69 results about "Visual patterns" patented technology

A method, apparatus, non-transitory computer readable medium, and system for image processing include obtaining a pattern prompt and a text image, where the pattern prompt describes a visual pattern and the text image depicts text, generating a pattern image based on the pattern prompt, where the pattern image depicts the visual pattern, and generating a patterned text image based on the pattern image and the pattern prompt.

The invention relates to the technical field of power equipment state monitoring, and discloses a cable state fault prediction method based on multi-source data fusion. According to the method, a digital twinborn model is constructed by acquiring multi-source data, and a theoretical health baseline changing along with working conditions is simulated and calculated in real time. And comparing the base line with the measured data to generate a thermoelectric coupling matrix with quantitative deviation, and diagnosing the degradation state according to the thermoelectric coupling matrix. And when the deviation exceeds a threshold value, performing attribution analysis by using a gradient boosting decision tree model, and generating a visual spectrogram associated with the partial discharge and the root cause. And finally, the defect type is judged by intelligently identifying a high-risk visual mode on the spectrogram, and the high-risk state is predicted. According to the method, working condition interference is filtered through dynamic reference, causal diagnosis is performed, the problem of high false alarm rate in the prior art is solved, and the prediction accuracy is remarkably improved.

The invention provides a full-slice image classification method and system based on multi-branch attention and random instance masks and a medium, and the method comprises the steps: obtaining a full-slice digital pathological image, carrying out the image segmentation, and obtaining a plurality of instances; performing feature extraction based on a feature extraction network to obtain an instance feature sequence; carrying out parallel analysis on the instance feature sequence based on a multi-branch attention mechanism, executing a random Top-K instance mask operation, and carrying out weighted summation on the instance feature sequence based on the attention distribution of each attention branch; aggregating the packet level feature representations of all the attention branches, generating a comprehensive feature representation of the full-slice digital pathological image, and obtaining a classification prediction result; by setting a plurality of parallel attention branches, different branches are promoted to actively learn and capture a plurality of different visual modes existing in the full-slice digital pathological image, so that tumor heterogeneity can be effectively represented, and the generalization ability of full-slice digital pathological image classification is improved.

The embodiment of the invention provides an agent-based task interaction method and device, equipment and a storage medium. The method comprises the steps that in response to a received request for indicating a target task, an execution plan of the target task is determined through an intelligent agent, and the execution plan indicates at least one subtask; during execution of the target task according to the execution plan, presenting a content object representing an intermediate execution result of the target task, the content object having an at least partially structured visual pattern, the intermediate execution result comprising at least a portion of a corresponding execution result of the at least one sub-task; in response to receiving feedback for at least a portion of the content object, a response to the feedback is provided with the agent.

A flotation froth video characterization method based on bubble cluster visual pattern statistical modeling comprises the following steps: firstly, performing feature extraction on continuous T frames of froth images by using an OfficientNet encoder to generate a feature map corresponding to each frame; the method comprises the following steps of: firstly, dividing each frame of feature map into a cluster mark sequence with the length of K by using a Patchy module, dynamically grouping cluster marks by using a mask with a U-Net structure to generate a T * K grouping index matrix, and finally introducing a double-branch additive attention Transform module, and the global branch is used for extracting an interactive dependence and overall evolution mode between the bubble clusters. According to the dynamic additive attention strategy provided by the invention, the modeling capability for the long-distance dependency relationship is effectively reserved while the calculation complexity is remarkably reduced, and more accurate visual analysis is provided for monitoring of the flotation process.

