61results about How to "Filter implementation" patented technology

The invention discloses a macro-block filtering method, which is used for filtering 9-bit or 10-bit coding source data. The method comprises the following steps: searching an ALPHA table and a BETA table which are preset and used for supporting 10-bit range data for a value corresponding to a quantification parameter; and filtering a boundary with a filtering intensity which is determined according to the type of a prediction mode, macro-block coding information and the position of the boundary in a macro block if the difference of the values of pixels on two sides of the boundary is confirmed to be less than the value corresponding to the quantification parameter, wherein the values of the pixels on each component comprise 9-bit or 10-bit data. The invention also discloses a device for realizing the method.

A preparation method of sub-wavelength continuous surface microstructure based on negative refractive index lens: (1) quantify and decompose the target structure, and make a half-tone photolithography mask with graphic feature size on the order of nanometers; (2) The surface of the sheet is coated with photoresist, and then a spacer layer is coated on the surface of the photoresist substrate, and then a metal layer with a thickness of about 10nm-120nm is evaporated on the surface of the spacer layer; (3) a spacer layer is coated on the surface of the metal layer, The prepared halftone photolithographic mask is tightly pressed on the surface of the gap layer; (4) the incident light is vertically irradiated on the surface of the halftone mask to realize the exposure of the photoresist; (5) after the exposure is completed, the gap layer and the metal layer, and develop the photoresist to obtain the desired continuous planar structure. The invention realizes the preparation of the sub-wavelength continuous surface structure, and can be widely used in the preparation of various new functional devices based on surface plasmons.

The present invention provides a hearing aid having a near field resonant parasitic element, including: a microphone; a signal processor for processing input audio signal to compensate a hearing loss of a user of the hearing aid; a speaker connected to an output of the signal processor for providing a sound signal to the user; a transceiver connected to the signal processor. A circuit system comprises at least one element and/or further electric element in the signal processor, the transceiver, interconnecting transmission lines and an antenna structure, the circuit system extends over an area of a support substrate. The hearing aid also comprises a resonant element within a near field of the circuit system to terminate and dissipate unwanted electromagnetic radiation from at least a part of the area. The resonant element realizes a notch filter filtering the unwanted electromagnetic radiation, and can be positioned outside any signal paths of the circuit system.

The present invention provides a filter comprising a substrate, a dielectric layer on the substrate, a first metal layer in the dielectric layer, a trough-line resonant cavity formed in the first metal layer to expose the dielectric layer under the first metal layer, an input feed line and an output feed line. According to the technical scheme of the invention, the trough-line resonant cavity is formed in the metal layer, so that signals can be transmitted via trough-lines. Therefore, the signal filtering effect is realized. The filter is compact in structure and can be easily manufactured through the integrated circuit process.

The invention provides a lung CT image classification method based on feature migration. According to the technical scheme, the depth and the number of freezing layers of the pre-trained GoogLeNet network are determined through preprocessing, data enhancement and feature visualization of the CT image data, intrinsic information of the image is effectively extracted, and the classification accuracyand the classification efficiency of the model are improved. According to the invention, an adaptive CT image filtering algorithm is adopted to filter images acquired by different devices; contrast stretching is carried out on the CT image, image information is enriched, and image features are better extracted; a selection is provided for network depth determination of the model by using a feature map visualization technology, and hyper-parameter adjustment is reduced. By applying the method, the feature extraction problem of the label-free image can be solved, the problem of dimensionality disasters faced by high-dimensional image data classification is relieved, and the accuracy of image classification is remarkably improved.

The embodiment of the invention discloses a power signal filtering method and system using Dantzig total gradient minimization. The method comprises the following steps: step 101, acquiring a signal sequence S acquired according to a time sequence; step 102, solving the total gradient sparsity p of the Dantzig; 103, solving a delay matrix D; step 104, solving a Dantzig total gradient vector gamma;and step 105, solving a signal sequence Snew after the noise is filtered out.

