124 results about "Noise generator" patented technology

The invention provides an adversarial digital camouflage generation method, device and equipment, and the method comprises the steps: carrying out the encoding of an obtained target background image based on an encoder, and forming a first hidden variable; performing denoising processing on the first hidden variable in a hidden space based on a denoising diffusion model to form a second hidden variable; processing the second hidden variable based on a decoder to generate a digital camouflage image; processing the first hidden variable based on an adversarial noise generator to generate an adversarial noise image; and obtaining an adversarial digital camouflage image corresponding to the target background image based on the combination of the camouflage image and the adversarial noise image. According to the adversarial digital camouflage generation method provided by the invention, the adversarial digital camouflage image can be quickly and accurately generated, so that the camouflage effect is good and the adversarial digital camouflage image is not easy to find when facing a neural network classification model.

A system comprises a plurality of regions, wherein ones of the plurality of regions are partitioned from others of the plurality of regions and at least one of the plurality of regions is a region under test. The system comprises at least one noise generator configured to generate noise in at least the region under test, and at least one noise monitor configured to monitor one or more effects of the noise generated in the region under test. The system comprises a test controller configured to: cause the at least one noise generator to generate the noise in at least the region under test; receive information from the at least one noise monitor indicative of the one or more effects of the noise generated in the region under test; and determine one or more conditions based on at least a portion of the received information.

The invention provides an active information security encryption control method and system for a space-ground integrated micro-grid group, and relates to the field of electric power information security, and the method comprises the steps: abstracting a distributed power supply into an inverter in the micro-grid group, and carrying out the modeling of the inverter to generate a micro-grid group state-space equation; establishing a security encryption controller equation based on differential privacy; establishing a noise generator of a discrete time sequence; introducing a random approximation method to generate a distributed protocol; generating a self-adaptive feedback mechanism; updating a security encryption controller equation according to a noise generator, an adaptive feedback mechanism and a distributed protocol; and solving the second coefficient matrix in the current security encryption controller equation according to the first coefficient matrix in the microgrid group state space equation to generate a target security encryption controller equation. Through a dynamic adjustment mechanism, balance is found between privacy protection and control precision, so that the encryption and decryption capability can be matched with the actual operation condition of the microgrid group.

The invention relates to an electromagnetic signal transmission method and system and a computer readable storage medium, and the method comprises the steps: inputting an obtained electromagnetic signal to a first noise generator, carrying out the noise addition of the electromagnetic signal based on the first learning interference resistance noise of the first noise generator, and generating a noise interference electromagnetic signal; the first noise generator inputs the obtained first training data into the initial neural network model, injects the first preset interference noise into each network level in the initial neural network model, and performs multi-level fusion processing on the first preset interference noise and the first training data through each network level; iteration training generation is carried out; the first anti-learning interference noise is obtained by performing minimum learning on the first preset interference noise in the iterative training process of the initial neural network model; and transmitting the noise interference electromagnetic signal to a signal receiving device. According to the invention, the problem of low security of electromagnetic signal transmission is solved.

The invention relates to a semantic segmentation method, a model training method, a model and a system. The semantic segmentation model comprises an image encoder, a noise generator, a text encoder, a splicing module, an attention module and a submerged space decoder, wherein the output end of the image encoder is connected with the first input end of the splicing module; the output end of the noise generator is connected with the second input end of the splicing module; the output end of the text encoder is connected with the first input end of the attention module; the output end of the splicing module is connected with the second input end of the attention module; the output end of the attention module is connected with the input end of the submerged space decoder; wherein the hidden space decoder is used for decoding the denoising mask potential features, and generating a prediction target mask used for segmenting a target object in the image. According to the invention, the accuracy of image segmentation is improved.

The invention belongs to the field of environmental noise monitors, and particularly relates to an environmental noise monitor remote automatic calibration device based on wireless communication, which is assembled on a mounting box, the mounting box is provided with a movable opening, the movable opening is provided with a traction block, the traction block is provided with a noise generator, the traction block is provided with an operation block, and the operation block is provided with an operation opening. The operation block is provided with an operation motor, the operation motor is provided with a rotating rod, the rotating rod is provided with a traction gear, the operation block is provided with a fixing rod, the fixing rod is provided with a limiting plate and a guide block, the guide block is provided with a pressing rod, the guide block is provided with a vertical block, the vertical block is provided with a traction rack, and the traction rack and the traction gear are assembled. According to the environmental noise monitor remote automatic calibration device based on wireless communication, the problem that noise data of an existing noise monitor is abnormal is solved through cooperation of the pressing rod and the traction rack.