The invention discloses a traffic flow prediction method based on cross-modal interactive geographic image coding, and relates to the technical field of intelligent traffic systems and data processing. The method comprises the following steps: firstly, acquiring historical traffic data and a road network geographic image, and respectively constructing an embedded representation containing space-time periodicity and extracting node-level visual features; secondly, through a hybrid cross attention mechanism, utilizing a learnable global visual token as an abstract agent, compressing visual features and performing cross-modal alignment with the dynamic space-time representation to generate an enhanced space-time representation; meanwhile, a visual relation mode is constructed based on local visual patches between the nodes, and refining is carried out through a general relation matrix; and finally, dynamically fusing time, space and visual relation characteristics by using an adaptive gating mechanism, and predicting future traffic flow. Visual modes such as road geometry can be effectively captured, the problem of dynamic and static heterogeneous mode alignment is solved, and the accuracy of traffic prediction is remarkably improved while the calculation efficiency is guaranteed.

The invention discloses a visual language model generation type semantic detection method based on hierarchical feature optimization. Cross-modal semantic alignment and high-quality image synthesis of a text and an image are achieved. Image features are decomposed into multi-scale representation, the decomposition rule is constructed based on a multi-scale token graph, the features are mapped to different resolutions through adjustment operation, and finally a layered feature set matched with the resolution of an original feature graph is formed; the optimization process is limited in a semantically meaningful feature subspace through quantization loss, the feature is prevented from exceeding an effective potential space of a decoder, the quantization loss is realized by calculating a distance mean value between the feature and the quantized feature, and gradient propagation to the quantization step is avoided by adopting a stop gradient operation in the quantization process; the gradient of an image quality evaluation model is introduced, the optimization process is guided to approach a visual mode conforming to human aesthetic appreciation, and the quality evaluation loss is calculated through the mean value of evaluation results of a plurality of quality evaluation models.

The invention discloses an interactive hypnotherapeutic computer-implemented system based on artificial intelligence, comprising a microphone for receiving voice commands from a user; a display device for displaying dynamic pictorial representations generated by the system; a computing device operatively connected to the display device and the microphone. The computing device is configured to execute a computer program implementing a hypnotherapeutic interaction with the user based on received voice commands, whereby the computing device comprises: a word recognition module configured to recognize predefined induction words or phrases from the received voice commands; an artificial intelligence, AI, response generation module trained to respond to the recognized induction words or phrases by generating corresponding dynamic pictorial representations, created by moving particles, on the display device, and configured to record interactions and optimize responses based on user-specific feedback. The system transitions through three stages represented by distinct visual patterns of dynamic pictorial representations, induced by the predefined induction commands, wherein the dynamic pictorial representations vary in parameters defining the behavior of pictorial representation's particles, in particular vary in color and / or speed and / or movement pattern and / or scattering of representations' particles; and wherein the visual representations generated by the system are semiotically inconsistent and do not match the semantic signification of the user's voice commands. The invention also discloses a computer program product comprising instructions that cause the computer system to: a. recognize specific induction words or phrases from voice commands received via a microphone; b. in response to the recognized induction words or phrases, generate corresponding dynamic pictorial representations, comprising particles, on a display device using an artificial intelligence response generation module, the pictorial representations being distinct for three stages of the system and varying in color, speed, movement pattern and scattering of representations' particles; c. record interactions and optimize responses based on user-specific feedback; wherein the visual representations are semiotically inconsistent and do not match the semantic signification of the user's voice commands.

This invention discloses a method and apparatus for identifying auroral substorms based on visual eye-tracking patterns and deep learning, primarily addressing the problems of low accuracy and lack of basis for model design in existing auroral substorm identification methods. The implementation scheme includes: collecting eye-tracking data from multiple space physicists to generate expert eye-tracking gaze pattern maps for various types of auroral substorms; establishing an auroral substorm eye-tracking dataset; further designing a visual eye-tracking pattern prediction module to learn the experts' visual gaze patterns; and setting up a substorm sequence recognition module embedded with visual patterns, incorporating space physics knowledge and expert visual gaze patterns. Finally, a deep learning model suitable for auroral substorm sequence recognition is constructed and trained, and the trained model is used to identify auroral sequences. This invention not only enhances the representation of auroral substorm features but also improves the accuracy and rationality of auroral substorm image sequence recognition, and can be used for automated space physics event recognition.

