with respectively with the camera, microphone, speaker and infrared sensor of electric connection of control mainboard, network module is connected on control mainboard, and this network module is connected with AI cloud end server communication, the utility model discloses through the integration camera, microphone, speaker and infrared sensor, can realize the interaction with the user, realizes the AI intelligent vision detection, replaced the traditional light box test vision's mode, makes the test more efficient.

The embodiment of the disclosure provides a cross-EDA software interaction method, applied to a first host running first EDA software, and the method comprises the following steps: calling an external command in the first EDA software to start a second EDA software; obtaining connection identification information of a first target host; wherein the first target host is a host running the second EDA software; opening an interaction switch in the first host, and obtaining interaction information files corresponding to the first target host based on the connection identification information, wherein the interaction information files comprise a host number and a connection port number; creating a first communication interface, and sending a connection request to the first target host based on the interaction information files, and establishing a communication connection with the first target host through the first communication interface.

The application relates to an aviation laser point cloud semantic segmentation method and device based on a multi-level context feature fusion network, which comprises the following steps: S1, data down-sampling preprocessing is performed on acquired target region point cloud; S2, the processed aviation laser point cloud data is input into a pre-established multi-level context feature fusion network coding layer, and each layer feature is input into a cross-layer attention module to obtain cross-layer connection features; S3, the coding features are up-sampled to the size of the point cloud to be segmented through a decoding layer, multi-level feature fusion modules are used for fusing the features of all levels, and then a Softmax function is used to obtain the final segmentation result of the aviation laser point cloud; and S4, a hybrid loss function is constructed to obtain the error between the segmentation result and the true value, and according to the error value, a random gradient descent method is used for back propagation training, model parameters are dynamically adjusted, and a stable segmentation result is obtained. The method and device can effectively enhance the association of point feature expression and spatial information, improve the information perception ability of multi-scale context features of the point cloud, and obtain a more fine semantic segmentation result.

This invention belongs to the field of drug metabolism prediction technology, specifically relating to a method and system for predicting drug metabolites based on conditional generative adversarial networks (GANs). The method includes: extracting initial features of drug molecules and masking them, inputting them into a shared encoder to obtain a first encoding result; incorporating atomic position information into the first encoding result, decoding it, and then predicting the structure of the drug molecule; inputting reactant molecules into the shared encoder, inputting the obtained reactant features into a shared decoder, and predicting chemical reaction products; inputting substrate molecules into the shared encoder for encoding; fusing the substrate encoding result with the corresponding enzyme's amino acid sequence and inputting it into the shared decoder to predict metabolites; constructing a generator using the shared encoder and shared decoder, building a multi-task discriminator, and performing adversarial training on the generator and discriminator to obtain a prediction model for predicting metabolites. This invention improves prediction accuracy by fusing information from drug structure and enzymes to construct a prediction model.

This invention relates to a semi-supervised method for removing metal artifacts in multi-energy spectral CT images. Based on a semi-supervised average student-teacher model, the U-net network structure incorporates a slice feature attention module, enhancing the perception of deep feature information between channels and improving the feature connections between different monoenergy maps of multi-energy spectral CT slices. Simultaneously, in unsupervised learning, the student model, combined with voxel-level contrastive learning, makes the model more focused on artifact regions, enabling artifact information interaction between low-energy and high-energy maps. This allows for better removal of artifacts in low-energy maps while preserving more tissue detail information. Consequently, it provides high-quality CT images for subsequent medical diagnostic tasks such as image classification and segmentation.

The application belongs to the technical field of electromagnetic launching aircraft, and relates to an online speed measurement feedback control method and system for an electromagnetic launching aircraft. The method comprises the following steps: acquiring electromagnetic launcher parameters and aircraft parameters, constructing an output speed equation of the electromagnetic launcher, constructing an output speed equation of the aircraft, and constructing a speed state equation of the electromagnetic launching aircraft; obtaining a speed observation vector according to the output speed equation of the electromagnetic launcher and the output speed equation of the aircraft; obtaining a speed fusion equation according to the speed observation vector and the speed state equation of the electromagnetic launching aircraft, constructing a Kalman filter, taking the speed observation vector as a filter observation, estimating a speed error of the electromagnetic launching aircraft, and performing speed feedback according to the estimation result. The method can meet the needs of different launching speeds of various aircrafts and improve the control precision.

This invention provides a dynamically reconfigurable binning method, comprising: acquiring logic units in a reconfigurable region and a static region respectively; detecting interaction signals connecting the static region and the reconfigurable region; and creating input logic units and output logic units with interface connections between the static region and the reconfigurable region based on the interaction signals. By parsing the input and output ports of the reconfigurable region module, a data interaction interface is generated to realize interaction between the static region and the reconfigurable region. This allows for the rational layout and routing of individually packaged reconfigurable logic unit blocks and the generation of dynamically reconfigurable code streams, overcoming the difficulty of achieving dynamic local reconfigurability and improving system flexibility and compilation efficiency. The reconfigurable module and FPGA chip provided by this invention have corresponding advantages.

The application provides an alternator, which comprises a motor shell, a stator assembly arranged in the motor shell, a rotor assembly arranged in the stator assembly and connected with the motor shell, a rectification assembly arranged on one side of the motor shell, a heat dissipation fan arranged on the other side of the motor shell, and a middle shaft for connecting the rectification assembly, the rotor assembly and the heat dissipation fan, characterized in that the motor shell is symmetrically provided with a rear end cover and a front end cover on both sides, respectively, and a belt transmission wheel is fixedly arranged at one end of the middle shaft and extends to the outside of the front end cover. The filter mechanism is arranged, so that the particulate matter is blocked on the surface of the annular blocking net when the heat dissipation fan rotates to send the external air into the motor shell, and the particulate matter blocked on the annular blocking net does not completely cover and accumulate on the surface due to the annular structure of the annular blocking net, so that the heat dissipation fan can interact with the external air to the maximum extent.

This invention discloses a method for interactive emotional expressions with a voice robot, comprising the following steps: Step S100: capturing audio information; Step S200: identifying the emotion type of the audio information; Step S300: outputting response information corresponding to the emotion type of the audio information according to a preset voice robot persona; Step S400: identifying the emotion type of the response information and outputting facial expression action instructions corresponding to the emotion type of the response information; Step S500: executing facial expression actions according to the facial expression action instructions. This invention achieves multi-dimensional interaction between the voice robot and the user's emotions, providing the user with a pleasant experience of having a close friend in the car.