56 results about "Data truncation" patented technology

Different combinations of global motion parameters are estimated for a current frame interpolated to derive local motion vectors for individual image blocks. Image blocks in a reference frame identified by the local motion vectors are compared to the image blocks in the current frame. The estimated global motion parameters that provide the best match between the image blocks in the current frame and the reference frame are selected for encoding the current frame. Selected sub regions of temporally consecutive image frame s can be used in order to release the computational burden for global motion estimation and provide more robust global motion estimation results. A data truncation method can also be used to remove bias caused by foreground moving objects.

The invention relates to a three-dimensional digital imaging method of the large view field cone-beam X-ray tilting scanning of a biased detector, belonging to the technical field of X-ray computerized tomography (CT). In the three-dimensional digital imaging method, the area array detector is placed to be biased, and an X-ray source is used for generating cone-beam X-rays which irradiate a member imaging area with a certain angle in a penetrating and tilting way relative to the length and width surface of a member; in the scanning process, the X-ray source and the area array detector are static, the member rotates in an angle of 360 degrees in an equal angle and step length way surrounding a rotating shaft, and the area array detector acquires an X-ray signal modulated by the member under each rotation angle. A three-dimensional computerized tomography image of a scanning area can be reestablished and obtained by a data truncation smoothing preprocessing method and a filtering back projection reestablishing arithmetic according to data obtained by the area array detector under the scanning angle of 360 degrees. Compared with the traditional tilting scanning method, the three-dimensional digital imaging method can double the tilting scanning imaging view field without changing the system hardware and the scanning speed and has high reestablishment quality, simple process and high efficiency.

The invention discloses a time series data intelligent fault diagnosis method based on deep learning. The method comprises the following steps: 1) collecting one-dimensional time series health state data of electromechanical equipment, wherein fault data is the fault data arranged at a preset measuring point of the electromechanical equipment; 2) constructing a one-dimensional time series originalfault data set according to the acquired N health state data; 3) preprocessing the data, wherein the preprocessing includes normalization and data truncation; 4) constructing a two-dimensional feature map fault data set; 5) carrying out data set division; 6) constructing a deep learning fault diagnosis model; 7) training the deep learning fault diagnosis model to obtain parameters; and 8) performing fault diagnosis on the input to-be-diagnosed sample data by using the deep learning fault diagnosis model, and outputting a final fault diagnosis result. By constructing the two-dimensional feature map and improving a traditional convolutional neural network model structure, a fault diagnosis speed and fault diagnosis accuracy are improved.

The invention discloses a rotary machine fault diagnosis method based on an improved convolutional neural network, and the method comprises the following steps: (1) collecting one-dimensional time sequence fault data, and obtaining an original fault data set; (2) performing preprocessing operation on the acquired data of the original fault data set, wherein the preprocessing comprises standardization, data truncation and data reconstruction; (3) dividing the preprocessed samples of each type of faults into a training set, a verification set and a test set; (4) an improved convolutional neuralnetwork fault diagnosis model is established, the model comprises an input layer, a feature extraction layer, a dimension reduction and parameter reduction layer and a softmax classification output layer, and the dimension reduction and parameter reduction layer comprises a 1 * 1 transition convolution layer and a global mean pooling layer; (5) training and testing the model; wherein the diagnosismodel can automatically extract features and diagnose the fault data without any manual feature extraction operation, so that people can diagnose the fault of the rotating machine more conveniently and quickly.

The invention discloses a sample selection method based on Weibull distribution and right truncation data estimation reliability. The sample selection method comprises the following steps of 1) determining that a research object obeys the three-parameter Weibull distribution and the shape parameters; 2) estimating the position parameters and the scale parameters in the three-parameter Weibull distribution by using the relation between the bit rank and the sample size in the product failure probability; and 3) establishing a right truncation data truncation principle of the product life probability distribution by utilizing the shape parameters, the position parameters and the scale parameters of the three-parameter Weibull distribution according to the specific condition of the research object. According to the method, a shorter observation sample is discarded; the reliability analysis efficiency of estimating the product within the observation life time is improved to the maximum extent; the number of the practical application samples is reduced, the reliability analysis efficiency of the products based on the right truncation data is improved, and the method is especially suitable for the reliability analysis of the small-sample right truncation data products, is simple, has strong operability, has no special requirements for product types of the specific research objects, and has wide applicability.

The invention discloses a multi-channel time series data fault diagnosis method based on a convolutional neural network. The method comprises the following steps: (1) collecting multi-channel one-dimensional time series fault data of a monitored object; (2) constructing a multi-channel one-dimensional time sequence original fault data set, and carrying out normalization and data truncation preprocessing; (3) constructing a multi-channel two-dimensional feature map fault data set; (4) dividing into a training set, a verification set and a test set; (5) constructing a multi-channel deep learningfault diagnosis model which comprises an input layer, a feature extraction layer, a dimension reduction and parameter reduction layer, a softmax classification layer and a support vector machine output layer, wherein the dimension reduction and parameter reduction layer comprises a transition convolution layer with a convolution kernel of 1 * 1 and a global mean pooling layer; and (6) training, verifying and testing a multi-channel deep learning fault diagnosis model, and finally automatically outputting a diagnosis result by a support vector machine, so that people can diagnose faults of theelectromechanical equipment more intelligently and conveniently.