The application discloses a kind of unmanned plane vision enhancement system and method based on binocular rotating polarizer, it is related to computer vision and optical imaging technical field, system includes binocular camera module, rotatable polarizer assembly that can be integrally moved into / out of optical path, mechanism for driving its lifting and rotation and main control unit.The method comprises: after moving polarizer into optical path, in a single acquisition cycle, it is driven to rotate at high speed to multiple preset angles and synchronously acquires binocular image sequence;Process image sequence to calculate Stokes parameters and construct three-dimensional polarization feature space in combination with depth information;The feature space is input into multimodal fusion neural network with original image, and enhanced image, semantic segmentation map and optimized depth map are output.The application realizes the quick and flexible switching of polarization mode and ordinary vision mode, and significantly improves the real-time vision perception and enhancement capability of unmanned plane in underwater, strong reflection and other complex environments through software and hardware depth collaboration.

A system comprising an output device, a brain oscillation monitor, and one or more processors. The output device is configured to output an audio signal for audio stimulation of the patient, and a visual pattern for video stimulation of the patient. The brain oscillation monitor is configured to generate feedback indicative of a response to the audio stimulation and the visual stimulation. The one or more processors are configured to determine a target frequency for the audio stimulation and the visual stimulation. The one or more processors are further configured to cause constructive interference of the audio stimulation and the visual stimulation output by the output device at the target frequency, by modifying at least one of a phase of the visual pattern or an amplitude of onsets for the visual pattern, relative to the audio signal, according to the feedback from the brain oscillation monitor.

The invention discloses a vehicle visual mode switching method, device and equipment and a computer program product. The method comprises the steps that the real-time vehicle distance, the threat fuzzy coefficient, the real-time vehicle speed, the traffic jam condition and the dynamic balance level of a current vehicle and the driving habit style of a driver are obtained; inputting the real-time vehicle distance, the threat fuzzy coefficient, the real-time vehicle speed, the traffic jam condition, the dynamic balance level and the driving habit style into a pre-trained visual mode decision model to obtain a target visual mode; switching the visual mode of the current vehicle to a target visual mode, and controlling a corresponding target visual sensor to operate; wherein the target visual mode is one of a vector-level visual mode, a contour-level visual mode and a high-precision visual mode. On the premise of ensuring the driving safety of the vehicle, the computing power resource consumption of the vehicle is optimized, so that the performance of vehicle hardware is improved, the service life of the vehicle hardware is prolonged, and the method can be applied to the technical field of intelligent driving.

The invention discloses an unmanned aerial vehicle building monitoring system and method based on binocular vision and deep learning, and the system is characterized in that an unmanned aerial vehicle carries a flight control system, an image collection module and a wireless communication module; the ground computing equipment comprises a task planning module, an image recognition and three-dimensional reconstruction module and a displacement analysis and alarm module; after the standardized visual pattern is arranged, the task planning module automatically generates a flight route according to the position of the pattern, the unmanned aerial vehicle executes the flight route and collects a synchronous image, the image is sent into the deep learning model to recognize the pattern, a three-dimensional coordinate is reconstructed, displacement analysis is performed after comparison with an initial state, and an alarm is triggered. According to the invention, the unmanned aerial vehicle platform carries the high-resolution camera for dynamic acquisition, and the deep learning algorithm is used for analysis, so that the depth perception capability of binocular stereo vision, the dynamic flexibility of the unmanned aerial vehicle platform and the environment robustness of deep learning are fused, and non-contact, full-dimension and high-economy building deformation monitoring is realized.

A tamper-evident item authentication label may include a substrate layer that includes a first substrate side and a second substrate side, a first adhesive layer, a second adhesive layer, a first unique indicium, a second unique indicium, and a plurality of cuts across the first unique indicium and the second unique indicium extending through at least a portion of a thickness of the substrate layer. The adhesive layers may be deposited on the substrate sides and arranged to fuse the tamper-evident item authentication label with a first and second surface. The first unique indicium may be deposited on the substrate layer and encoding a first unique identifier that links the tamper-evident item authentication label to one or more records in a remote database system. The second unique indicium may be deposited on the substrate layer. The second unique indicium may include a stochastically created visual pattern.