A system including a transmitter and a receiver. The transmitter has a human machine interface and a plurality of magnets. The receiver has a noise generator and a vibration motor. The receiver has a first state and a second state. The receiver includes a proximity sensor. The receiver is configured to change from the first state to the second state in response to detection of a magnetic field by the proximity sensor. The transmitter is pairable with the receiver when the receiver is in the second state. The noise generator and the vibration motor are each operable by the human machine interface when the transmitter and the receiver are paired. The transmitter is magnetically couplable to a user device by the plurality of magnets.

A sound enrichment system for provision of tinnitus relief, the sound enrichment system includes a noise generator, at least one signal modulator for random or pseudo-random modulation of a noise signal that is obtained using the noise generator, and an output transducer for conversion of the modulated noise signal to an acoustic signal for presentation to a user. A method of providing a noise enriched sound signal for provision of relief of tinnitus includes generating a randomly or pseudo-randomly modulated noise signal, generating an acoustic noise signal using the modulated noise signal, and presenting the acoustic noise signal to a tinnitus suffering person.

The invention provides a lip synchronization method and device based on diffusion Transform, equipment, a medium and a program product, and relates to the technical field of artificial intelligence processing. The method comprises the following steps: acquiring a source video frame, a target text and a target audio; performing noise adding processing on the source video frame by using a stream matching noise generator to generate a noisy video frame; performing global lip shape synchronization processing based on the noisy video frame, the target text, the target audio and the trained lip synchronization diffusion model, and outputting a lip synchronization video frame; wherein the trained lip synchronous diffusion model is obtained by maskless training based on a video sample, a text sample and an audio sample. According to the embodiment of the invention, noise is added through the stream matching noise generator, the inter-frame jitter phenomenon is reduced, the lip synchronization video frame can be smoothly generated through the trained lip synchronization diffusion model, and the cross-language adaptation capability is improved.

The invention discloses a language learning evaluation method based on a deep learning algorithm, and relates to the technical field of language processing, and the method comprises the steps: receiving a voice signal and original text content, constructing a noisy voice sample through an adversarial disturbance generator, inputting a pre-trained voice recognition model, and outputting a voice recognition text and an anti-noise confidence score; constructing a context semantic graph in the semantic analysis path, generating an edge attention weight, calculating a semantic coherence index, and matching a predefined grammar rule base in the activated grammar analysis path to generate a grammar error tag; and dynamically generating feedback content based on the multi-dimensional evaluation matrix, and updating a preset corresponding threshold value and a rule base weight according to the adoption of the user on the feedback content. According to the method, collaborative quantitative analysis of pronunciation, semantics and grammar problems is realized by constructing a multi-dimensional evaluation matrix of an anti-noise confidence score, a semantic coherence index and a grammar error tag.

To provide a gustation improvement device which does not hinder an eating person from eating.SOLUTION: A gustation improvement device comprises: a first electrode to be attached to a first part being the vicinity of a mandibular branch in the face; a second electrode to be attached to a second part being the vicinity of the first part and being the vicinity of the mandibular branch in the face; a noise generator; and a current stimulation circuit which applies a noise generated by the noise generator, between the first electrode and the second electrode as a predetermined noise stimulation.SELECTED DRAWING: Figure 1

The utility model relates to the technical field of enterprise teaching and training, in particular to a noise generator for teaching and training, which comprises a framework, two groups of supporting pieces, a sliding piece, an air blowing pipe, a connecting hose, a regulating valve and a sound production block, the two groups of supporting pieces are arranged on the framework, the sliding piece is slidably connected with the two groups of supporting pieces, one end of the air blowing pipe penetrates through the sliding piece, and the other end of the air blowing pipe penetrates through the connecting hose. One end, far away from the air blowing pipe, of the connecting hose is connected with the adjusting valve, the sound production block is also arranged on the framework, and a first sound production hole, a second sound production hole and a third sound production hole are formed in the sound production block, so that the problems that in the prior art, most employees, especially new employees, have fuzzy concepts on noise levels, and the noise level is poor are solved. The technical problem that the risk of occupational diseases is increased due to the fact that sufficient noise protection knowledge is lacked and effective measures are difficult to take to protect a user is solved.

