42results about How to "Improve feature extraction accuracy" patented technology

The invention relates to a multi-scale analysis method for powerhouse vibration, which is characterized in that it comprises the following steps: Step 1): Obtain the horizontal vibration velocity signals of the wind cover + Y in the middle layer of the powerhouse of a hydropower station under different working conditions; Step 2): adopt The self-adaptive iterative filtering method decomposes the horizontal vibration velocity signal of the wind cover + Y in the middle layer of the hydropower plant under each working condition, and obtains several stationary components corresponding to the horizontal vibration velocity signal of the wind cover + Y in the middle layer under each working condition ; Step 3): based on the percentile method, using several stationary components corresponding to the horizontal vibration velocity signal of the wind cover +Y of the middle layer of the hydropower plant building under each working condition, constructing the standard eigenvector of the plant vibration, which is used to identify the structure of the plant The multi-scale analysis of powerhouse vibration is completed, and the invention can be widely used in the field of operation and maintenance of powerhouse structures of hydropower stations.

The application discloses an image feature extraction method, a computer readable storage medium and a computer terminal. Wherein, the method includes: acquiring a target image; using a feature extraction model to perform feature extraction on the target image to obtain target features of the target image, wherein the target feature and the target image have variability such as rotation, and the feature extraction model includes: sequentially connected feature rotation Equivariant layer and at least one feature fusion layer, at least one feature fusion layer has different output resolutions, each feature fusion layer includes: rotation equivariant sublayer and multiple fusion sublayers, multiple fusion sublayers have the same structure . The present application solves the technical problem of low accuracy in feature extraction for target images corresponding to targets with variable angles in the related art.

The present application provides a method and system for extracting and classifying ECG signal features. The main steps of the method include: (1) sequentially locating the R wave, Q wave, S wave, T wave and P wave of the ECG signal according to the adaptive threshold, (2) Taking the segmentation point between cardiac cycles as the starting point to slide and intercept 1-second ECG signals, the window length of the sliding window is 1 second, and the step size is one cardiac cycle. domain feature, obtain the morphological feature for the complete cardiac cycle signal included in the 1-second signal, (4) use rough set to do attribute reduction, and filter effective features as the input of discriminant network; the sliding window length and variable step size of the present invention The settings ensure that the complete cardiac cycle signal will not be truncated, combined with fixed-length and single-cycle methods to simultaneously obtain time-frequency domain features and morphological features in one signal detection, increasing the dimension of the feature set; using rough sets to attribute the feature set Reduction, improve the efficiency and accuracy of classification and discrimination.