A digital image radial pattern decoding system is described. In one example, an unfolded digital image is formed by the radial pattern decoding system by unfolding a radial pattern in a digital image. An inflated digital image is then generated by the radial pattern decoding system by upsampling the unfolded radial pattern. A grid pattern is determined by the radial pattern decoding system based on the inflated digital image. A radial pattern cell is then generated based on a reverse transform of the grid pattern. A visual pattern is generated by the radial pattern decoding system based on the radial pattern cell.

An eye exam can be performed using an electronic device in a virtual environment to determine vision corrective measures based on vision correction simulation. The electronic device can execute a visual assessment application for displaying a user interface to create a 3D virtual environment corresponding to a field of view of a user associated with the electronic device. The electronic device can render a visual pattern in the field of view and apply a vision correction filter to the visual pattern. The electronic device can obtain a set of user response data captured by a plurality of sensors in response to the visual pattern and determine whether the set of user response data satisfy a response quality criterion. Filter parameters of the vision correction filter can be dynamically adjusted based on the set of user response data until the set of user response data satisfy the response quality criterion.

A patient's visual health can be evaluated via a virtual reality (VR) system, which can include a VR headset in electronic communication with a computing device. The computing device can cause a phenomenon (such as a visual test or an image) to be displayed on the screens of the VR headset. Using varying combinations of sensors, cameras, probes, and microphones, the VR headset collects data about the patient as she perceives and responds to the phenomenon. Optionally, the computing device can alter the phenomenon and analyze the patient's perception and responses to evaluate the patient for macular degeneration. Optionally, the computing device can alter the phenomenon based on the patient's perception and responses.

Real time pattern preview generation and capture techniques are described. In an example, a live stream of digital images is displayed in a user interface by a computing device. A real time preview of visual patterns in the user interface is then generated and displayed based on the digital images. The visual patterns, for instance, are generated by the computing device in real time using a combination of shape extraction and pattern generation. An option is also made available to convert the real time preview into a vector image, such as a vector pattern tile.

The invention discloses automatic packaging equipment which comprises a belt lifting conveyor, a vertical film rolling bag-making packaging machine and an intelligent counting and feeding device. The intelligent counting and feeding device comprises a first belt conveyor, a second belt conveyor, a quantitative transferring mechanism used for receiving plastic parts output by the first belt conveyor and transferring the plastic parts to the second belt conveyor, and a visual figure counting system used for shooting and calculating the plastic parts conveyed by the second belt conveyor. The visual pattern counting system comprises image acquisition equipment, an industrial computer for storing and processing acquired images, and a PLC (programmable logic controller) communicated with the industrial computer, a software program built in the industrial computer can identify and calculate the number of plastic parts of the images, and control instructions are sent to the PLC according to the number of the plastic parts. The PLC controls starting, stopping and the working speed of the first belt conveyor, the second belt conveyor and the quantitative transfer mechanism according to the received control instruction. The plastic part counting and packaging device has the advantages that the effect of automatically and accurately counting and packaging plastic parts is achieved.

The utility model provides a small-size folding light path visual training device which comprises a front shell, a middle frame and a rear shell which are connected in sequence, the front shell is provided with a lens device, and a visual training module, a left darkroom, a right darkroom, a flexible FPC and a reflection lens are arranged between the middle frame and the rear shell; the visual training module, the reflecting lens, the left darkroom, the right darkroom and the lens device jointly form a light path system. According to the visual training device with the small-size folding light path, the problems that an existing device is large in size, close in imaging position and poor in experience are solved, the visual training device is heavy to wear, the left eye and the right eye can see the same visual pattern through the spirally-arranged LED light sources, the left eye and the right eye share one visual cylinder, the size is saved, and the imaging distance is increased.