This invention discloses a method for establishing a network for accurate target tracking in noisy environments, comprising the following steps: setting the number of training iterations for adversarial noise training, the type and magnitude of noise generated by the noise generator, and the balance coefficient of the tracker's loss function; feeding a pair of images into the noise generator, which adds noise to the image pairs according to the set parameters, and then feeding the processed image pairs into the backbone network for feature extraction; classifying and regressing the response map obtained after feature extraction to obtain the tracking result of the current frame, and then iteratively running this process until all video frames of all video sequences in the test dataset are traversed; recording and saving the tracking result of each frame, and quantitatively analyzing the tracking accuracy and tracking success rate. This invention, while maximizing the BCE loss value, minimizes the final tracker loss value through training, thereby enabling the tracker to handle tracking in various noisy environments while maintaining tracking performance on clean videos.

A system for customization of on-road sound production for a vehicle includes one or more noise generators supported relative to the vehicle and configured to produce a soundscape in an environment of the vehicle. The system further includes memory storing instructions that, when executed by one or more processors, cause the processor(s) to generate the soundscape from a sound file based on an indication of a user-preferred sound file of multiple files, an environmental context surrounding the vehicle, or both. The instructions cause the processor(s) to determine whether an operational status of the vehicle is suitable for on-road operation. The instructions cause the processor(s) to further produce, in response to a determination that the operational status of the vehicle is suitable for on-road operation, the soundscape utilizing the noise generator(s) while the operational status of the vehicle is suitable for on-road operation.

The utility model discloses a multi-mode fusion safety protection device in a non-shielding mode. The multi-mode fusion safety protection device comprises double box bodies, a drawer, a white noise generator, a plurality of ultrasonic transmitters, a double-door structure and an incoming call sensing assembly. The double box bodies comprise an outer box body and a middle box body, the middle box body is slidably connected into the outer box body, and an air interlayer is arranged between the outer box body and the middle box body; the drawer is slidably connected into the middle box body through a sliding assembly. The double-door structure is installed at the front end of the drawer and used for sealing the outer box body and the middle box body. Wherein the inner walls of the middle box body and the drawer are respectively provided with a flexible sound insulation layer, and sealing assemblies are respectively arranged between the double-door structure and the front side of the outer box body and between the front sides of the middle box body. White noise and ultrasonic waves can be emitted into the drawer, the noise isolation effect is remarkably improved, information can be effectively prevented from being stolen, and the normal communication function of a mobile phone is guaranteed.

The invention discloses a novel air compressor energy-saving device, and belongs to the field of air compressors, the novel air compressor energy-saving device comprises an air compressor and a machine case, the air compressor is installed in the machine case, a supporting column is fixedly installed on the machine case, a noise generator and a storage battery are installed in the machine case, and electric quantity generated by the noise generator is stored in the storage battery; sound-absorbing cotton is arranged on the inner wall of the machine box, and the machine box is provided with a replacing assembly used for replacing the sound-absorbing cotton. A damping assembly used for reducing vibration of the air compressor is arranged in the machine box. The device has the effects of reducing noise, improving the working environment of workers and maintaining the physical and psychological health of the workers.

The invention relates to the technical field of computer security, in particular to a feature space backdoor attack method, system and device based on fusion data transaction model parameter constraint and a storage medium. Obtaining an original training data set, constructing a data transaction model, calculating malicious sample loss values, and screening samples with smaller loss values as target samples; poisoning features are randomly extracted from the original data set, discrete noise is generated through a noise generator to form a trigger, the trigger is added to a target sample, a label is turned over, and an attack data set is formed; original training data are sampled, and a toll-snow information matrix of each layer is calculated; using the original model parameters to initialize the backdoor model, adding the L1 norm of the parameter disturbance difference value and the parameter difference constraint term based on the Feisnow information matrix into a loss function, and training the backdoor model; performance fluctuation caused by random sample selection is avoided through sample screening based on loss values, so that the backdoor model maintains the classification capability of clean samples while learning trigger features.