The invention relates to a nuclear medicine heart image quantification technology, in particular to to a technological method for multi-pinhole SPECT or SPECT/CT dynamic image quantitative reestablishment and myocardial blood flow absolute quantification measurement and application of the technological method in myocardial blood flow state evaluation. The technological method includes the specific steps of nuclide physical decay correction, in-scanning patient movement correction, scattering correction, geometric warping correction, data truncation compensation, tissue attenuation correction, noise removal, pixel value conversion, myocardial blood flow quantitative calculation and blood flow state evaluation. With the technological means, a quantitative multi-pinhole SPECT dynamic heart image can be generated, myocardial blood flow absolute quantification calculation can be carried out through the quantitative multi-pinhole dynamic image, and thus quantitative measurement of the myocardial blood flow is achieved; meanwhile, a blood flow state diagram is established with the three indexes of resting blood flow, load blood flow and blood flow reserve and is put into the practical application of evaluating the myocardial blood flow state, the purpose of carrying out quantitative measurement of the myocardial blood flow with multi-pinhole SPECT and SPECT/CT dynamic development is achieved, and the technological method can be applied to evaluation of the myocardial blood flow state.

The invention provides a method for enlarging an SFDR (spurious free dynamic range) of a signal after data truncation. A uniformly distributed random integer signal sequence in a certain amplitude range is generated through design, the bit width of the random integer signal sequence is identical with the truncated bit width, the random integer signal sequence is stacked on a signal before truncation, then bit width truncation is performed, and a random signal after truncation is obtained, so that the harmonic amplitude of the truncation error is decayed rapidly, harmonic distortion caused by direct truncation is eliminated, and the dynamic range of the random signal after truncation is enlarged.

The invention discloses a frequency and amplitude adaptive algorithm for a transient electrical signal. The algorithm includes the following steps: performing cubic spline interpolation on an electrical signal of a sample at a time domain; performing data truncation according to sampling intervals after obtaining an interpolation sequence; constructing reconstructed samples; and estimating the frequency and amplitude of each reconstructed sample by adopting a Prony algorithm. Through the verification of experimental results, the algorithm can rapidly and accurately realize the synchronous adaptive tracking of the frequency and amplitude of a power grid; and the algorithm is suitable for the synchronous detection of electrical parameters under the circumstances that the frequency and amplitude of a power grid signal are simultaneously changed with time, so that the algorithm is high in accuracy and can provide reliable and effective technical support for the accurate measuring and analysis of the three elements of the power grid.

The invention discloses an urban VOCs total pollution time sequence prediction method and system, a storage medium and equipment, and the method comprises the steps: preprocessing total VOCs monitoring data, embedding the data for projection, wherein the data comprises position information including year, month and day information, and also including the setting information of holidays and festivals; performing convolution pooling on the obtained data, reducing redundant data, generating an attention probability matrix through a sparse full-self-attention mechanism in an encoder to obtain a feature map, inputting the feature map once repeatedly to obtain the feature map, and transmitting the feature map to a decoder; cutting half of the data, inputting the data, performing the same operation, splicing the feature maps obtained twice to form a finally obtained feature map, and transmitting the finally obtained feature map to the decoder; and masking the data, inputting the data into the decoder, and obtaining a prediction result by the decoder according to the masked input and the prediction result obtained in the previous step. A longer time sequence can be processed, and the information of the time sequence can be extracted more effectively.

The invention provides a data processing method and a data processing system. The data processing method comprises the steps of receiving data transmitted by a client, wherein the data comprises header information, length information and actual data, the header information comprises encapsulation identification information, and the length information comprises the preset length of each data packet sequence; carrying out de-capsulation on the actual data according to the encapsulation identification information so as to acquire original data; and segmenting the original data according to the preset length of each data packet sequence so as to acquire at least one data packet sequence. According to the invention, de-capsulation is carried out on the encapsulated actual data according to the encapsulation identification information in the header information so as to acquire the original data, and segmentation is carried out on the original data according to the preset length of each data packet sequence in the length information so as to acquire at least one data packet sequence, thereby achieving the effects of data adhesion and data truncation.

The invention belongs to the technical field of Internet communication and network security, and particularly relates to an IPv6 (Internet Protocol Version 6)-based encodable and traceable network transmission rule verification method. According to the method, HASH operation is carried out on a digital object, and an IPv6 address is generated by an obtained code; meanwhile, the IPv6 address is used as a routing address for accessing the data; therefore, a network with data content as the center is formed, and transmission of digital objects can be controlled and traced from a network layer. Verifying the network transmission rule of the data IPv6; the network transmission rule verification comprises a data out-of-domain rule, a pre-audited data out-of-domain rule and data truncation of cross-border platform management. The digital object can be managed, controlled and traced from a network layer according to a network transmission rule.

The invention relates to a system and method for predicting a truncated image, a method for preparing data, and a medium therefor. The method for preparing the data comprises a virtual simulation stepused for carrying out virtual simulation on image data so as to obtain a virtual distorted image with data truncation and a virtual fine standard image without data truncation at the same time. The method for predicting the truncated image predicts the truncated image based on the trained learning network, and the trained learning network is obtained by performing data training based on a data set composed of the virtual distorted image and the virtual fine standard image which are obtained through the method for preparing the data. Also provided are a system corresponding to the method and arecording medium in which the recorded instruction can implement the method.