A system for producing an image to design a product can include a processor and a memory. The memory can store a regularizing module, a blending module, a denoising module, and a communications module. The regularizing module can produce a regularized image of a denoised image of an interpolation of a first diffused image and a second diffused image. The regularized image can be regularized with respect to a visual pattern. The blending module can: (1) determine a blending weight and (2) produce, based on the blending weight, a blended image of the denoised image and the regularized image. The denoising module can denoise the blended image to produce the image to design the product. The communications module can cause the image to be sent to a computer-aided design system to design the product.

The invention relates to an unmanned aerial vehicle intelligent path planning method and system based on visual signals, and belongs to the field of path planning. The method comprises the following steps: optimizing a neural network through progressive pruning; fuzzy coordinates and scene types are provided for the unmanned aerial vehicle, the unmanned aerial vehicle flies to a fuzzy coordinate area, and when the unmanned aerial vehicle is close to the fuzzy coordinate area, the unmanned aerial vehicle is switched to a pure vision mode to perform environment perception through a neural network, and a three-dimensional semantic map is constructed; performing real-time path planning based on the three-dimensional semantic map to obtain a flight path; the method comprises the following steps: performing autonomous flight along a flight path, when approaching an operation target, performing autonomous operation to obtain real-time monitoring data, storing the real-time monitoring data in airborne storage equipment, and returning the real-time monitoring data after returning. And an unmanned aerial vehicle intelligent path planning system with high robustness, high autonomy and high task adaptability is constructed.

The invention relates to the technical field of training equipment, and discloses a multi-modal stimulation mouse rich environment training device, which comprises a box body, a training cage and a visual stimulation assembly, the training cage and the visual stimulation assembly are arranged in the box body; the visual stimulation assembly comprises a face gear and a plurality of visual moving assemblies, the face gear is provided with a meshing end face facing the training cage, and the outer periphery of the training cage is meshed with the meshing end face; the multiple visual moving assemblies are arranged on the meshing end face at intervals. The visual movement assembly is movably arranged in the radial direction of the meshing end face, and the visual stimulation assembly further comprises a resistance adjusting mechanism used for adjusting resistance of the visual movement assembly during radial movement. The visual moving assembly is movably arranged in the radial direction of the face gear, and visual patterns seen by a mouse not only translate, but also continuously jump, shake or pulse, so that the cognitive load of training is greatly increased.

The application belongs to the technical field of computer vision, and provides a visual Transformer self-supervised learning method and system based on multi-dimensional relationship modeling; the method uses self-relationship modeling on spatial dimensions and channel dimensions, uses different image transformation to process images to obtain different views of the images; the different views of the images are processed by a teacher network and a student network respectively to obtain feature maps; the feature map extracted by the student network is further processed by a convolution layer; a self-relationship matrix of the feature map in the spatial dimensions and the channel dimensions is calculated by point multiplication; the difference between the relationship matrices extracted by the teacher network and the student network is calculated as a loss function, and the derivative of the loss function with respect to the network parameters is used to update the network parameters; compared with the existing self-supervised learning method which only considers the features of visual patterns, the method simultaneously considers the correlation of visual patterns in the spatial and channel dimensions, and can significantly improve the accuracy of image recognition, semantic segmentation, target detection, instance detection and the like.

An image processing system includes: an industrial machine; a visual sensor; a visual pattern arranged on a part to be detected, for identifying a relative positional relationship between the visual sensor and the visual pattern; an image processing unit that determines the relative positional relationship between the visual sensor and the visual pattern on the basis of an image of the visual pattern captured by the visual sensor; and a focusing operation control unit that, with the visual pattern within the field of view of the visual sensor, causes the visual sensor to capture the visual pattern while changing the relative positions of the visual sensor and the visual pattern in a predetermined direction with reference to the visual pattern, on the basis of the determined relative positional relationship to obtain a focus degree of the visual pattern, whereby the visual sensor is focused on the visual pattern.