A random number generator includes a pre-processing circuit and a chaos random number generator (CRNG). The pre-processing circuit includes a perturbation source and a control circuit, wherein the control circuit is coupled to the perturbation source. The CRNG, which is coupled to the pre-processing circuit, includes a chaos system and a post-processing circuit, wherein the post-processing circuit is coupled to the chaos system.

Some embodiments are directed to a brain scanning system including one or more noise generators configured to generate an electromagnetic noise field, the noise generator being arranged for positioning in proximity to the subject's head, wherein during operation of the brain scanning system the generated electromagnetic noise field interacts with the brain's electromagnetic field, resulting in an interfered noise signal.

The invention provides a weak signal noise shaping accumulation detection system and method, and the system comprises a noise generator which is used for generating noise with a set amplitude; the first input end of the summator is used for receiving a weak signal to be detected, the second input end of the summator is connected with the noise generator, and the summator is used for superposing the weak signal and the noise to obtain a synthetic signal; the input end of the analog-to-digital converter is connected with the output end of the summator, and the analog-to-digital converter is used for converting the synthesized signal into a digital signal; the input end of the signal accumulation and detection module is connected with the output end of the analog-to-digital converter, and the signal accumulation and detection module is used for sampling and accumulating the digital signals for multiple times so as to extract information of weak signals. According to the monitoring system provided by the invention, the detection of weak signals lower than the ADC quantization threshold is realized by superposing the noise reservation information, and the structure is simple and clear and easy to realize.

The invention belongs to the technical field of metal stretching testing, and particularly relates to a metal stretching control system and method.The system comprises a double-end synchronous stretching module used for applying an axial stretching load to a sample; the multi-frequency vibration excitation module integrates a mechanical vibration unit and an ultrasonic vibration unit, superposes vibration energy through a sweep frequency excitation technology, and is mechanically coupled with the middle area of the sample; the composite environment simulation module comprises a noise generator and a temperature adjusting unit and is used for simulating mechanical noise and temperature control compensation; the multi-source data fusion module is used for synchronously acquiring a stress-strain curve, morphological change and thermal field distribution through a laser vibration meter and an infrared camera, and eliminating signal leakage interference through wavelet transformation; and an intelligent control terminal. According to the invention, the data precision can be further improved, and the data reliability of subsequent resonance frequency analysis and temperature compensation and the active regulation and control of material performance are ensured.

The present invention belongs to the technical field of vibration testing of structural components, and specifically relates to a method for monitoring abnormal vibration of an aircraft, which solves the technical problems of low sampling accuracy and inability to achieve real-time online monitoring in existing methods. The method comprises building a sensor signal acquisition system, which includes a vibration measurement module, a multi-channel switching matrix, a noise generator, an ADC sampling chip, an FPGA controller, a USB controller, an eMMC memory, an RS422 transceiver, and a power management and power distribution unit; the FPGA controller compares the sensor data with a set threshold value, and if the threshold value is exceeded, it is considered that an abnormal vibration event has occurred at the corresponding position. At this time, a total of 500ms of data before and after the vibration occurs is stored in the eMMC memory, the FPGA controller analyzes the abnormal vibration data, calculates the position of the vibration source, and frames the vibration source position and sends it to a host computer.

The invention provides an automatic driving scene generation method based on multi-target antagonism optimization, which comprises the following steps of: 1, constructing a noise generator, taking a current noisy scene xt, a high-definition map code m and a time step t as input, and taking predicted noise epsilon theta (xt, t) as output; step 2, designing a multi-target evaluator Ltotal; wherein L < total > is a weighting function of a plurality of potential functions; the independent variable of the potential function is obtained from an estimated denoising scene; step 3, reverse sampling of adversarial guidance is executed; and step 4, outputting a final scene x0. On the premise of strict scene authenticity, the key confrontation scene with the highest test value for the automatic driving system is efficiently and controllably generated, and the test coverage rate and efficiency are greatly improved.