The invention provides an acquisition circuit of airborne vibration signals of an engine. A capacitor C4 used for filtering extremely-low frequency interference signals is added in front of a current limiting resistor of a negative input end of an operational amplifier of a single-ended electric charge amplification circuit. A feedback resistor R1 of the operational amplifier is connected to a joint between the capacitor C4 and the current limiting resistor. A feedback capacitor C3 is connected to the front end of the capacitor C4. The capacitance of the capacitor C4 is determined by working characteristics of the engine during change of rotation speed. According to the invention, complex interference signals in engine vibration signals of the engine can be effectively filtered and precise acquisition of the electric-charge type engine vibration signals is achieved.

The invention discloses a real-time filtering method for multi-frame joint detection and tracking algorithm output trace point sequences, and belongs to the technical field of radar target detection and tracking. In order to solve the problem that existing processing methods lack real-time performance in improving the tracking accuracy of a target, and the correlation between the trace point sequences output by a tracking algorithm before detection is difficult to calculate, a track point set belonging to the same target at the same time is first obtained according to the track point sequencesoutput by the tracking algorithm before detection in batches, then trace points in the track point set are separately subjected to Kalman filtering to obtain corresponding updated state estimate andposterior covariance estimate, and then a minimum value of the updated posterior covariance estimate is obtained based on the determinant minimization criterion to obtain corresponding weight values;and finally, the obtained weight values are used for obtaining the state estimate and the posterior covariance estimate after weighted sum, and the filtering of the track point sequences is completed.According to the real-time filtering method for the multi-frame joint detection and tracking algorithm output trace point sequences, the tracking accuracy of the tracking algorithm before detection is significantly improved.

The embodiment of the invention discloses a power signal filtering method and system utilizing an energy scale. The method comprises the following steps: step 101, acquiring a signal sequence S acquired according to a time sequence; step 102, solving an energy scale coefficient sequence; step 103, solving an edge energy scale coefficient; step 104, solving a direction energy scale coefficient; step 105, solving a direction scale threshold value; step 106, solving a high-density energy scale coefficient; and step 107, solving a signal sequence after noise filtering.

The invention belongs to the field of electromagnetic compatibility, relates to an anti-counterattack interference power supply filter for a direct-current system and aims to effectively suppresses reverse counterattack interference on a direct-current power supply bus by current surge generated in the operation of electrical equipment taking a high-power IGBT element as a core. The filter is mainly applied to a direct-current power supply system of a Dongfeng diesel locomotive, and is used for suppressing counterattack interference on a power supply bus caused by current surge generated during the operation of air conditioner frequency converter equipment connected with the power supply bus of the same direct-current system so as to improve the reliability and safety of other vehicle-mounted electrical equipment in the same direct-current system. The power supply filter provided by the invention comprises a set of full-balance structure pi-type low-frequency filter circuit module mainly used for suppressing electromagnetic interference of low-frequency components, a set of high-frequency filter circuit module which is composed of a full-balance four-order Butterworth low-pass filter circuit and is mainly used for suppressing the electromagnetic interference of medium-high frequency components, and a set of RDC damping absorption circuit module which is designed in a redundantmode and aims at suppressing high-amplitude direct current surge. According to the low-frequency filter circuit module and the high-frequency filter circuit module provided by the invention, the full-balance circuit structure form is adopted, so that surge of counterattack interference can be suppressed when fleeing into a bus from a positive electrode end or a negative electrode end.

The embodiment of the invention discloses a power signal filtering method and system using a conversion learning algorithm. The method comprises the following steps: step 101, acquiring a signal sequence S acquired according to a time sequence, step 102, solving the number of signal segmentation segments, step 103, solving a segmentation sequence, step 104, solving a sparse conversion control factor, step 105, initializing iteration process parameters, step 106, updating iteration process parameters, step 107, judging the value of the iteration control parameter and ending the iteration process, and step 108, solving a filtered signal sequence.