The invention discloses an unmanned aerial vehicle vision enhancement system and method based on a binocular rotating polarizer, and relates to the technical field of computer vision and optical imaging, and the system comprises a binocular camera module, a rotatable polarizer assembly capable of integrally moving in / out of a light path, a mechanism for driving the rotatable polarizer assembly to lift and rotate, and a main control unit. The method comprises the following steps: after moving a polarizer into a light path, driving the polarizer to rotate at a high speed to a plurality of preset angles in a single acquisition period and synchronously acquiring a binocular image sequence; processing the image sequence to calculate a Stokes parameter and constructing a three-dimensional polarization feature space in combination with depth information; and inputting the feature space and the original image into a multi-modal fusion neural network, and outputting an enhanced image, a semantic segmentation image and an optimized depth image. According to the invention, rapid and flexible switching between a polarization mode and a common visual mode is realized, and the real-time visual perception and enhancement capability of the unmanned aerial vehicle in complex environments such as underwater and strong reflection is significantly improved through deep cooperation of software and hardware.

The invention provides a portable printing device and a portable automatic printer using the same. The portable printing device comprises a frame, a drawing type paper storage mechanism, a thermal printing mechanism and a paper feeding mechanism. The frame is provided with a paper storage cavity and a mounting cavity which are communicated and sequentially arranged in the paper feeding direction. The drawing type paper storage mechanism is configured to extend out of the paper storage cavity or retract into the paper storage cavity in the paper feeding direction and is used for being matched with the frame to contain printing paper. The thermal printing mechanism is arranged in the mounting cavity and is used for printing to form a visual pattern; the paper feeding mechanism comprises a paper pickup assembly located between the drawing type paper storage mechanism and the thermal printing mechanism, the paper pickup assembly comprises a first paper pickup wheel, a second paper pickup wheel and a paper separation wheel, the first paper pickup wheel drives a plurality of pieces of printing paper to move to the position between the second paper pickup wheel and the paper separation wheel, and the second paper pickup wheel drives one piece of printing paper to enter the thermal printing mechanism. And the paper separating wheel stops redundant printing paper through frictional resistance. The portable automatic printer comprises an embedded control unit and a portable printing device.

The invention discloses a man-machine cooperation continuous image compression method and system based on a hybrid expert adapter, and the technical scheme is that the method comprises the steps: generating and preprocessing a general feature data set and a downstream task data set, and building a basic coding and decoding network; constructing a hybrid expert adapter containing a public expert module and a task expert module in parallel, wherein each expert internally comprises a bottleneck type lightweight structure and spatial frequency domain double branches; a man-machine cooperation decoupling architecture is constructed, two-stage training is adopted, and a parameter isolation strategy is implemented in the second stage; the modes can be dynamically switched during reasoning, wherein the mixed expert adapter needs to be removed in the human eye vision mode, and the corresponding experts are dynamically activated according to task identification in the machine vision mode. According to the method, the human eye vision quality and the machine vision task performance are considered, multi-task zero-forgetting continuous learning is supported, and the downstream machine vision task analysis precision is improved.

According to the universal autonomous charging method and system for the mobile robot and the storage medium, a positioning target is constructed into a fixed geometrical shape, a reflective substrate is installed, an observable dark visual pattern is attached to the surface of the positioning target, and the type of an environment sensing sensor running on the current robot is obtained; selecting corresponding data calculation and processing modes for the acquired data according to different sensor types; and finally, generating a charging motion control instruction according to the calculated space coordinates and course angle data or pixel error coordinates to control the mobile robot to move to a specified position above the charging pile for charging. The technical effect that mobile robots with different sensor configurations can be quickly adapted without replacing a positioning target and charging software is achieved, and the development and deployment cost of the robot charging system is effectively reduced.