The embodiment of the invention provides a federal large model learning method based on confrontation feature aggregation and a multi-party computing system, and relates to the field of artificial intelligence. The method is applied to a multi-party computing system, the multi-party computing system comprises a plurality of participants and a central service party, and the participants are in communication connection with the central service party. The method comprises the following steps: for any one of a plurality of participants, the participants perform adversarial feature generation processing on local data based on a noise generator to obtain adversarial features, and the noise generator is used for generating noise data; uploading the confrontation features to the central service party; and the central service party performs model training on a specific task through an adversarial feature cluster to obtain a target model, and the adversarial feature cluster is obtained by aggregating the adversarial features issued by the plurality of participants. According to the method provided by the invention, the communication overhead between each participant and the server is reduced, and the privacy protection of local data is improved.

The application discloses a 12-lead electrocardiosignal denoising method and system based on a dual generative adversarial network, which comprises the following steps: obtaining time-adjacent clean electrocardiosignal segments and noisy electrocardiosignal segments; constructing a noise generative adversarial network, wherein the generator takes a random noise vector as input and simulates noise as output, the discriminator takes the noisy electrocardiosignal segment as a positive sample and a synthetic signal composed of the clean electrocardiosignal segment and the simulated noise as a negative sample for adversarial training, and a noise generator snapshot library is generated; performing dynamic noise injection on the clean electrocardiosignal segment to obtain 'noisy-clean' paired data; constructing a denoising network comprising a denoiser; obtaining the 12-lead electrocardiosignal containing noise and inputting it into the denoiser to output the denoised 12-lead electrocardiosignal. The application aims to overcome the defects in the prior art, such as the lack of real 12-lead electrocardiosignal denoising paired data and the poor model generalization ability caused by excessive dependence on the Gaussian noise assumption.

The embodiment of the invention discloses an adversarial attack method and device for improving digital human posture estimation, and the method comprises the steps: obtaining a to-be-processed image which is a monocular image in live broadcast; based on a dual heterogeneous noise generator, generating image-based noise beta1 and text-image integration-based noise beta2 corresponding to the to-be-processed image; optimizing the noise beta 1 based on the image and the noise beta 2 based on the text-image integration through an adversarial loss function and mean square error loss, and generating an adversarial noise beta; the adversarial noise beta is iteratively enhanced by adopting a PGD method to generate enhanced noise beta ', and an adversarial sample is generated based on the enhanced noise beta' and the image to be processed and is used for improving adversarial attack of digital human attitude estimation.

The invention relates to a method for establishing a bidirectional communication connection (8) by means of a media system (9) of an at least partially automated motor vehicle (10), in particular an electric vehicle, wherein the media system (9) has at least one microphone, preferably an external microphone, and at least one external loudspeaker (7) in the form of an AVAS, vehicle warning noise generator. In a first step, an audio command (3) uttered by a user (2) is received via the at least one microphone, the user (2) being outside the motor vehicle (10), the audio command (3) describing a hands-free function. In a second step, the received audio command (3) is authorized by the media system (9) to check its admissibility.In the case of a positive authorization, in a third step at least one corresponding control unit of the motor vehicle (10) is controlled by means of the media system (9) to establish a communication connection (8), whereby a bidirectional communication between the user (2) and an addressee (6) is enabled.

A device may generate, using a random telegraph signal (RTS) noise generator, a simulated RTS noise as input to a transistor included in an electronic circuit. The device may determine, based on the simulated RTS noise input to the transistor, a simulated output signal from the transistor. The device may automatically modify the electronic circuit based on the simulated output signal.

An approach for generating time-series dynamic-system training data. The approach may comprise providing a plurality of dynamic systems to a dynamic system dictionary. Where the dynamical system dictionary may comprise a library of functions. The approach may further comprise classifying each of the plurality of dynamic systems. Where classifying may comprise, generating a hierarchical dynamic system data, based on constraint learning, with an encoder and noise generator. The approach may further comprise training a diffusion decoder to generate a time-series segment, based on the classified plurality of dynamical systems. Further, the approach may comprise providing a first time series data segment and generating time-series dynamical training data based on the diffusion decoder using the first time-series data segment as input.

An aerosol delivery device (500) comprising: a controller (520); and a noise generator (580) comprising one or more ultrasonic transducers, the noise generator (580) being controllable by the controller (520) and being configured to generate an ultrasonic beam directed towards an intended position of a user of the aerosol supply device (500) such that an audible deterrent noise is generated at the intended position.