9results about How to "Solve the noise problem" patented technology

The utility model relates to the field of dimming drive circuits, and particularly discloses an LED analog-digital hybrid dimming drive circuit, which comprises an MCU controller and a dimming drive chip, the low-side output end of the dimming drive chip is electrically connected with a power switch element, the power switch element is electrically connected with the negative electrode of an LED lamp through an inductance element, and the power switch element is electrically connected with the negative electrode of the LED lamp through the inductance element. The PWM output end of the MCU controller outputs digital signals to the PWM input end of the dimming driving chip, a digital-to-analog conversion integrating circuit is arranged between the IO output end of the MCU controller and the ADJ input end of the dimming driving chip, and the IO output end of the MCU controller outputs digital signals to the digital-to-analog conversion integrating circuit. The digital-to-analog conversion integrating circuit converts the digital signal into an analog signal and transmits the analog signal to the ADJ input end of the dimming driving chip.

The application discloses a false information detection method, device and equipment based on a propagation forest, and the method comprises the following steps: acquiring propagation data of each source blog in a blog propagation in a social network for a source blog set to be detected, wherein the propagation data comprises propagation content and propagation relationship; extracting text features of the propagation content; calculating a propagation prototype and constructing a propagation tree after dividing the source blog set into K class clusters according to the text features; constructing a propagation forest based on the propagation prototype and the propagation tree; aggregating node neighborhood representations in the propagation forest to obtain embedding representations of the nodes; calculating enhanced embedding representations of source blog nodes and subsequent propagation blog nodes according to corresponding text features and embedding representations respectively; and classifying the enhanced embedding representations of the source blog nodes and the subsequent propagation blog nodes by applying a detection model to obtain false detection results of each source blog in the source blog set. The application can improve the performance of a false information detection task.

This invention discloses a smart contract similarity detection method that integrates global and local features. It extracts bytecode function blocks using a contract function block extraction tool, encodes these function blocks into feature vectors using the SimHash algorithm, and calculates the similarity matrix of the function blocks using the Hamming distance algorithm, thus obtaining local similarity values. Next, it extracts the contract's program dependency graph by adding control flow and data flow analysis, and then encodes the program dependency graph into feature vectors using a graph embedding model encoder based on the Graph2vec algorithm. These feature vectors are used as input to a similarity calculation function to obtain the global similarity value of the contract. Finally, the similarity result of the contract is calculated by fusing global and local features. This invention's method of integrating global and local features considers multi-dimensional contract code features, achieving more accurate detection results and possessing good versatility and practical value.

This utility model discloses a low-noise water heater, comprising: a heating element disposed at the bottom of a cavity; the heating element includes a heating tube and a heat spreader plate, the heating tube and the heat spreader plate being thermally connected, the heat spreader plate being attached to the bottom of the cavity, and the heating tube being spirally arranged; the ratio of the thickness of the heat spreader plate to the surface load of the heating tube is greater than or equal to 40%, and the thickness of the heat spreader plate is 3.5-5mm; wherein, the effective heat transfer area ratio of the heating tube and the heat spreader plate is greater than or equal to 50%, and the heating tube and the heat spreader plate cooperate to homogenize the heat generated by the heating tube; a first through hole is provided on the lower surface of the cavity, and a water inlet pipe passes through the spiral gap of the heating tube and the first through hole in sequence; a bracket is provided on the inner surface of the cavity, the bracket being disposed on the first through hole, and the bracket being used to buffer the water flow of the water inlet pipe. This device fundamentally solves the noise problem caused by uneven heating and bubble bursting during the heating process.

This application relates to the field of intelligent transportation and proposes a calibration method, device, equipment, and storage medium for roadside radar. The method includes: acquiring first positioning information and speed information of a vehicle, and second positioning information of a connected vehicle; clustering the first positioning information to generate a vehicle trajectory; determining the vehicle trajectory matching the second positioning information based on the relative positional relationship of the vehicle trajectories and pre-determined lane information of the connected vehicle; aligning the second positioning information and third positioning information in the vehicle trajectory according to time, and performing coordinate transformation calculations on the aligned positioning information to determine the calibration information of the roadside radar. Because this application generates vehicle trajectories by clustering based on the first positioning information and speed information, it can effectively solve the noise and discontinuity problems of roadside radar, and it eliminates the need to close roads when calibrating roadside radar on public roads.

The invention discloses a stomach intestinal metaplasia grading method and system based on noise robustness and an open set, and belongs to the technical field of medical image processing. Processing the stomach endoscope image based on a pre-trained intestinal metaplasia grading model, and outputting an intestinal metaplasia grading result; the intestinal metaplasia grading model comprises a noise identification module, a comparison learning module, a label correction module and a classification output module; the pre-training process comprises the following steps: obtaining a clean sample set and a noise sample set based on a noise identification module; calculating class prototype similarity based on a comparative learning module, and discriminating valid noise and invalid noise; calculating an effective noise confidence coefficient based on a label correction module, and generating a differentiated pseudo label; and performing multi-module cooperative training based on the joint optimization total loss function, and outputting an intestinal metaplasia classification result through a classification output module. According to the method, the robustness of the model in noise and open set scenes is effectively improved, and the problem of 0-2-level fine-grained classification of intestinal metaplasia is solved.

This invention discloses an air conditioner for use in the home appliance industry, comprising: an energy storage device for containing energy storage material; a jet driving device assembled in the energy storage device; and a flow distribution device connected to the energy storage device via the jet driving device. When the jet driving device acts on the energy storage device, the energy storage device jets the energy storage material into the flow distribution device, and the jetted energy storage material is dispersed and emitted in the flow distribution device to release heat or cold energy. This invention solves the noise problem of air conditioners.

The utility model relates to pipeline sound attenuation technical field, concretely relates to a kind of rhombic small hole pipeline silencer, including outer sleeve and sound attenuation component, sound attenuation component includes top sealing plate, bottom sealing plate, air outlet, air inlet, sound attenuation cylinder and sound attenuation cylinder cover plate, air inlet and air outlet are installed with pipeline, only need to reserve the installation length and space of overall silencer on site, sound attenuation cylinder is steel structure type, without sound attenuation cotton setting, without replacing sound attenuation cotton, reduce vibration, noise through the structure design of prismatic sound attenuation hole, improve sound attenuation effect, sound attenuation cylinder cooperates multiple prismatic sound attenuation hole, and the sound attenuation effect of circular small hole silencer is good, and compared with circular small hole, noise reduction is improved by 3-5dB, to generate local eddy current at rhombic corner, and the attenuation of pipeline low-frequency vibration and airflow pulse vibration plays an active role, to solve the noise problem generated by low-frequency pipeline airflow vibration.

The application relates to the technical field of intelligent application of automobiles, in particular to an automatic driving perception method and device of a vehicle, a vehicle and a storage medium, wherein the method comprises the following steps: acquiring initial picture data of the surrounding environment of the vehicle; inferring at least one target task according to the initial picture data, and detecting whether preset adversarial samples exist in the initial picture data in the inference process; if the initial picture data does not have the preset adversarial samples, obtaining a perception result of the initial picture data after the inference is completed; otherwise, the preset adversarial samples of the initial picture data are denoised by using a preset denoising strategy, first picture data after denoising is obtained, and the perception result of the initial picture data is obtained again based on the first picture data. Therefore, the problems that all data are usually denoised in the related art, the workload of data processing is large, the perception efficiency is low, and noise of input data of a perception task cannot be effectively avoided are solved.